Establishment of a novel cellular model for myxofibrosarcoma heterogeneity

Human cancers frequently display substantial intra-tumoural heterogeneity in virtually all distinguishable phenotypic features, such as cellular morphology, gene expression, and metastatic potential. In order to investigate tumour heterogeneity in myxofibrosarcoma, we established a novel myxofibrosarcoma cell line with two well defined sub-clones named MUG-Myx2a and MUG-Myx2b. The parental tumour tissue and both MUG-Myx2 cell lines showed the same STR profile. The fact that MUG-Myx2a showed higher proliferation activity, faster migration and enhanced tumourigenicity was of particular interest. NGS mutation analysis revealed corresponding mutations in the FGFR3, KIT, KDR and TP53 genes. In contrast, the MUG-Myx2a cell lines showed an additional PTEN mutation. Analysis of CNV uncovered a highly aberrant karyotype with frequent losses and gains in the tumour sample. The two MUG-Myx2 cell lines share several CNV features of the tumour tissue, while some CNVs are present only in the two cell lines. Furthermore, certain CNV gains and losses that are exclusive to either MUG-Myx2a or MUG-Myx2b, distinguish the two cell lines. As it is currently not possible to purchase two different sarcoma cell lines derived from the same patient, the novel myxofibrosarcoma cell lines MUG-Myx2a and MUG-Myx2b will be useful tools to study pathogenesis, tumour heterogeneity and treatment options.

Myxofibrosarcoma (MFS) is the most common sarcoma in elderly patients and is characterised histologically by a multinodular growth pattern and variably prominent myxoid stroma. The tumour is mainly composed of spindle cells with variable cytologic atypia accentuated along curvilinear vessels 1 . Clinically, increasing grades and stages of the tumours are frequently seen in MFS after local recurrences, which may eventually lead to metastatic diseases. Recurrence has been shown to occur despite repeated surgery involving wide-ranging local excisions and negative surgical margins [1][2][3] . MFS is characterized by non-specific cytogenetic aberrations that increase with grade, suggesting a multistep tumour progression model due to acquired genetic instability 4 . The mechanisms responsible for the coexistence of distinct sub-clones and the biological consequences of this coexistence remain poorly understood. MALDI imaging mass spectrometry demonstrated a high grade of intra-tumour heterogeneity in MFS 5 . Despite their histological resemblance, imaging mass spectrometry demonstrated that intra-tumour heterogeneity was a consistent feature in each of the MFS tumours studied, and revealed that different nodules exhibited high grade-/low grade-like biomolecular signatures. Hence, cytogenetic heterogeneity and clonal evolution seem to be common in myxoid sarcoma 5,6 . In addition, large-scale sequencing analyses of other solid cancers have identified extensive heterogeneity both between individual tumors and genetic intra-tumour heterogeneity [7][8][9] . Genomic instability generates a high level of intercellular genetic heterogeneity and has been linked to both drug resistance and poor prognosis in cancer, because therapeutic procedures rely on single tumour biopsy samples 10-12 . In this context the availability of appropriate in vitro cell systems is of particular importance. Permanent cell lines derived from primary sarcomas offer the opportunity to study functional alterations in sarcoma biology. However, the supply of commercial sarcoma cell lines is extremely inadequat. In order to investigate this intra-tumour heterogeneity of MFS in vitro, we established two sub-clones of a primary MFS, named MUG-Myx2a and MUG-Myx2b. The new cell lines provide a model for tumour heterogeneity and allow comprehensive genetic and epigenetic characterization. Furthermore, because of their tumourigenicity, the cell lines MUG-Myx2a cells show higher proliferation, migration, and tumourigenicity. The growth behaviour of the two clones was investigated with the MTS assay after 1 to 6 days. The statistical comparability of the two clones revealed highly significant variations at all time points (day 1: p = 0.00064; day 2: p = 1.87E-06; day Scientific RepoRts | 7:44700 | DOI: 10.1038/srep44700 3: 2.82E-17; day 6: 6.23E-06; n = 12) (Fig. 1C). In order to evaluate the real-time growth pattern and migration potential, cells were detected in quadruplicate with the xCELLigence System. Using the RTCA 1.2.1 software, the population doubling time of the cells was calculated. MUG-Myx2a cells displayed a lower doubling time (20 h vs 22 h) and a higher migration potential than MUG-Myx2b cells (Fig. 1D,E). After treatment with 0-25 μ M doxorubicin, 0-25 μ M Verinostat, and 0-25 nM bortezomib for 48 h, cell viability was measured by the MTS assay. Chemotherapeutic drugs inhibited cell growth in a dose-dependent manner in both cell lines ( Fig. 2A-C). Nevertheless MUG-Myx2a cells showed a significant higher sensitivity to all three therapeutic drugs.
