SULF2 overexpression positively regulates tumorigenicity of human prostate cancer cells

SULF2 is a 6-O-endosulfatase which removes 6-O sulfate residues from N-glucosamine present on heparan sulfate (HS). The sulfation pattern of HS influences signaling events mediated by heparan sulfate proteoglycans (HSPGs) located on cell surface, which are critical for the interactions with growth factors and their receptors. Alterations in SULF2 expression have been identified in the context of several cancer types but its function in cancer is still unclear where the precise molecular mechanism involved has not been fully deciphered. To further investigate SULF2 role in tumorigenesis, we overexpressed such gene in prostate cancer cell lines. The normal prostate epithelial cell line RWPE-1 and the prostate cancer cells DU-145, and PC3 were transfected with SULF2-expressing plasmid pcDNA3.1/Myc-His(−)-Hsulf-2. Transfected cells were then submitted to viability, migration and colony formation assays. Transfection of DU-145 and PC3 prostate cancer cells with SULF2 resulted in increased viability, which did not occur with normal prostate cells. The effect was reverted by the knockdown of SULF2 using specific siRNAs. Furthermore, forced expression of SULF2 augmented cell migration and colony formation in both prostate cell lines. Detailed structural analysis of HS from cells overexpressing SULF2 showed a reduction of the trisulfated disaccharide UA(2S)-GlcNS(6S). There was an increase in epithelial-mesenchymal transition markers and an increase in WNT signaling pathway. These results indicate that SULF2 have a pro-tumorigenic effect in DU-145 and PC3 cancer cells, suggesting an important role of this enzyme in prostatic cancer metastasis.


Background
Cancer is the second leading cause of death worldwide, accounting for over 8 million deaths annually. Among men, prostate, lung and bronchus, and colorectal cancer accounts for about 50% of all newly diagnosed tumors; prostate cancer alone accounts for 28% of incidents among men [1][2][3]. Its incidence rate is about six times higher in developed countries when compared to the developing ones [4,5] being the estimated death count in the United States 29,720 in 2013 [1].
The current screening method to diagnose prostate cancer is based on a measurement of serum prostate specific antigen (PSA) levels and a digital rectal examination, while the decisive diagnosis is based on the results of transrectal, ultrasound-guided prostate biopsies [6][7][8]. The current therapeutic approaches for the advanced stages of prostate cancer are palliative rather than therapeutic [9]. Thus, determining the molecular pathways that lead to the development and progression of the disease is a challenge and critical for improved therapeutic approaches.
Searching for a better understanding of cancer, as well as for tumor markers, proteoglycans (PGs) have gained ground among the molecules involved in tumorigenesis. PGs are high molecular weight compounds, formed by a protein skeleton to which glycosaminoglycans (GAGs) chains are covalently bound [10,11]. They are located predominantly in the extracellular matrix (ECM) or associated with cell surface of most eukaryotic cells [12,13].
The PGs interact with numerous proteins and modulate their activity, influencing biological processes such as embryonic development and cell proliferation [11,13]. Suhovskih et al. [14] reported that in prostate tumors, complex changes occur in PGs, with decreased expression of decorin and lumican, an overall increase in syndecan-1 and glypican-1 in tumor stroma, along with the disappearance of agrecan in tumor epithelial cells. All changes result in the expression patterns of highly individual PGs in different prostate tumors, which may be potentially useful as molecular markers for the diagnosis of prostate cancer and personalized treatment.
HSPGs consist of macromolecules presenting one or more heparan sulfate (HS) chains covalently bound to the protein backbone [15][16][17][18] and are present on the cell surface and ECM of all tissues of animals with tissue organization [19][20][21][22]. Among its many roles, membrane HSPGs can bind to cytokines, chemokines and growth factors, protecting them from proteolysis. These interactions provide a reservoir of regulatory factors that may be released by selective degradation of HS chains [15,17,20]. HSPGs can also cooperate with integrins and other cell adhesion receptors to facilitate cell-ECM adhesion, and cell motility [16][17][18][19]. Finally, they can also act as coreceptors for a variety of growth factors, lowering its activation threshold or changing the duration of the signaling reactions [15][16][17][18].
In general, HS chain biosynthesis initiate by alternating actions of various glycosyltransferases which add residues of D-glucuronic acid (GlcA) and N-acetyl-D-glucosamine (GlcNAc). Subsequently, the chain undergoes a series of polymeric modifications reactions: N-deacetilation/Nsulfation, the epimerization of the β-Dglucurcnic acid to α-L-iduronic acid, and O-sulfation at different positions [23]. Each product is a reaction substrate for the next enzyme [22].
Recent studies have shown that after the synthesis, the HS can also be structurally and functionally modified in the extracellular compartment where 6-O-endossulfatases 1 and 2 (SULFs) are extracellular enzymes that remove 6-O-sulfate groups selectively, modulating their biological activities [24][25][26]. Recent studies revealed that different types of tumors present an increase in SULFs expression, including: hepatocellular carcinoma [27], pancreatic cancer [28], squamous cell carcinoma of the head and neck [29], gastric cancer [30], lung adenocarcinoma and squamous cell carcinoma of the lung [31] for SULF1 and hepatocellular carcinoma [32], lung adenocarcinoma and lung squamous cell carcinoma [31] for SULF2.
Zhao et al. [33] reported that SULF1 is present in prostatic stromal cells in the transition regions but not in benign prostatic hyperplasia. Ciampa et al. [34] identified that SULF2 chromosome locus is associated to prostate cancer susceptibility regions. However, the literature is ambiguous about the function of SULFs in cancer, and the enzymes are reported both as anti and as protumorigenic [25].
Thus, this study aimed to analyze the effects of the overexpression of SULF2 in prostate cancer cell lines via analyzing their viability, proliferation, migration and colony formation capabilities. Finally epithelial-mesenchymal transition markers were also assessed.

