The chromosomal translocation t(X;18)(p11.2;q11.2) is tightly linked to the tumorigenesis of synovial sarcoma. Through this translation the SYT gene on chromosome 18 is fused with a testis/cancer antigen gene on the X chromosome, generating either a SYT-SSX1, SYT-SSX2, or less often a SYT-SSX4 fusion gene. It has been anticipated that the individual synovial sarcoma carries only one of these variants, however, in this study we demonstrated that SYT-SSX1 and SYT-SSX2 co-exist in a significant proportion of the cases. From 121 SYT-SSX positive primary tumors, co-expression of SYT-SSX1 and SYT-SSX2 was seen in 12 cases (10%), which were characterized in further detail both at the RNA, DNA and chromosomal level. In all 12 cases the SYT-SSX1 and SYT-SSX2 fusions resulted in identical SYT-SSX fusion transcripts. However, at the genomic level the translocations were different, and most likely occurred between variable intronic sites in the target genes. By interphase FISH analyses of 10 cases SYT-SSX2 translocations were found to be the most abundant in all but one of the cases, in which SYT-SSX1 was predominating. The findings reveal a new heterogenous feature of synovial sarcoma, accounting for approximately 10% of all cases, which may shed light on the molecular genetic mechanisms behind translocations in general, and on the etiology of synovial sarcoma in particular.
The constant inheritance of the malignant phenotype from the mother cell to the daughter cells is one of the hallmarks of cancer, implying that cancer is a genetic disease which could be explained and tackled on the genetic level. This simple truth is applicable to all types of malignancies, and has been the driving force for the extensive search of genetic alterations associated with tumor development. At present, the role of cytogenetic and molecular analyses have been established in basic and clinical investigation of hematological malignancies, and specific genetic alterations are used in clinical practice for diagnostic, prognostic and therapeutic purposes (O'Dwyer and Druker, 2000). In solid tumors, a similar situation is presently evolving for sarcomas, where several specific chromosomal alterations, mostly reciprocal translocations, have been associated with distinct histopathological entities (Sreekantaiah, 1998). In some situations, different oncogenic fusion genes are associated with a single type of cancer, while in other situations one gene can fuse to different partner genes, resulting in distinct neoplastic phenotypes (Åman, 1999; Ladanyi and Bridge, 2000).
The fusion genes identified in synovial sarcoma represent a third category in which several members of a gene family are fused to a single partner gene and associated with the same malignant disease. The t(X;18)(p11.2;q11.2) has been tightly linked to the development of synovial sarcoma (Turc-Carel et al., 1987). Through translocations, the SYT gene on chromosome 18 fuses with a member of the SSX gene family, all located on chromosome X, resulting in the generation of either a SYT-SSX1, a SYT-SSX2, or rarely a SYT-SSX4 fusion gene (Clark et al., 1994; Crew et al., 1995; de Leeuw et al., 1995; Skytting et al., 1999). The alteration is detected in more than 90% of the cases, it frequently occurs as a single abnormality, and the resulting SYT-SSX fusion is always retained during tumor progression while the reciprocal SSX-SYT fusion is frequently eliminated (Panagopoulos et al., 2001). Consequently, the SYT-SSX fusion products are expected to play important roles in the etiology of this tumor, possibly by interfering with the regulation of gene expression, similar to what happens in many other sarcomas. Recently, a formal proof of the oncogenic activity of SYT-SSX1 was provided, by the demonstration of induction of tumor formation in nude mice (Nagai et al., 2001). The SYT gene is widely expressed in human tissues and its product is a nuclear transcriptional activating protein (Brett et al., 1997), which has also been shown to be involved in the active control of cell adhesion (Eid et al., 2000). The SSX gene family contains at least six highly homologous members, which normally show a tissue specific expression restricted to testis and thyroid but are expressed in a variety of human cancers (Gure et al., 1997; Fligman et al., 1995). The SSX proteins have two transcriptional repression domains, the Krüppel associated box (KRAB) repression domain and a novel repression domain (SSXRD). Through the fusions to SYT, the KRAB domain is removed while the SSXRD domain is retained (Crew et al., 1995; Lim et al., 1998).
