Fascin-1 enhances experimental osteosarcoma tumor formation and metastasis and is related to poor patient outcome

Osteosarcoma and normal bone tissue specimens were collected between June 1990 and December 2005 from 67 patients during primary tumor resection in accordance with the regulations of the local Swiss Ethic committee. The clinical characteristics of the OS patients are presented in Table 1. Among the 67 patients’ cohort, 53 received complete neoadjuvant chemotherapy and the subsequent response was determined histologically on resected tumor specimens according to Salzer-Kuntschik [17]. Salzer-Kuntschik grades I, II, and III were considered to be a good response, whereas grades IV, V, and VI were classified as non-responders. The TMA was constructed and processed as previously described [18, 19], stained with anti-Fascin-1 mouse monoclonal antibody (1400; DAKO), and counterstained with hematoxylin. The grading of Fascin-1 immunostaining was determined based on the intensity and the percentage of the stained area. The percentage of staining was calculated with a custom made MATLAB (v2009b, Mathworks Inc) program as described [20]. Positive stained (brown) and non-stained area (blue) were separated by color deconvolution algorithm [21]. The area percentage of the stain was defined as the positive stained area (number of brown pixels) over total tissue area (number of blue and brown pixels). Grade 1 (non or weak staining) was considered as negative and grades 2 and 3 (moderate and intense staining) were considered as positive for Fascin-1 expression. A Kaplan-Meier analysis was used to correlate Fascin-1 expression with overall survival of the patients.

Human SaOS-2 (ATCC® HTB85™) and 143B (ATCC® CRL-8303™) OS cells, purchased from American Type Culture Collection (ATCC; Manassas, USA), were cultured in DMEM (4.5 g/l glucose)/HamF12 (1:1) medium (Invitrogen) supplemented with 10% heat inactivated fetal bovine serum (FBS, Gibco) at 37 °C in a humidified atmosphere of 5% CO₂. SaOS-2 and 143B cells were stably transduced with a LacZ gene and selected to obtain SaOS-2-LacZ and 143B-LacZ cell lines as previously reported [22]. SaOS-2-LacZ and 143B-LacZ cells stably over-expressing Fascin-1 (SaOS-2/Fascin-1 and 143B/Fascin-1) were then obtained by transduction with pLenti6/V5-DEST-FASCIN (a gift from Lynda Chin; Addgene plasmid # 31207) and subsequent selection with Blasticidin (Invitrogen, Carlsbad, CA, USA). Control SaOS-2-LacZ or 143B-LacZ cells were obtained by transduction with empty pLentiV5-DEST.

To achieve Fascin-1 silencing, we tested 5 different shRNA sequences cloned into the pLKO.1 retroviral vector (FSCN1 MISSION shRNA; Sigma-Aldrich). After transduction of SaOS-2-LacZ and 143B-LacZ cells, Puromycin resistant cells were selected. The efficiency of Fascin-1 silencing was examined on Western blot, and the most efficient oligonucleotide sequence 5’-CCGGCCCTTGCCTTTCAAACTGGAACTCGAGTTCCAGTTTGAAAGGCAAGGGTTTTTG-3′ was selected and used in the subsequent experiments. pLKO.1 retroviral vector containing a scrambled oligonucleotide sequence was used as control. All generated stable cell lines were authenticated by short tandem repeat DNA profiling (Microsynth) with a PowerPlex®16HS system (Promega) and by comparison with the German Collection of Microorganisms and Cell Cultures database (DSMZ).

Cells grown on glass coverslips were fixed with 3.7% formaldehyde in phosphate-buffered saline (PBS) at room temperature (RT) for 15 min and permeabilized with 0.2% Triton X-100 for 10 min. After incubation at RT for 30 min with blocking solution (0.1% FBS in PBS), the cells were incubated with specific primary antibodies in blocking solution at RT for 1 h. After 3 washing steps with PBS, cells were incubated with fluorescently-labelled secondary antibodies for 45 min at RT. Primary antibodies used were directed against Fascin-1 (1:400; DAKO) and V5 antibody (1:5000; Invitrogen). Filamentous actin was stained by means of Alexa-633-phalloidin and nuclei using NucBlue (Invitrogen). Secondary antibodies were Alexa-Fluor488 conjugated donkey anti-mouse or Alexa-Fluor488 conjugated donkey anti-rabbit (Life Technologies). Immunofluorescent cells were visualized using a spinning disc confocal microscope (iMic, TILL Photonics) equipped with a 60X N.A. 1.35 lens. For the morphological analysis, we used the Cell Profiler software to estimate the perimeter and spreading area of each individual cell [23]. We then used these values to calculate the ratio between them using the formula: R = perimeter * 100 / area.

