|Home | About | Journals | Submit | Contact Us | Français|
Synovial sarcoma (SS) is the most frequent nonrhabdomyosarcomatous soft tissue sarcoma encountered in adolescents and young adults, and despite advances in the treatment of local disease, metastases remain the main cause of death. The aim of this study was to characterize a single-center series of pediatric SS molecularly to seek any biomarkers or pathways that might make suitable targets for new agents. Seventeen cases of pediatric SS showing the SYT-SSX fusion transcript were screened immunohistochemically, biochemically, molecularly, and cytogenetically (depending on the available material) to investigate any expression/activation of epidermal growth factor receptor, platelet-derived growth factor receptor alpha (PDGFRα), PDGFRβ, Akt, and deregulated Wnt pathway. The most relevant outcome was the finding of activated epidermal growth factor receptor, PDGFRα, and PDGFRβ, which activated Akt in both the monophasic and biphasic histologic subtypes. Consistently, Akt activation was completely abolished in an SS cell line assay when stimulated by PDGF-AA and treated with the phosphatidylinositol 3-kinase inhibitor LY294002. Our results also showed the nuclear localization of β-catenin and cyclin D1 gene products in monophasic SS and the movement of β-catenin into the cytoplasm in the glandular component of the biphasic subtype. Although they need to be confirmed in larger series, these preliminary data suggest that therapeutic strategies including specific inhibitors of the phosphatidylinositol 3-kinase/Akt pathway might be exploited in SS.
Synovial sarcoma (SS) is one of the most common mesenchymal malignancies and accounts for approximately 8% to 10% of all soft tissue sarcomas; it is also reported to be the most frequent nonrhabdomyosarcomatous soft tissue sarcoma encountered in adolescents and young adults (15–20% of cases). It is characterized by the specific chromosomal translocation t(X;18) (p11;q11) that fuses the SYT gene from chromosome 18 with the SSX1 (approximately 2/3 of cases), SSX2 (approximately 1/3 of cases), or SSX4 gene (rare cases) from the X chromosome. Although it is thought that SYT-SSX plays a central part in the development of SS, the mechanism of tumor initiation is still unknown.
Gene array and immunohistochemistry (IHC) studies have recently identified high epidermal growth factor receptor (EGFR), Her-2/neu, IGF2 and HGFR gene expression in SS [1,2], but the correlation between this and the activation of specific cascades [such as phosphatidylinositol 3-kinase (PI3K)/Akt] has not been fully investigated.
Akt is an intracellular serine/threonine kinase, which, once activated by PI3K, moves from the cell membrane to the cytoplasm and/or nucleus, where it controls survival (by inhibiting pro- and activating antiapoptotic factors), proliferation (by means of direct p21 and p27 phosphorylation), and other activities essential to tumor progression, such as angiogenesis, invasion, and metastasis. It is a key activator of the mammalian target of rapamycin that induces the expression of proangiogenic genes by stabilizing the hypoxia-inducible factor. In addition to direct GSK3B inactivation, it has also been shown that Akt directly phosphorylates the β-catenin Ser552 residue in epithelial cancer cells  leading to the nuclear shift/activation of β-catenin.
In cell adhesion and transcription functions, β-catenin has the appropriate selection of which is crucial for normal development and the avoidance of cancer. It is well known that there is a striking cytoplasmic and nuclear accumulation of β-catenin in most SS, which is consistent with the recently reported presence of a transcriptionally active nuclear complex containing SYT-SSX2 and β-catenin  and supports the idea that the sarcoma chimeric protein contributes to cancer formation by activating one of the β-catenin-targeted programs. However, because the accumulation of β-catenin in SS does not apparently depend on canonical Wnt activation and mutations in APC, β-catenin, and E-cadherin are rare [5,6], it may be that β-catenin is stabilized through its phosphorylation by receptor tyrosine kinases (RTKs) .
Bearing this in mind, after making a preliminary immunophenotypic analysis, we investigated 17 cases of pediatric SS—all with an SYT-SSX fusion transcript—using molecular biochemical methods suited to the type of material available (formalin-fixed or frozen) to seek any potential biomarkers or pathways that might be suitable targets for licensed drugs, such as the expression of EGFR, platelet-derived growth factor receptor alpha (PDGFRα), PDGFRβ, Akt, and deregulated Wnt pathways.
