Synovial sarcomas account for 5–10% of all malignant soft tissue tumors. They have been shown to express different membranous growth factor receptors, many of them signaling via intracellular kinase cascades. In our study, the functional role of PI3K/AKT signals in synovial sarcoma is analyzed with regard to tumor biology and therapeutic applicability. Immunohistochemical stainings of (Ser473)-phosphorylated (p)-AKT, its targets p-(Ser9)-GSK-3β and p-(Ser2448)-mTOR and the cell cycle regulators Cyclin D1 and p27KIP1 were performed in 36 synovial sarcomas. The PIK3CA gene was screened for mutations. In vitro, four synovial sarcoma cell lines were treated with the PI3K inhibitor LY294002. Phosphorylation of AKT, GSK-3β and mTOR was assessed, and cellular proliferation and apoptosis were analyzed to functionally characterize the effects of PI3K inhibition. Finally, coincubations of LY294002 with cytotoxic drugs were performed. Most tumors showed significant expression levels of p-AKT, p-GSK-3β and p-mTOR, indicating activation of the PI3K/AKT signaling cascade in synovial sarcomas; Cyclin D1 and p27KIP1 were differentially expressed. Mutations in the PIK3CA gene could be excluded. In vitro, PI3K inhibition diminished synovial sarcoma cell growth accompanied by reduced phosphorylation of AKT, GSK-3β and mTOR. Mechanistically, PI3K pathway inhibition lead to enhanced apoptosis and decreased cellular proliferation linked to reduced Cyclin D1 and increased p27KIP1 levels. Simultaneous treatment of synovial sarcoma cell lines with LY294002 and cytotoxic drugs resulted in additive effects. In summary, PI3K signaling plays an essential role in growth control of synovial sarcomas and might be successfully targeted in multimodal therapeutic strategies.
Synovial sarcomas account for 5–10% of all malignant soft tissue tumors. They are consistently characterized by a reciprocal t(X;18) translocation, which juxtaposes the SS18 gene on chromosome 18 to either the SSX1, the SSX2 or rarely the SSX4 gene on chromosome X. The SS18 protein functions as a transcriptional coactivator and is associated with the SWI/SNF complex, whereas the SSX proteins act as transcriptional corepressors and are associated with the polycomb complex.1 The chimeric SYT/SSX proteins constitute aberrant transcriptional regulators,1, 2 which have been shown to be involved in the activation of oncogenic pathways.3, 4 Currently, the therapeutic outcome of synovial sarcomas is mainly determined by the efficiency of surgery because a high tendency for local relapse is seen. Standardized chemotherapy and radiotherapy are further therapeutic options; however, specifically targeted therapies are currently not available.5
Several receptor tyrosine kinases have been shown to be expressed in synovial sarcomas, including the EGF receptor,6 the HER2 receptor,7 the PDGF receptor beta, the stem cell factor receptor c-KIT,8 the hepatocyte growth factor receptor c-MET,9 members of the FGF receptor family10 and the IGF-IR.11 However, apart from few exceptions, most approaches to inhibit synovial sarcoma cell growth in vitro by specifically targeting members of this group of cell membrane receptors only rarely showed effects.12, 13
One of the crucial pathways in the intracellular mediation of receptor tyrosine kinase signals is the PI3K/AKT signaling cascade. On growth factor stimulation, PI3K catalyzes the generation of phosphatidylinositol-3,4,5-triphosphate (PIP3) from phosphatidylinositol-4,5-triphosphate (PIP2). Functioning as second messenger, PIP3 recruits AKT/PKB, a 57 kDa Ser/Thr-kinase, to the plasma membrane where the latter is activated by phosphorylation through a 3-phosphoinositol-dependent protein kinase. Subsequently, AKT/PKB itself phosphorylates diverse growth-controlling effectors, such as MDM2, mTOR or GSK-3β, and (e.g., via GSK-3β or mTOR) regulates synthesis, stability or subcellular localization of cell cycle regulators such as Cyclin D1 and p27KIP1.14–17 D-type cyclins are growth factor sensors whose activity is essential for the progression through the G1 phase of the cell cycle. In contrast, cyclin-dependent kinase (CDK) inhibitors as p27KIP1 block CDK activity and thereby prevent the transition from G1 to S phase.18, 19
It has been shown in several tumors that PI3K signaling may gain independence from upstream signaling pathways by genetic alterations in different components of the pathway. Particularly, oncogenic mutations affecting the PI3KCA gene, which encodes the p110 alpha phosphatidylinositol-3′-kinase catalytic subunit, have been described.20, 21
Our study was carried out to analyze the biological relevance of the PI3K/AKT pathway in synovial sarcomas. As PI3K/AKT signaling constitutes a common effector of diverse receptor tyrosine kinases expressed in these tumors, we investigated if this pathway represents a useful target for specific therapeutic interventions.
