- Email: [email protected]
Extraskeletal myxoid chondrosarcoma: A clinicopathologic, immunohistochemical, and molecular analysis of 18 cases
Extraskeletal Myxoid Chondrosarcoma: A Clinicopathologic, Immunohistochemical, and Molecular Analysis of 18 Cases SUMIKA OKAMOTO, MD, MASANORI HISAOKA, MD, TSUYOSHI ISHIDA, MD, TETSUO IMAMURA, MD, HIROAKI KANDA, MD, SHYOHEI SHIMAJIRI, MD, AND HIROSHI HASHIMOTO, MD Extraskeletal myxoid chondrosarcoma (EMCS) is an uncommon clinicopathologically well-defined tumor, but its pathogenesis and biologic behavior are poorly understood. We reviewed 18 cases of EMCS to verify clinicopathologic features and immunohistochemical profiles together with molecular detection of the tumor-specific fusion genes. The tumors were located mainly in the proximal extremities and limb girdles (72%). Two tumors arose at unusual anatomic sites: the finger and the hip joint. Nine of the 17 followed-up patients were alive and disease free, 4 were alive with recurrences and/or metastases, and 4 died of the tumor. Fifteen tumors showed typical features of EMCS, and 3 had hypercellular areas in addition to conventional EMCS areas. The tumors were variably immunoreactive for S-100 protein (50%), NSE (89%), peripherin (60%), and synaptophysin (22%). Chromogranin A and some epithelial markers (AE1/ AE3, CAM5.2, and epithelial membrane antigen) were entirely negative. Frequent expressions of the neural/neuroendocrine markers suggest possible neural/neuroendocrine differentiation in at least some EMCSs, in addition to chondroid differentiation. In a reverse-
transcription polymerase chain reaction (RT-PCR) assay using paraffin-embedded specimens, EWS-CHN or TAF2N-CHN fusion gene transcripts characteristic of EMCS could be detected in 15 (83%) of the 18 cases: EWS-CHN type 1 in 11 cases, EWS-CHN type 2 in 1, and TAF2N-CHN in 3. Three fusion-negative cases included 2 conventional EMCSs and 1 considered a “cellular” variant of the tumor. None of 30 other soft tissue and bone tumors with myxoid or chondroid morphology that we examined contained these fusion genes. Thus, RT-PCR detection of EWS-CHN or TAF2N-CHN fusion gene using archival paraffin-embedded tissue is a feasible and useful ancillary technique for the diagnosis of EMCS. HUM PATHOL 32: 1116-1124. Copyright © 2001 by W.B. Saunders Company Key words: extraskeletal myxoid chondrosarcoma, fusion gene, reverse-transcription polymerase chain reaction. Abbreviations: EMCS, extraskeletal myxoid chondrosarcoma; RTPCR, reverse-transcription polymerase chain reaction; EMA, epithelial membrane antigen; PGK, phosphoglycerate kinase; PBGD, porphobilinogen deaminase.
Extraskeletal myxoid chondrosarcoma (EMCS) is a rare malignant soft tissue tumor described as a distinct clinicopathologic entity by Enzinger and Shiraki in 1972.1 The tumor usually develops in deep parts of the proximal extremities and limb girdles in middle-aged adults. EMCS has a prolonged and indolent clinical course, with a high rate of local recurrences and distant metastases; tumor-related death often occurs after a long survival period.2-5 The tumor is histologically characterized by multinodular growth of a lacelike arrangement of primitive chondroid cells in an abundant myxoid matrix. In addition to such a typical feature, cellular areas devoid of a myxoid matrix with or without epithe-
lioid or rhabdoid cells or showing nuclear pleomorphism are occasionally observed.1-7 Meis-Kindblom et al,4 noted that histologic grading was of no prognostic value in their large series of EMCS, whereas other investigators have reported that cases of high-grade EMCS have a truly adverse clinical outcome, as shown by aggressive tumor growth and death after a short period.5-7 Although chondroid differentiation of the constituent cells of EMCS has been supported by ultrastructural and histochemical observations,1,8-13 the histogenesis of this type of tumor is still controversial. In some cases of EMCS, a characteristic ultrastructural finding of long, parallel microtubules within rough endoplasmic reticulum is seen,12,14-18 but its significance is still unproven. More recently, the histochemical, immunohistochemical, and ultrastructural evidence of neural/ neuroendocrine differentiation of the tumor has been documented.5,19-21 Cytogenetic and molecular studies of EMCS have shown that approximately 75% of the tumors have a characteristic t(9;22)(q22;q12) resulting in a fusion of the EWS gene at 22q12 and the CHN gene (also referred to as TEC, NOR-1, or MINOR) at 9q22.6,14,22-29 EWS-CHN fusion proteins may induce tumorigenesis in EMCS by activating the expression of CHN target genes, but no putative target genes have been identified.30 More recently, a novel TAF2N-CHN fusion gene resulting from t(9;17)(q22;q11.2) has been identified. The TAF2N (also known as hTAFII68 or RBP56) gene
From the Department of Pathology and Oncology and the Department of Pathology and Cell Biology, School of Medicine, University of Occupational and Environmental Health, Kitakyushu; the Department of Pathology, Faculty of Medicine, University of Tokyo Hospital, Tokyo; the Department of Surgical Pathology, School of Medicine, Teikyo University, Tokyo; and The Cancer Institute, Japanese Foundation for Cancer Research, Tokyo, Japan. Accepted for publication May 15, 2001. Supported in part by grants-in-aid from the Ministry of Education, Science, Sports and Culture, Japan (12770102 and 12670184), and by a grant from the Fukuoka Cancer Society, Japan, 2000. Address correspondence and reprint requests to Hiroshi Hashimoto, MD, Department of Pathology and Oncology, School of Medicine, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807-8555, Japan. Copyright © 2001 by W.B. Saunders Company 0046-8177/01/3210-0014$35.00/0 doi:10.1053/hupa.2001.28226
1116
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
has sequence homology to the EWS gene,31-33 and the TAF2N-CHN fusion gene’s function is considered identical to that of the EWS-CHN fusion gene. We reviewed immunohistochemical and molecular features of 18 cases of EMCS to verify the expressions of neural/neuroendocrine markers and the diagnostic usefulness of reverse-transcription polymerase chain reaction (RT-PCR) detection of the EWS-CHN or TAF2NCHN fusion gene using archival formalin-fixed, paraffin-embedded tumor specimens. MATERIALS AND METHODS Materials Eighteen cases of EMCS were selected from the files of soft tissue tumors of the authors (H.H., T. Ishida, T. Imamura, and H.K.), and the original hematoxylin and eosin–stained slides of each case were reviewed. In 1 of the 18 cases, the clinicopathologic, ultrastructural, and cytogenetic findings had been reported previously (case 5).25 Initially, 21 cases were originally diagnosed as EMCS, but 3 of them were excluded from the present study because they were reclassified as other tumor entities after the critical review. One of these 3 cases was grade 2 chondrosarcoma of bone with prominent myxoid change, involving the adjacent soft tissue. Another was considered malignant myoepithelioma (mixed tumor) of soft parts with rhabdoid features rather than EMCS on the basis of positive immunoreactivity to cytokeratins, epithelial membrane antigen (EMA), actins, and S-100 protein. In the remaining case, a possible diagnostic consideration of low-grade myxofibrosaroma/myxoid malignant fibrous histiocytoma emerged. The other 30 soft tissue and bone tumors were included in this study as negative controls: 3 myxoid liposarcomas, 3 myxoid malignant fibrous histiocytomas, 3 malignant peripheral nerve sheath tumors, 1 lowgrade fibromyxoid sarcoma, 1 extraskeletal mesenchymal chondrosarcoma, 3 intramuscular myxomas, 2 juxtaarticular myxomas, 3 nerve sheath myxomas, 1 ossifying fibromyxoid tumor of soft parts, 1 extraskeletal chondroma, 3 chondrosarcomas with myxoid change, 3 chordomas, and 3 enchondromas.
Light Microscopy Each tumor tissue fixed in 10% formalin was routinely processed and embedded in paraffin. Three-m-thick sec-
tions were recut and stained with hematoxylin and eosin. Periodic acid–Schiff staining with and without diastase digestion was also performed to verify the presence of intracytoplasmic glycogen granules in tumor cells. Sections were stained immunohistochemically by a streptoavidin-biotin-peroxidase complex technique using the commercially available primary antibodies listed in Table 1.
Molecular Analysis RNA Extraction. Five 4-m-thick sections sliced from paraffin blocks were deparaffinized with xylene and ethanol. After drying, 200 L of lysis buffer (20 mmol/L Tris-HCL pH 8.0, 20 mmol/L EDTA, 2% sodium dodecyl sulfate) was added to each tissue pellet. Each pellet was homogenized with a hand homogenizer, 10 L of proteinase K (100 mg/mL) was added, and then it was incubated overnight at 55°C. One milliliter of Trizol reagent (Gibco BRL, Gaithersburg, MD) was added, and total RNA was extracted according to the manufacturer’s instructions. The RNA pellet was resuspended in 10 L of DNase/RNase-free water and then treated with DNase I (Gibco BRL) for 15 minutes at 37°C. RT-PCR. Ten microliters of dissolved RNA was reverse transcribed into complementary DNA (cDNA) using 1 L of random primer (Gibco BRL) and 200 U reverse transcriptase (SuperScript II; Gibco BRL). The cDNA was then incubated with 60 U RNase H (Takara, Ohtsu, Japan) for 20 minutes at 37°C. The integrity of the RNA for RT-PCR was confirmed by amplification of the phosphoglycerate kinase (PGK) transcript (247 base pairs [bp]) and porphobilinogen deaminase (PBGD) transcript (127 bp) using the following primer sequences: PGK forward, 5⬘-CAGTTTGGAGCTCCTGGAAG-3⬘; PGK reverse, 5⬘-TGCAAATCCAGGGTGCAGTG-3⬘; PBGD forward, 5⬘-TGTCTGGTAACGGCAATGCGGCTGCAAC-3⬘; and PBGD reverse, 5⬘-TCAATGTTGCCACCACACTGTCCGTCT-3⬘. PCR was carried out to amplify EWS-CHN and TAF2N-CHN fusion gene transcripts using specific sets of primers (Table 2). Three variations of the EWS-CHN fusion gene have previously been identified: EWS exon 12–CHN exon 3 (EWS-CHN type 1), EWS exon 7–CHN exon 2 (EWSCHN type 2), and EWS exon 11–CHN exon 2 (EWS-CHN type 3).6 Three sets of the primer for these EWS-CHN fusion gene transcripts were designed by the Oligo Primer Analysis Software (Molecular Biology Insight, Cascade, CO). The previously described primer sets for TAF2N-CHN fusion gene transcripts31 were used. The PCR was performed with 2.5 U Taq DNA polymerase (AmpliTaq Gold; Perkin Elmer, Norwalk, CT) 1.5 mmol/L MgCl2, 10⫻ PCR buffer (Perkin
TABLE 1. Primary Antibodies and Sources Antibody
Source
Dilution
S-100 protein NSE (BBS/NC/VI-H14) Peripherin* Synaptophysin (SY38) PGP9.5 (13C4/31A3) Chromogranin A (CAK-A3) Vimentin (V9) AE1/AE3† CAM5.2† EMA (E29) ␣–Smooth muscle actin (1A4) HHF35 Desmin (D33)†
Dako (Glostrup, Denmark) Dako Chenicon (Temecula, CA) Dako UltraClone (Rossiters Farmhouse, Wellow, Isle of Wight, England) Dako Dako Dako Becton Dickinson (Mountain View, CA) Dako Dako Enzo (Farmingdale, NY) Dako
1:200 1:200 1:500 1:100 1:400 1:100 1:30 1:50 Postdilluted 1:100 1:150 1:50 1:100
*For peripherin immunohistochemistry, microwave pretreatment (3 ⫻ 5 minutes) in 10 mmol/L citrate buffer (pH 6.0) was performed. †Sections were pretreated with protease (0.15%) for 20 minutes at 37°C for these markers.