Synovial sarcomas are sporadically occurring, highly malignant tumors, which preferentially affect young adults and children. Based on the histopathological characteristics, two main subtypes are recognized, including the biphasic type composed of epithelial and spindle cells, and the monophasic type composed of spindle cells only. Several studies have demonstrated a relationship between the type of SYT-SSX fusion and the tumor morphology. Biphasic synovial sarcomas typically express the SYT-SSX1 fusion transcript, whereas in monophasic tumors SYT-SSX1 and SYT-SSX2 fusion are equally frequent (Kawai et al., 1998; Nilsson et al., 1999; dos Santos et al., 2001). It has also been shown that SYT-SSX1 is significantly associated with poor metastasis-free survival, and a higher tumor cell proliferation rate have been demonstrated in some studies but not in others (Kawai et al., 1998; Nilsson et al., 1999; dos Santos et al., 2001). The individual synovial sarcoma is generally believed to carry only a single type of SYT-SSX fusion. However, in our RT–PCR screening for SYT-SSX variants we have recurrently detected positive signals for both SYT-SSX1 and SYT-SSX2 transcripts in the same tumors. In the present study a series of such cases were investigated in detail with the aim to characterize the nature, relative abundance, and distribution of the two translocations in the individual tumors.
Detection of tumors with both SYT-SSX1 and SYT-SSX2 fusions
From a collection of 121 SYT-SSX positive synovial sarcomas (B Skytting et al., in preparation), 12 tumors were identified which express both SYT-SSX1 and SYT-SSX2, and which were therefore characterized in further detail on both the RNA, DNA, and chromosomal level (Table 1). In the initial screening for SYT-SSX fusion transcripts seminested RT–PCR analyses were applied. In all positive tumors a single product of the expected size was obtained, and in the subsequent sequencing the 151 bp product was confirmed to correspond to the SYT-SSX1 fusion and the 190 bp product to SYT-SSX2. Repetition of the RT–PCR analyses on new RNA preparations confirmed the results of the initial analysis. In these cases, the sequencing was performed both on products generated directly with the specific primers (Figure 1), and on products generated with the consensus primers by which cDNA of both fusion transcripts were amplified without distinction. Using the latter primers, sequences of both fusions were evident in case 1 (Figure 1), while the SYT-SSX2 sequence was obtained in cases 2–8 and 10–12, and SYT-SSX1 was detected in case 9.
SYT-SSX1 fusion products were amplified from genomic DNA in eight of the 11 double fusion tumors, and in seven of these cases the translocations were detected by nested PCR, while in case 9 the gene fusion was demonstrated in the primary PCR. Similarly, the SYT-SSX2 translocation was detected at the genomic level in all of the 11 tumors analysed (Table 1). The genomic PCR products were obtained in the first PCR for 10 of the cases, and after nested PCR in case 9. A variety of combinations of primers, which were derived from the breaking introns and the closest flanking exons of the involved genes, were used. As expected, the resulting products were of varying sizes in the different tumors (Table 1). Furthermore, even when the same primer combination was used PCR products of varying sizes were seen (Figure 2). Taken together, these findings suggest that, although the respective SYT-SSX1 and SYT-SSX2 fusions have resulted in identical SYT-SSX fusion transcripts, the translocation breakpoints are different at the genomic level.
The genomic PCR products obtained by primary or nested PCR were partially sequenced, which revealed the presence of SYT sequences at the 5′-ends and SSX1 and SSX2 sequences at the 3′-ends. Furthermore, for the SYT-SSX1 fusion, insertions of unknown sequences were identified in three of the tumors. In cases 4 and 5 the SSX1 part of the fusion introns were replaced by an apparently identical sequence of unknown origin at nucleotide −491, and in case 6 by another unknown sequence at position −617 (data not shown).