Material and Methods
Patients, tumors and cell lines
Thirty-six cases of synovial sarcoma were analyzed comprising 23 monophasic and 13 biphasic tumors. Approval of the study by the Ethical Committee of the University of Bonn Medical Center was obtained. Fluorescence in situ hybridization (FISH) or PCR analysis were used to confirm the diagnosis of synovial sarcoma revealing a t(X;18) translocation as described before.22 The synovial sarcoma cell lines CME-1, 1273/99, SYO-1 and Fuji have been described before, and all carry a SS18/SSX translocation, which was confirmed by PCR using primers specific for the translocation subtype (last confirmation may 2010).23–26
DNA extraction and mutational analysis of the PI3KCA gene
DNA from paraffin-embedded cases of synovial sarcoma (of which sufficient material was available) was extracted using the Ambion RecoverAll total nucleic acid isolation kit (Ambion, Austin Texas) according to the manufacturer's instructions. PCR fragments covered the coding sequence of the previously described mutational hot spot regions (Exon 1, Exon 9 and Exon 20) of the PIK3CA gene. Primers and PCR conditions were described before.27 All PCRs were run in a UNO Thermoblock cycler (Biometra, Göttingen, Germany). PCR was carried out in a final volume of 10 μl containing 10–50 ng of DNA, 5 pmol of each primer, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.0–1.5 mM MgCl2, 200 mM of each deoxynucleotide and 0.25 units of Taq polymerase (Life Technologies, Karlsruhe, Germany). PCR products were run on 2% agarose gels and purified using the QIAquick Gel extraction kit (Qiagen, Hilden, Germany). Sequencing was performed by Entelechon GmbH (Regensburg, Germany) and Eurofins MWG Operon (Ebersberg, Germany).
Tissue specimens were fixed in 4% buffered formaldehyde and embedded in paraffin. After antigen retrieval (10 mM sodium citrate buffer, pH 6.0, microwave 600 W, 10 min) immunohistochemical stainings were performed on 4 μm sections with an Autostainer (DAKO, Hamburg, Germany; Cyclin D1, p27KIP1) or manually (p-(Ser473)-AKT, p-(Ser9)-GSK-3β and p-(Ser2448)-mTOR). For Cyclin D1 and p27KIP1, the antigen–antibody binding was visualized by means of the avidin–biotin complex (ABC-method) using AEC (3-amino-9-ethylcarbazol) as chromogen. For p-AKT, p-GSK-3β and p-mTOR stainings, the catalyzed signal amplification system (CSA II; DAKO) was used according to the manufacturer's instructions using DAB as chromogen. p-AKT, p-GSK-3β and p-mTOR antibodies (Cell Signaling Technology, Frankfurt, Germany) were used at a dilution of 1:200, p27KIP1 (Abcam, Cambridge, UK) at a dilution of 1:50 and Cyclin D1 (DCS, Hamburg, Germany) at a dilution of 1:25. Positive controls and negative control stainings using appropriate mouse IgG (DAKO) and rabbit IgG isotypes (DCS) were included. For p-(Ser473)-AKT, p-(Ser9)-GSK-3β and p-(Ser2448)-mTOR, cytoplasmic and membranous immunoreactivity was assessed using a semiquantitative score (negative, weak, moderate and strong) defining the staining intensity in the positive control (invasive ductal breast carcinoma) as strong. Nuclear Cyclin D1 expression was characterized as negative, weak (nuclear reaction in <5% of the tumor cells) or strong (nuclear reaction in >5% of the tumor cells). Based on its higher basal expression level, p27KIP1 immunoreactivity was characterized as absent, weak (nuclear reaction in <50% of the tumor cells) or strong (nuclear reaction in >50% of the tumor cells).