1117
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
TABLE 2. Primer Sequences and PCR Product Sizes Primer
Nucleotide Sequence
EWS-CHN type 1 EWS exon 12 CHN exon 3 EWS-CHN type 2 EWS exon 7 CHN exon 2 EWS-CHN type 3 EWS exon 11 CHN exon 2 TAF2N-CHN TAF2N exon 6* CHN exon 3*
PCR Product (bp)
5⬘-GCGATGCCACAGTGTCCTATG-3⬘ 5⬘-ATATTGGGCTTGGACGCAGGG-3⬘
89
5⬘-CTCCAAGTCAATATAAGCCAAC-3⬘ 5⬘-GGACGTCCGGCGAGGCGAAGC-3⬘
146
5⬘-TCTGGCAGACTTCTTTAAGCA-3⬘ 5⬘-GGACGTCCGGCGAGGCGAAGC-3⬘
133
5⬘-GAGCAGTCAAATTATGATCAGCAGC-3⬘ 5⬘-CCTGGAGGGGAAGGGCTAT-3⬘
137
* Data from Sjo¨gren et al.31
Elmer), 200 mol/L deoxynucleotide triphosphates, and 0.2 mol/L for each set of primers. The amplification profile of the PCR consisted of 40 cycles of denaturation at 94°C for 1 minute, annealing at 55°C for 45 seconds, and elongation at 72°C for 50 seconds, followed by a final extension at 72°C for 10 minutes. In the PCR procedure, a reaction mixture of the reagents devoid of template cDNA and a no–reverse transcription control were included as negative controls. For visualization, 5 L of PCR product was stained with ethidium bromide and electrophoresed on a 3% agarose gel. To confirm the sequence of the fusion gene transcripts, the PCR products were cloned in a pCR2.1 vector (TA Cloning Kit; Invitrogen, San Diego, CA) and sequenced using an automated sequencer system, ALF express DNA sequencer (Pharmacia Biotech, Uppsala, Sweden).
RESULTS Clinical Findings Clinical details of the 18 patients are summarized in Table 3. There were 10 men and 8 women, whose ages ranged from 26 to 75 years (median, 45 years; mean, 47.4 years). The tumors were located in the thigh (8 cases), upper arm (2 cases), lower leg (2 cases), buttock (1 case), shoulder (1 case), back (1 case), foot (1 case), finger (1 case), and hip joint (1 case). The duration of a symptom before diagnosis ranged from 2 months to 25 years (median, 4 years; mean, 5.1 years). All patients underwent surgical treatments: wide local resection in 13 patients,
TABLE 3. Clinicopathologic Features of EMCSs Age (yr)/ Sex
Specimen
Location
Tumor Size (cm)
Preoperative Duration
1st recurrence [primary] Primary Primary Primary
Upper arm
2 3 4
74/M [59] 46/M 26/M 35/F
12 ⫻ 8 [5] 4.5 ⫻ 3.5 ⫻ 2.5 2 9⫻7⫻4
7 yr
Lower leg Hip joint Thigh
5
58/M
Primary
Thigh
6 7 8 9
75/F 51/F 58/F 34/F
Primary Primary Primary 1st recurrence
Thigh Buttock Thigh Groin
10 11 12 13
54/F 45/M 58/M 41/M
1st recurrence Primary Primary Primary
Back Thigh Finger Upper arm
14 15
37/M 34/F
Primary Primary
16 17 18
42/F 45/F 41/M
Primary Primary Primary
Case 1
Treatment
Follow-up Multiple recurrences; alive at 2 yr
1 yr 1 yr 5 yr
Intralesional resection [marginal resection] Wide resection Total hip arthroplasty Wide resection
8 ⫻ 7.8 ⫻ 5
4 yr
Wide resection
11 ⫻ 8 ⫻ 6.5 11 ⫻ 8 ⫻ 8 7⫻7⫻6 5
25 yr 3 yr 1 yr 4 mo NA
Wide Wide Wide Wide
2.2 ⫻ 1.0 10 ⫻ 7.5 ⫻ 5.5 6⫻4 5
2 mo 5 yr 1 yr 6 mo 2 yr
Wide resection Wide resection Marginal resection Wide resection
Thigh Thigh
24 ⫻ 11 ⫻ 8 8 ⫻ 5.5 ⫻1
5 yr 14 yr
Lower leg Shoulder Foot
3⫻2⫻1 4 8 ⫻ 7.5 ⫻ 7
5 yr 1 yr 6 yr
Abbreviations: NED, no evidence of disease; NA, information not available.
1118
resection resection resection resection
Wide resection Radiation, chemotherapy, wide resection Wide resection Amputation Below-knee amputation
NED; alive at 4 yr NED; alive at 10 yr Multiple recurrences and metastases; died at 15 yr Lung metastasis at 9 yr, resection; alive at 9 yr NED; alive at 7 mon NA Died at 8 yr 6 mo Recurrences, resection; alive at 17 yr NED; alive at 4 yr 6 mo NED; alive at 3 yr NED; alive at 1 yr 8 mo Recurrence at 5 yr, resection, multiple recurrences; died at 18 yr NED; alive at 1 yr 4 mo Lung metastases at presentation; alive at 11 yr NED; alive at 4 yr NED; alive at 33 yr Popliteal and lung metastases at 3 yr; died at 6 yr
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
amputation in 2, marginal/intralesional resection in 2, and total arthroplasty in 1. One patient received preoperative radiotherapy and chemotherapy (case 15). Follow-up information was available for 17 patients, and the follow-up period ranged from 7 months to 33 years. Nine patients were alive without any evidence of disease 7 months to 33 years after diagnosis, and 2 were alive and disease-free after surgery for lung metastasis (case 5) and local recurrences (case 9), respectively. Another 2 patients had multiple recurrences or metastases and were alive with disease (case 1 and 15). The remaining 4 patients died from the tumor at 6, 8.5, 15, and 18 years after surgery. Pathologic Findings Tumors ranged from 2 to 24 cm in size (median, 6 cm; mean, 7.8 cm). Most tumors were deep-seated and arose within the skeletal muscle (cases 1, 2, 4 through 11, 13 through 15, 17, and 18), tendinous structures of the phalanx (case 12), synovial tissue involving the adjacent femoral neck (case 3), and subcutaneous adipose tissue (case 16). In 1 tumor arising in the skeletal muscle, the adjacent humeral cortex was involved (case 13). All 18 tumors had a distinct multinodular configuration delineated by fibrous connective tissue. Most of the tumors had expansive growth, but 4 had infiltrative growth to the surrounding soft tissues (cases 1, 6, 12, and 16). Most of the tumor cells had abundant intracytoplasmic diastase-digested periodic acid–Schiff–positive granules, probably of glycogen. None of the tumors formed discernible cartilaginous structures. Fifteen of the 18 tumors had typical features of EMCS: a lobular proliferation of spindle and oval cells arranged in cords or strands or in an lacelike pattern and embedded in an abundant myxoid matrix, which was often basophilic (cases 1 through 15; Fig 1A). Some also had areas of small nests or whorls of the tumor cell (cases 5, 6, and 13; Fig 1B). The individual tumor cells had spindle-shaped or oval hyperchromatic nuclei that occasionally contained small nucleoli and showed mild to moderate nuclear atypia, and they also had a variable amount of eosinophilic cytoplasm (Fig 1C). Monovacuolated or multivacuolated tumor cells, resembling lipoblasts, were present in 3 cases (cases 2, 3, and 14; Fig 1D). Small clusters of rhabdoid cells were seen in 1 case (case 11). In 1 patient who had received preoperative radiotherapy and chemotherapy, the tumor showed areas with tumor giant cells that had pronouncedly pleomorphic or bizarre nuclei (case 15; Fig 1E). Mitotic figures were rarely encountered (mean 0 to 1 per 10 high-power fields [hpf]; magnification ⫻400). Hemorrhagic areas or aggregates of siderophages (9 of 15 cases) and tumor necroses (7 of 15 cases) were often seen. The other 3 tumors had multiple lobules consisting of a cellular proliferation of tumor cells embedded in a fibrocollagenous or fibromyxoid stroma, together with areas of conventional EMCS (cases 16 through 18; Fig 2A). In 1 of the 3 tumors, constituent tumor cells in cellular areas were round to oval and vacuolated and frequently had indented nuclei and perinuclear empty
spaces resembling lacunae (case 16; Fig 2B). In the cellular portions of another tumor (case 17), uniform and rather small tumor cells with indistinct cell borders were diffusely distributed (Fig 2C). Areas of a fascicular proliferation of atypical spindle cells were observed in cellular areas of case 18, which also showed small foci with atypical epithelioid or rhabdoid cells (Fig 2D). Tumor necrosis and hemorrhage were present in 2 (cases 16 and 18). Mitotic figures were few (0 to 1 per 10 hpf) in 2 tumors (cases 16 and 17), and more frequent (4 per 10 hpf) in 1 (case 18). Immunohistochemical Findings Immunohistochemical staining results of each case are listed in Table 4. S-100 protein was expressed in half of the tumors: strong and diffuse or focal immunoreactivity in 6 cases, and weak and focal in 3. NSE was expressed in most of the tumors (89%) in 10% to 80% of tumor cells. Most of the tumor cells in approximately half of the tumors (53%) were positively immunoreactive to peripherin (Fig 3A). Four (22%) of the 18 tumors had diffuse positive immunoreactivity for synaptophysin (Fig 3B). Diffuse expression of PGP9.5 was observed in 1 tumor (case 18). None of the tumors expressed chromogranin A. Diffuse and strong vimentin immunoreactivity was seen in most of the tumors examined (94%). Some tumors were focally immunostained for ␣–smooth muscle actin (cases 5,16, and 17; 17%) and HHF35 (cases 5 and 16; 11%), but all tumors were negative for desmin. No immunoreactivity for AE1/AE3, CAM5.2, and EMA was detected in any of the tumors. Molecular Findings Results of RT-PCR assay are summarized in Table 4. In the RT-PCR assay, the fusion gene transcripts could be amplified in 15 (83%) of the 18 paraffinembedded tumors (Fig 4). EWS-CHN type 1 fusion gene transcripts were detected in 11 (60%) cases. EWSCHN type 2 and TAF2N-CHN fusion gene transcripts were seen in 1 (6%) and 3 (17%) cases, respectively. Each type of the fusion genes was further confirmed by a subsequent sequence analysis (Fig 5). EWS-CHN type 3 fusion was not detected in any of the tumors. Three EMCSs were negative for the fusion gene, although the presence of adequate RNA was confirmed by the PCR amplification of the PBGD and/or PGK gene transcripts (cases 10, 11, and 18). The EWS-CHN or TAF2NCHN fusion gene transcripts were not amplified in any of the 30 other soft tissue and bone tumors used as negative controls or the 3 tumors reclassified into other types of tumor than EMCS. DISCUSSION EMCSs account for approximately 2.5% of soft tissue sarcoma,9 commonly occur in middle-aged adults, and arise mainly in the proximal extremities and limb girdles. Generally, EMCS is considered a tumor showing chondroid differentiation. Chondroid differ-
1119
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
FIGURE 1. (A) Low-power view of EMCS showing multilobular growth of tumor cells embedded in an abundant myxoid matrix. (Case 12; hematoxylin and eosin, original magnification ⫻10.) (B) Whorl arrangements of tumor cells. (Case 6; original magnification ⫻33.) (C) Individual tumor cells showing mild to moderate nuclear atypia. (Case 2; original magnification ⫻100.) (D) Monovacuolated or multivacuolated tumor cells, reminiscent of lipoblasts. (Case 14; original magnification ⫻100.) (E) Tumor cells showing pronounced nuclear pleomorphism in a postradiotherapeutic and chemotherapeutic state. (Case 15; original magnification ⫻25.)
