1. Introduction
Synovial sarcoma is a rare mesenchymal tumour type that accounts for 8–10% of all soft tissue sarcomas [
1]. As with many other sarcoma subtypes, synovial sarcoma can occur in any anatomical site, but is most commonly found in the extremities [
2]. It typically arises in adolescents and young adults although can affect patients of any age [
3,
4]. Synovial sarcoma is characterised by a pathognomonic translocation between chromosome X and 18 (X:18), which results in the expression of fusion proteins including SS18-SSX1, SS18-SSX2, and SS18-SSX4 [
5,
6,
7]. Extensive work has shown that these fusion proteins play important roles in driving sarcomagenesis, such as regulating the biology of the SWI/SNF chromatin remodelling complex [
8,
9,
10]. Patients with synovial sarcoma have poor outcomes with a 5-year survival of 50–60% [
4,
11,
12] and have limited treatment options in the advanced/metastatic setting, including anthracyclines, ifosfamide, trabectedin, and pazopanib [
13].
In order to identify new therapeutic strategies and gain a better understanding of the biology of this disease, it is necessary to develop preclinical models of synovial sarcoma. Given its rarity, there are very few synovial sarcoma cell line models available in public repositories, with only five identified in the most recent census of sarcoma cell line models [
14]. Established cancer cell lines have been cultured on plastic for decades and due to genomic and phenotypic drift during serial passaging have diverged from the tumours for which they were derived [
15]. This has contributed to poor reproducibility in preclinical findings which is likely to have played a role in the high failure rate of translating therapeutic discoveries into oncology clinical trials and regulatory approvals [
16,
17].
To bridge this gap, patient-derived xenografts (PDX), where tumours obtained from patients are serially passaged in mice, have emerged as a valuable tool for preclinical research as they retain much of the molecular and histopathological features of the original human tumour [
18,
19]. Several sarcoma PDX collections have been described encompassing multiple subtypes including synovial sarcoma [
20,
21,
22]. However, PDX models have certain limitations including high costs required for animal maintenance, slow tumour growth, and variable engraftment rate, which collectively diminish the feasibility of undertaking large-scale genetic and drug screens. These limitations can be overcome by the establishment of matched PDX-derived cell lines which retain many of the genomic features of the PDX tumours, are easier to grow and manipulate in culture, and are amenable to high-throughput screens [
18]. The development of matched PDX-derived cell lines is non-trivial due to the relatively high failure rate associated with establishing cell lines from tumour specimens [
23,
24] and there are only a few reported studies demonstrating success using this approach, including in breast, colorectal, and pancreatic cancers [
21,
25,
26,
27]. In sarcomas, matched PDX-derived lines have been established in osteosarcoma, Ewing’s sarcoma, clear cell sarcoma, and
CIC::DUX4 sarcoma [
23,
28,
29,
30].
In this study, we established a novel synovial sarcoma cell line, ICR-SS-1, which was derived from a PDX model from The Jackson Laboratory biorepository. We have further undertaken a comparative analysis of the drug response profiles of ICR-SS-1 versus two established commercially available synovial sarcoma cell lines (SYO-1 and HS-SY-II). To our knowledge, this is the first study to report a matched PDX-derived cell line for synovial sarcoma.
2. Materials and Methods
2.1. Patient-Derived Xenograft Model
ICR-SS-1 was established from a publicly available PDX model (J000104314) deposited in The Jackson Laboratory biorepository (The Jackson Laboratory,
http://tumor.informatics.jax.org/mtbwi/pdxDetails.do?modelID=J000104314, accessed on 17 June 2022). This PDX was derived from a tumour obtained from a 21-year-old male diagnosed with a grade 3 metastatic synovial sarcoma. This PDX model has been shown by The Jackson Laboratory to harbour the
SS18::SSX1 fusion gene. PDX tumours were serially passaged in NOD scid gamma (NSG) mice and tumour volume calculated by ½ × length × width
2 (
Figure 1A). Histology showed a hypercellular neoplasm composed of sheets of uniform spindle and ovoid cells, with nuclear overlapping, scanty cytoplasm, and minimal surrounding stroma. There was no discernible pleomorphism. This is in keeping with monophasic synovial sarcoma (
Figure 1B). These tumours maintain the same histological features as the histology images deposited in The Jackson Laboratory biorepository database.