Culture and treatment of synovial sarcoma cells, MTT cell proliferation assay
Synovial sarcoma cell lines were cultured as described before.12 For proliferation assays, a comparative approach of serum-supplemented (10% FBS) and serum-reduced (2.0% FBS) growth conditions was chosen to test dependence of the cells on external growth factor stimulation. For the other experiments, cells were grown in serum-reduced medium (2.0% FBS). Treatments were done with different amounts of LY294002 (3.125–50 μM; Cell Signaling Technology). DMSO was used as control. For MTT proliferation assays, cells were cultured in 96-well dishes (Nunc, Wiesbaden, Germany) in a volume of 100 μl and a concentration of 5 × 104 cells/ml. Assays were performed for 72 hr, using a minimum of four replicates. For coincubation with conventional chemotherapeutic drugs, CME-1 and SYO-1 cells were treated with varying concentrations of doxorubicin, vincristine and actinomycin D (0.03–1000 ng/ml) without or with 1.8 μM (CME-1)/4.9 μM (SYO-1) LY294002, corresponding to doses resulting in ∼30% growth inhibition in these cells. Proliferation assays were performed using the MTT cell proliferation kit (Roche, Mannheim, Germany) according to the manufacturer's instructions. The formazan dye was quantified using a scanning multiwell spectrophotometer. Synergy was evaluated by the fractional product method.28 A difference of more than 10% between observed and predicted effect was considered to signify synergistic activity between LY294002 and the cytotoxic drug.
For flow cytometry and DAPI stainings, cells were cultured as described above in 25 cm2 cell culture flasks (Greiner, Frickenhausen, Germany) with a preincubation of 18 hr and treatment for 24 hr with 5 and 15 μM LY294002 (flow cytometry) or individual doses corresponding to a concentration of the substance leading to 50 or 75% growth inhibition (GI50 and GI75) as calculated from MTT assays (DAPI stainings). Appropriate controls were included.
For immunocytochemical analysis of synovial sarcoma cells, cells were treated for 24 hr, harvested, pelleted, fixed in 3.7% buffered paraformaldehyde and embedded in paraffin after standard dehydration. Semiquantitative analysis of nuclear Cyclin D1 and p27KIP1 expression was performed by counting strongly immunoreactive cells in at least five high power fields using a Leica DMLB microscope (X 400) (Leica, Solms, Germany).
Knockdown of PIK3CA by RNAi
1273/99 and CME-1 synovial sarcoma cells were cultured in 25 cm2 flasks in a volume of 4 ml serum-supplemented medium. At a density of 60%, cells were transfected with 30 pmol PIK3CA Stealth RNAi (PIK3CAHSS10800 4-6, Invitrogen) or nontargeting control siRNA (Invitrogen) using Lipofectamine RNAiMAX (Invitrogen) according to the instructions of the manufacturer. After 24 hr, cells were trypsinized, reseeded as described, and MTT assays were performed essentially as described above. To document p110α knockdown, 5 × 104 transfected cells were plated in 12-well dishes in medium supplemented with 2% FBS and cultured for 72 hr. Protein extraction and western blotting were performed as described below using antibodies against p110α (Cell Signaling Technology) and β-actin (clone AC-15, Sigma) as internal control.