entiation has been suggested by the results of histochemical or immunohistochemical studies showing the presence of chondroitin-4 and 6-sulphate in an intercellular matrix in EMCS9,11,13 and S-100 protein positivity,3,13 by the ultrastructural features of constituent cells of EMCS bearing some resemblance to chondroblasts,1,8 –12 and by morphologic similarities between EMCS and embryonic chondrogenesis.34 Furthermore, a recent report showing the capability of the tumor cells to synthesize type II collagen has provided additional evidence for chondroid differentiation of this type of tumor.20 On the other hand, recently, several investigators have described neural/neuroendocrine
differentiation of EMCS by immunohistochemical and ultrastructural observations.5,19 –21 Oliveira et al5 reported that synaptophysin was expressed in 13 (72%) of 18 cases of EMCS, and chromogranin A and Leu-7 were expressed in 1 case each. Harris et al20 showed a unique example of EMCS presenting unequivocal features of both chondroid and neuroendocrine differentiation. Ultrastructurally, this case showed tumor cells containing neuroendocrine granules in addition to positive immunoreactivities to synaptophysin and chromogranin A. Cummings et al21 documented that 3 of 5 cases of EMCS were positively immunoreactive to peripherin, a type III intermediate filament of 56 kd, was
1120
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
FIGURE 2. (A) Low-power view of a cellular variant EMCS consisting of lobules of both myxoid and cellular components. (Case 16; original magnification ⫻13.) (B) Cellular area of round tumor cells with indented nuclei and lacuna-like spaces in the cytoplasm embedded in a fibrocollagenous matrix. (Case 16; original magnification ⫻100.) (C) Transition between myxoid and cellular areas. The latter consists of a haphazard proliferation of rather small tumor cells showing indistinct cell borders set in a fibromyxoid matrix. (Case 17; original magnification ⫻50.) (D) Area of a fascicular proliferation of atypical spindle cells. (Case 18; original magnification ⫻50.)
involved in the growth and development of the peripheral nervous system and was expressed in the neural crest and neural tube– derived cells.35 Our immunohistochemical results support these previous reports and suggest neural/neuroendocrine differentiation in EMCSs. Pluripotential precursor cells of the neural crest can give rise to many derivatives, including neurons, and cranial skeletal components such as cartilage, bone, and dermis.36 The divergent differentiation in EMCS may imply a neuroectodermal derivation of the tumor, although chondrocytes at other sites, such as the extremities, vertebrae, and ribs, are thought to be derived from the somatic mesoderm.37 In the 18 cases of EMCS in the current study, 2 cases occurred in the finger and the hip joint, respectively, where EMCS rarely develops.3,11,16,38 In addition to conventional morphologic appearances of EMCS, the detection of the EWS-CHN fusion gene transcripts could confirm the diagnosis in these cases. Three of our 18 tumors had cellular areas intermingled with conventional myxoid areas of EMCS and were considered a cellular variant of EMCS.4 The histologic variation of this type of tumor, including the cellular variant,
raises a potential diagnostic problem. We could detect the EWS-CHN fusion gene transcripts in 2 of the cellular variants. Even in 1 case in our series showing nuclear pleomorphism, probably caused by preoperative radiotherapeutic effects,39 the TAF2N-CHN fusion gene transcripts were identified. These results indicate an important role of molecular detection of the specific fusion genes in the diagnosis of EMCS, particularly in clinicopathologically unusual cases. Previous molecular studies have shown that the EWSCHN fusion gene was associated with approximately 75% of EMCSs.6,27 Recently, a novel TAF2N-CHN fusion gene was found in EMCS.31–33 In 3 of the current cases, including 1 cellular form, the EWS-CHN and TAF2N-CHN fusion gene transcripts were not detected. A hitherto unknown variation of the gene might be involved in cases negative for both kinds of fusion genes, although negative results do not always indicate absence of the fusion gene in RT-PCR using formalin-fixed, paraffin-embedded tumor tissues. This speculation could be partially supported by the presence of some karyotypic profiles of EMCS other than t(9;22)(q22;q12) and t(9;17)(q22;q11.2), such as t(9;17;15)(q22;q11;q22), t(2;13)(q32;p12), and t(11;
1121
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
TABLE 4. Immunohistochemical and Molecular Findings of EMCSs RT-PCR Case S-100 NSE Peripherin Synaptophysin PGP9.5 Chromogranin A Vimentin AE1/AE3 CAM5.2 EMA ␣-SMA HHF35 Desmin PBGD/PGK Fusion Gene ⫺
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
2
⫹* ⫹
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
3
⫹ ⫹
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
4
⫺ ⫹*
⫺
⫺
⫺
⫺
⫹*
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
5
⫹ ⫹
⫺
⫹
⫺
⫺
ND
⫺
⫺
⫺
⫹
⫹
⫺
⫹/⫹
6
⫹ ⫹
⫺
⫺
ND
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
7
⫹ ⫹
⫹
⫹
ND
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
8
⫹ ⫹
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
9 10 11
⫺ ⫹ ⫺ ⫺ ⫹* ⫹
⫹ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫹ ⫺ ⫹
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫹/⫺ ⫹/⫹ ⫹/⫺
1
⫺
12
⫹ ⫹
⫺
⫺
ND
⫺
⫹*
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
13 14 15
⫹* ⫹ ⫺ ⫹ ⫺ ⫹
⫹ ⫹ ND
⫺ ⫺ ⫹
⫺ ⫺ ND
⫺ ⫺ ⫺
⫹ ⫹* ND
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ND
⫺ ⫺ ⫺
⫹/⫺ ⫹/⫺ ⫹/⫺
16
⫺ ⫹
⫺
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫹
⫹
⫺
⫹/⫺
17 18
⫺ ⫹ ⫺ ⫹
⫺ ⫹
⫺ ⫹
⫺ ⫹
⫺ ⫺
⫹ ⫹
⫺ ⫺
⫺ ⫺
⫺ ⫺
⫹ ⫺
⫺ ⫺
⫺ ⫺
⫹/⫹ ⫹/⫹
TAF2N-CHN EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 not detected not detected EWS-CHN type 2 EWS-CHN type 1 TAF2N-CHN TAF2N-CHN EWS-CHN type 1 EWS-CHN type 1 Not detected
Abbreviations: ␣-SMA, ␣–smooth muscle actin; ND, not done. *Focal immunoreactivity.
22)(q11;p11).23,24,40 Regarding the frequency of the fusion types, Antonescu et al6 found the EWS-CHN type 1 in 3 (33%) of 9 EMCSs examined, the EWS-CHN type 2 in 2 (22%), and the EWS-CHN type 3 in 1 (11%), and Labelle et al27 described that type 1 was detected in 3 (38%) of 8 cases of EMCS investigated and type 2 in 2 (25%). Although the results of our molecular assay also showed that most of the tumors had the EWS-CHN type 1 fusion gene, the number of cases examined is too small to draw a conclusion. According to the study by Meis-Kindblom et al,4 adverse prognostic factors of a patient with EMCS were larger tumor size, proximal tumor location, older pa-
tient age, and metastasis. They mentioned that highgrade histologic features, including tumor necrosis, increased cellularity, high mitotic rate, nuclear pleomorphism, epithelioid cells, rhabdoid cells, and spindle cell foci, did not affect prognosis. In contrast, some other reports have shown that tumor cellularity predicted the prognosis of a patient with an EMCS,1,12,13 and Oliveira et al5 documented that large tumor size, high cellularity, presence of anaplasia or rhabdoid features, high mitotic activity (⬎2 per 10 hpf), and high Ki-67 labeling index were statistically associated with a poor prognosis. In our 17 cases of EMCS available for follow-up information, 2 of the 3 patients with the
FIGURE 3. (A) Most of the tumor cells were positively immunostained for peripherin. (Case 1; original magnification ⫻100.) (B) Diffuse synaptophysin immunoreactivity. (Case 5, original magnification ⫻100.)