Preferential abundance of SYT-SSX2 in synovial sarcomas with double fusions
The occurrence of two different SYT-SSX fusions in synovial sarcomas could theoretically be explained by the existence of two translocations being present in the same cells, or alternatively the tumors could be genetically heterogenous with the different types of translocations in different cells. In order to characterize the tissue distribution and the relative abundance of SYT-SSX1 and SYT-SSX2 carrying cells in these tumors, interphase FISH analyses were performed. The chromosomal remodelings resulting from the SYT-SSX translocations are illustrated in Figure 3, together with the locations of the PAC probes used for their detection. The results from the interphase FISH analyses are detailed in Table 1 and illustrated in Figure 4. When the SYTprobeB, which covers the SYT breakpoint, was used in combination with the SSX1 or SSX2 probes, the SYTprobeB signal was split into two smaller signals in cells where a SYT-SSX translocation was present (Figure 4c). However, hybridization with the SYTprobeA located distal to the SYTprobeB always resulted in two signals only (Figure 4d,g and h). Therefore, the finding of three signals with the SYTprobeB was not a consequence of a possible trisomy 18, but should instead be interpreted as indicating a split at the SYT locus. In this analysis, we found no nuclei with four split signals, implicating that the double transcripts could not result from two translocations being present in the same cell (Figure 4c).
The SYT-SSX1 translocation was detected as a fusion signal after co-hybridization with the SYT probeA and the SSX1probe (Figures 3 and 4). The SYT-SSX1 control and case 9 were found to have very high proportions of SYT-SSX1 carrying cells, with more than 85% positive nuclei. However, in the other nine double fusion tumors analysed, the proportion of SYT-SSX1 positive nuclei were low, ranging from 6 to 14% (Table 1). The detected frequencies were all above the background level of 2% false positives determined from the control expressing SYT-SSX2 only. In six of the nine cases the differences were supported by statistical analyses. Thus, for cases 3–5, 7, 10 and 12 the detection of 9–14% SYT-SSX1 positive nuclei represented significantly higher levels than the 2% falsely positive nuclei detected in the negative control (P-values 0.006–0.01).
To detect the SYT-SSX2 translocation, the SYT probeA was co-hybridized with the SSX2probe. Positive nuclei with fusion signals were found in very high proportions in the SYT-SSX2 control, and in nine of the 10 double fusion tumors analysed (67–90%, Table 1). In the remaining case 9, which showed a majority of SYT-SSX1 positive cells, the presence of the SYT-SSX2 translocation was determined from the frequency of split signals occurring at co-hybridization with the SSX1probe and the SSX2probe. This analysis revealed 16% positive nuclei (Figure 4d) which was clearly above the background level of 3% false positives seen in the negative control expressing SYT-SSX1 only.
Taken together, the FISH results indicate that the SYT-SSX2 fusion was predominant in 10 of the tumors, while SYT-SSX1 was predominant in one case. Although detection methods based on PCR amplifications are not quantitative per se, it can be noted that the findings by FISH are in agreement with those obtained by the initial sequencing of the RT–PCR products, and with the genomic PCR.
It is often suggested that in tumors characterized by a specific oncogenic genetic alteration, the occurrence of this abnormality is sufficient for tumor development, while the tumor progression would result from other, secondary events. In line with this assumption the demonstration of co-existing SYT-SSX1 and SYT-SSX2 fusions in a significant proportion of primary synovial sarcomas is surprising, and as in all situations where unexpected findings are made, the methodological aspects of the study were scrutinized. In most of the cases studied the double SYT-SSX translocations were demonstrated by three independent methods, each associated with its own advantages and disadvantages. RT–PCR based detection of the SYT-SSX fusions were repeatedly performed, with identical results on independent RNA preparations, to minimize the risk of cross-contaminations. Similarly, the sequencing performed on the products generated by three different semi-nested and direct RT–PCR assays makes the detection of false positives from mispriming less likely. The demonstration of fusion products of different sizes in the genomic PCRs, even using the same primer combinations, implies that the translocations occurred between different sites within the target genes. This finding also provides a strong support of the methodological accuracy in detecting double SYT-SSX fusions in these tumors. By interphase FISH analyses, the presence of SYT-SSX2 in the majority of tumor cells was undoubtedly shown in all but one of the tumors, which instead was clearly positive for SYT-SSX1. In the SYT-SSX2 predominant cases the presence of SYT-SSX1 ranged from 6 to 14%. The lowest figures obtained here may represent a random variation of background false positives, and is compatible with the presence of SYT-SSX1 only in a very small proportion of tumor cells in these cases. However, for the SYT-SSX1 predominant case 9, the presence of SYT-SSX2 in a subpopulation of cells was well supported by the demonstration of split signals between the closely located SSX1 and SSX2 genes in 16% of the nuclei. The scoring of split signals is usually straightforward and highly reliable, and the frequency of positive cells obtained for this tumor is well above the background level of 3% false positives.