Methods
Cell culture RWPE-1, PC3 and DU-145 cell lines were purchase from ATCC (American Type Culture Collection, Manassas, VA, USA). PC3, prostate adenocarcinoma derived from bone metastatic site, and DU-145, prostate carcinoma derived from brain metastatic site, were grown in Roswell Park Memorial Institute medium (RPMI, Gibco, Life Technologies, CA, USA) supplemented with 10% (v/v) fetal bovine serum (FBS, Cultilab, Campinas, Brazil), penicillin (100 units/ml) and streptomycin (100 μg/ml, Invitrogen, Life Technologies, CA, USA) at 37°C in a humidified atmosphere of 5% CO 2 . RWPE-1, a normal prostate epithelial cell line, was grown in Keratinocyte Serum Free Medium supplied with bovine pituitary extract and human recombinant epidermal growth factor (Gibco, Life Technologies, CA, USA) at 37°C in a humidified atmosphere of 5% CO 2 .
Transfection and expression of SULF2 in culture Cells were cultured in 24-well plates and transfected with 5 μg of cDNA coding SULF2, cloned into the vector pcDNA3.1/Myc-His(−)-HSulf-2 (Addgene plasmid 13004). This plasmid was kindly donated by Prof. Dr. Steven D. Rosen [19]. For transfection FuGENEHD® reagent (Promega Corporation, WI, USA) was used according to manufacturer's instructions. The DNA was diluted in OptiMEM (Invitrogen, Life Technologies Corporation, CA, USA) combined with FuGENE and incubated for 20 min at room temperature. After incubation, the complex was added to the respective culture medium of each cell line. The cells were cultured for 20 days in the presence of geneticin (Promega Corporation, WI, USA) and clonally selected in accordance to the level of SULF2 overexpression.
Knockdown of SULF2 using siRNA SULF2 gene silencing was performed with siRNA preset by the manufacturer (Life Technologies Corporation, CA, USA). Three siRNAs were used for the gene, in addition to the positive (GAPDH) and negative (scramble sequence) controls: human SULF2 siRNA1 F: GGAC AACACGUACAUCGUAtt and R: UACGAUGUACGU GUUGUCCag; human SULF2 siRNA2 F: GGUGCUAC AUCCUAGAGAAtt and R: UUCUCUAGGAUGUAGC ACCga; human SULF2 siRNA3 F: GGACAGCUUUCU UCGGGAAtt and R: UUCCCGAAGAAAGCUGUCCgg. Cells were plated in 24-well plates so that they were 60-80% confluent by the time of transfection, according to the manufacturer's instructions. On the test day, the siRNA was added to the Lipofectamine RNAiMAX (Life Technologies Corporation, CA, USA) and incubated for 20 min. Finally, the solution was added to the cell media without FBS and no antibiotics. After 8 hours of incubation, the medium was replaced by the respective culture medium of each cell line. Viability assays and cell migration were performed at different times to analyze the consequences of silencing SULF2.