1122
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
FIGURE 4. RT-PCR detection of EWS-CHN or TAF2N-CHN fusion gene and 2 ubiquitous genes (PGK and PBGD) in formalin-fixed, paraffin-embedded EMCS tumor tissues. Numbers of the lanes are identical to the case numbers in Table 3. M, size marker (100-bp DNA ladder); N, negative control (distilled water).
cellular variant EMCS were alive without any recurrence or metastasis, and 1 died of the tumor. On the other hand, tumors in 3 of the 4 patients whose cause of death was the tumor showed conventional morpho-
FIGURE 5. Partual nucleotide sequences of (A) EWS-CHN type 1, (B) EWS-CHN type 2, and (C) TAF2N-CHN in the detected fusion transcripts.
logic features of EMCS. Thus, we could not find any relationship between tumor cellularity and clinical outcome in patients with EMCS in the current study. However, the number of cases in our study is too small to evaluate the biologic behavior of the tumor. Biologic significance in heterogeneity of the fusion gene is unknown, although in the current study 3 of the 4 patients who died of the tumor had the EWS-CHN type 1 fusion gene. Further systematic studies of a larger series are needed to elucidate this point. EMCS should be distinguished from a variety of myxoid or cartilage-forming, benign, or malignant, and soft tissue or bone tumors. For example, intramuscular myxoma, nerve sheath myxoma/neurothekeoma, extraskeletal chondroma, myoepithelioma (mixed tumor) of soft parts, parachordoma, metastatic adenocarcioma, myxoid liposarcoma, low-grade myxofibrosarcoma/myxoid malignant fibrous histiocytoma, epithelioid sarcoma, malignant peripheral nerve sheath tumor, and extraskeletal or skeletal chondrosarcoma with prominent myxoid change. Among them, parachordoma and myoepithelioma of soft parts are occasionally difficult to differentiate from EMCS. Recognition of the histologic features of conventional EMCS that are present, at least in part, is most important for a differential diagnosis. Ancillary approaches such as immunohistochemistry and electron microscopy are also valuable. Epithelial markers, such as cytokeratins and EMA, are rarely expressed in EMCS,3-5,41 in contrast to frequent expression of epithelial markers in both parachordoma and myoepithelioma. Myoepithelioma also has positive immunoreactivities to actin in addition to S-100 protein, whereas coexpression of these proteins is usually lacking in EMCS. In summary, the immunohistochemical results of the current study indicate that at least some EMCSs have features of neural/neuroendocrine differentiation. Despite the morphologic and immunohistochemical diversities, molecular analysis targeting the EWS-CHN or TAF2N-CHN fusion gene should provide useful and reliable information for the diagnosis of EMCS. Acknowledgment. The authors thank the following pathologists for providing case material and clinical information to this study: Dr S. Teshima, Douai Memorial Hospital; Dr K. Mizuguchi, Mizonokuchi Hospital, Teikyo University School of Medicine; Dr T. Kasuga, Nakano General Hospital; Drs H.
1123
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
Horiuchi and S. Matsuya, Kanto Medical Center NTT EC; Dr K. Hara, Aichi Medical University; Dr K. Iwasaki, Sasebo Municipal General Hospital; Dr M. Takeshita, National Hospital Kyushu Medical, Center; Dr S. Era, Beppu National Hospital; Dr K. Nakanishi, Iizuka Hospital. They also thank A. Tanaka and M. Katayama for their technical assistance.
REFERENCES 1. Enzinger FM, Shiraki M: Extraskeletal myxoid chondrosarcoma: An analysis of 34 cases. HUM PATHOL 3:421-435, 1972 2. Saleh G, Evans HL, Ro JY, et al: Extraskeletal myxoid chondrosarcoma: A clinicopathologic study of ten patients with long-term follow-up. Cancer 70:2827-2830, 1992 3. Abramovici LC, Steiner GC, Bonar F: Myxoid chondrosarcoma of soft tissue and bone: A retrospective study of 11 cases. HUM PATHOL 26:1215-1220, 1995 4. Meis-Kindblom, Bergh P, Gunterberg B, et al: Extraskeletal myxoid chondrosarcoma: A reappraisal of its morphologic spectrum and prognostic factors based on 117 cases. Am J Surg Pathol 23:636650, 1999 5. Oliveira AM, Sebo TJ, McGrory JE, et al: Extraskeletal myxoid chondrosarcoma: A clinicopathologic, immunohistochemical, and ploidy analysis of 23 cases. Mod Pathol 13:900-908, 2000 6. Antonescu CR, Argani P, Erlandson RA, et al: Skeletal and extraskeletal myxoid chondrosarcoma: A comparative clinicopathologic, ultrastructural, and molecular study. Cancer 83:1504-1521, 1998 7. Lucas DR, Fletcher CDM, Adsay NV, et al: High-grade extraskeletal myxoid chondrosarcoma: A high-grade epithelioid malignancy. Histopathology 35:201-208, 1999 8. Weiss SW: Ultrastructure of the so-called “chordoid sarcoma”: Evidence supporting cartilaginous differentiation. Cancer 37:300-306, 1976 9. Tsuneyoshi M, Enjoji M, Iwasaki H, et al: Extraskeletal myxoid chondrosarcoma: A clinicophathologic and electron microscopic study. Acta Pathol Jpn 31:439-447, 1981 10. Martinez-Tello FJ, Navas-Palacios JJ: Ultrastructural Study of conventional chondosarcomas and myxoid-and mesenchymal-chondrosarcomas. Virchows Archiv [A] 396:197-211, 1982 11. Kindblom LG, Angervall L: Myxoid chondrosarcoma of the synovial tissue: A clinicopathologic, histochemical, and ultrastructural analysis. Cancer 52:1886-1895, 1983 12. Dardic I, Lgace´ R, Carlier MT, et al: Chordoid sarcoma (extraskeletal myxoid chondrosarcoma): A light and electron microscope study. Virchows Arch [A] 399:61-78, 1983 13. Fletcher CDM, Powell G, McKee PH: Extraskeletal myxoid chondrosarcoma: A histochemical and immunohistochemical study. Histopathology 10:489-499, 1986 14. Hinrichs SH, Jaramillo MA, Gumerlock PH, et al: Myxoid chondrosarcoma with a translocation involving chromosomes 9 and 22. Cancer Genet Cytogenet 14:219-226, 1985 15. Deblois G, Wang S, Kay S: Microtubular aggregates within rough endoplasmic reticulum: An unusual ultrastructural feature of extraskeletal myxoid chondrosarcoma. HUM PATHOL 17:469-475, 1986 16. Wolford JF, Bedetti CD: Skeletal myxoid chondrosarcoma with microtubular aggregates within rough endoplasmic reticulum. Arch Pathol Lab Med 112:77-81, 1988 17. Suzuki T, Kaneko H, Kojima K, et al: Extraskeletal myxoid chondrosarcoma characterized by microtubular aggregates in the rough endoplasmic reticulum and tubulin immunoreactivity. J Pathol 156:51-57, 1988 18. Iwata J, Numoto S, Hayashi K, et al: Two cases of extraskeletal myxoid chondrosarcoma with special reference to microtubular aggregates. J Clin Electron Microsc 23:383-392, 1990 19. Chhieng DC, Erlandson RA, Antonescu C, et al: Neuroendocrine differentiation in adult soft tissue sarcoma with features of extraskeletal myxoid chondrosarcoma: Report of seven cases. Mod Pathol 11:8A, 1998 (abstr)
20. Harris M, Coyne J, Tariq M, et al: Extraskeletal myxoid chondrosarcoma with neuroendocrine differentiation: A pathologic, cytogenetic, and molecular study of a case with a novel translocation t(9;17)(q22;q11.2). Am J Surg Pathol 24:1020-1026, 2000 21. Cummings TJ, Shea CR, Reed JA, et al: Expression of the intermediate filament peripherin in extraskeletal myxoid chondrosarcoma. Cutan Pathol 27:141-146, 2000 22. Turc-Carel C, Dal Cin P, Rao U, et al: Recurrent breakpoints at 9q31 and 22q12.2 in extraskeletal myxoid chondrosarcoma. Cancer Genet Cytogenet 30:145-150, 1988 23. Mark-Bendel E, Terrier P, Turc-Carel C: Rearrangement of 9q22: A crucial event in extraskeletal myxoid chondrosarcoma? Cancer Genet Cytogenet 52:267, 1991 24. Stenman G, Andersson H, Mandahl N, et al: Translocation t(9;22)(q22;q12) is a primary cytogenetic abnormality in extraskeletal myxoid chondrosarcoma. Int J Cancer 62:398-402, 1995 25. Hirabayashi Y, Ishida T, Yoshida MA, et al: Translocation (9; 22)(q22;q12): A recurrent chromosome abnormality in extraskeletal myxoid chondrosarcoma. Cancer Genet Cytogenet 81:33-37, 1995 26. Sciot R, Dal Cin P, Fletcher C, et al: t(9;22)(q22-31;q11-12) is a consistent marker of extraskeletal myxoid chondrosarcoma: Evaluation of three cases. Mod Pathol 8:765-768, 1995 27. Labelle Y, Zucman J, Stenman G, et al: Oncogenic conversion of a novel orphan nuclear receptor by chromosome translocation. Hum Mol Genet 4:2219-2226, 1995 28. Clark J, Benjamin H, Gill S, et al: Fusion of the EWS gene to CHN, a member of the steroid/thyroid receptor gene superfamily, in a human myxoid chondrosarcoma. Oncogene 12:229-235, 1996 29. Brody RI, Ueda T, Hamelin A, et al: Molecular analysis of the fusion of EWS to an orphan nuclear receptor gene in extraskeletal myxoid chondrosarcoma. Am J Pathol 150:1049-1058, 1997 30. Labelle Y, Bussie`res J, Courjal F, et al: The EWS/TEC fusion protein encoded by the t(9;22) chromosomal translocation in human chondrosarcomas is a highly potent transcriptional activator. Oncogene 18:3303-3308, 1999 31. Sjo¨gren H, Meis-Kindblom JM, Kindblom LG, et al: Fusion of the EWS-related gene TAF2N to TEC in extraskeletal myxoid chondrosarcoma. Cancer Res 59:5064-5067, 1999 32. Panagopoulos I, Mencinger M, Dietrich CU, et al: Fusion of the RBP56 and CHN genes in extraskeletal myxoid chondrosarcomas with translocation t(9;17)(q22;q11). Oncogene 18:7594-7598, 1999 33. Attwooll C, Tariq M, Harris M, et al: Identification of a novel fusion gene involving hTAFII68 and CHN from a t(9;17)(q22;q11.2) translocation in an extraskeletal myxoid chondrosarcoma. Oncogene 18:7599-7601, 1999 34. Vilanova JR, Simon-Marin R, Burgos-Bretones J, et al: Nonconventional chondrosarcomas and chondrogenesis. Histopathology 9:719-728, 1985 35. Portier MM, De Nechaud B: Peripherin, a new member of the intermediate filament protein family. Dev Neurosci 6:335-344, 1983-1984 36. Baker CVH, Bronner-Fraser M, Le Douarin NM, et al: Early and late-migrating cranial neural crest populations have equivalent developmental potential in vivo. Development 124:3077-3087, 1997 37. Stefansson K, Wollman RL, Moore BW, et al: S-100 protein in human chondrocytes. Nature 295:63-64, 1982 38. Kilpatrick S, Inwards CY, Fletcher CDM, et al: Myxoid chondrosarcoma (chordoid sarcoma) of bone; a report of two cases and review of the literature. Cancer 79:1903-1910, 1997 39. Fechner RE, Nichols GE: Iatrogenic lesions, in Silverberg SG, DeLellis RA, Frable WJ (eds): Principles and Practice of Surgical Pathology and Cytopathology (ed 3). New York, NY, Churchill Livingstone, 1997, pp 269-302 40. Bridge JA, Bhatia PS, Anderson JR, et al: Biologic and clinical significance of cytogenetic and molecular cytogenetic abnormalities in benign and malignant cartilaginous lesions. Cancer Genet Cytogenet 69:79-90, 1993 41. Dei Tos AP, Wadden C, Fletcher CDM: Extraskeletal myxoid chondrosarcoma: An immunohistochemical reappraisal of 39 cases. Appl Immunohistochem 5:73-77, 1997
1124
transcription polymerase chain reaction (RT-PCR) assay using paraffin-embedded specimens, EWS-CHN or TAF2N-CHN fusion gene transcripts characteristic of EMCS could be detected in 15 (83%) of the 18 cases: EWS-CHN type 1 in 11 cases, EWS-CHN type 2 in 1, and TAF2N-CHN in 3. Three fusion-negative cases included 2 conventional EMCSs and 1 considered a “cellular” variant of the tumor. None of 30 other soft tissue and bone tumors with myxoid or chondroid morphology that we examined contained these fusion genes. Thus, RT-PCR detection of EWS-CHN or TAF2N-CHN fusion gene using archival paraffin-embedded tissue is a feasible and useful ancillary technique for the diagnosis of EMCS. HUM PATHOL 32: 1116-1124. Copyright © 2001 by W.B. Saunders Company Key words: extraskeletal myxoid chondrosarcoma, fusion gene, reverse-transcription polymerase chain reaction. Abbreviations: EMCS, extraskeletal myxoid chondrosarcoma; RTPCR, reverse-transcription polymerase chain reaction; EMA, epithelial membrane antigen; PGK, phosphoglycerate kinase; PBGD, porphobilinogen deaminase.