In this study co-existence of SYT-SSX1 and SYT-SSX2 fusion genes was demonstrated in 10% of SYT-SSX positive synovial sarcomas based on a tumor collection of more than 100 cases, implying that this situation may represent a recurrent event of relevance for the tumor development. The molecular mechanisms underlying this phenomenon are unknown, but several feasible explanations may be offered. The presence of independently arising subclones being present in the same tumor, each carrying either SYT-SSX1 or SYT-SSX2 represents a first and distinct possibility. If the SYT-SSX fusion is really the initiating genetic key event such a tumor would be expected to have a polyclonal origin, or consist of two independently occurring primary tumors. Alternatively, it could have a monoclonal origin resulting from a yet unidentified key event and where the SYT-SSX fusions represent secondary events occurring in subclones during the tumor progression. In agreement with this hypothesis, there are good examples of solid tumors in which the most frequent cytogenetic abnormality is not the genetic key event, such as retinoblastoma which is characterized by an isochromosome 6p but where the tumors result from inactivation of the retinoblastoma tumor suppressor gene located in 13q. Applied to the present study, both models require that the SYT-SSX1 fusion would have occurred before SYT-SSX2 in case 9 which is predominant for SYT-SSX1, while in the other cases the tumor development would have been the reverse. Considering that SYT-SSX1 tumors are more common in the general population, a majority of SYT-SSX2 predominant tumors in this series is unexpected. Furthermore, the strong association between the SYT-SSX translocation and the development of synovial sarcoma is generally interpreted as if one SYT-SSX fusion is sufficient for tumor development, which would make this latter possibility both unlikely and unappealing. However, the presence of a SYT-SSX translocation in more than 90% of synovial sarcomas also implies that this genetic rearrangement occurs with a high probability, e.g. as a result of a close physical location of the corresponding chromosomal regions during a critical phase of mitosis. Consequently, if one SYT-SSX fusion would be sufficient for a synovial sarcoma to develop, this tumor form would be expected to be commonly occurring in the general population. Therefore, towards these ends, and considering the rarity of synovial sarcoma, it seems reasonable to speculate about a yet unidentified initiating event.
This presumptive first event should stimulate cell proliferation, and in turn facilitate a second alteration to take place, which in the case of SYT-SSX translocation would give rise to a tumor with the morphological characteristics of a synovial sarcoma. The evolution of a subclone carrying one type of SYT-SSX fusion from a cell with the other SYT-SSX fusion variant represents a third and likely possibility. This type of tumor would theoretically have either a monoclonal origin where the tumor growth results from a SYT-SSX fusion only, or have a polyclonal origin involving a first unknown event which was followed by a SYT-SSX translocation. In no case in the present study were the two fusions found to be present in the same cell, as demonstrated by the finding of one split SYT signal in the vast majority of cells, and the absolute lack of nuclei with split of both SYT loci using the FISH probe covering the SYT gene locus. This distribution of SYT-SSX translocations excludes the possibility that a second SYT-SSX translocation have taken place in a cell already carrying one form of SYT-SSX fusion. However, it is possible that a transition through, for example, a deletional remodeling have occurred in a tumor cell carrying one type of SYT-SSX fusion, leading to a population of cells with the other type of fusion only. A similar mechanism has already been described in other types of tumors, for example in mouse plasmocytomas (Kovalchuk et al., 1997). In the cases with double fusion transcripts, SYT-SSX2 was predominating in all but one of the tumors, in which SYT-SSX1 was more abundant. This situation might appear somewhat in contrast to the findings that the SYT-SSX1 transcript is associated with a more aggressive tumor behavior in the form of distant metastasis, and possibly also with a higher proliferation rate (Kawai et al., 1998; Nilsson et al., 1999). However, considering that a transition probably occurs with a low frequency, and that it is only possible to go from SYT-SSX2 to SYT-SSX1 but not in the reverse direction (Figure 3), the remodeling is expected to preferentially occur late in tumor progression, and the resulting tumor cell populations to be low abundant. Co-cultivation of SYT-SSX1 and SYT-SSX2 cell lines in various proportions would represent a good model for how tumor populations predominant in either of SYT-SSX2 or SYT-SSX1 could occur in vitro.