For flow cytometric immunophenotyping, 1 × 106 cells were fixed on ice in ice-cold 2% paraformaldehyde for 10 min. They were then washed in PBS, collected by centrifugation, resuspended and incubated in ice-cold PBS with 0.25% Triton X-100 for 5 min on ice. After another washing step, cells were resuspended in 100 μl PBS/0.5% BSA containing an Alexa Fluor 647 labeled phospho-(Ser10)-histone H3 antibody (Cell Signaling Technology; 1:20) and a phycoerythrin-labeled cleaved-poly(ADP-ribose)-polymerase (PARP; Asp214) antibody (BD Biosciences, Heidelberg, Germany; 1:5) and incubated for 30 min at room temperature. After a further washing step, 500 μl PBS containing 10 μg/ml DAPI (Sigma, Taufkirchen, Germany) was added to stain DNA, and cells were incubated for an additional 30 min at room temperature. Analysis was performed using a three-laser LSRII analytical flow cytometer (BD Biosciences). Data were analyzed using Flowjo (Tree Star, OR, USA) analysis software. At least 30.000 events were recorded per experiment. Only single cells were included in the analysis. Each experiment was carried out at least in duplicate.
Cell transfection and expression vectors
The constitutively active myristoylated AKT construct (myr-AKT) and its control empty vector (pUSE) were obtained from Upstate (Lake Placid, NY). CME-1 cells were cultured in 25 cm2 flasks (Greiner) in a volume of 4 ml RPMI 1640-containing 10% FBS. At a density of 50%, cells were transfected with 1.25 μg of the myr-AKT or the pUSE control plasmid together with 1.25 μg of the pmaxGFP vector (Amaxa, Cologne, Germany) using lipofectamine (Invitrogen) according to the manufacturer's instructions. After 24 hr, cells were trypsinized and seeded into 75 cm2 flasks (Greiner). After another 24 hr, 5 μM LY294002 or DMSO as control were added for 24 hr. Analysis was performed by flow cytometry essentially as described above including only transfected, strongly GFP-fluorescence positive cells in the analysis.
Analysis of apoptosis by 4′,6-diamidino-2-phenylindole staining
Cells were treated as described above. They were harvested and washed in PBS, fixed in 3.7% paraformaldehyde for 10 min at room temperature and washed again. After incubation with 1 μg/ml 4′,6-diamidino-2-phenylindole (Sigma) for 10 min and two further washing steps cells were mounted on appropriate slides using Fluoromount-G medium (Southern Biotechnologies Associates, Birmingham, AL). Nuclei were visualized and photographed using a Leica DMLB fluorescence microscope. Apoptotic cells were morphologically defined by chromatin condensation and fragmentation. For each assay at least 300 cells were analyzed in triplicate.
Western blot analysis
Cells were cultured as described above in six-well dishes (Nunc) in a volume of 2 ml for 18 hr before treatment. Four hours before treatment, cells were washed twice with TBS and medium was replaced by serum-free medium. Different doses of LY294002 were added immediately after the application of 200 ng/ml recombinant human IGF-II. Incubation was performed for 5 min (AKT) or 30 min (GSK-3β, mTOR). For 1273/99 and CME-1, experiments were additionally performed under serum-supplemented conditions (2% FBS, 10% FBS) without addition of IGF-II. Cell lysis and Western blots were performed as described before.29 Filters were incubated with p-(Ser473)-AKT, p-(Ser9)-GSK3β, p-(Ser2448)-mTOR, AKT, GSK3β and mTOR antibodies (Cell Signaling Technology) according to the instructions of the manufacturer. Secondary antibody labeling and filter development were performed using the ECL kit (Amersham, Buckinghamshire, UK) as described before. Densitometric analysis of the western blots was performed using the ImageJ software (http://rsb.info.nih.gov/ij).