MUG-Myx2a (p49) and MUG-Myx2b (p52) cells successfully formed tumours in 3 of 5 transplanted mice. Small nodules were palpable 4 weeks after inoculation. After 8 weeks the animal were dissected. Tumour formation in mice was significantly higher when MUG-Myx2a cells were injected, in comparison to MUG-Myx2b (Fig. 3A). The success rate of growing MUG-Myx2a/b cells in nu/nu Foxn1 mice was 60%. Successful engraftment was determined by pathological examination of formalin-fixed, paraffin-embedded (FFPE) material from the tumour samples. Hematoxylin and eosin (HE)-stained FFPE mice tumours showed variation concerning the morphology of MUG-Myx2a (Fig. 3B) and MUG-Myx2b (Fig. 3C). In particular, MUG-Myx2b showed the characteristic prominent myxoid stroma of MFS.
The high mitotic rate and high proliferative index were confirmed by IHC with the proliferation marker Ki-67. Using the ImageScope software, Ki-67 positive and negative cells were quantified after IHC staining. MUG-Myx2a tumours from all mice displayed an increased proliferation level, compared to MUG-Myx2b tumours (Ki-67 positivity: 0.937 ± 0.033 vs 0.604 ± 0.185; n = 3). A representative staining of one pair of tumours is shown and summarized in Fig. 3D (MUG-Myx2a) and 3E (MUG-Myx2b).

Mutation Analysis by Next-Generation Sequencing (NGS). Mutation Analysis of 50 genes by NGS
revealed that the primary patient's tumour tissue and MUG-Myx2b showed identical mutations in the FGFR3, KIT, KDR, and TP53 genes. Conversely, the MUG-Myx2a cell line showed the same mutations as MUG-Myx2b and the primary patient tumour, but had an additional mutation in the PTEN gene ( Table 2).

Chromosomal Copy Number Variations (CNVs).
Low coverage whole genome sequencing was used to determine chromosomal gains and losses in the primary tumour tissue as well as in both MUG-Myx2 cell lines. The tumour tissue showed widespread gains and losses for several chromosomes (Fig. 4, outer ring). The MUG-Myx2 cell lines both showed gains for chromosomes 3,6,8,14, and 17 as well as losses of chromosomes 3, 6, 9, and 13. Additionally, both cell lines showed gains for chromosomes 1, 3, and 6 and losses of chromosomes 8 and X. These alterations were not evident in the primary tumour tissue. MUG-Myx2a (Fig. 4, middle ring) showed losses of chromosome 4 and gains of chromosome 12, but these characteristics were not shared with the primary tumour, or the MUG-Myx2b cell line. On the other hand, MUG-Myx2b (Fig. 4, innermost ring) exhibited losses of chromosome 5, which were exclusive to this cell line. In addition to these substantial CNVs, we also determined the copy number state of 50 cancer genes, which were also tested for the presence of mutations. Both MUG-Myx2 cell lines showed a similar pattern of CNVs in these genes and shared the distinct loss of CDKN2a and MLH1, as well as a gain of FGFR1, with the primary tumor tissue ( Fig. 5; Table 3). However, gain of the JAK3 and NRAS locus was only seen in the MUG-Myx2 cell lines.

Discussion
Myxofibrosarcoma is a malignant neoplasm with variably prominent myxoid stroma, cellular pleomorphism, and a distinct curvilinear vascular pattern. It represents the most common sarcoma in elderly patients, with a slight    predominance in males 1 . In order to enable the investigation of tumour heterogeneity on cellular and molecular level, it is highly desirable to establish new human primary cell lines. Clonal heterogeneity may be evident within single samples, but can also be observed between different tumour regions within the same primary site, or even between primary and metastatic sites, so-called "regional heterogeneity" 13,14 .    and migration ability than MUG-Myx2b. Additionally, compared to MUG-Myx2b, MUG-Myx2a cells showed a higher sensitivity to commonly used chemotherapeutic drugs. Both cell lines successfully formed tumours in nu/nu Foxn1 mice, although the MUG-Myx2a cell line again showed higher proliferation capacity and thereby confirmed the growth analysis data. A nu/nu Foxn1 mouse model demonstrated, that both cell lines showed equally aggressive in vivo growth behaviour. Using an alternative mouse model, such asNSG mice, might show increased tumourigenicity.