The correct recognition of the different entities and subentities of soft tissue sarcomas are essential for their proper clinical handling. For synovial sarcomas the SYT-SSX1 fusion is associated with a more aggressive clinical behavior, and in addition there is a tight link between the type of fusion and the histopathological subtype. In this study, we found that two of the SYT-SSX2 predominant double transcript tumors were biphasic. This fact provides further evidence for co-existence of the SYT-SSX1 and SYT-SSX2 fusion genes in the same tumors. Since only the SYT-SSX1 variant has been associated with the biphasic phenotype (Kawai et al., 1998; Nilsson et al., 1999), our findings represent an important first observation which should be tested in a different sample set and especially in the rare cases of biphasic and SYT-SSX2 positive tumors reported.
Materials and methods
Tumor specimens were collected in liquid nitrogen directly at the surgical theatre from patients operated on for synovial sarcoma, the clinical data of which will be published separately (B Skytting et al., in preparation). The present study focused on the 12 tumors identified from the collection of 146 cases, which were found to express fusion transcripts of both SYT-SSX1 and SYT-SSX2 (Table 1). In addition two tumors expressing only SYT-SSX1 and SYT-SSX2, respectively, were included as controls in the analyses. The study of the tumor material was approved by a national ethical committee.
Reverse transcription PCR (RT–PCR) and cDNA sequencing
The sequences and positions of all synthetic oligonucleotides used in the study are detailed in Table 2. Total RNA was isolated with the RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany) and used for RT–PCR based detection of the SYT-SSX1 and SYT-SSX2 fusion transcripts essentially as described (Nilsson et al., 1999; Brodin et al., 2001). In short, cDNA was synthesized with SuperScript II RNase H- Reverse Transcriptase (GIBCO BRL Life Technologies, USA) in the presence of random primers for 1 h at 42°C. In the first of the two subsequent PCRs, both translocations were amplified with Oligo 1+2 using the following thermocycling conditions: an initial denaturation at 94°C for 3 min was followed by 35 step cycles of denaturation at 94°C for 30 s, annealing at 63°C for 30 s, and elongation at 72°C for 30 s, and a final extension at 72°C for 10 min. To distinguish the two fusion products, seminested PCRs were then performed with Oligo 1+3 for SYT-SSX1 and Oligo 1+4 for SYT-SSX2. The amplified products were detected by ethidium bromide staining after size separation in 2% agarose gels and visualized in a Fluor-S MultiImager System (BioRad). All PCR products generated with the specific primer combinations were directly sequenced by cycle sequencing with dye-labeled terminators (Big Dye Terminators, Applied Biosystems, Foster City, USA) and PCR primers, and subsequently analysed on an ABI PRISM 377XL sequencer (PE Applied Biosystems, USA).
In the initial analysis of the 146 synovial sarcomas, the seminested approach was applied to screen for the SYT-SSX1 and SYT-SSX2 fusions respectively. This identified 81 tumors which expressed SYT-SSX1 and 52 which were positive for SYT-SSX2 (B Skytting et al., in preparation). Among the totally 121 positive tumors, 12 were found to express both the SYT-SSX1 and SYT-SSX2 fusion transcripts. Parallel amplification of control samples without the addition of RNA or without RT served as negative controls according to the previously described method (Nilsson et al., 1999). In the remaining 25 tumors, the results were always negative on the three occasions the analyses were repeated, while amplification of β-actin was positive thus providing an independent control of RNA presence and quality. For the 12 tumors with double fusion transcripts and the two controls, additional RT–PCR and sequencing analyses were performed on RNA preparations from new tissue samples. In these cases sequencing was performed on the products from PCR with the consensus primers (Oligos 1+2) and on products generated directly with the specific primers (Oligos 1+3 and 1+4 respectively).