Our findings support the expression and activation of EGFR, PDGFRα, and PDGFRβ, which may activate Akt. These albeit preliminary data suggest that therapeutic strategies including specific inhibitors of the PI3K/Akt pathway might be exploited in SS.
We analyzed specimens from 17 patients with SS (nine males and eight females aged 7–18 years; median age, 11 years), all but one of whom (BSS8 in Table 1) were treated at the Pediatric Oncology Unit of the Fondazione IRCCS, Istituto Nazionale Tumori, Milan, Italy. All of the specimens came from primary tumors and had been obtained before any treatment had been given, and representative samples obtained from formalin-fixed material were immunophenotyped. All of the biochemical and molecular analyses were made using frozen sections after the tumoral component had been carefully dissected under a microscope to avoid contamination by normal or necrotic tissue.
Written informed consent was obtained from all of the patients and/or their parents or legal guardians.
The samples were immunophenotyped using the following antibodies and dilutions: PDGFRα (C-20), sc-338 rabbit polyclonal IgG (Santa Cruz Biotechnology Inc., Santa Cruz, CA) diluted 1:200; PDGFRβ (P-20), sc-339 rabbit polyclonal IgG (Santa Cruz Biotechnology) diluted 1:100; KIT, rabbit polyclonal antihuman CD117 (code A 4502; Dako, Carpinteria, CA) diluted 1:50; HER2/NEU, c-erb-2 rabbit polyclonal antihuman Oncoprotein (code A 0485; Dako) diluted 1:2000; cyclin D1 rabbit monoclonal antibody clone SP4 (code RM-9104-S1; Lab Vision, Fremont, CA) diluted 1:100; β-catenin mouse monoclonal antibody (clone 14, code C19220; Transduction Laboratories, Lexington, KY) diluted 1:2000.
The antigens for all antibodies were retrieved using 5 mM citrate buffer (pH 6) in an autoclave at 95°C for 6 minutes, except for cyclin D1 and β-catenin, which took 30 minutes. All of these antibodies were developed as described elsewhere .
Epidermal growth factor receptor was immunostained using the EGFR-pharmDx kit (code K1492; Dako) following the manufacturer's instructions. The EGFR-immunostained cells were quantitatively evaluated, and their staining intensity was scored as described elsewhere . The following positive controls were used in each IHC experiment: one case of fibromatosis for PDGFRα and PDGFRβ, one case of KIT-mutated GIST for KIT, one case of ductal salivary gland carcinoma for HER2/NEU, one case of colon carcinoma for β-catenin, and one case of mantle cell lymphoma for BCL1.
Protein extraction and immunoprecipitation/Western blot analysis. The EGFR, PDGFRα, and PDGFRβ proteins were extracted, immunoprecipitated, and blotted as described elsewhere . The A431 cell line (American Type Culture Collection, Manassas, VA) was used as a positive control for the EGFR protein expression/phosphorylation experiments, and the NIH3T3 cell line (American Type Culture Collection) for the PDGFRα protein expression/phosphorylation experiments.
Twenty micrograms of cytoplasmic total protein extract was used with anti-phospho-Akt Ser 473 polyclonal antibody (#9271; Cell Signaling Technology, Beverly, MA) diluted 1:1000 in the Akt Western blot (WB) experiments, and the filters were subsequently stripped and incubated with anti-Akt polyclonal antibody (#9272; Cell Signaling) diluted 1:1000; the NIH3T3 cell line was used as a positive control. Thirty micrograms cytoplasmic total protein extract was used with the anti-phospho-β-catenin Y142 (clone ab27798; Abcam, Cambridge, UK) diluted 1:1000 in the β-catenin WB experiments, and the filters were subsequently stripped and incubated with anti-β-catenin (clone BDI109; Abcam) diluted 1:1000; the A431 cell line was used as a positive control.