2.2. PDX Dissociation
To generate a cell suspension from the xenograft tumour, the tissue was minced and digested for 2 h at 37 °C in DMEM/Ham’s F12 1:1 with 15 mM HEPES, 0.1× insulin-transferrin selenium A (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 1× penicillin/streptomycin, 10 ng/mL EGF (Peprotech, London, UK), 10 µg/mL hydrocortisone (Sigma Aldrich, St. Louis, MO, USA), 0.5 mg/mL collagenase (Sigma Aldrich), 0.1 mg/mL hyaluronidase (Sigma Aldrich), 100 units/mL DNase I (Sigma Aldrich), 10 µM Y-27632 (LC Laboratories, Woburn, MA, USA), and 5% FBS (Gibco). Red blood cells were lysed using RBC lysis buffer (Invitrogen, Waltham, MA, USA) and remaining cells were incubated with 0.05% trypsin-EDTA (Gibco) at 37 °C. After trypsinisation, cells were treated with 1 mg/mL DNase I (Sigma Aldrich) at 37 °C, before passing through a 70 µm strainer. Mouse cell depletion beads (Miltenyi Biotec, Surrey, UK) were used to remove contaminating murine cells.
2.3. Cell Culture
Dissociated and mouse cell depleted PDX tumour cells were cultured in DMEM/Ham’s F12 1:1 with 15 mM HEPES, 1× penicillin/streptomycin, 2.4 mM L-glutamine, 5 µM Y-27632 (LC Laboratories), 5 µg/mL insulin (Sigma Aldrich), 400 ng/mL hydrocortisone (Sigma Aldrich), 10 ng/mL EGF (Peprotech), 250 ng/mL amphotericin B (Thermo Fisher Scientific), 9.62 ng/mL cholera toxin (Sigma Aldrich), and 10% FBS (Gibco). Following successful continuous growth for >10 passages in culture, the cell line was designated ICR-SS-1. HS-SY-II (from RIKEN BioResource Centre, Kyoto, Japan), SYO-1 (obtained from Dr Chris Lord, Institute of Cancer Research, London, UK), and NIH-3T3 (obtained from Dr Matilda Katan, University College London, London, UK) cells were cultured in DMEM supplemented with 1× penicillin/streptomycin and 10% FBS (Gibco). SK-UT-1 cells (obtained from Dr Priya Chudasama, German Cancer Research Centre, Heidelberg, Germany) were cultured in MEM supplemented with 1× penicillin/streptomycin and 10% FBS (Gibco). Cells were grown at 37 °C with 5% CO2. Medium was replenished twice weekly.
For spheroid formation, ICR-SS-1 cells (1000/well) were seeded into 96-well, round-bottom, ultra-low attachment plates (Corning). Plates were spun at 1000× g and grown at 37 °C with 5% CO2 for 3 days before imaging.
2.4. Cell Proliferation
ICR-SS-1 cells (1000/well) were seeded into 96-well, flat-bottom, black-walled plates (Greiner Bio-One, Frickenhausen, Germany). Plates were fixed in 10% neutral-buffered formalin solution (Sigma Aldrich) at respective timepoints and stained with 5 µg/mL Hoechst 33342 (Tocris, Tocris Bioscience, Bristol, UK). Plates were scanned and cells were counted using a Celigo imaging cytometer. Cell counts were normalised to day 1 and an exponential growth equation was fitted to the data using GraphPad Prism (GraphPad, v8.2.1) in order to determine doubling time.
2.5. SS18::SSX Fusion PCR
RNA was extracted from J000104314 tissue and ICR-SS-1 cells using RNeasy mini and QIAshredder kits (Qiagen, Germantown, MD, USA), following the manufacturer’s instructions, and used to generate cDNA via a Superscript III kit (Invitrogen). The HS-SY-II cell line was used as a positive control. PCR mixtures were made containing the common
SS18 forward primer and one of
SSX1,
SSX2, or
SSX4 reverse primers.
ACTB was used as a loading control. Primer sequences are provided in
Table 1 below.
Thermocycler conditions used for
SS18::SSX or
ACTB amplification are presented in
Table 2 and
Table 3 below. PCR products were run on 2% agarose gels with SYBR safe DNA gel stain (Invitrogel, Invitrogen) or ethidium bromide (Sigma Aldrich) and visualised using UV illumination.