Normal sequences of the SYT, SSX1 and SSX2 breaking introns
Partial genomic sequences of the human SYT gene were available in GenBank (accession # AC021302.4), in the form of a working draft with 18 unordered fragments. By alignment with the SYT cDNA sequence, fragment 113753–140203 was found to contain the 14 019 bp long intron where breakages occur plus the flanking exons. From this sequence, Oligos 5–10 were designed for subsequent genomic PCR of the SYT-SSX translocations.
The breaking intron 4 of the SSX1 and SSX2 genes were generated by sequencing of constitutional DNA from two reference individuals. PCR products of intron 4 were amplified with primers located in exons 4 and 5 of the respective genes, i.e. Oligos 11+12 for SSX1 and Oligos 14+15 for SSX2. The products were partially sequenced from the 3′-ends, and the obtained sequences of approximately 400 nt (GenBank accession #s AF351583 and AF351584) were used to design Oligos 13 and 16 for the detection of SYT-SSX1 and SYT-SSX2, respectively.
Genomic PCR and sequencing of the SYT-SSX1 and SYT-SSX2 translocations
High molecular weight DNA was isolated with the QIAmp Mini Kit (QIAGEN GmbH) from 11 of the synovial sarcomas expressing both fusion transcripts (cases 2–12, Table 1). Genomic PCR was then performed with various primer combinations to detect the SYT-SSX translocations. In the first round of amplifications each of Oligos 5, 7 and 9 in SYT were combined with Oligo 12 in exon 5 of SSX1 and Oligo 15 in exon 5 of SSX2. When necessary, secondary nested PCR was performed with either of Oligos 6, 8, or 10 in SYT and Oligos 13 or 16 in intron 4 of SSX1 and SSX2, respectively. All amplifications were carried out using the Expand Long Template PCR System (Roche Diagnostics, Germany) with the PCR conditions: denaturation at 94°C for 2 min; 10 step cycles of 94°C for 10 s, 65°C for 30 s, 68°C for 3 min; 25 step cycles of 94°C for 10 s, 65°C for 30 s, 68°C for 3 min+20 s/cycle; and a final extension at 68°C for 7 min.
All generated PCR products detectable by agarose gel electrophoresis were subjected to sequencing, which was carried out in both directions with BigDye Terminators and the PCR primers.
Isolation of PAC clones
Genomic clones for FISH analyses were isolated from the total human PAC library RPCI6 created by Pieter de Jong at the Roswell Park Cancer Institute. DNA pools of the library were purchased from the Resource Center of the Human Genome Project and screened by PCR amplification with Oligos 17+18 specific for D18S1253, Oligos 19+20 for DXS255, Oligos 21+22 for DXS14, and Oligos 23+24 and 25+26 for SYT. For each of the four loci, one positive PAC clone was selected including N193/SYTprobeA at D18S1253, L0721/SSX1probe at DXS255, A02177/SSX2 probe at DXS14 and N2252/SYTprobeB at the SYT gene locus. The SYTprobeB was identified with two primer pairs located on either side of the breaking intron, i.e. Oligo 23+24 in the sixth exon upstream of the breaking intron, and Oligos 25+26 in the last exon of SYT. All four PAC clones were verified by PCR amplifications with the primer pairs used for their initial detection, and their chromosomal origins were confirmed by mapping on normal metaphases prepared from cultured lymphocytes of normal healthy individuals. Simultaneous hybridization of the differentially labeled SSX1 and SSX2 probes gave a fusion signal on metaphase X chromosomes, and two discrete signals on interphase nuclei.