The CME-1 cell line was cultured as previously described . For the biochemical analyses, the CME-1 cells were serum-starved for 16 hours, and then stimulated with 50 ng/ml of PDGF-AA (catalogue 100-13A; PeproTech, Princeton, NJ) for 5, 15, 30, and 60 minutes. The PI3K inhibitor LY294002 (kindly provided by Dr. L. Lanzi, Experimental Oncology Unit, Fondazione IRCCS, Istituto Nazionale Tumori, Milan, Italy) was used at volumes of 0.5, 5, and 50 µM in CME-1 serum-starved cells stimulated with 50 ng/ml of PDGF-AA for 15 minutes.
The WB analyses were made as described above for the primary specimens.
Total RNA was extracted from formalin-fixed materials and reverse-transcribed. All of the samples were tested for cDNA integrity and DNA contamination by amplifying the β-actin and HPRT (hypoxanthine guanine phosphoribosyl transferase) housekeeping genes.
SYT/SSX fusion transcripts were detected by polymerase chain reaction (PCR) as described in detail elsewhere . Briefly, goodquality RNA was obtained from all 17 samples, all of which showed the SYT-SSX gene fusion transcript: 11 (65%: 5/8 monophasic cases and 6/9 biphasic cases) carried SYT-SSX1 and six (35%: 3/8 monophasic and 3/9 biphasic cases) carried SYT-SSX2 (Table 2).
In all cases, the presence of an SYT-SSX translocation was confirmed by fluorescence in situ hybridization (FISH) analysis using BAC probes RP11-38O23 and RP11-344N17 for SSX1, RP11-552J9 and RP13-77O11 for SSX2, and RP11-737G21, RP11-786F14, and RP11-399L5 for SYT as previously described .
These were made using previously described probes .
Epidermal growth factor receptor, PDGFRα, PDGFRβ, PDGFA, TGFA, and glyceraldehyde-3-phosphate dehydrogenase cDNAs were relatively quantified by means of real-time quantitative PCR (ABI PRISM 5700 PCR Sequence Detection Systems; Applied Biosystems, Foster City, CA) using a TaqMan-based analysis and following the manufacturer's instructions. The relative changes in gene expression were calculated using the 2-ΔΔCt method .
DNA was extracted from selected paraffin-embedded sections of six MSS showing nuclear β-catenin, and β-catenin exon 3 was amplified as previously described . No activating mutations were found in these samples.
Receptor tyrosine kinases: EGFR, PDGFRα, PDGFRβ, KIT, and HER2/NEU. Epidermal growth factor receptor immunoreactivity was high  in the cytoplasm of almost all of the tumor cells in five of eight MSS but was restricted to the spindle cell component in all of the BSS samples. Her2/Neu was negative in all of the MSS but showed mild cytoplasmic reactivity in the glandular component of three of nine BSS. Platelet-derived growth factor receptor α was reactive in the cytoplasm of most of the MSS tumor cells and in the spindle and glandular cells of the BSS; the same was true of PDGFRβ, but the intensity of the staining was less in half the cases.
There was no Kit decoration in any of the MSS, and it was restricted to the cytoplasm of the glandular component in five of nine BSS, as previous reported .
β-Catenin and cyclin D1 products (bcl1). Because cyclin D1 can be considered a marker of β-catenin activation, we evaluated the immunoreactivitiy of both proteins.
Most tumor cells in six of eight MSS cases showed nuclear or nuclear/cytoplasmic β-catenin immunoreactivity, whereas cytoplasmic and cell membrane immunostaining was restricted to the glandular component in all of the BSS.
Nuclear bcl1 immunoreactivity was observed in 50% to 80% of the MSS tumor cells and decorated all of the nuclei of the glandular component in the BSS.
Using a pool of normal mesenchymal-derived tissues as calibrators, we observed an increase in the transcripts of EGFR (median, 1.7 x 101; range, 1 x 100 to 1.9 x 102) and PDGFRα (median, 5 x 100; range, 1 x 100 to 6.4 x 101); the median level of the PDGFRβ transcript was similar to that of the calibrator (median, 1 x 100; range, 1 x 10-1 to 4 x 100). These results support the high IHC EGFR and PDGFRα scores in both the MSS and BSS.