A distinctive feature of cancer cells compared to normal cells is chromosome instability, which is proposed to be critical for the initiation of tumourigenesis 15 . For this reason, we investigated the genomic integrity of the MUG-Myx2 cell lines by low density whole genome sequencing. Gains in gene copy number can drive the expression of oncogenes, whereas decreased gene dosage by hemizygous and/or homozygous deletion may inactivate tumour suppressor genes 16 . The MUG-Myx2 cell lines match some, but not all chromosomal aberrations of the primary tumour. Interestingly, many features seem to be more pronounced in the cell lines, consistant with the notion, that the primary tumour contains several sub-clonal populations of cells which cumulatively contribute to the CNV phenotype of the primary tumour. However, the two cell lines might represent single sub-clones which share many features with the primary tumour, but have also accrued additional changes. In the CNV analysis of individual genes, the copy number state of the CDKN2a, MLH1, and FGFR1 genes is identical in the primary tumour and the cell lines. CDKN2a, also known as p16, has been described as frequently deleted in osteosarcomas 17 whereas we found changes which are present in both the primary tumour and MUG-Myx2a/b. MLH1 is part of the DNA mismatch repair system and its loss could contribute to the widespread genetic changes seen in this tumour. FGFR1 amplification is a common feature of several tumour types and activates the cell cycle via the RAS pathway.
The recent development of NGS as a diagnostic tool in the clinical setting has enabled rapid, targeted sequencing of tumours to determine causative mutations. When combined with various selective capture approaches, NGS has allowed for the efficient simultaneous genetic analysis of a large number of candidate genes. Here, we applied polymerase chain reaction (PCR) based NGS to identify alterations to oncogenes and tumour suppressor genes. Our NGS data revealed, that both MUG-Myx2a and MUG-Myx2b cells possessed the p53 mutation R213Q, whereas, only MUG-Myx2a showed the PTEN R173H mutation.
p53 and PTEN are the two most frequently mutated tumour suppressors in human cancer. p53 plays a major role in regulating the response of mammalian cells to stress and damage, partially through transcriptional activation of genes involved in cell cycle control, DNA repair, senescence, angiogenesis, and apoptosis 18 . Furthermore, mutations in p53 have also been identified as the most common genetic alterations in sarcoma 19 . PTEN is involved in basic cellular functions including adhesion, migration, proliferation and cell survival 20 . Somatic mutations in PTEN are now known to cause tumourigenesis in a number of human tissues 21 . Both p53 and PTEN are involved in sustaining cellular homeostasis and complex regulatory interactions 22,23 . The combined inactivation pf p53 and PTEN could greatly accelerate tumour development [24][25][26] . Mutational analysis of p53 and PTEN in STS tissue samples revealed an incidence of 25.6% (22 out of 86) p53 mutations but only 2.3% (2 out of 86) PTEN mutations 27 . The PTEN R173H mutation is located in one of the three hotspot sites of the PTEN gene and has been reported in patients with PTEN associated tumors 28 . This mutation reduces PTEN activity by 50% 29 . The absence of this mutation from the primary tumor as well as the MUG-Myx2b cell line can either be explained by tumor heterogeneity, with MUG-Myx2a cells originating from a tumor sub-clone not present in the paraffin block sample used for analysis of the primary tumor, or by a secondary mutation during cell culture.
KDR (kinase insert domain receptor) is the human gene encoding vascular endothelial growth factor receptor 2 (VEGFR-2). It functions as the main mediator of VEGF-induced endothelial proliferation, survival, migration, tubular morphogenesis, and sprouting. Several studies have reported the genetic polymorphism of the KDR gene and the implicated risk of coronary artery diseases 30,31 . However, the clear role of individual KDR SNPs and their physiological functions in cancer progression and prognosis remains unknown. Mutations in the KIT receptor tyrosine kinase, which are commonly present in gastrointestinal stromal tumours (GISTs; 70-80% of all cases) are clustered in four exons 32 . The FGFR3 F386L polymorphism has been reported in association with low-grade tumours and early disease stage in prostate cancer 33 .