Interphase FISH analyses
Imprints for interphase FISH analyses were made from 10 of the double transcript tumors (cases 1–5, 7 and 9–12) and from the two single transcript tumors used as controls (Table 1). A small piece of each tumor was gently touched onto a glass slide, air dried, fixed in methanol:acetic acid (3 : 1), and dehydrated in an ethanol series.
Dual-color FISH was performed using standard methods. The PAC clones were labeled with digoxigenin or biotin (Roche Diagnostics, Boehringer Mannheim, Germany) by nick-translation, mixed together with human Cot-1 DNA (Gibco, BRL), denatured, and hybridized onto denatured slides with interphase nuclei. The biotin labeled probes were detected with Fluorescein Avidin DCS (Vector Laboratories Inc., USA), and the digoxigenin labeled probes with anti-digoxigenin conjugated with rhodamine (Roche Diagnostics). In the FISH analyses the digoxigenin labeled SYTprobes A and B were used in combination with the biotin-labeled SSX1 and SSX2 probes to detect the SYT-SSX translocations. The interphase FISH results were scored by an independent analyser who was without knowledge of the molecular status of the tumor samples. The results were evaluated in a Zeiss Axioplan 2 Imaging epifluorescence microscope. Nuclei in which the two probes were fused, touched or close to each other (distance <=1 probe signal) were scored as positive for gene fusions. From each case 100 non-overlapping nuclei were included in the scoring without other selection criteria. All discrete hybridization signals were counted and manual adjustment was applied to detect signals in slightly different focal planes.
The specificity and sensitivity of the dual color FISH assays were determined by analyses of the control samples. Co-hybridization of the SYTprobeA and the SSX1 probe to the control tumor expressing only SYT-SSX1 gave the expected positive fusion signal in 87 of the 100 analysed nuclei, and was only falsely positive in 2 out of 100 nuclei of the control expressing only SYT-SSX2. Co-hybridization of the SYTprobeA and the SSX2probe to the SYT-SSX2 control was positive in 78 out of 100 nuclei as expected, but in the SYT-SSX1 control the frequency of false positive nuclei was very high (69 out of 100). Hence, since this combination could not be used to accurately detect the SYT-SSX2 translocation in cells with the SYT-SSX1 translocation, it is not applicable to tumors with a high proportion of SYT-SSX1 positive nuclei. This methodological difficulty is related to the fact that if the SSX1 gene is broken and fused with SYT, a FISH fusion signal will be detected with both the SYTprobeA+SSX1 probe and with the SYTprobeA+SSX2 probe, since the SSX1 and SSX2 probes will remain closely located also after the SSX1 break have occurred. Therefore, in tumors with a large proportion of SYT-SSX1 positive nuclei, the presence of SYT-SSX2 was indirectly determined by co-hybridization with the SSX1 and SSX2 probes. Using this split FISH assay the SSX1 and SSX2 probes are only separated if the SSX2 gene is broken, but not if the SSX1 gene is broken. In case 9 (Table 1) the frequency of split signals (distance >2 probe signals) was scored in 100 nuclei. Furthermore, the split assay was also used as a positive control to validate the demonstration of SYT-SSX2 using the fusion assay (SYTprobeA+SSX2 probe) in new tumor samples from the double fusion tumors, cases 2 and 4. This gave the proportions of positive/negative nuclei of 95/5 and 89/11, which are well in agreement with the proportions determined by the fusion assay (Table 1).
The proportions of positive and negative interphase nuclei detected at interphase FISH were compared between the synovial sarcomas using χ2 test of Fisher's exact test with the SAS 6.12 statistical package (SAS Institute, USA). Probabilities of less than 0.05 were accepted as significant.