The migratory properties of OS cells were assessed in a wound healing migration assay. Cells were seeded into 24-well plates. At confluency, wounds measuring between 0.5 and 1 mm in width and approximately 1 cm in length were created with a sterile pin. After 3 washing steps with PBS to remove detached cells and cell debris, the wounds were marked with a Nikon object marker attached to a Nikon Diaphot microscope. Photos of the marked wounds were taken immediately after wounding and after incubation at 37 °C for 16 h. The width of the marked wound areas were measured with the ImageJ Software and the migration rate (μm/h) was then calculated as described [24].

The gelatinase activity of MMP2 and MMP9 was analyzed by means of collagen zymography as previously described [25]. For western blot analysis, cells were lysed in ice-cold 50 mM Tris, pH 7.5 containing 250 mM NaCl, 5 mM EDTA, 0.5% NP-40, 50 mM NaF, 0.2 mM Na3VO4, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulphonyl fluoride (PMSF), and 10 mg/ml aprotinin. Cell lysates were cleared by centrifugation and protein concentrations determined with a Bio-Rad assay (Bio-Rad Laboratories, Hercules, CA, USA). Endogenous and V5-tagged Fascin-1 protein and GAPDH were detected using mouse monoclonal Fascin-1 (1:1000; DAKO), V5 antibody (1:5000; Invitrogen), and rabbit polyclonal anti-GAPDH antibody (1:5000; Santa Cruz Biotechnologies) and corresponding horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnologies). Peroxidase activity was visualized with the Immobilon chemiluminescence substrate (Millipore, Billerica, MA, USA) and detected with Image Lab System (Bio-Rad, Hercules, CA, USA).

All animal experiments were performed according to the institutional guidelines and approved by the Ethics Committee of the Veterinary Department, Canton of Zurich, Switzerland. Specific pathogen-free female SCID mice obtained from Charles River (Sulzfeld, Germany) were housed for at least 2 weeks before commencement of the study and then randomly divided into four groups of 12 mice each. As described previously [19], after anesthesia of the mice with a mixture of Fentanyl, Midazolam, and Medetomidin, 10 μl of PBS containing 10⁵ cells was injected into the medullar cavity of the left tibia. The health status of the mice was monitored three times a week and primary tumor growth was monitored weekly by X-ray. The tumor volume was calculated from caliper measurements of the tumor-bearing leg and the healthy control leg by the equation: length × (width)² / 2 of the tumor bearing tibia minus length × (width)² / 2 of the control tibia. At the end of the study, the mice were anaesthetized with a mixture of Ketamine, Xylazine, and Acepromazine. Then, the mice were sacrificed by removal of the blood from the cardiovascular system using PBS perfusion into the right ventricle of the beating heart, and the lungs were prepared. For the LacZ gene expression, the fixed organs were stained with 5-bromo-4-chloro-3-indolyl-β-D-galactoside (X-Gal) solution at 37 °C for at least 3 h as described [22]. Metastases in individual lungs, visible as blue-stained foci, were then quantified by manual counting under the microscope.

To identify effector molecules involved in OS metastasis, we performed combined immunohistochemistry-based TMA and Kaplan-Meier survival analysis. Human OS TMA sections including tumor specimens collected from 67 patients during primary tumor resection were analyzed immunohistochemically for Fascin-1 expression, among them twenty-four patients had metastatic disease (Table 1). Representative images of tissue microarray sections with non-detectable, weak, moderate and intense Fascin-1 immunostaining are presented in Fig. 1a.

A Kaplan-Meier survival analysis showed that, irrespective of local or metastatic disease, patients with Fascin-1 positive expression in tumor tissues had a significantly (p < 0.01) shorter overall survival than those with non-detectable Fascin-1 expression (Fig. 1b).

A second Kaplan-Meier survival analysis distinguishing between patients with local or with metastatic disease, revealed that patients with metastases and positive immunostaining for Fascin-1 exhibited a significantly shorter survival than those with metastases and Fascin-1 negative tumors (p < 0.05) (Fig. 1c). In the cohort with local disease no significant difference in the survival time was present between patients with Fascin-1 positive tumors and those that were Fascin-1 negative. In a further Kaplan-Meier analysis correlating Fascin-1 expression with the response to neoadjuvant therapy, non-responders who were Fascin-1-positive tended to have a shorter overall survival than Fascin-1-negative non-responders although the difference was not statistically significant (p = 0.1069, Additional file 1: Figure S1). Since the lack of statistical significance was likely related to the small number of patients, this finding nevertheless implicate that Fascin-1 might be associated with worse outcome in subsets of OS patients who responded poorly to chemotherapy. All findings taken together imply that the expression of Fascin-1 in OS tumors correlates with a poor prognosis for the patients, and consequently might be considered as new potential predictor for OS outcome.

In order to assess the function of Fascin-1 in OS and to account for the heterogeneity in phenotypes of OS [26], we used one osteoblastic (SaOS-2) and one osteolytic cell line (143B) and generated from those cell lines stably overexpressing Fascin-1 or with silenced Fascin-1 as described in the methods section. The levels of V5-Fascin-1 expression and the silencing efficiency were determined by western blot (Fig. 2a).