Real-time PCR revealed the presence of the cognate ligand of PDGFRα (PDGF-AA) and EGFR (TGFA) in all cases. Given the lack of appropriate calibrators for the relative quantification of PDGF-AA and TGFA, we compared their median Ct values (26.5; range, 21.9–29.7 for PDGF-AA; and 31.2; range, 27.3–34.8 for TGFA) with that of the Gapdh housekeeping gene (25.9; range, 21.1–29).
Because the IHC results suggested that the RTKs expressed in most of the SS samples were EGFR, PDGFRα and, to a lesser extent, PDGFRβ, we investigated their activation in immunoprecipitation (IP)/WB analysis experiments.
Frozen samples of four MSS (#3, #4, #5, and #8) and three BSS (#1, #4, and #8) were available for biochemical analysis (frozen sections from both BSS1 and BSS4 had revealed an extensive glandular component on histologic examination). After specific IP (with anti-EGFR, anti-PDGFRα, and anti-PDGFRβ) and WB, all of the samples expressed activated (i.e., phosphorylated) EGFR (Figure 3A and Table 3), PDGFRα (Figure 3B and Table 3), and PDGFRβ (not shown).
To confirm RTK activation further, we investigated the presence and activation of the shared downstream effector Akt, the phosphorylation of which was evaluated directly by means of WB. This experiment revealed phosphorylated Akt in six of seven SS samples (Figure 3C and Table 3).
Time-course studies of Akt activation (in the form of the phosphorylation of Akt Ser 473 in response to PDGF-AA stimulation) showed Akt activation in serum-starved CME-1 cells as early as 5 minutes after stimulation with 50 ng/ml PDGF-AA; this peaked after 15 minutes and returned to baseline values after 1 hour (not shown).
We next examined the effects of the established PI3K inhibitor LY294002 on Akt activation in serum-starved and LY294002-treated CME-1 cells and observed that Akt Ser 473 phosphorylation was completely abolished by the addition of 50 µg/ml LY294002 (Figure 4).
Because gene amplification is one of the mechanisms responsible for RTK activation, we investigated EGFR, PDGFRα, and PDGFRβ gene copy numbers by means of FISH in two MSS (#1 and #6) and three BSS (#3, #5, and #7), all of which showed a normal disomic hybridization pattern for all RTK receptors, which is consistent with the absence of gene amplification.
As a further step, we extracted cytoplasmic β-catenin and assessed its phosphorylation in tyrosine 142 (Y142). Western blot analysis showed that all of the tested BSS samples (BSS1, BSS4, and BSS8) had cytoplasmic β-catenin (thus confirming the IHC results), and β-catenin Y142 phosphorylation was observed in BSS1 and BSS4 (Figure 5 and Table 3). Despite the definite prevalence of nuclear β-catenin as IHC decoration in our MSS samples, Y142-phosphorylated cytoplasmic β-catenin was also observed by WB in MSS3 and MSS8 (Figure 5 and Table 3), thus suggesting that IHC is not sensitive enough to indicate β-catenin localization precisely.
In light of the above data, it can be assumed that the cytoplasmic localization of β-catenin very often parallels its phosphorylation in Y142. Because β-catenin Y142 phosphorylation can be achieved by means of the formation of β-catenin/RTK complexes or the activation of RTKs, and our IHC and IP/WB results suggested EGFR and PDGFRα activation, the filters obtained after IP/WB analysis with anti-EGFR and anti-PDGFRα were stripped and incubated with the anti-β-catenin antibody. No β-catenin/EGFR or β-catenin/PDGFRα coimmunoprecipitation was observed in either the MSS or the BSS specimens (not shown).
The most relevant finding of this comprehensive investigation of a small, single-center series of pediatric SS cases is the presence of activated (i.e., phosphorylated) EGFR, PDGFRα, and Akt, which was supported by the biochemical results on snap-frozen material; we have previously reported the activation of PDGFRβ in a series of cryop-reserved SS specimens taken from adults . It is also worth noting the IHC evidence of nuclear β-catenin expression coupled with a cyclin D1 product (bcl1) in paraffin-embedded tissue; as previously reported [16,17], these results confirm that cyclin D1 can be considered one of the targets of activated β-catenin in SS.