Synovial sarcomas display significant expression of p-(Ser473)-AKT, p-(Ser9)-GSK-3β and p-(Ser2448)-mTOR
Immunohistochemistry revealed strong expression of p-AKT in 54% (19/35) of the cases. Four cases showed moderate, 12 cases displayed weak expression of p-AKT. p-GSK-3β was strongly expressed in 41% (14/34) of the cases, 9 tumors showed moderate and 11 cases displayed weak p-GSK-3β expression levels. p-mTOR was strongly expressed in 45% (15/33 cases), 10 cases moderately and 7 cases weakly expressed p-mTOR, 1 case did not show any mTOR expression. In 8 cases, simultaneous strong expression of all three phosphorylated proteins was detected. Thirty-three percent of the cases showed strong expression of Cyclin D1 protein was found in 33% of the tumors; in this subgroup all cases but one simultaneously displayed strong p-AKT expression. Weak Cyclin D1 expression was found in 38% of the cases, 7 cases did not show Cyclin D1 expression. Seventy percent of the cases displayed weak nuclear p27KIP1, 27% of the cases showed strong p27KIP1 expression, one case was negative. There was an overlap of five cases between the strongly Cyclin D1 and p27KIP1 expressing subgroups. Regarding all markers analyzed, no significant difference was observed between monophasic and biphasic histology or the different SS18/SSX translocations (Fig. 1).
The PI3KCA gene does not harbor mutations in synovial sarcomas
Twenty-three cases of synovial sarcoma were included in a mutational analysis of the previously described hot spot regions of the PI3KCA gene.20 No mutations were detected in these regions.
PI3K/AKT signals promote growth of synovial sarcoma cells
The synovial sarcoma cell lines CME-1, 1273/99, SYO-1 and Fuji were cultured with increasing concentrations of the PI3K inhibitor LY294002. All cell lines showed a significant dose-dependent growth inhibition on treatments with LY294002. Under serum-reduced growth conditions (compared to growth in the presence of full serum supplement), maximal cellular growth was decreased and the response to specific doses of LY294002 was increased, documenting dependence of the cells on external growth factor stimulation. The 50% (GI50) growth inhibition levels under reduced serum supplement were calculated using Prism Version 4.0a for Macintosh. GI50 was 4.1 (CME-1), 4.5 (Fuji), 28.0 (1273/99) and 12.5 μM (SYO-1; Fig. 2).
Inhibition of PI3K signaling in synovial sarcoma cells lines is associated with reduced phosphorylation of AKT, GSK-3β and mTOR
To elucidate the role of downstream signaling pathways affected by PI3K, IGF-II stimulated synovial sarcoma cells were treated with different concentrations of LY294002 (3–50 μM) and analyzed for phosphorylation of AKT (Ser473), GSK-3β (Ser9) and mTOR (Ser2448). IGF-II-induced phosphorylation of AKT and GSK-3β ranged from 1.61- to 5.07-fold in all cell lines, and a dose-dependent decrease of phospho-protein levels on treatment with LY294002 was observed. Phosphorylation of mTOR was induced by IGF-II in 1273/99 and SYO-1 (1.75- and 1.45-fold) but unchanged in Fuji and CME-1. A dose-dependent decrease of phospho-mTOR levels on treatment with LY294002 was detected in all cell lines (Fig. 3, Supporting Information Table 1). Comparable results were obtained for experiments performed under serum-supplemented growth conditions without addition of IGF-II. Under reduced serum levels, treatments were associated with an enhanced decrease of phospho-protein levels (data not shown).
Immunocytochemical stainings of Fuji cells treated with different doses of LY294002 for 24 hr revealed a dose-dependent reduction of nuclear Cyclin D1 (5 μM: 54.1 ± 5.1% of control, t-test p < 0.001; 10 μM: 19.6 ± 1.4% of control, t-test p < 0.001) and an increase of nuclear p27KIP1 protein levels (5 μM: 165.3 ± 7.7 % of control, t-test p < 0.001; 10 μM: 308.2 ± 6.9% of control, t-test p < 0.001; Fig. 4c). Comparable results were obtained for CME-1 cells (data not shown).