In conclusion, the well-characterized novel MFS cell lines MUG-Myx2a and MUG-Myx2b will be a useful tool to gain further insights into the pathogenesis and tumour heterogeneity of MFS and explore new therapy options. Invitrogen, Darmstadt, Germany), containing 10% foetal bovine serum (FBS), 1% L-glutamine, 100 units/ml penicillin, 100 μ g/ml streptomycin and 0.25 μ g amphotericin B (all Invitrogen). From this fraction the sub-clone MUG-Myx2b was grown. On the other hand, the viscous colloidal supernatant was collected and cultured as sub-clone MUG-Myx2a in the above-mentioned culture medium. During cultivation time, the cells were regularly cryopreserved. Cells grew to be adherent as a monolayer and were passaged more than 70 times during the period of 24 months. Even from the beginning of cultivation MUG-Myx2a and MUG-Myx2b differed regarding their morphology and growth behaviour (see Fig. 1). All cell cultures were periodically checked for mycoplasma by PCR.

Methods
Cell line identification Power Plex ® 16 System. Frozen tumour tissue from the patient was dissected into small pieces and re-suspended in 180 μ l ATL buffer (Qiagen, Hilden, Germany). Cell pellets (5 × 10 5 ) from both clones (passage 45 and passage 48) were re-suspended in 200 μ l PBS; subsequently 20 μ l Proteinase K and 200 μ l AL Buffer (Qiagen) were added. DNA preparations were performed using the QIAamp DNA Mini kit (Qiagen) in accordance with the manufacturer's protocol. After normalising the DNA, 1 ng DNA of each sample was amplified using the Power Plex ® 16 System (Promega, Vienna, Austria) in a 10 μ l reaction. One μ l of the product was mixed with Hi-Di formamide (Applied Biosystems Inc., Foster City, US) and Internal Lane Standard (ILS600), denatured and fractionated on an ABI 3730 Genetic Analyser (Applied Biosystems Inc.). The resulting data were processed and evaluated using ABI Genemapper 4.0 (Applied Biosystems Inc.).

MUG-Myx2a/b morphology.
For IF analysis, cells were seeded at a concentration of 1 × 10 4 cells on polystyrene culture slides (BD Biosciences, San Diego, CA). Slides were washed with PBS and fixed by exposure to formalin 4% for 10 minutes. After drying and rehydration, the slides were treated with Large Volume UltraV-Block (ThermoScientific, Waltham, US) for 10 min at room temperature to block nonspecific binding, incubated with the primary monoclonal mouse anti-Vimentin antibody (Dako) for 30 min and, after several washing steps, incubated with the Cy2 conjugated sheep anti-mouse IgG secondary antibody (Jackson Immunoresearch, Suffolk, UK) at a dilution of 1:200 for 30 min. Nuclei were counterstained with DAPI (Invitrogen).
Mice tissue. IHC studies using the streptavidin-biotin peroxidase complex method were carried out on histological slides from MUG-Myx2a and MUG-Myx2b mice tumours, employing a rabbit monoclonal primary antibody against the anti-Ki-67 (clone 30-9) (Ventana Medical Systems) using the BenchMark Ultra instrument (Ventana Medical Systems). Slides were photographed using an Olympus BX51 microscope with an Olympus DP71 microscope digital camera. The stained slides were scanned digitally and positive and negative cells were quantified using the ImageScope software (ImageScope Virtual Slide, version 6.25, Aperio Technol., Vista, US). Positivity was determined by assessing the number of positive cells / number of total cells. Cell proliferation analysis. xCELLigence System. The xCELLigence RTCA-DP device from Roche Diagnostics (OLS, Bremen, Germany) was used to monitor cell proliferation in real-time. Respectively 5 × 10 3 MUG-Myx2a/b cells were seeded in electronic microtiter plates (E-Plate ™ , OLS) and measured for 72 h with the xCELLigence system according to the instructions in the user's manual. Cell density measurements were performed in quadruplicate with a programmed signal detection every 20 min. Data acquisition and analyses were performed with the RTCA software (version 1.2, OLS).
Drug sensitivity assay. MUG-Myx2a and MUG-Myx2b were adjusted to a density of 3 × 10 3 cells and incubated in 96-well microplates. The cells were exposed to various concentrations of chemotherapeutic drugs (the DNA topoisomerase II inhibitor doxorubicin hydrochloride, the histone deacetylase inhibitors inhibitor Verinostat, and the proteasome inhibitor Bortezomib; Selleckchem, Houston, TX) for 48 h. Drug sensitivity was determined by the MTS assay following the manufacturers' instructions in quadruplicate using a photometer (Spektramax; BMG Labtech., Offenburg, Germany) at the wavelength of 490 nm. Results were expressed as the mean from three independent experiments (n = 3, measured in biological quadruplicates) and error bars represent the S.D.