Extraskeletal myxoid chondrosarcoma (EMCS) is a rare malignant soft tissue tumor described as a distinct clinicopathologic entity by Enzinger and Shiraki in 1972.1 The tumor usually develops in deep parts of the proximal extremities and limb girdles in middle-aged adults. EMCS has a prolonged and indolent clinical course, with a high rate of local recurrences and distant metastases; tumor-related death often occurs after a long survival period.2-5 The tumor is histologically characterized by multinodular growth of a lacelike arrangement of primitive chondroid cells in an abundant myxoid matrix. In addition to such a typical feature, cellular areas devoid of a myxoid matrix with or without epithe-
lioid or rhabdoid cells or showing nuclear pleomorphism are occasionally observed.1-7 Meis-Kindblom et al,4 noted that histologic grading was of no prognostic value in their large series of EMCS, whereas other investigators have reported that cases of high-grade EMCS have a truly adverse clinical outcome, as shown by aggressive tumor growth and death after a short period.5-7 Although chondroid differentiation of the constituent cells of EMCS has been supported by ultrastructural and histochemical observations,1,8-13 the histogenesis of this type of tumor is still controversial. In some cases of EMCS, a characteristic ultrastructural finding of long, parallel microtubules within rough endoplasmic reticulum is seen,12,14-18 but its significance is still unproven. More recently, the histochemical, immunohistochemical, and ultrastructural evidence of neural/ neuroendocrine differentiation of the tumor has been documented.5,19-21 Cytogenetic and molecular studies of EMCS have shown that approximately 75% of the tumors have a characteristic t(9;22)(q22;q12) resulting in a fusion of the EWS gene at 22q12 and the CHN gene (also referred to as TEC, NOR-1, or MINOR) at 9q22.6,14,22-29 EWS-CHN fusion proteins may induce tumorigenesis in EMCS by activating the expression of CHN target genes, but no putative target genes have been identified.30 More recently, a novel TAF2N-CHN fusion gene resulting from t(9;17)(q22;q11.2) has been identified. The TAF2N (also known as hTAFII68 or RBP56) gene
From the Department of Pathology and Oncology and the Department of Pathology and Cell Biology, School of Medicine, University of Occupational and Environmental Health, Kitakyushu; the Department of Pathology, Faculty of Medicine, University of Tokyo Hospital, Tokyo; the Department of Surgical Pathology, School of Medicine, Teikyo University, Tokyo; and The Cancer Institute, Japanese Foundation for Cancer Research, Tokyo, Japan. Accepted for publication May 15, 2001. Supported in part by grants-in-aid from the Ministry of Education, Science, Sports and Culture, Japan (12770102 and 12670184), and by a grant from the Fukuoka Cancer Society, Japan, 2000. Address correspondence and reprint requests to Hiroshi Hashimoto, MD, Department of Pathology and Oncology, School of Medicine, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807-8555, Japan. Copyright © 2001 by W.B. Saunders Company 0046-8177/01/3210-0014$35.00/0 doi:10.1053/hupa.2001.28226
1116
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
has sequence homology to the EWS gene,31-33 and the TAF2N-CHN fusion gene’s function is considered identical to that of the EWS-CHN fusion gene. We reviewed immunohistochemical and molecular features of 18 cases of EMCS to verify the expressions of neural/neuroendocrine markers and the diagnostic usefulness of reverse-transcription polymerase chain reaction (RT-PCR) detection of the EWS-CHN or TAF2NCHN fusion gene using archival formalin-fixed, paraffin-embedded tumor specimens. MATERIALS AND METHODS Materials Eighteen cases of EMCS were selected from the files of soft tissue tumors of the authors (H.H., T. Ishida, T. Imamura, and H.K.), and the original hematoxylin and eosin–stained slides of each case were reviewed. In 1 of the 18 cases, the clinicopathologic, ultrastructural, and cytogenetic findings had been reported previously (case 5).25 Initially, 21 cases were originally diagnosed as EMCS, but 3 of them were excluded from the present study because they were reclassified as other tumor entities after the critical review. One of these 3 cases was grade 2 chondrosarcoma of bone with prominent myxoid change, involving the adjacent soft tissue. Another was considered malignant myoepithelioma (mixed tumor) of soft parts with rhabdoid features rather than EMCS on the basis of positive immunoreactivity to cytokeratins, epithelial membrane antigen (EMA), actins, and S-100 protein. In the remaining case, a possible diagnostic consideration of low-grade myxofibrosaroma/myxoid malignant fibrous histiocytoma emerged. The other 30 soft tissue and bone tumors were included in this study as negative controls: 3 myxoid liposarcomas, 3 myxoid malignant fibrous histiocytomas, 3 malignant peripheral nerve sheath tumors, 1 lowgrade fibromyxoid sarcoma, 1 extraskeletal mesenchymal chondrosarcoma, 3 intramuscular myxomas, 2 juxtaarticular myxomas, 3 nerve sheath myxomas, 1 ossifying fibromyxoid tumor of soft parts, 1 extraskeletal chondroma, 3 chondrosarcomas with myxoid change, 3 chordomas, and 3 enchondromas.
Light Microscopy Each tumor tissue fixed in 10% formalin was routinely processed and embedded in paraffin. Three-m-thick sec-
tions were recut and stained with hematoxylin and eosin. Periodic acid–Schiff staining with and without diastase digestion was also performed to verify the presence of intracytoplasmic glycogen granules in tumor cells. Sections were stained immunohistochemically by a streptoavidin-biotin-peroxidase complex technique using the commercially available primary antibodies listed in Table 1.
Molecular Analysis RNA Extraction. Five 4-m-thick sections sliced from paraffin blocks were deparaffinized with xylene and ethanol. After drying, 200 L of lysis buffer (20 mmol/L Tris-HCL pH 8.0, 20 mmol/L EDTA, 2% sodium dodecyl sulfate) was added to each tissue pellet. Each pellet was homogenized with a hand homogenizer, 10 L of proteinase K (100 mg/mL) was added, and then it was incubated overnight at 55°C. One milliliter of Trizol reagent (Gibco BRL, Gaithersburg, MD) was added, and total RNA was extracted according to the manufacturer’s instructions. The RNA pellet was resuspended in 10 L of DNase/RNase-free water and then treated with DNase I (Gibco BRL) for 15 minutes at 37°C. RT-PCR. Ten microliters of dissolved RNA was reverse transcribed into complementary DNA (cDNA) using 1 L of random primer (Gibco BRL) and 200 U reverse transcriptase (SuperScript II; Gibco BRL). The cDNA was then incubated with 60 U RNase H (Takara, Ohtsu, Japan) for 20 minutes at 37°C. The integrity of the RNA for RT-PCR was confirmed by amplification of the phosphoglycerate kinase (PGK) transcript (247 base pairs [bp]) and porphobilinogen deaminase (PBGD) transcript (127 bp) using the following primer sequences: PGK forward, 5⬘-CAGTTTGGAGCTCCTGGAAG-3⬘; PGK reverse, 5⬘-TGCAAATCCAGGGTGCAGTG-3⬘; PBGD forward, 5⬘-TGTCTGGTAACGGCAATGCGGCTGCAAC-3⬘; and PBGD reverse, 5⬘-TCAATGTTGCCACCACACTGTCCGTCT-3⬘. PCR was carried out to amplify EWS-CHN and TAF2N-CHN fusion gene transcripts using specific sets of primers (Table 2). Three variations of the EWS-CHN fusion gene have previously been identified: EWS exon 12–CHN exon 3 (EWS-CHN type 1), EWS exon 7–CHN exon 2 (EWSCHN type 2), and EWS exon 11–CHN exon 2 (EWS-CHN type 3).6 Three sets of the primer for these EWS-CHN fusion gene transcripts were designed by the Oligo Primer Analysis Software (Molecular Biology Insight, Cascade, CO). The previously described primer sets for TAF2N-CHN fusion gene transcripts31 were used. The PCR was performed with 2.5 U Taq DNA polymerase (AmpliTaq Gold; Perkin Elmer, Norwalk, CT) 1.5 mmol/L MgCl2, 10⫻ PCR buffer (Perkin
TABLE 1. Primary Antibodies and Sources Antibody
Source
Dilution
S-100 protein NSE (BBS/NC/VI-H14) Peripherin* Synaptophysin (SY38) PGP9.5 (13C4/31A3) Chromogranin A (CAK-A3) Vimentin (V9) AE1/AE3† CAM5.2† EMA (E29) ␣–Smooth muscle actin (1A4) HHF35 Desmin (D33)†
Dako (Glostrup, Denmark) Dako Chenicon (Temecula, CA) Dako UltraClone (Rossiters Farmhouse, Wellow, Isle of Wight, England) Dako Dako Dako Becton Dickinson (Mountain View, CA) Dako Dako Enzo (Farmingdale, NY) Dako
1:200 1:200 1:500 1:100 1:400 1:100 1:30 1:50 Postdilluted 1:100 1:150 1:50 1:100
*For peripherin immunohistochemistry, microwave pretreatment (3 ⫻ 5 minutes) in 10 mmol/L citrate buffer (pH 6.0) was performed. †Sections were pretreated with protease (0.15%) for 20 minutes at 37°C for these markers.
