Higher cMET dependence of sacral compared to clival chordoma cells: contributing to a better understanding of cMET in chordoma

Chordomas are rare slow growing, malignant bone tumors of the axial skeleton with no approved medical treatment. As the majority of chordomas express cMET and its ligand, HGF, and crosstalks between EGFR and MET-signaling exist, we aimed to explore cMET activity in chordoma cell lines and clinical samples. We investigated nine chordoma patients and four chordoma cell lines for cMET expression. Two clival and two sacral chordoma cell lines were tested for chromosomal abnormalities of the MET gene locus; we studied the influence of HGF on the autocrine secretion and migration behavior, as well as protein expression and phosphorylation. Two MET/ALK inhibitors were investigated for their effects on cell viability, cell cycle, cyclin alterations, apoptosis, and downstream signaling pathways. Moderate and strong expression of membrane and cytoplasmic cMET in chordoma patients and cell lines used, as well as concentration-dependent increase in phospho cMET expression after HGF stimulation in all four chordoma cell lines was shown. U-CH2, MUG-Chor1, and UM-Chor1 are polysomic for MET. Chordoma cell lines secreted EGF, VEGF, IL-6, and MMP9 upon HGF-stimulation. Sacral cell lines showed a distinct HGF-induced migration. Both inhibitors dose-dependently inhibited cell growth, induce apoptosis and cell-cycle arrest, and suppress downstream pathways. Heterogeneous responses obtained in our in vitro setting indicate that cMET inhibitors alone or in combination with other drugs might particularly benefit patients with sacral chordomas.

www.nature.com/scientificreports/ phosphorylation, cytokine secretion and migration potential were analyzed. Furthermore, the effects of the MET/ ALK inhibitors crizotinib and cabozantinib were investigated in terms of viability assessment, cell cycle arrest, cyclin alterations, apoptosis and analyses of downstream signaling pathways. These studies contribute to a better understanding of cMET and HGF in chordoma biology.

Results
Immunohistochemical evaluation of cMET expression in chordoma patients. To figure out the cMET expression, immunohistochemical staining was performed on different chordoma cell lines as well as on patient samples. All four chordoma cell lines showed a moderate or strong expression of membrane and cytoplasmic cMET immunostaining (Fig. 1A). cMET was also detected in all patient samples (5 sacral, 4 clival). Staining intensity varied from moderate to strong. Due to the small number of samples, no significant differences in cMET expression was found between clival and sacral samples (Fig. 1B).
Karyotyping and MET amplification. Molecular karyotyping of the chordoma cell lines was carried out using a multicolor-Fluorescence in situ hybridisation (FISH) technique ( Fig. 2A). While MUG-Chor1 and MUG-CC1 presented a hypodiploid and hyperhaploid genome with variable translocations, respectively, UM-Chor1 showed a hypertriploid karyotype with complex rearranged chromosomes. Detailed karyotypes are shown in Fig. 2. Genomewide copy number profiling and ploidy estimates established from shallow whole genome sequencing (sWGS) were highly concordant (Fig. 3). cMET status was assessed using multicolour FISH and from sWGS data. For MUG-Chor1 a total of 3 MET locus specific and centromeric FISH signal were detected, which confirmed the polysomy 7 from karyotyping and sWGS (Fig. 2B). In U-CH2 cells the combination of chromosome 7 centromeric probes and the MET locus specific FISH in metaphases showed three centromeric signals and 4 MET specific signal cluster from which two were with stronger signals indicating a possible amplification of the region. In interphase different levels of polyploidization were observed resulting between 3 and 11 centromeric signals and 6-19 MET locus specific signals between the different cells. This was in line with the sWGS sequencing data, which identified a focal amplification on an already gained chr7q. MUG-CC1 showed  www.nature.com/scientificreports/ no MET amplification, and the combination of chromosome 7 centromeric probes and the MET locus specific FISH showed the presence of two copies of each signals. In accordance with the karyotyping results, one signal was unchanged, but the second was rearranged on chromosome 7. Interestingly, focal amplification calling indicated an amplification of MET for UM-Chor1 that was not confirmed by FISH. UM-Chor1 cell line showed the presence of 6 copies of chromosome 7 centromeric signals and a consistent 4 copies of MET locus specific signals in line with the karyotyping results (Table 1).

Hepatocyte growth factor (HGF) stimulates cMET signalling in chordoma cells.
To determine the appropriate concentration for HGF stimulation, MUG-Chor1 cell line was treated with 0, 1, 5, 10, 25, and 50 ng/ml HGF. A concentration dependent increase of phospho-cMET expression was observed (Fig. 4A). Based on preliminary data, cells were stimulated in all subsequent experiments with 50 ng/ml HGF for 60 min. We analysed the protein expression of cMET and phospho-cMET under HGF stimulation in all four chordoma cell lines by western blot analysis (Fig. 4B). A ratio of phospho-cMET versus β-actin was calculated (mean ± SD; n = 3). Highest cMET expression and phosphorylation was seen in U-CH2 (-HGF: 0. HGF effect on chordoma cell migration correlates with cMET protein levels. To assess the role of HGF on chordoma cell migration, we studied the cell lines´ migration behaviour upon stimulation with HGF using the real-time xCELLigence system. We found that HGF distinctly enhanced the migration of two sacral chordoma cell lines MUG-Chor1 and U-CH2 cells, moderately attracted the clival UM-Chor1 cells, and did not affect MUG-CC1 cells (Fig. 4C). Stimulation with HGF leads to secretion of the epidermal growth factor (EGF), the vascular endothelial growth factor (VEGF), interleukin-6 (IL-6), and the matrix metalloproteinase 9 (MMP9).

