Twenty-three consecutive primary EWS tumor samples were obtained by surgeons from the Department of Biomechanics, Medicine and Rehabilitation of the Locomotor System of the Clinics University Hospital (Ribeirão Preto School of Medicine-University of São Paulo) between May 2005 and September 2015. The survival analysis was followed until June 2016. No local or systemic treatment had been conducted in these patients before the surgery. All samples were obtained with informed consent and the research approved by the Ethics Committee of the University of São Paulo (no. 43619215.9.0000.5407). Tissues were included in paraffin by the Pathology department of the Clinics University Hospital (Ribeirão Preto School of Medicine, University of São Paulo, Ribeirão Preto, SP, Brazil).

The EWS cell lines SK-ES-1 and RD-ES were acquired from the Rio de Janeiro Cell Bank (Rio de Janeiro, Brazil). Before the experiments each cell line authentication was conducted in the laboratory of Biochemical Genetics-FMRP/USP, by examining the CSF1PO, D13S317, D16S539, D5S818, D7S820, THO1, TPOX, vWA, and AMEL polymorphic loci for Short tandem repeat profiling (STR) under the supervision of Professor Dr Aguinaldo Luiz Simões. Cells were grown in McCoy's or RPMI medium (Gibco; Grand Island, NY, USA) supplemented with 10% of fetal bovine serum and an antibiotic mixture (100 units/ml penicillin, and 100 µg/ml streptomycin) and maintained in an incubator at 37°C with 5% CO₂ in a humidified atmosphere.

The drugs, hydroxifasudil (pan-ROCK inhibitor), GSK429286 and SR3677 (ROCK1 and ROCK2 inhibitor respectively) were purchased from Sigma-Aldrich (St. Louis, MO, USA). For all experiments, the drugs were added on the culture medium immediately before applied to cells. Corresponding control cultures received equal volumes of solvent dimethyl sulfoxide (DMSO).

Total RNA from cell lines was extracted using TRIzol Reagent (Invitrogen, Karlsruche, Germany) following the manufacturer's protocol. RNA samples from 9 osteoblast primary cultures were kindly provided by Professor Adalberto Luiz Rosado from the School of Dentistry of Ribeirão Preto, University of São Paulo, and used as controls. The concentration and quality of the RNA was accessed using a ND-1000 NanoDrop spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Wilmington, DE, USA). cDNA was synthetized using the High Capacity kit (Applied Biosystems; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's instructions. The qRT-PCR was performed using Taqman^® gene assays [ROCK1 (Hs01127699-m1), ROCK2 (Hs00178154-m1)], according to the manufacturer's protocol on the 7500 Real Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). As reference genes, GUS was used to normalize expression levels. The MRC5 (normal fibroblast) cell line were used as calibrator. Relative expression was calculated by 2^−ΔΔCT analysis method (11).

Total protein was extracted from EWS cell lines with RIPA buffer (Thermo Scientific, Inc.) according to manufacture instructions. Equal amounts of heat-denatured protein samples (50 mg per lane) were separated on 10% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The antibodies included primary rabbit monoclonal anti-ROCK1 (Ab45171; dilution, 1:500; Abcam, Cambridge, MA, USA), primary rabbit monoclonal anti-ROCK2 (Ab125025; dilution, 1:10,000; Abcam) and rabbit monoclonal anti-GAPDH antibody (AbEPR6256; dilution, 1:10,000; Abcam). The immunoblots were developed using goat anti-rabbit secondary antibody (Ab6721; dilution, 1:5,000; Abcam) followed by detection with the ECL Western Blotting Substrate kit (Abcam, Cambridge, UK) and visualized in a ChemiDoc Bioimaging System (Bio-Rad, Hercules, CA, USA). Expression levels were quantified using ImageJ^® software and normalized to loading controls.