1117
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
TABLE 2. Primer Sequences and PCR Product Sizes Primer
Nucleotide Sequence
EWS-CHN type 1 EWS exon 12 CHN exon 3 EWS-CHN type 2 EWS exon 7 CHN exon 2 EWS-CHN type 3 EWS exon 11 CHN exon 2 TAF2N-CHN TAF2N exon 6* CHN exon 3*
PCR Product (bp)
5⬘-GCGATGCCACAGTGTCCTATG-3⬘ 5⬘-ATATTGGGCTTGGACGCAGGG-3⬘
89
5⬘-CTCCAAGTCAATATAAGCCAAC-3⬘ 5⬘-GGACGTCCGGCGAGGCGAAGC-3⬘
146
5⬘-TCTGGCAGACTTCTTTAAGCA-3⬘ 5⬘-GGACGTCCGGCGAGGCGAAGC-3⬘
133
5⬘-GAGCAGTCAAATTATGATCAGCAGC-3⬘ 5⬘-CCTGGAGGGGAAGGGCTAT-3⬘
137
* Data from Sjo¨gren et al.31
Elmer), 200 mol/L deoxynucleotide triphosphates, and 0.2 mol/L for each set of primers. The amplification profile of the PCR consisted of 40 cycles of denaturation at 94°C for 1 minute, annealing at 55°C for 45 seconds, and elongation at 72°C for 50 seconds, followed by a final extension at 72°C for 10 minutes. In the PCR procedure, a reaction mixture of the reagents devoid of template cDNA and a no–reverse transcription control were included as negative controls. For visualization, 5 L of PCR product was stained with ethidium bromide and electrophoresed on a 3% agarose gel. To confirm the sequence of the fusion gene transcripts, the PCR products were cloned in a pCR2.1 vector (TA Cloning Kit; Invitrogen, San Diego, CA) and sequenced using an automated sequencer system, ALF express DNA sequencer (Pharmacia Biotech, Uppsala, Sweden).
RESULTS Clinical Findings Clinical details of the 18 patients are summarized in Table 3. There were 10 men and 8 women, whose ages ranged from 26 to 75 years (median, 45 years; mean, 47.4 years). The tumors were located in the thigh (8 cases), upper arm (2 cases), lower leg (2 cases), buttock (1 case), shoulder (1 case), back (1 case), foot (1 case), finger (1 case), and hip joint (1 case). The duration of a symptom before diagnosis ranged from 2 months to 25 years (median, 4 years; mean, 5.1 years). All patients underwent surgical treatments: wide local resection in 13 patients,
TABLE 3. Clinicopathologic Features of EMCSs Age (yr)/ Sex
Specimen
Location
Tumor Size (cm)
Preoperative Duration
1st recurrence [primary] Primary Primary Primary
Upper arm
2 3 4
74/M [59] 46/M 26/M 35/F
12 ⫻ 8 [5] 4.5 ⫻ 3.5 ⫻ 2.5 2 9⫻7⫻4
7 yr
Lower leg Hip joint Thigh
5
58/M
Primary
Thigh
6 7 8 9
75/F 51/F 58/F 34/F
Primary Primary Primary 1st recurrence
Thigh Buttock Thigh Groin
10 11 12 13
54/F 45/M 58/M 41/M
1st recurrence Primary Primary Primary
Back Thigh Finger Upper arm
14 15
37/M 34/F
Primary Primary
16 17 18
42/F 45/F 41/M
Primary Primary Primary
Case 1
Treatment
Follow-up Multiple recurrences; alive at 2 yr
1 yr 1 yr 5 yr
Intralesional resection [marginal resection] Wide resection Total hip arthroplasty Wide resection
8 ⫻ 7.8 ⫻ 5
4 yr
Wide resection
11 ⫻ 8 ⫻ 6.5 11 ⫻ 8 ⫻ 8 7⫻7⫻6 5
25 yr 3 yr 1 yr 4 mo NA
Wide Wide Wide Wide
2.2 ⫻ 1.0 10 ⫻ 7.5 ⫻ 5.5 6⫻4 5
2 mo 5 yr 1 yr 6 mo 2 yr
Wide resection Wide resection Marginal resection Wide resection
Thigh Thigh
24 ⫻ 11 ⫻ 8 8 ⫻ 5.5 ⫻1
5 yr 14 yr
Lower leg Shoulder Foot
3⫻2⫻1 4 8 ⫻ 7.5 ⫻ 7
5 yr 1 yr 6 yr
Abbreviations: NED, no evidence of disease; NA, information not available.
1118
resection resection resection resection
Wide resection Radiation, chemotherapy, wide resection Wide resection Amputation Below-knee amputation
NED; alive at 4 yr NED; alive at 10 yr Multiple recurrences and metastases; died at 15 yr Lung metastasis at 9 yr, resection; alive at 9 yr NED; alive at 7 mon NA Died at 8 yr 6 mo Recurrences, resection; alive at 17 yr NED; alive at 4 yr 6 mo NED; alive at 3 yr NED; alive at 1 yr 8 mo Recurrence at 5 yr, resection, multiple recurrences; died at 18 yr NED; alive at 1 yr 4 mo Lung metastases at presentation; alive at 11 yr NED; alive at 4 yr NED; alive at 33 yr Popliteal and lung metastases at 3 yr; died at 6 yr
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
amputation in 2, marginal/intralesional resection in 2, and total arthroplasty in 1. One patient received preoperative radiotherapy and chemotherapy (case 15). Follow-up information was available for 17 patients, and the follow-up period ranged from 7 months to 33 years. Nine patients were alive without any evidence of disease 7 months to 33 years after diagnosis, and 2 were alive and disease-free after surgery for lung metastasis (case 5) and local recurrences (case 9), respectively. Another 2 patients had multiple recurrences or metastases and were alive with disease (case 1 and 15). The remaining 4 patients died from the tumor at 6, 8.5, 15, and 18 years after surgery. Pathologic Findings Tumors ranged from 2 to 24 cm in size (median, 6 cm; mean, 7.8 cm). Most tumors were deep-seated and arose within the skeletal muscle (cases 1, 2, 4 through 11, 13 through 15, 17, and 18), tendinous structures of the phalanx (case 12), synovial tissue involving the adjacent femoral neck (case 3), and subcutaneous adipose tissue (case 16). In 1 tumor arising in the skeletal muscle, the adjacent humeral cortex was involved (case 13). All 18 tumors had a distinct multinodular configuration delineated by fibrous connective tissue. Most of the tumors had expansive growth, but 4 had infiltrative growth to the surrounding soft tissues (cases 1, 6, 12, and 16). Most of the tumor cells had abundant intracytoplasmic diastase-digested periodic acid–Schiff–positive granules, probably of glycogen. None of the tumors formed discernible cartilaginous structures. Fifteen of the 18 tumors had typical features of EMCS: a lobular proliferation of spindle and oval cells arranged in cords or strands or in an lacelike pattern and embedded in an abundant myxoid matrix, which was often basophilic (cases 1 through 15; Fig 1A). Some also had areas of small nests or whorls of the tumor cell (cases 5, 6, and 13; Fig 1B). The individual tumor cells had spindle-shaped or oval hyperchromatic nuclei that occasionally contained small nucleoli and showed mild to moderate nuclear atypia, and they also had a variable amount of eosinophilic cytoplasm (Fig 1C). Monovacuolated or multivacuolated tumor cells, resembling lipoblasts, were present in 3 cases (cases 2, 3, and 14; Fig 1D). Small clusters of rhabdoid cells were seen in 1 case (case 11). In 1 patient who had received preoperative radiotherapy and chemotherapy, the tumor showed areas with tumor giant cells that had pronouncedly pleomorphic or bizarre nuclei (case 15; Fig 1E). Mitotic figures were rarely encountered (mean 0 to 1 per 10 high-power fields [hpf]; magnification ⫻400). Hemorrhagic areas or aggregates of siderophages (9 of 15 cases) and tumor necroses (7 of 15 cases) were often seen. The other 3 tumors had multiple lobules consisting of a cellular proliferation of tumor cells embedded in a fibrocollagenous or fibromyxoid stroma, together with areas of conventional EMCS (cases 16 through 18; Fig 2A). In 1 of the 3 tumors, constituent tumor cells in cellular areas were round to oval and vacuolated and frequently had indented nuclei and perinuclear empty
spaces resembling lacunae (case 16; Fig 2B). In the cellular portions of another tumor (case 17), uniform and rather small tumor cells with indistinct cell borders were diffusely distributed (Fig 2C). Areas of a fascicular proliferation of atypical spindle cells were observed in cellular areas of case 18, which also showed small foci with atypical epithelioid or rhabdoid cells (Fig 2D). Tumor necrosis and hemorrhage were present in 2 (cases 16 and 18). Mitotic figures were few (0 to 1 per 10 hpf) in 2 tumors (cases 16 and 17), and more frequent (4 per 10 hpf) in 1 (case 18). Immunohistochemical Findings Immunohistochemical staining results of each case are listed in Table 4. S-100 protein was expressed in half of the tumors: strong and diffuse or focal immunoreactivity in 6 cases, and weak and focal in 3. NSE was expressed in most of the tumors (89%) in 10% to 80% of tumor cells. Most of the tumor cells in approximately half of the tumors (53%) were positively immunoreactive to peripherin (Fig 3A). Four (22%) of the 18 tumors had diffuse positive immunoreactivity for synaptophysin (Fig 3B). Diffuse expression of PGP9.5 was observed in 1 tumor (case 18). None of the tumors expressed chromogranin A. Diffuse and strong vimentin immunoreactivity was seen in most of the tumors examined (94%). Some tumors were focally immunostained for ␣–smooth muscle actin (cases 5,16, and 17; 17%) and HHF35 (cases 5 and 16; 11%), but all tumors were negative for desmin. No immunoreactivity for AE1/AE3, CAM5.2, and EMA was detected in any of the tumors. Molecular Findings Results of RT-PCR assay are summarized in Table 4. In the RT-PCR assay, the fusion gene transcripts could be amplified in 15 (83%) of the 18 paraffinembedded tumors (Fig 4). EWS-CHN type 1 fusion gene transcripts were detected in 11 (60%) cases. EWSCHN type 2 and TAF2N-CHN fusion gene transcripts were seen in 1 (6%) and 3 (17%) cases, respectively. Each type of the fusion genes was further confirmed by a subsequent sequence analysis (Fig 5). EWS-CHN type 3 fusion was not detected in any of the tumors. Three EMCSs were negative for the fusion gene, although the presence of adequate RNA was confirmed by the PCR amplification of the PBGD and/or PGK gene transcripts (cases 10, 11, and 18). The EWS-CHN or TAF2NCHN fusion gene transcripts were not amplified in any of the 30 other soft tissue and bone tumors used as negative controls or the 3 tumors reclassified into other types of tumor than EMCS. DISCUSSION EMCSs account for approximately 2.5% of soft tissue sarcoma,9 commonly occur in middle-aged adults, and arise mainly in the proximal extremities and limb girdles. Generally, EMCS is considered a tumor showing chondroid differentiation. Chondroid differ-
1119
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
FIGURE 1. (A) Low-power view of EMCS showing multilobular growth of tumor cells embedded in an abundant myxoid matrix. (Case 12; hematoxylin and eosin, original magnification ⫻10.) (B) Whorl arrangements of tumor cells. (Case 6; original magnification ⫻33.) (C) Individual tumor cells showing mild to moderate nuclear atypia. (Case 2; original magnification ⫻100.) (D) Monovacuolated or multivacuolated tumor cells, reminiscent of lipoblasts. (Case 14; original magnification ⫻100.) (E) Tumor cells showing pronounced nuclear pleomorphism in a postradiotherapeutic and chemotherapeutic state. (Case 15; original magnification ⫻25.)