LY294002-dependent growth reduction in synovial sarcomas involves the induction of apoptosis and the reduction of proliferation
To analyze if the effect of LY294002 on synovial sarcoma cells was predominantly proapoptotic or antiproliferative, a flow cytometric analysis of cleaved PARP (Asp214) as an indicator of apoptosis and phospho-(Ser10)-Histone H3 (pHH3), a marker of mitosis, was performed in CME-1 and SYO-1 cells. In both cell lines, LY294002-induced growth reduction was significantly associated with an increase of the cell population positive for cleaved PARP (CME-1: 5 μM: 156% of control, t-test, p < 0.01; 15 μM: 318% of control, t-test, p < 0.001. SYO-1: 5 μM: not significantly different compared to control; 15 μM: 267 % of control, t-test, p < 0.001; Fig. 4a, Supporting Information Table 2). With increasing dosages, CME-1 cells additionally showed a significant decrease of the pHH3-positive subpopulation (5 μM: 92% of control, t-test, p < 0.01; 15 μM: 86% of control, t-test, p < 0.01), whereas mitotic activity was not significantly affected in SYO-1 (Supporting Information Table 2). Correspondingly, microscopic analysis of DAPI-stained cells showed a significant increase of nuclei displaying chromatin fragmentation or condensation in all four cell lines (Fig. 4b).
RNAi-mediated PI3KCA knockdown mimics and constitutive AKT activation counteracts LY294002-effected PI3K inhibition in synovial sarcoma cells
As LY294002 represents a first-generation broad-spectrum PI3K inhibitor, which also affects other PI3K family members, e.g., mTOR,30 additional experiments were performed to provide evidence for the specifity of the effects. Aiming at a methodically independent confirmation of our observations, 1273/99 and CME-1 synovial sarcoma cells were transfected with PI3KCA siRNAs. Cell growth as measured by MTT assays was reduced to 66.4 ± 0.56% (1273/99) and 69 ± 1.1% (CME-1) compared to the mock-transfected controls, and downregulation of p110α protein was documented in western blots (Fig. 5a). To find out if an activated AKT signal is able to counteract the inhibitory effects of LY294002, we overexpressed myr-AKT in CME-1 cells. In the control experiment using the empty pUSE vector, treatment with 5 μM LY294002 led to an increase of the fraction positive for cleaved PARP by 43.7 ± 0.01%, whereas in myr-AKT expressing CME-1 cells, the increase of the apoptotic rate due to LY294002 was significantly lower (3.1 ± 0.01%, t-test: p < 0.01; Fig. 5b). The decrease of the mitotic phospho(Ser10)-Histone H3 positive fraction was similar in pUSE- and myr-AKT transfected cells (pUSE: 24.3 ± 0.02%; myr-AKT 21.4 ± 0.01%).
Inhibition of PI3K signaling sensitizes synovial sarcoma cells for chemotherapeutic treatment
CME-1 and SYO-1 synovial sarcoma cells were coincubated with increasing doses of doxorubicin (DOXO), vincristine (VCR) and actinomycin D (ACT-D), chemotherapeutic drugs used in current therapeutic protocols in the treatment of synovial sarcoma and doses of the PI3K-Inhibitor LY294002 leading to ∼30% growth reduction. Fractional product method analysis characterized the interaction of PI3K inhibitior and chemotherapeutic drugs as additive in both cell lines (Fig. 6).
Aberrant activation of the PI3K/AKT signaling pathway has been described in many different tumors including musculoskeletal sarcomas occurring preferentially in childhood and adolescence, i.e., rhabdomyosarcoma, osteosarcoma and Ewing's sarcoma.16, 31–33 Activation of PI3K/AKT signaling in this group of tumors appears to be mainly due to autocrine or paracrine loops of growth factors and their corresponding receptors, whereas mutations in effectors of the pathway are rare. One of the main contributors to PI3K/AKT signals in Ewing's sarcoma and rhabdomyosarcoma is IGF-IR whose ligands have been reported to be transcriptionally activated by the respective specific fusion oncogenes.34–36 Consistently, recent preclinical therapeutic approaches targeting the IGF-IR yielded promising results.31, 32, 37–39
Although only little data on the role of PI3K/AKT signaling in synovial sarcomas exists, it has recently been shown that, similar to the situation in pediatric sarcomas, the oncoproteins encoded by the different t(X;18) fusion transcripts characteristic for synovial sarcomas transcriptionally activate the IGF2 gene.3, 4 Own experimental observations show that this transcriptional induction is predominantly mediated through IGF2 promotors P2 and P4 (M. Trautmann & W. Hartmann, unpublished data). IGF-II protein is a ligand for the receptor tyrosine kinase IGF-IR, which transduces its signal mainly via the PI3K/AKT signaling pathway. In vitro, synovial sarcoma cells were highly sensitive to a small molecular weight kinase inhibitor of the IGF-IR, and the inhibitory effect on cellular proliferation and survival was associated with a reduction of AKT phosphorylation.12 However, though the IGF-IR obviously plays an important role in synovial sarcomas, it may not be as outstanding as in Ewing's sarcoma and rhabdomyosarcoma because several further growth factor receptors signaling via the PI3K/AKT axis have been shown to be expressed at relevant levels.