Discussion
In the pathogenesis of various epithelial tumors, receptor tyrosine kinases such as cMET or EGFR, have been implicated and molecular-targeted therapies are being administered 11,12 . HGF, the only known ligand for the cMET receptor, is mainly expressed in cells of mesenchymal origin, although some epithelial cancer cells appear to express both HGF and MET 13,14 .
To better understand the role of cMET in chordomas, we studied whether the cMET gene-locus was polysomic or amplified in two sacral and two clival chordoma cell lines via single cell COBRA FISH analysis and copy number profiling. Comparing the number of allele/chromosomes within these cell lines, MUG-CC1 presented with 33 chromosomes, which comprises a nearly-haploid (< 30 chromosomes) karyotype; this cell line showed monosomy of chromosome 7. In contrast, the other chordoma cell lines displayed a varying polysomy of chromosome 7. Thus, heterogeneity between the karyotypes of these cell lines existed, which particularly affected chromosome 7, on which both, the cMET and HGF genes, are located 15 . Comparing chordoma cell lines after HGF stimulation, U-CH2 showed the strongest phospho-cMET-which is the activated form of cMET-expression, and MUG-CC1 the lowest. We found that sacral chordoma cell lines showed a stronger migration towards an  Our viability data clearly showed that, after HGF-stimulation, sacral chordoma cell lines responded better to treatment with cMET inhibitors than clival cell lines. In accordance with these data, treatment with cMET inhibitors caused a reduction of cells in the G1 phase, resulting in G2/M arrest of the cell cycle. It is well known that the cyclin B1/CDK1 complex together with cyclin A/CDK2 promotes G2/M transition in the eukaryotic cell cycle 20 . Compared to control cells, the expression of cyclin B1 and its corresponding cyclin-dependent kinase CDK1 was clearly reduced in MUG-Chor1 and U-CH2 cells. Thus, the checkpoint for the transition between the cell cycle phases was not being formed and the cells were arrested. Aberrant HGF/cMET axis activation, which is closely related to cMET gene polysomy and amplification, promotes tumor development and progression by stimulating the MAPK and other signalling pathways 21,22 . Our data revealed that the sacral chordoma cell lines showed a significant reduction in phosphorylation of Akt, Erk, and Stat3 after treatment with cMET inhibitors.
Remarkably, UM-Chor1 showed no changes in these phosporylation patterns.

Conclusion
For a future treatment of chordoma patients the origin of the primary tumor seems to play a crucial role. A high grade of heterogeneity between individual cell lines was observed. cMET inhibitors effected the growth behavior and cMET phosphorylation state of chordoma cells in different ways. Based on our in vitro cell culture analyses, it can be concluded that eventual treatment with cMET inhibitors would be more appropriate in sacral patients than in clival patients. Further investigation into chordoma heterogeneity will allow for a more precise treatment of individual patients and will pave the way towards a personalized chordoma treatment management.

Methods
The authors confirm that all used methods were performed in accordance with the relevant guidelines and regulations. Cobra fish. Cells were harvested using the chemically induced chromosome condensation technique published 25 . Slides with metaphase chromosomes were hybridized using a multicolor FISH approach, known as COBRA-FISH. The 43-color FISH staining of every chromosome arm in a different color combination, digital imaging and analysis were performed as described 26 . Chromosomal breakpoints were assigned using inverted images counterstained with 4' ,6-diamidino-2-phenylindole (DAPI; Downers Grove, IL, USA) together with the information derived from the short-and long-arm specific hybridization during COBRA-FISH. Karyotypes were described according to ISCN 2009. Two-color Fluorescence In Situ Hybridization (FISH) was performed as previously described 27 . BAC clone (BACPAC Resources Center) RP11-95I10 was selected to analyze the cMET locus. As reference DNA of centromere 7 was used. RP11-95I10 and DNA of centromere 7 were labeled Cy3-dUTP and Fluorescein-12-dCTP respectively using a nick translation labeling reaction 28 .

Immunohistochemistry (IHC).
Copy number profiling. Genome wide copy number alterations (CNA) were established using shallow whole genome sequencing (sWGS). Shotgun libraries were prepared using the TruSeq DNA LT Sample preparation Kit (Illumina, San Diego, CA, USA) following the manufacturer´s instructions. Briefly, after fragmentation and concentrating the volume to 50 µl end repair, A-tailing and adapter ligation were performed following the manufacturer's instructions. Libraries were quality checked on an Agilent Bioanalyzer using a DNA 7500 Chip (Agilent Technologies, Santa Clara, CA, USA) and quantified using qPCR with a commercially available PhiX library (Illumina) as a standard. Libraries were pooled equimolarily and sequenced on an Illumina MiSeq in a 150 bp single read run. Copy number profiling from sWGS data was performed using the ichorCNA algorithm applying After incubation supernatant was stored at -80 °C for further analysis. The epidermal growth factor (EGF), the vascular endothelial growth factor (VEGF), interleukin 6 (IL-6), and the matrix metalloproteinase 9 (MMP9) levels were determined using the commercially available Affymetrix eBioscience lot number (ln) 107311000 (eBender MedSystems, Vienna, Austria) on a Bioplex200 system (Biorad, Hercules, CA) in combination with Bio-Plex Manager software, version 4.1 (Bio-Rad Laboratories, Hercules, CA, USA), with 5-parametric curve fitting. Standard range and sensitivity for the respective cytokines: EGF (ln 95592000) 2.93-12000; VEGF-A (ln 107198005) 5.37-22000 pg/ml; HGF (ln 99618101): 10-41700 pg/ml; IL-6 (ln 102903000) 9.57-9800 pg/ml; MMP9 (ln 97975000) 1.06-4350 pg/ml. Complete growth medium was used as background control.