entiation has been suggested by the results of histochemical or immunohistochemical studies showing the presence of chondroitin-4 and 6-sulphate in an intercellular matrix in EMCS9,11,13 and S-100 protein positivity,3,13 by the ultrastructural features of constituent cells of EMCS bearing some resemblance to chondroblasts,1,8 –12 and by morphologic similarities between EMCS and embryonic chondrogenesis.34 Furthermore, a recent report showing the capability of the tumor cells to synthesize type II collagen has provided additional evidence for chondroid differentiation of this type of tumor.20 On the other hand, recently, several investigators have described neural/neuroendocrine
differentiation of EMCS by immunohistochemical and ultrastructural observations.5,19 –21 Oliveira et al5 reported that synaptophysin was expressed in 13 (72%) of 18 cases of EMCS, and chromogranin A and Leu-7 were expressed in 1 case each. Harris et al20 showed a unique example of EMCS presenting unequivocal features of both chondroid and neuroendocrine differentiation. Ultrastructurally, this case showed tumor cells containing neuroendocrine granules in addition to positive immunoreactivities to synaptophysin and chromogranin A. Cummings et al21 documented that 3 of 5 cases of EMCS were positively immunoreactive to peripherin, a type III intermediate filament of 56 kd, was
1120
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
FIGURE 2. (A) Low-power view of a cellular variant EMCS consisting of lobules of both myxoid and cellular components. (Case 16; original magnification ⫻13.) (B) Cellular area of round tumor cells with indented nuclei and lacuna-like spaces in the cytoplasm embedded in a fibrocollagenous matrix. (Case 16; original magnification ⫻100.) (C) Transition between myxoid and cellular areas. The latter consists of a haphazard proliferation of rather small tumor cells showing indistinct cell borders set in a fibromyxoid matrix. (Case 17; original magnification ⫻50.) (D) Area of a fascicular proliferation of atypical spindle cells. (Case 18; original magnification ⫻50.)
involved in the growth and development of the peripheral nervous system and was expressed in the neural crest and neural tube– derived cells.35 Our immunohistochemical results support these previous reports and suggest neural/neuroendocrine differentiation in EMCSs. Pluripotential precursor cells of the neural crest can give rise to many derivatives, including neurons, and cranial skeletal components such as cartilage, bone, and dermis.36 The divergent differentiation in EMCS may imply a neuroectodermal derivation of the tumor, although chondrocytes at other sites, such as the extremities, vertebrae, and ribs, are thought to be derived from the somatic mesoderm.37 In the 18 cases of EMCS in the current study, 2 cases occurred in the finger and the hip joint, respectively, where EMCS rarely develops.3,11,16,38 In addition to conventional morphologic appearances of EMCS, the detection of the EWS-CHN fusion gene transcripts could confirm the diagnosis in these cases. Three of our 18 tumors had cellular areas intermingled with conventional myxoid areas of EMCS and were considered a cellular variant of EMCS.4 The histologic variation of this type of tumor, including the cellular variant,
raises a potential diagnostic problem. We could detect the EWS-CHN fusion gene transcripts in 2 of the cellular variants. Even in 1 case in our series showing nuclear pleomorphism, probably caused by preoperative radiotherapeutic effects,39 the TAF2N-CHN fusion gene transcripts were identified. These results indicate an important role of molecular detection of the specific fusion genes in the diagnosis of EMCS, particularly in clinicopathologically unusual cases. Previous molecular studies have shown that the EWSCHN fusion gene was associated with approximately 75% of EMCSs.6,27 Recently, a novel TAF2N-CHN fusion gene was found in EMCS.31–33 In 3 of the current cases, including 1 cellular form, the EWS-CHN and TAF2N-CHN fusion gene transcripts were not detected. A hitherto unknown variation of the gene might be involved in cases negative for both kinds of fusion genes, although negative results do not always indicate absence of the fusion gene in RT-PCR using formalin-fixed, paraffin-embedded tumor tissues. This speculation could be partially supported by the presence of some karyotypic profiles of EMCS other than t(9;22)(q22;q12) and t(9;17)(q22;q11.2), such as t(9;17;15)(q22;q11;q22), t(2;13)(q32;p12), and t(11;
1121
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
TABLE 4. Immunohistochemical and Molecular Findings of EMCSs RT-PCR Case S-100 NSE Peripherin Synaptophysin PGP9.5 Chromogranin A Vimentin AE1/AE3 CAM5.2 EMA ␣-SMA HHF35 Desmin PBGD/PGK Fusion Gene ⫺
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
2
⫹* ⫹
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
3
⫹ ⫹
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
4
⫺ ⫹*
⫺
⫺
⫺
⫺
⫹*
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
5
⫹ ⫹
⫺
⫹
⫺
⫺
ND
⫺
⫺
⫺
⫹
⫹
⫺
⫹/⫹
6
⫹ ⫹
⫺
⫺
ND
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
7
⫹ ⫹
⫹
⫹
ND
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
8
⫹ ⫹
⫹
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫺
9 10 11
⫺ ⫹ ⫺ ⫺ ⫹* ⫹
⫹ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫹ ⫺ ⫹
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫹/⫺ ⫹/⫹ ⫹/⫺
1
⫺
12
⫹ ⫹
⫺
⫺
ND
⫺
⫹*
⫺
⫺
⫺
⫺
⫺
⫺
⫹/⫹
13 14 15
⫹* ⫹ ⫺ ⫹ ⫺ ⫹
⫹ ⫹ ND
⫺ ⫺ ⫹
⫺ ⫺ ND
⫺ ⫺ ⫺
⫹ ⫹* ND
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ⫺
⫺ ⫺ ND
⫺ ⫺ ⫺
⫹/⫺ ⫹/⫺ ⫹/⫺
16
⫺ ⫹
⫺
⫺
⫺
⫺
⫹
⫺
⫺
⫺
⫹
⫹
⫺
⫹/⫺
17 18
⫺ ⫹ ⫺ ⫹
⫺ ⫹
⫺ ⫹
⫺ ⫹
⫺ ⫺
⫹ ⫹
⫺ ⫺
⫺ ⫺
⫺ ⫺
⫹ ⫺
⫺ ⫺
⫺ ⫺
⫹/⫹ ⫹/⫹
TAF2N-CHN EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 EWS-CHN type 1 not detected not detected EWS-CHN type 2 EWS-CHN type 1 TAF2N-CHN TAF2N-CHN EWS-CHN type 1 EWS-CHN type 1 Not detected
Abbreviations: ␣-SMA, ␣–smooth muscle actin; ND, not done. *Focal immunoreactivity.
22)(q11;p11).23,24,40 Regarding the frequency of the fusion types, Antonescu et al6 found the EWS-CHN type 1 in 3 (33%) of 9 EMCSs examined, the EWS-CHN type 2 in 2 (22%), and the EWS-CHN type 3 in 1 (11%), and Labelle et al27 described that type 1 was detected in 3 (38%) of 8 cases of EMCS investigated and type 2 in 2 (25%). Although the results of our molecular assay also showed that most of the tumors had the EWS-CHN type 1 fusion gene, the number of cases examined is too small to draw a conclusion. According to the study by Meis-Kindblom et al,4 adverse prognostic factors of a patient with EMCS were larger tumor size, proximal tumor location, older pa-
tient age, and metastasis. They mentioned that highgrade histologic features, including tumor necrosis, increased cellularity, high mitotic rate, nuclear pleomorphism, epithelioid cells, rhabdoid cells, and spindle cell foci, did not affect prognosis. In contrast, some other reports have shown that tumor cellularity predicted the prognosis of a patient with an EMCS,1,12,13 and Oliveira et al5 documented that large tumor size, high cellularity, presence of anaplasia or rhabdoid features, high mitotic activity (⬎2 per 10 hpf), and high Ki-67 labeling index were statistically associated with a poor prognosis. In our 17 cases of EMCS available for follow-up information, 2 of the 3 patients with the
FIGURE 3. (A) Most of the tumor cells were positively immunostained for peripherin. (Case 1; original magnification ⫻100.) (B) Diffuse synaptophysin immunoreactivity. (Case 5, original magnification ⫻100.)
1122
EXTRASKELETAL MYXOID CHONDROSARCOMA (Okamoto et al)
FIGURE 4. RT-PCR detection of EWS-CHN or TAF2N-CHN fusion gene and 2 ubiquitous genes (PGK and PBGD) in formalin-fixed, paraffin-embedded EMCS tumor tissues. Numbers of the lanes are identical to the case numbers in Table 3. M, size marker (100-bp DNA ladder); N, negative control (distilled water).
cellular variant EMCS were alive without any recurrence or metastasis, and 1 died of the tumor. On the other hand, tumors in 3 of the 4 patients whose cause of death was the tumor showed conventional morpho-
FIGURE 5. Partual nucleotide sequences of (A) EWS-CHN type 1, (B) EWS-CHN type 2, and (C) TAF2N-CHN in the detected fusion transcripts.
logic features of EMCS. Thus, we could not find any relationship between tumor cellularity and clinical outcome in patients with EMCS in the current study. However, the number of cases in our study is too small to evaluate the biologic behavior of the tumor. Biologic significance in heterogeneity of the fusion gene is unknown, although in the current study 3 of the 4 patients who died of the tumor had the EWS-CHN type 1 fusion gene. Further systematic studies of a larger series are needed to elucidate this point. EMCS should be distinguished from a variety of myxoid or cartilage-forming, benign, or malignant, and soft tissue or bone tumors. For example, intramuscular myxoma, nerve sheath myxoma/neurothekeoma, extraskeletal chondroma, myoepithelioma (mixed tumor) of soft parts, parachordoma, metastatic adenocarcioma, myxoid liposarcoma, low-grade myxofibrosarcoma/myxoid malignant fibrous histiocytoma, epithelioid sarcoma, malignant peripheral nerve sheath tumor, and extraskeletal or skeletal chondrosarcoma with prominent myxoid change. Among them, parachordoma and myoepithelioma of soft parts are occasionally difficult to differentiate from EMCS. Recognition of the histologic features of conventional EMCS that are present, at least in part, is most important for a differential diagnosis. Ancillary approaches such as immunohistochemistry and electron microscopy are also valuable. Epithelial markers, such as cytokeratins and EMA, are rarely expressed in EMCS,3-5,41 in contrast to frequent expression of epithelial markers in both parachordoma and myoepithelioma. Myoepithelioma also has positive immunoreactivities to actin in addition to S-100 protein, whereas coexpression of these proteins is usually lacking in EMCS. In summary, the immunohistochemical results of the current study indicate that at least some EMCSs have features of neural/neuroendocrine differentiation. Despite the morphologic and immunohistochemical diversities, molecular analysis targeting the EWS-CHN or TAF2N-CHN fusion gene should provide useful and reliable information for the diagnosis of EMCS. Acknowledgment. The authors thank the following pathologists for providing case material and clinical information to this study: Dr S. Teshima, Douai Memorial Hospital; Dr K. Mizuguchi, Mizonokuchi Hospital, Teikyo University School of Medicine; Dr T. Kasuga, Nakano General Hospital; Drs H.