Our biochemical data relating to cryopreserved samples and IHC data relating to formalin-fixed material (all showing overexpressed and activated EGFR, PDGFRα, and PDGFRβ), together with the demonstration of each receptor's cognate ligand by real-time PCR, support the idea that Akt is activated in response to multiple signals acting through the receptors through an autocrine/paracrine-mediated loop. This was confirmed by the absence of the activating mutation in EGFR  or PDGFRα  and the fact that the FISH data ruled out any gene alteration in the involved RTKs. In line with these findings, the Akt activation induced by PDGF-AA in the CME-1 SS cell line was completely abolished by the specific PI3K inhibitor LY294002, thus confirming that the PI3K/Akt pathway is biologically active in SS cell lines and is a potential therapeutic target in SS patients.
We also found a nuclear β-catenin expression in MSS and the increased expression of cyclin D1 (a β-catenin target gene) in both MSS and BSS. These findings not only indicate that this gene is activated but also suggest a possible mechanism underlying the β-catenin activation. The absence of β-catenin gene mutations in the six MSS showing nuclear β-catenin immunoreactivity and the absence of RTK/β-catenin complexes at IP analysis rule out the possibility that either are responsible for the nuclear localization of β-catenin; however, because our model highlighted the presence of activated Akt (a serine/threonine kinase downstream target of EGFR, PDGFRα, and PDGFRβ), this could presumably assist the nuclear translocation of β-catenin through its serine 552 phosphorylation or the direct inhibition of GSK3B . Unfortunately, the lack of protein extract from our frozen material prevented us from analyzing whether all these cases were also phosphorylated on serine residues.
The detection of cytoplasmic β-catenin Y142 phosphorylation coupled with a cytoplasmic decoration indicates that this residue may be important for the cytoplasmic localization. This is an intriguing point because, as previously observed in BSS , we found that β-catenin expression was restricted to the BSS glandular component, thus indicating a shift from the nucleus to the cytoplasm and/or cytoplasm membrane. In this light, the detection of β-catenin phosphorylation on a Tyr 142 residue in two of our BSS supports the possibility that the change in the localization of β-catenin occurs during the mesenchymal-epithelial transition, particularly because there was no correlation between the transcript type and the type of residue activation.
A clinical trial of geftinib (an EGFR inhibitor) in SS patients with locally advanced or metastatic disease has been conducted in Europe , despite the rare occurrence of EGFR mutations  and the gain in gene copy number (confirmed by our results). Inhibiting a downstream effector (such as Akt) to block multiple upstream activated RTKs (EGFR, PDGFRα, and PDGFRβ) seems to be a further promising therapeutic possibility as we found that more than one receptor is activated at the same time and direct Akt inhibitors have recently been developed.
It is also worth noting that a previous IHC-based study of non-small cell lung carcinoma  found that patients whose surgical specimens showed high levels of both EGFR and Akt activation were more sensitive to RTK inhibitors. Although the results of this study must be considered with caution as the antibody has now been withdrawn from production, the clinical evidence of a response seems to be reliable. We are faced with the problem of the reliability of IHC phospho-specific antibodies every day and find that they do not assure reproducible results, which is why they were not used in the present study.
Our comprehensive investigation of a small, single-center series of SS in children and adolescents adds molecular details to the scanty published findings concerning the biology of pediatric SS. They may also be of interest to clinicians as it is still debated whether SS has the same clinical behavior (and the same biology) in different age groups . This is important because SS are not always treated using the same strategy in different ages, and there is still disagreement between pediatric and adult medical oncologists concerning the role of chemotherapy.
In conclusion, although very preliminary and needing confirmation in a larger series, our finding that multireceptor-mediated Akt activation (and its inhibition by a PI3K inhibitor in the CME-1 SS cell line) may stabilize β-catenin in SS strongly suggests that inhibiting the PI3K/Akt pathway may be a useful adjunctive therapeutic strategy.
1Supported by grants from Associazione Italiana per la Ricerca sul Cancro to S.P.