Immunohistochemistry for ROCK1 and ROCK2 was performed in 23 and 22 available tissue samples, respectively (Anti-ROCK1; Ab45171, dilution, 1:250; Anti-ROCK2; Ab125025; dilution, 1:200; Abcam). The reactions were performed with the EXPOSE Mouse and Rabbit Specific HRP/DAB Detection IHC kit (ab80436; Abcam) according to the manufacturer's recommendations. Briefly, histological sections were submitted to xylol and alcohol baths for complete dewaxing and subsequent hydration. Antigen retrieval was performed in steam cooker for 40 min in Tris-EDTA pH 9.0 buffer. Blocking of endogenous peroxidase and nonspecific binding was performed, respectively, with hydrogen peroxide (15 min) and blocking solution of the kit for 10 min. The primary antibodies were diluted as recommended by the manufacturer and incubated at room temperature for ~2 h. After washing, the sections were incubated with the kit complement solution for 10 min and then with the HRP conjugate for 15 min. Finally, the slides were incubated with diaminobenzidine solution (DAB) for a standard time for each antibody and counterstained with Harris Hematoxylin. A colon adenocarcinoma biopsy was used as a positive control for the anti-ROCK1 antibody, whereas a non-neoplastic kidney sample was used as a control for ROCK2. For negative control of the reactions the primary antibody was replaced with PBS buffer. For evaluation of the immunostaining, the slides were scanned with an Olympus BX61VS Slide Scanner system (Olympus Optical do Brasil Ltda, São Paulo, Brasil) and at least five regions representative of the tumor were analyzed using the IHC Profiler plugin, according to Varghese et al (12). For subsequent statistical analysis, the samples were classified into two groups: Negative or positive (low positive, positive or high positive).

Cell survival assays were performed through the XTT kit (XTT II; Roche Molecular Biochemicals, Mannheim, Germany). In summary, 2,000 cells were seeded in 96-well plates and allowed to attach overnight. Subsequently, cells were treated with different concentrations of each drug and incubated for 48, 72 and 96 h (h). After treatment, the culture medium was replaced with medium containing 10 µl of XTT dye (3 mg/ml) in each well. The plates were incubated for 4 h at 37°C and the formazan product was measured at 455 and 650 nm by using an iMarkmicroplate reader (Bio-Rad). As a control, cells treated with the same concentration of drug vehicule, DMSO (Sigma-Aldrich) were used. Each experiment was performed at least in triplicate wells and repeated in three sets of tests.

Colony formation assays were performed according to Franken et al (13). Briefly, single cell suspensions of 1,000 cells were seeded in 6-well plates and treated with several concentration of each drug for 48 h. After this period, the culture medium were replaced with a drug-free medium and cells incubated at 37°C for 7 to 15 days. Then, the colonies were fixed with methanol and stained with Giemsa 3%. Only colonies containing more than 50 cells were scored. Assays were performed in duplicate in three independent sets of tests.

Cells treated with each drug at different concentrations for 24 h were detached by trypsin, fixed in 100% ethanol and stained with propidium iodide for analysis on a Guava Personal Cell Analysis system (Guava Technologies, Hayward, CA, USA) according to the standard protocol provided by the manufacturer. The percentages of cells in G0/G1, S and G2/M phase were analyzed using the GUAVA Cytosoft software version 4.2.1 (Guava Technologies). All cell cycle assays were performed in triplicate in three independent sets of tests.

Wound healing assays were performed according to Liang et al (14). with minor modifications. Succinctly, cells were grown to confluence on 12-well plates, and scratch wounds were created using a pipet tip (200 µl) and photographed at time zero. Then, cells were treated with different concentrations of each drug and cultured for 24 h in medium with only 1% of fetal bovine serum. After that period, cells were photographed. The cell-free area was measured with the Motic Images Plus v2.0 software (Motic China Group Co., Ltd., Xiamen, China). Cell migration rates were calculated as the distance travelled by the cells in this area over time. Assays were performed in duplicate in three independent sets of tests.

Cell invasion was measured by migration of cells through gel-coated Transwell inserts. EWS cells were harvested, re-suspended in serum-free medium, treated with different concentrations of each drug and seeded on the top of Matrigel-coated invasion 8 µm pore size chambers (Becton Dickinson & Co., Franklin Lakes, NJ, USA; 5×10⁵ cell/insert). Bellow the insert, the well was filled with medium containing 10% fetal bovine serum. Cells were then allowed to migrate for 24 h in an incubator and after that period non-invasive cells were removed from the membrane upper surface with swabs. The ones attached to the lower side of the membrane were fixed with 100% methanol and stained with Giemsa 3%. The membranes were removed from the inserts, placed on microscope slides with Entelan (Merk, NY, USA) and counted with ImageJ^® software in ten random fields at magnification, ×20. Assay was performed in three sets of independent experiments.