2.6. Human and Mouse PTGER2 PCR
DNA was extracted from ICR-SS-1, NIH-3T3, and SK-UT-1 cells using a DNeasy blood and tissue kit (Qiagen), following the manufacturer’s instructions. NIH-3T3 was used as a mouse positive control and SK-UT-1 was used as a human positive control. PCR mixtures were made containing either a human- or mouse-specific
PTGER2 forward primer and a common reverse
PTGER2 primer. Primer sequences are provided in
Table 4 below.
Thermocycler conditions used for human or mouse
PTGER2 amplification are presented in
Table 5 below. PCR products were run on 1.5% agarose gels with SYBR safe DNA gel stain (Invitrogel) and visualised using UV illumination.
2.7. Short Tandem Repeat (STR) Analysis
Genomic DNA was harvested from J000104314 PDX tissue and ICR-SS-1 cells using a DNeasy blood and tissue kit (Qiagen), following the manufacturer’s instructions. DNA was quantified using a Qubit dsDNA HS assay kit, and STR profiles were then analysed via Eurofins cell line authentication service.
2.8. Proteomic Analysis
ICR-SS-1, HS-SY-II, and SYO-1 cells were seeded in T25 culture flasks and incubated for 72 h in order to reach 80% confluence. Cells were then lysed in 8 M urea and 0.1 M ammonium bicarbonate (ABC), and protein concentration was measured by bicinchoninic acid (BCA) assay. For each cell line, 40 ug of total protein was reduced with 10 mM dithiothreitol at 56 °C for 40 min and alkylated with 55 mM iodoacetamide at 25 °C for 30 min in the dark. After dilution to final concentration of 2 M urea and 0.1 M ABC, each sample was digested with 0.4 g of trypsin (Thermo Scientific) at 37 °C overnight. The resulting digest was acidified to pH < 4 by trifluoroacetic acid (TFA), desalted on Pierce C18 Spin Columns (Thermo Scientific) according to the manufacturer’s protocol, and dried in a SpeedVac.
Dried samples were resuspended in mobile phase A (2% acetonitrile, 0.1% formic acid), spiked with iRT peptides (Biognosys AG, Schlieren, Switzerland), and 2 ug of total peptides were loaded onto a 2 cm × 0.1 mm trap column self-packed with ReproSil Pur C18AQ (120 Å, 10 µm) beads. Sequential window acquisition of all theoretical mass spectra (SWATH)-mass spectrometry data were acquired using the same instrument parameters as previously described in Milighetti et al. [
31]. The acquired data were integrated with a previously published dataset of synovial sarcoma, leiomyosarcoma, undifferentiated pleomorphic sarcoma, and dedifferentiated liposarcoma FFPE tissue samples from Milighetti et al. [
31], and all data were analysed by DIA-NN (v1.8) software (accessed on 17 June 2022) [
32] using a publicly available pan-human library [
33]. The default settings with “match between runs” and “unrelated runs” was used for data processing. Oxidation of methionine, carbamidomethylation of cysteines, and N-term methionine excision were included as possible amino acid modifications with a maximum number of 5 modifications per peptide sequence. Quantified protein data were log2 transformed and quantile normalised in R using the proBatch package [
34] and further processed in Perseus [
35]. Proteins with >75% values in at least one sarcoma subtype or across all cell lines were retained. Missing values were imputed using the “replace missing values from normal distribution” tool in Perseus using default settings. The imputed dataset was then median-centred across all samples and visualised using two-way unsupervised clustering based on Pearson’s correlation coefficient.
Proteomic profiles of the three synovial sarcoma cell lines and FFPE tissue samples were analysed by significance analysis of microarrays (SAM) using the samR package [
36] in R. For this, the log2 transformed and quantile normalised dataset was used. Data for cell lines and tissue samples were selected, technical replicates averaged, and proteins with more than 30% of missing values were removed. The resulting dataset was processed by samR using a delta score threshold of 0.77 to reach 1% FDR. Lists of positively and negatively regulated proteins were then separately subjected to over-representation analysis using the online tool g:Profiler [
37] with the g:GOSt module for functional profiling (g:Profiler,
https://biit.cs.ut.ee/gprofiler/gost, accessed on 17 June 2022) and the following setup: full list of identified proteins used as a background; Benjamini–Hochberg FDR method with 0.1 FDR threshold; GOBP and Hallmark gene set databases downloaded from MSigDB (Molecular Signatures Database, v7.5.1,
http://www.gsea-msigdb.org/gsea/msigdb, accessed on 17 June 2022) [
38,
39].