Åman P . 1999 Cancer Biol. 9: 303–318
Brett D, Whitehouse S, Antonson P, Shipley J, Cooper C, Goodwin G . 1997 Hum. Mol. Genet. 6: 1559–1564
Brodin B, Haslam K, Yang K, Bartolazzi A, Xie Y, Starborg M, Lundeberg J, Larsson O . 2001 Gene 268: 173–182
Clark J, Rocques PJ, Crew AJ, Gill S, Shipley J, Chan AM, Gusterson BA, Cooper CS . 1994 Nat. Genet. 7: 502–508
Crew AJ, Clark J, Fisher C, Gill S, Grimer R, Chand A, Shipley J, Gusterson BA, Cooper CS . 1995 EMBO J. 14: 2333–2340
de Leeuw B, Balemans M, Olde Weghuis D, Geurts van Kessel A . 1995 Hum. Mol. Genet. 4: 1097–1099
Eid JE, Kung AL, Scully R, Livingston DM . 2000 Cell 102: 839–848
Fligman I, Lonardo F, Jhanwar SC, Gerald WL, Woodruff J, Ladanyi M . 1995 Am. J. Pathol. 147: 1592–1599
Gure AO, Tureci O, Sahin U, Tsang S, Scanlan MJ, Jager E, Knuth A, Pfreundschuh M, Old LJ, Chen YT . 1997 Int. J. Cancer 72: 965–971
Kawai A, Woodruff J, Healey JH, Brennan MF, Antonescu CR, Ladanyi M . 1998 N. Engl. J. Med. 338: 153–160
Kovalchuk AL, Müller JR, Janz S . 1997 Oncogene 15: 2369–2377
Ladanyi M, Bridge JA . 2000 Hum. Pathol. 31: 532–538
Lim FL, Soulez M, Koczan D, Thiesen HJ, Knight JC . 1998 Oncogene 17: 2013–2018
Nagai M, Tanaka S, Tsuda M, Endo S, Kato H, Sonobe H, Minami A, Hiraga H, Nishihara H, Sawa H, Nagashima K . 2001 Proc. Natl. Acad. Sci. USA 98: 3843–3848
Nilsson G, Skytting B, Xie Y, Brodin B, Perfekt R, Mandahl N, Lundeberg J, Uhlen M, Larsson O . 1999 Cancer Res. 59: 3180–3184
O'Dwyer ME, Druker BJ . 2000 Curr. Opin. Oncol. 12: 594–597
Panagopoulos I, Mertens F, Isaksson M, Limon J, Gustafson P, Skytting B, Åkerman M, Sciot R, Dal Cin P, Samson I, Iliszko M, Ryoe J, Debiec-Rychter M, Szadowska A, Brosjö O, Larsson O, Rydholm A, Mandahl N . 2001 Genes Chromosomes Cancer 31: 362–372
dos Santos NR, de Bruijn DR, Guerts van Kessel A . 2001 Genes Chromosomes Cancer 30: 1–14
Skytting B, Nilsson G, Brodin B, Xie Y, Lundeberg J, Uhlen M, Larsson O . 1999 J. Natl. Cancer Inst. 91: 974–975
Sreekantaiah C . 1998 Cytogenet. Cell. Genet. 82: 13–29
Turc-Carel C, Dal Cin P, Limon J, Rao U, Li FP, Corson JM, Zimmerman R, Parry DM, Cowan JM, Sandberg AA . 1987 Proc. Natl. Acad. Sci. USA 84: 1981–1985
The authors would like to thank Dr Yaping Jin for help with the statistical analyses and Dr Armando Bartolazzi for valuable comments on the manuscript. This study was supported by The Swedish Cancer Foundation, the Milton Foundation, the Cornell Foundation, the Cancer Society in Stockholm, the Swedish Children Cancer Society, the Stockholm County Council, and the Torsten and Ragnar Söderberg Foundations.
About this article
Cite this article
Yang, K., Lui, W., Xie, Y. et al. Co-existence of SYT-SSX1 and SYT-SSX2 fusions in synovial sarcomas. Oncogene 21, 4181–4190 (2002). https://doi.org/10.1038/sj.onc.1205569
- synovial sarcoma
- double fusion
- double translocation
In silico and in vitro screening of small molecule Inhibitors against SYT-SSX1 fusion protein in synovial sarcoma
Computational Biology and Chemistry (2018)
Archives of Pathology & Laboratory Medicine (2018)
Current Opinion in Otolaryngology & Head and Neck Surgery (2018)
Multiple splice variants of EWSR1-ETS fusion transcripts co-existing in the Ewing sarcoma family of tumors
Cellular Oncology (2013)