Real-time chemotaxis assay. Chemotaxis assays were performed in cell invasion and migration (CIM-Plate ™ ; OLS) using an xCELLigence RTCA-DP instrument (OLS). A CIM-plate comprises two separate parts: an upper chamber with a microporous membrane (pore size = 8 mm) embedded with gold microelectrodes on its underside and a lower chamber. This migratory assay is based on a small and harmless electrical current applied between the electrodes. Cell adherence to this surface while migrating from the upper chamber toward 10% FBS as chemoattractant placed in the lower chamber is measured. The more cells accumulate on the underside of the membrane, the more the electrical current is impeded by the insulating properties of cellular plasma membranes. The overall cell-induced impedance is measured in real-time as the cell index (CI), a dimensionless parameter.
Scientific RepoRts | 7:44700 | DOI: 10.1038/srep44700 Tumour formation. Tumourigenicity of MUG-Myx2a/b clones: 7 week old female nu/nu Foxn1 mice (Charles River Laboratories, Sulzfeld, Germany) were xenotransplanted with MUG-Myx2a (p49) and MUG-Myx2b (p52) cells. Respectively 5 × 10 6 cells were suspended in 0.2 ml of serum-free medium and subcutaneously inoculated into the right and left flank of 5 mice. The mice were observed daily and the tumour growth was monitored. All animal work was done in accordance with a protocol approved by the institutional animal care and use committee at the Austrian Federal Ministry for Science and Research (BMWF) (vote 66.010/0160-II/3b/2012).

Mutation Analysis by Next-Generation Sequencing (NGS). Genomic DNA of MUG-Myx2a and
MUG-Myx2b was isolated on a Maxwell, MDxResearch System (Promega). Highly multiplexed PCR was used to generate amplicon libraries covering the coding region of 50 genes commonly implicated in cancer (Cancer Hotspot Panel v2, Cat.Nr. 4475346; Thermo Fisher Scientific, Waltham, MA). All analyses were performed in duplicates. Libraries were prepared using the Ion AmpliSeq Library Kit 2.0 (Cat.Nr. 4475345; Thermo Fisher Scientific) and sequencing was performed on an Ion Proton Sequencer (Thermo Fisher Scientific). Emulsion PCR and sequencing runs were performed with the appropriate kits (Ion One Touch Template Kit version 2 and Ion Proton 200 Sequencing Kit; Thermo Fisher Scientific) using Ion PI chips. Sequencing length was set to 520 flows and yielded reads ranging from 70 to 150 bp, consistent with the expected amplicon size range. NGS Data Analysis. Initial data analysis was performed using the Ion Torrent Suite Software version 4.1 Plug-ins (Thermo Fisher Scientific, open source, general public license). Briefly, this included base calling, alignment to the reference genome (HG19) using the Torrent Mapper and variant calling by a modified diBayes approach by taking into account the flow space information. All called variants were annotated using open source software (Annovar 34 , http://www.openbioinformatics.org/annovar/annovar download form.php, and SnpEff 35 , http://snpeff.sourceforge.net/download.html; both last accessed September 28, 2015). Coding, nonsynonymous mutation calls present in both duplicate analyses were further evaluated and visually inspected in Integrative Genomics Viewer. Variant calls resulting from technical read errors or sequence effects were excluded from the analysis. Mutations were subsequently assessed for biological relevance by computerized prediction (in silico prediction). Sorting Intolerant From Tolerant algorithm (SIFT) was used to generate SIFT scores.
Low density whole genome copy number variation (CNV). NGS libraries werer prepared using 500 ng genomic DNA and the Ion Plus Fragment Library Kit (Cat.Nr. 4471252, Thermo Fisher Scientific). DNA libraries were sequenced on the Ion Proton Sequencer yielding 8-10 million reads per library. CNV analysis was performed using the CNAnorm Bioconductor R package 36 . Briefly, reads were aligned to hg19 and summarized in windows of 10 kB size. CNAnorm was used to normalize for GC content and perform CNV analysis using coverage data from a non neoplastic reference sample. Circos plots were created using the circos software package 37 (http://circos.ca/). Ploidy analysis of single genes was calculated in R using gene coordinates from ensembl database and ploidy ratio information from CNAnorm.