To get an insight into the activation status of the PI3K/AKT pathway in synovial sarcomas, immunohistochemical stainings of activated PI3K-downstream targets AKT, mTOR and GSK-3β as well as the cell cycle regulators Cyclin D1 and p27KIP1 were performed in a set of synovial sarcomas. mTOR represents a key regulator of cellular growth and proliferation acting through the control of translation of essential cellular regulator proteins as Cyclin D1 and p27KIP1.17, 40 Cyclin D1 and p27KIP1 turnover is regulated by GSK-3β, which, itself, is subject to an inhibitory phosphorylation step by AKT.14, 18 Our immunohistochemical data document strong expression of phosphorylated AKT, GSK-3β and mTOR in a large group of synovial sarcomas, indicating an activation of the pathway, which is in good agreement with data published previously.41, 42 The presence of subsets of tumors expressing Cyclin D1 and p27KIP1 at different levels imply that therapeutic targeting of PI3K/AKT signaling might be successful through interference with the Cyclin/CDK inhibitor network—aiming at a shift in the balance of pro- and anti-proliferative signals. Clinical data showing that low p27KIP1 expression correlates with a poor outcome of synovial sarcoma patients support this concept.43
In the light of reports of activating mutations of the PIK3CA gene in different types of cancer, we wondered if PIK3CA gene mutations might be responsible for the activation of the pathway seen. We screened 23 synovial sarcoma samples for mutations in the hot spot regions (exons 1, 9 and 20) of the PI3KCA gene,20 but no mutations were detected. This result is in good agreement with the very recent finding of Barretina et al.44 who reported one single PIK3CA mutation in 23 synovial sarcomas. The low frequency of activating PI3KCA mutations or inactivating mutations in PTEN44, 45 in synovial sarcomas argues in favor of the concept that in these tumors autocrine and paracrine circuits (which are partly driven by the t(X;18)-oncoproteins) are mainly responsible for PI3K/AKT signaling activation. As relevant expression levels have been documented for the EGF receptor, the HER2 receptor, the PDGF receptor beta, the stem cell factor receptor c-KIT, the hepatocyte growth factor receptor c-MET, members of the FGF receptor family and the IGF-IR (all signaling via the PI3K/AKT cascade) it appears probable that PI3K/AKT signals result from the integration of inputs from several different receptors.6–11 In terms of therapeutic applicability, it would therefore be reasonable to target the pathway at a central point downstream of the receptors, one possible target being PI3K itself.
To get an insight into the biological relevance of PI3K/AKT signaling in synovial sarcomas, we performed in vitro experiments targeting PI3K enzymatic activity. In all cell lines, a significant and dose-dependent inhibition of cellular growth was observed upon treatment with LY294002 (Fig. 2) and dose-dependent decreases in phosphorylation of AKT-, GSK-3β- and mTOR documented specific mechanisms in LY294002-mediated PI3K inhibition on the protein level (Fig. 3). Cells grown under serum-reduced conditions were more sensitive to specific doses of LY294002, which was paralleled by an enhanced decrease of phospho-protein levels, implying dependence of the cells on external growth factor stimulation. Reduced cellular growth was also discernible in independent experiments employing siRNA-mediated knockdown of the PI3K p110α subunit, a finding, which indirectly implies specificity of the pharmacological effect seen (Fig. 5a).