Since Fascin-1 is associated with the initiation and formation of filopodia structures, we considered whether Fascin-1-silencing or overexpression in OS cells may affect filopodia formation. As expected, independently of the type of the cell line, silencing of Fascin-1 was accompanied by a slight decrease in the ratio of cell perimeter by spreading area, suggesting reduced protrusion potential (Fig. 2b and c). Interestingly, Fascin-1 overexpression did not cause detectable differences in comparison to the wild type.

Based on the results of the Kaplan-Meier analyses, we speculated that Fascin-1 expressed by OS tumors may be a stimulatory factor in the regulation of metastatic activity. Consequently, we compared wound-healing migration of SaOS-2 and 143B cells stably overexpressing V5-tagged Fascin-1 (SaOS-2/Fascin-1 and 143B/Fascin-1) with that of empty vector transfected SaOS-2 (SaOS-2/EV) and 143B (143B/EV) control cells with low expression of endogenous Fascin-1. The migration rate of SaOS-2/Fascin-1 cells was 2 times (p < 0.01) higher than that of control SaOS-2/EV cells, while silencing of Fascin-1 decreased moderately the migration rate (p < 0.05) (Fig. 3a). Similar results were obtained with 143B cells, where Fascin-1 overexpression increased the migration rate by 2.5 folds in comparison to control EV cells (p < 0.01), and silencing of Fascin-1 slightly decreased the migratory rate (p < 0.05) (Fig. 3b). It is noteworthy that Fascin-1 did not affect the proliferation of SaOS-2 and 143B cells in vitro. The calculated doubling times and the percentage of proliferating cells, assessed by immunostaining for the proliferation marker Ki67, were indistinguishable in cultures of cells overexpressing Fascin-1 and control cells (not shown).

Furthermore, we tested whether Fascin-1 enhances the activity of some mediators of metastasis such as MMP-2 and MMP-9, which are among the most well established proteases known to degrade the extracellular matrix and facilitate invasion and metastasis. In contrast to MMP-2, which shows equivalent activity in all studied conditions, up-regulation of Fascin-1 increased significantly the MMP-9 activity compared to the EV control cells (Fig. 3c).

To study the impact of Fascin-1 on tumor growth and spreading in a physiologically faithful microenvironment, we performed orthotopic transplantation of OS cell lines into the tibia of SCID mice. Orthotopic tumors derived from injected SaOS-2 cells are mainly osteoblastic, resembling the osteogenic tumor phenotype in humans [19, 27], while, upon intratibial injection of 143B cells, the resulting primary tumors are predominantly osteolytic, thereby representing the osteolytic OS cases seen in human patients [28]. Irrespectively of the phenotype of the used OS cells, the generated experimental tumors preferentially metastasize in lungs. Thus, these orthotopic mouse models reflect more closely the clinical pattern than intravenous injection or subcutaneous implantation of OS tumor cells [29]. To assess the effect of Fascin-1 on OS malignancy we compared the tumorigenic and metastatic potential of SaOS-2-LacZ and 143B-LacZ cells over- and under-expressing Fascin-1. As shown in Fig. 4, mice injected with SaOS-2/Fascin-1 cells developed fast growing primary tumors six weeks after injection and showed osteoblastic lesions in the bone and the surrounding soft tissue (Fig. 4a, lower panel). In contrast, in mice injected with SaOS-2/EV cells, primary tumors remained radiologically undetectable until week eight (Fig. 4a, upper panel). Mice of both SaOS-2/Fascin-1 and SaOS-2/EV groups were sacrificed after 10 weeks because the animals in the SaOS-2/Fascin-1 group became moribund. Accordingly to the X-ray monitoring, Fascin-1 xenografts with a final mean primary tumor volume of 145 ± 25.8 mm³ (SaOS-2/Fascin-1) were significantly (p < 0.05) larger than SaOS-2/EV (43.2 ± 3.5 mm³) xenografts (Fig. 4c). Similarly, the calculated mean primary tumor volume (356.4 ± 40.1 mm³) of 143B/Fascin-1 xenografts were significantly (p < 0.05) bigger (137.8 ± 34.6 mm³) than 143B/EV xenografts (Fig. 5a and c). Moreover, the ex vivo X-gal-staining of lung whole mounts revealed that SaOS-2/Fascin-1 and 143B/Fascin-1 cells generated an average number of metastatic lesions 2.5 and 2 times larger than wild type cells respectively (Fig. 4d and 5d). It is noteworthy that silencing of Fascin-1 in both SaOS-2 and 143B cell line systems did not significantly affect neither primary tumor growth (Fig. 4b/c and Fig. 5b/c respectively), nor metastasis formation (Fig. 4d and 5d respectively). However, although not statistically significant, the metastatic potential of ShFascin-1 in 143B cells was nearly 40% lower than control cell (Fig. 5d). Taken together, our data demonstrates for the first time that Fascin-1 enhances primary tumor formation and promotes lung metastasis in OS.