Associations between ROCK1 and ROCK2 protein expression and the clinical variables [age (<14 years vs. >14 years old); sex (male vs. female); EWS/FLI1 status (positive vs. negative); Huvos grade (<90% of necrotic areas-Huvos levels 1 and 2- vs. >90% necrotic areas-huvos levels 3 and 4; tumor volume (<200 cm³ vs. >200 cm³); tumor skeletal location (axial vs. appendicular); remission (incomplete vs. complete); metastasis (presence vs. absence); relapse (presence vs. absence); death (alive vs. deceased] were determined by two-tailed Fisher's exact test. Survival analysis was carried out based on Log-Rank tests represented on Kaplan-Meier curves. The functional assays data was statistically analyzed by Student's two-tailed t-test or One-Way Repeated Measures Analysis of Variance (ANOVA) followed by the Bonferroni Pairwise Multiple Comparison. All tests were carried out for α=0.05. All analyses were performed using the SPSS 21.0 software (IBM SPSS, Armonk, NY, USA) and expressed as the mean ± standard deviation.

The majority of tumor samples presented positive immunostaining for ROCK1 (17/23; 75%) and ROCK2 (14/23; 65%) proteins (Fig. 1A). However, no significant associations were observed between their expression and any of the relevant clinical features such as Huvos classification, FLI1/EWS status, relapse, metastasis, or death. Nonetheless, we observed a trend for poorer outcome in patients with positive samples, and significant higher risk of incomplete remission in patients with ROCK2 positive tumors (OR=35.0, 95% CI, 1.74–702.9; P=0.026). In addition, positivity for ROCK2 seems to indicate increased risk of larger tumor volume (OR=2.33, 95% CI, 0.22–25.24; P=0.58) (Table I). Moreover, ROCK1 and ROCK2 positivity was also suggestive of lower patient's survival, even though no significant differences were found (Fig. 1B). Event-free survival (EFS) for ROCK1 was estimated at 23.8±14.1% for positive samples vs. 50±25% for negative ones (P=0.925). EFS of ROCK2 positive patients was 11.9±11.1% vs. 60±21.9% (P=0.423) (Fig. 1C).

mRNA expression levels of ROCK1 and ROCK2 were evaluated in two EWS cell lines, SK-ES-1 and RD-ES through quantitative real-time PCR. As seen in Fig. 2A, ROCK1 did not show any significant difference in expression when compared to the control (nine primary osteoblast cell lines). Conversely, ROCK2 was found with significant higher expression (P=0.03) compared to controls (fold-change 1.97 for RD-ES and 1.68 for SK-ES-1). Protein expression levels of both kinases were found comparable (Fig. 2B).

To investigate the prospect of targeting ROCK1 and ROCK2 in EWS we evaluated the in vitro effects of three specific inhibitors on cell viability, clonogenicity and cell cycle in the SK-ES-1 and RD-ES EWS cell lines. For each experiment the drugs GSK924286, a specific ROCK1 inhibitor, SR3677, a specific ROCK2 inhibitor and hydroxyfasudil, a pan-ROCK inhibitor were used at different doses according with manufacturer's instructions. Firstly, doses comprising the IC₅₀ reported values were used, being 3.5, 7, 14 and 21 nM for GSK924286, 1.25 nM, 2.5, 5 and 10 nM for SR3677 and 0.3, 0.6, 1.2 and 2.4 µM for hydroxyfasudil. Cells were treated for 24, 48, 72 and 96 h though none of these treatments affected cell viability at any time (data not shown). Then, doses were increased to 25, 50, 100 and 150 nM for GSK924286, 25, 50, 100 and 200 nM for SR3677 and 2, 4 and 8 µM for hydroxyfasudil and cells treated for the same periods. Nonetheless, cell viability was not affected again (Fig. 3A). For the other functional assays, doses of 50 nM e 100 nM were chosen for GSK924286 and SR3677 and of 4 and 8 µM for hydroxyfasudil. Likewise, and cell cycle dynamics and the clonogenic capacity were not significantly affected by inhibition of ROCKs (Fig. 3B and C).

Moreover, the migration capacity of SK-ES-1 cell line was not significantly altered after treatment with any of the drugs as seen through the wound healing assay. Nonetheless, gap closure under treatment with the ROCK2-inhibitor and the pan-inhibitor was increased in ~50% (Fig. 4A). This effect on the migratory capacity of the cells was evinced by the invasion assays, where treatment with SR377 50 nM and hidroyfasudil 8 µM induced higher penetrance of cells through the Matrigel layer (P=0.0063 and P=0.0344, respectively) (Fig. 4B).