We wanted to know if it was rather cell death or cellular proliferation, which led to decreased signals in MTT assays. Flow cytometric analysis of cleaved PARP as a marker of apoptosis revealed a substantial dose-dependent increase of apoptotic cell death upon treatments with LY294002 in both cell lines tested; corresponding observations were made in the morphologic analysis of DAPI-stained nuclei in all cell lines. CME-1 cells additionally displayed a dose-dependent decrease of the mitotic, p-(Ser10)-histone H3-positive cell fraction, which was accompanied by decreased levels of Cyclin D1 and increased nuclear levels of p27KIP1 as detected by immunocytochemistry. The proapoptotic and antimitogenic effects of a reduced dose of LY294002 in CME-1 cells were largely reversible by overexpression of constitutively activated AKT, representing another indicator of specificity of the pharmacological effect of LY294002 (Fig. 5b).
The changes seen in Cyclin D1 and p27KIP1 protein levels certainly contribute to the growth-reducing and antimitogenic effect of PI3K inhibition. In a previous study, Xie et al. reported regulation of Cyclin D1 protein levels by the SS18/SSX fusion product, mainly through inhibition of its degradation, independently from PI3K signaling. Antisense interference with the SS18/SSX fusion gene led to diminished cellular growth rates, which were associated with a decrease in Cyclin D1 levels.23 According to our data, Cyclin D1 protein levels are highly responsive to PI3K/AKT/GSK-3β pathway inhibition, which is of particular interest since Cyclin D1 obviously represents an integration point of genuine fusion protein-driven signals, inputs from the wnt/β-catenin pathway and PI3K/AKT signals.46, 47 Possibly, as reported for childhood brain tumors, PI3K/AKT signals are essential for the maintenance of other pathways' inputs, including the wnt/β-catenin pathway, which makes it an outstanding therapeutic target structure.48
However, as discernible from the flow cytometric data, the common trait in both synovial sarcoma cell lines tested is the induction of apoptosis. There are several pathways linking PI3K/AKT signaling to antiapoptosis. One important interaction is the promotion of cell survival through phosphorylation of the proapoptotic protein Bcl-2-Antagonist of Cell Death (BAD) by AKT, thereby preventing its inhibitory binding to Bcl-2.49 Importantly, Bcl-2 has been shown to be of particular significance in synovial sarcomas.50
We finally wanted to get an insight into a possible role of PI3K inhibition in the treatment of synovial sarcomas. PI3K-inhibitors have been shown to increase the efficacy of conventional chemotherapeutic drugs in other tumor cells.32 To elucidate a potential therapeutic role of PI3K-inhibiton in future multimodal treatments of synovial sarcomas, we coincubated conventional cytotoxic drugs currently used in synovial sarcoma therapy with LY294002 and observed additive effects on cellular proliferation. This finding points to the option to use PI3K inhibitors in innovative therapeutic approaches, combining conventional cytotoxic drugs with specific inhibitory substances targeting an activated signaling pathway. Multimodality defined in this manner coupled with the possibility to reduce the doses of the individual compounds could help minimizing the toxicity of the individual compounds. As PI3K/AKT activation, at least in part, appears to be due to the action of the t(X;18)-oncoproteins, this approach represents a specific molecularly based therapeutic strategy for synovial sarcomas.
In summary, our data show that PI3K/AKT signaling is commonly activated in synovial sarcomas. Targeting PI3K in vitro results in increased apoptosis and inhibition of cellular proliferation of synovial sarcoma cells, and combination of PI3K inhibitors with conventional chemotherapeutic drugs results in additive effects. These findings argue in favor of PI3K as a potential therapeutic target in synovial sarcomas.
This study was supported by the Deutsche Krebshilfe (KoSar-Sarcoma Net) (WH, GM, PS, EW, RB) the BONFOR program of the Medical Faculty (WH, NF), University of Bonn, and grant HBFG-109-517 to the Flow Cytometry Core Facility at the Institute of Molecular Medicine, University of Bonn. No potential conflicts of interest were disclosed.