Proteomic profiles of ICR-SS-1 and the other two synovial sarcoma cell lines were compared by a two-tailed
t-test in Perseus. For this analysis, the log2 transformed, quantile normalised dataset was used. Technical replicates for individual cell lines were kept as separate samples and proteins with no valid values across the three cell lines were removed. A permutation-based FDR threshold of 0.05 and artificial within-group variance of 0.1 were applied in Perseus to identify significant differentially regulated proteins [
36]. Lists of identified up- and down-regulated proteins were further analysed by g:Profiler as described above.
2.9. Drug Screen and Dose Response Assays
ICR-SS-1 (2000/well), HS-SY-II (3000/well), and SYO-1 (2000/well) cells were seeded in clear 96-well, flat-bottom plates (Corning Inc., Corning, NY, USA). Plates were incubated for 24 h before replacing media with a panel of small molecule inhibitors at a concentration of 500 nM for all drugs except NVP-AUY922, which was at a concentration of 50 nM (details and source of inhibitors are shown in
Supplementary Table S1). After 72 h, cell viability was determined using CellTitre-Glo (Promega, Madison, WI, USA), following the manufacturer’s instructions. Dose response assays were conducted by seeding ICR-SS-1 (2000/well), HS-SY-II (3000/well), and SYO-1 (2000/well) cells in clear 96-well, flat-bottom plates (Corning). Plates were incubated for 24 h, after which the medium was replaced with increasing concentrations of doxorubicin hydrochloride (Sigma Aldrich) or pazopanib (LC Laboratories) at the indicated dose. Data points from dose response assays were used to fit a four-point non-linear regression curve via Graphpad Prism and the drug screen data were subjected to hierarchical clustering using Perseus software [
35] with Euclidean distance as the distance metric.
4. Discussion
Synovial sarcoma is a rare cancer type with poor outcomes in the advanced setting despite multidisciplinary clinical management. There is a lack of effective systemic therapies for these patients and therefore an urgent need to develop new approaches to tackle this disease. Key to the identification of novel agents is the availability of well-characterised preclinical cell line models. To date, only five synovial sarcoma cell lines are available in public biorepositories [
14], all of which have been cultured for decades. Here, we present and characterise a new synovial sarcoma cell line ICR-SS-1, which has been established from a PDX model held in The Jackson Laboratory biorepository. This cell line retains the
SS18::SSX1 fusion gene and faithfully recapitulates the molecular features of human synovial sarcoma tumours as shown by mass spectrometry analysis.
Despite being driven by a single translocation (the
SS18::SSX fusion gene), it is well-established that in synovial sarcoma patients there is a wide heterogeneity in observed clinical responses to systemic therapies. For instance, only a subset of advanced synovial sarcoma patients benefits from treatment with chemotherapeutic agents such as doxorubicin and trabectedin [
44,
45,
46]. The mechanistic basis for this heterogeneity is currently unknown and there are no predictive biomarkers available for stratification in order to rationally target the right drugs to the appropriate patient population. Notably, our data show that the ICR-SS-1 cell line is significantly more resistant to doxorubicin compared to two established synovial cell lines SYO-1 and HS-SY-II. Our proteomic analysis finds that when compared to the established cell lines, the ICR-SS-1 line shows significantly upregulated expression of proteins involved in EMT. In line with this observation, previous studies have shown that the induction of EMT drives doxorubicin resistance in multiple cancer types [
47]. This suggests that ICR-SS-1 could serve as a useful model to study the biology of anthracycline-resistant synovial sarcoma and identify new salvage therapies, such as those that target the EMT pathway, following the failure of doxorubicin treatment.
By subjecting ICR-SS-1 to a targeted small molecule inhibitor panel and comparing the drug responses to the two other synovial sarcoma cell lines, we identified agents from three different drug target classes which are effective in all three cell lines. Of these, both SYO-1 and HS-SY-II have previously been shown to be sensitive to PLK-1 inhibition [
48] while BET bromodomain inhibitors [
49]. In addition, we show that the dual PI3K-mTOR inhibitor NVP-BEZ235 (also known as dactolisib) reduces the cell viability of ICR-SS-1 and the two other synovial sarcoma cell lines. Our data are consistent with previous reports of short-term patient-derived sarcoma modelling studies which demonstrate that tumour cells from synovial sarcoma patients are sensitive to drugs that block this pathway [
50,
51]. Taken together, our data suggest that targeting the PI3K-mTOR pathway may have utility particularly in the context of chemotherapy-resistant synovial sarcoma.