1123
HUMAN PATHOLOGY
Volume 32, No. 10 (October 2001)
Horiuchi and S. Matsuya, Kanto Medical Center NTT EC; Dr K. Hara, Aichi Medical University; Dr K. Iwasaki, Sasebo Municipal General Hospital; Dr M. Takeshita, National Hospital Kyushu Medical, Center; Dr S. Era, Beppu National Hospital; Dr K. Nakanishi, Iizuka Hospital. They also thank A. Tanaka and M. Katayama for their technical assistance.
REFERENCES 1. Enzinger FM, Shiraki M: Extraskeletal myxoid chondrosarcoma: An analysis of 34 cases. HUM PATHOL 3:421-435, 1972 2. Saleh G, Evans HL, Ro JY, et al: Extraskeletal myxoid chondrosarcoma: A clinicopathologic study of ten patients with long-term follow-up. Cancer 70:2827-2830, 1992 3. Abramovici LC, Steiner GC, Bonar F: Myxoid chondrosarcoma of soft tissue and bone: A retrospective study of 11 cases. HUM PATHOL 26:1215-1220, 1995 4. Meis-Kindblom, Bergh P, Gunterberg B, et al: Extraskeletal myxoid chondrosarcoma: A reappraisal of its morphologic spectrum and prognostic factors based on 117 cases. Am J Surg Pathol 23:636650, 1999 5. Oliveira AM, Sebo TJ, McGrory JE, et al: Extraskeletal myxoid chondrosarcoma: A clinicopathologic, immunohistochemical, and ploidy analysis of 23 cases. Mod Pathol 13:900-908, 2000 6. Antonescu CR, Argani P, Erlandson RA, et al: Skeletal and extraskeletal myxoid chondrosarcoma: A comparative clinicopathologic, ultrastructural, and molecular study. Cancer 83:1504-1521, 1998 7. Lucas DR, Fletcher CDM, Adsay NV, et al: High-grade extraskeletal myxoid chondrosarcoma: A high-grade epithelioid malignancy. Histopathology 35:201-208, 1999 8. Weiss SW: Ultrastructure of the so-called “chordoid sarcoma”: Evidence supporting cartilaginous differentiation. Cancer 37:300-306, 1976 9. Tsuneyoshi M, Enjoji M, Iwasaki H, et al: Extraskeletal myxoid chondrosarcoma: A clinicophathologic and electron microscopic study. Acta Pathol Jpn 31:439-447, 1981 10. Martinez-Tello FJ, Navas-Palacios JJ: Ultrastructural Study of conventional chondosarcomas and myxoid-and mesenchymal-chondrosarcomas. Virchows Archiv [A] 396:197-211, 1982 11. Kindblom LG, Angervall L: Myxoid chondrosarcoma of the synovial tissue: A clinicopathologic, histochemical, and ultrastructural analysis. Cancer 52:1886-1895, 1983 12. Dardic I, Lgace´ R, Carlier MT, et al: Chordoid sarcoma (extraskeletal myxoid chondrosarcoma): A light and electron microscope study. Virchows Arch [A] 399:61-78, 1983 13. Fletcher CDM, Powell G, McKee PH: Extraskeletal myxoid chondrosarcoma: A histochemical and immunohistochemical study. Histopathology 10:489-499, 1986 14. Hinrichs SH, Jaramillo MA, Gumerlock PH, et al: Myxoid chondrosarcoma with a translocation involving chromosomes 9 and 22. Cancer Genet Cytogenet 14:219-226, 1985 15. Deblois G, Wang S, Kay S: Microtubular aggregates within rough endoplasmic reticulum: An unusual ultrastructural feature of extraskeletal myxoid chondrosarcoma. HUM PATHOL 17:469-475, 1986 16. Wolford JF, Bedetti CD: Skeletal myxoid chondrosarcoma with microtubular aggregates within rough endoplasmic reticulum. Arch Pathol Lab Med 112:77-81, 1988 17. Suzuki T, Kaneko H, Kojima K, et al: Extraskeletal myxoid chondrosarcoma characterized by microtubular aggregates in the rough endoplasmic reticulum and tubulin immunoreactivity. J Pathol 156:51-57, 1988 18. Iwata J, Numoto S, Hayashi K, et al: Two cases of extraskeletal myxoid chondrosarcoma with special reference to microtubular aggregates. J Clin Electron Microsc 23:383-392, 1990 19. Chhieng DC, Erlandson RA, Antonescu C, et al: Neuroendocrine differentiation in adult soft tissue sarcoma with features of extraskeletal myxoid chondrosarcoma: Report of seven cases. Mod Pathol 11:8A, 1998 (abstr)
20. Harris M, Coyne J, Tariq M, et al: Extraskeletal myxoid chondrosarcoma with neuroendocrine differentiation: A pathologic, cytogenetic, and molecular study of a case with a novel translocation t(9;17)(q22;q11.2). Am J Surg Pathol 24:1020-1026, 2000 21. Cummings TJ, Shea CR, Reed JA, et al: Expression of the intermediate filament peripherin in extraskeletal myxoid chondrosarcoma. Cutan Pathol 27:141-146, 2000 22. Turc-Carel C, Dal Cin P, Rao U, et al: Recurrent breakpoints at 9q31 and 22q12.2 in extraskeletal myxoid chondrosarcoma. Cancer Genet Cytogenet 30:145-150, 1988 23. Mark-Bendel E, Terrier P, Turc-Carel C: Rearrangement of 9q22: A crucial event in extraskeletal myxoid chondrosarcoma? Cancer Genet Cytogenet 52:267, 1991 24. Stenman G, Andersson H, Mandahl N, et al: Translocation t(9;22)(q22;q12) is a primary cytogenetic abnormality in extraskeletal myxoid chondrosarcoma. Int J Cancer 62:398-402, 1995 25. Hirabayashi Y, Ishida T, Yoshida MA, et al: Translocation (9; 22)(q22;q12): A recurrent chromosome abnormality in extraskeletal myxoid chondrosarcoma. Cancer Genet Cytogenet 81:33-37, 1995 26. Sciot R, Dal Cin P, Fletcher C, et al: t(9;22)(q22-31;q11-12) is a consistent marker of extraskeletal myxoid chondrosarcoma: Evaluation of three cases. Mod Pathol 8:765-768, 1995 27. Labelle Y, Zucman J, Stenman G, et al: Oncogenic conversion of a novel orphan nuclear receptor by chromosome translocation. Hum Mol Genet 4:2219-2226, 1995 28. Clark J, Benjamin H, Gill S, et al: Fusion of the EWS gene to CHN, a member of the steroid/thyroid receptor gene superfamily, in a human myxoid chondrosarcoma. Oncogene 12:229-235, 1996 29. Brody RI, Ueda T, Hamelin A, et al: Molecular analysis of the fusion of EWS to an orphan nuclear receptor gene in extraskeletal myxoid chondrosarcoma. Am J Pathol 150:1049-1058, 1997 30. Labelle Y, Bussie`res J, Courjal F, et al: The EWS/TEC fusion protein encoded by the t(9;22) chromosomal translocation in human chondrosarcomas is a highly potent transcriptional activator. Oncogene 18:3303-3308, 1999 31. Sjo¨gren H, Meis-Kindblom JM, Kindblom LG, et al: Fusion of the EWS-related gene TAF2N to TEC in extraskeletal myxoid chondrosarcoma. Cancer Res 59:5064-5067, 1999 32. Panagopoulos I, Mencinger M, Dietrich CU, et al: Fusion of the RBP56 and CHN genes in extraskeletal myxoid chondrosarcomas with translocation t(9;17)(q22;q11). Oncogene 18:7594-7598, 1999 33. Attwooll C, Tariq M, Harris M, et al: Identification of a novel fusion gene involving hTAFII68 and CHN from a t(9;17)(q22;q11.2) translocation in an extraskeletal myxoid chondrosarcoma. Oncogene 18:7599-7601, 1999 34. Vilanova JR, Simon-Marin R, Burgos-Bretones J, et al: Nonconventional chondrosarcomas and chondrogenesis. Histopathology 9:719-728, 1985 35. Portier MM, De Nechaud B: Peripherin, a new member of the intermediate filament protein family. Dev Neurosci 6:335-344, 1983-1984 36. Baker CVH, Bronner-Fraser M, Le Douarin NM, et al: Early and late-migrating cranial neural crest populations have equivalent developmental potential in vivo. Development 124:3077-3087, 1997 37. Stefansson K, Wollman RL, Moore BW, et al: S-100 protein in human chondrocytes. Nature 295:63-64, 1982 38. Kilpatrick S, Inwards CY, Fletcher CDM, et al: Myxoid chondrosarcoma (chordoid sarcoma) of bone; a report of two cases and review of the literature. Cancer 79:1903-1910, 1997 39. Fechner RE, Nichols GE: Iatrogenic lesions, in Silverberg SG, DeLellis RA, Frable WJ (eds): Principles and Practice of Surgical Pathology and Cytopathology (ed 3). New York, NY, Churchill Livingstone, 1997, pp 269-302 40. Bridge JA, Bhatia PS, Anderson JR, et al: Biologic and clinical significance of cytogenetic and molecular cytogenetic abnormalities in benign and malignant cartilaginous lesions. Cancer Genet Cytogenet 69:79-90, 1993 41. Dei Tos AP, Wadden C, Fletcher CDM: Extraskeletal myxoid chondrosarcoma: An immunohistochemical reappraisal of 39 cases. Appl Immunohistochem 5:73-77, 1997
1124