Real-time PCR
The expression of SULF2 was analyzed before and after the transfection of cell lines. Total RNA was extracted from cell lines (2.10 6 cells) using Trizol® reagent (Invitrogen, Life Technologies Corporation, CA, USA). The primers used for the amplification reaction were designed from the research database sequences and data already published: human SULF2 F: CTGTGGGAAGGCTGGGAAGG and R: TGAGAGTGCGTGCTTGCTTTC; human beta-actin F: ACCAACTGGGACGACATGGAGAAA and R: TAGC ACAGCCTGGATAGCAACGTA; human GAPDH F: TC GACAGTCAGCCGCATCTTCTTT and R: ACCAAATCC GTTGACTCCGACCTT. The Real-Time PCR reaction was performed using SYBR®-Green PCR Master Mix, including AmpliTaq-GOLD polymerase (Applied Biosystems, USA) on ABI PRISM 7500 Real Time PCR System (Applied Biosystems, USA). All reactions were performed in triplicate.

Western blotting
To verify the overexpression of SULF2, cellular proteins were extracted from both the cell extract and the culture medium. The adherent cells were removed from Petri dishes using cell lysis buffer (Cell Signaling, MA, USA) containing protease inhibitor cocktail (Roche, Mannheim, Germany) and then exposed to sonication. The collected conditioned medium was concentrated on Centricon centrifugal filter units (Millipore, Merck, MA, USA). 100 μg of samples resuspended in non-reducing sample buffer (Tris-HCl 100 mM pH 6.8, 4% SDS, 0.02% Blue bromophenol, 20% glycerol) were applied to 7.5% polyacrylamide gel and subjected to SDS-PAGE (80 V for 2 h). After electrophoresis, the proteins were transferred from the gel to a nitrocellulose membrane, incubated overnight at 4°C with primary anti-human SULF2 produced in rabbit (H-80, Santa Cruz Biotechnology, CA, USA) and human anti-beta-actin produced in goat (1:500-1000) (Santa Cruz Biotechnology, CA, USA) diluted in TBS with 1% BSA, and then incubated with IgG secondary antibody conjugated with peroxidase (1:2000). The membrane was incubated with the SuperSignal West Pico chemiluminescent substrate (Thermo Fischer Scientific, IL, USA). The chemiluminescent signal was detected using the gel documentation system G:BoxChemi HR16573 (Syngene, Frederick, MD, USA). Densitometric analysis of bands was performed using ImageJ (http://rsb.info.nih.gov/ij/) software, using beta-actin as a control for each sample.

Incorporation of sodium [ 35 S]-sulfate for Structural Analysis of Glycosaminoglycans (GAGs)
Cells transfected or not with pcDNA3.1/Myc-His-(−)-HSulf-2 were subjected to metabolic labeling with [ 35 S]-sulfate in a final concentration of 100 μCi/ml. After 24 h, the medium was collected, the cells were removed from the plate with 1 mL of 0.025% EDTA and lysed with 1 ml of 3.5 M urea in Tris-HCl 10 mM, pH 8.0. The extracellular matrix was removed with 5% trypsin, 4% EDTA. Cell extract, medium and extracellular matrix were subjected to proteolysis with maxatase (4 mg/ml in 50 mM Tris-HCl, pH 8.0 containing 1.5 mM NaCl) at 60°C for 24 h. After proteolysis, GAGs were precipitated with 3 volumes of ethanol at −20°C for 24 h. GAGs were analyzed by electrophoresis in agarose gel in PDA buffer (0.05 M 1,3-diaminepropane-acetate) [35]. GAGs were precipitated by 0.2% cetyltrimethylammonium bromide (Merck, Darmstadt, Germany) for 1 h. The gels were exposed to a radiosensitive film Multipurpose (Packard Instruments Co.) for 24 h, identified in Cyclone® system (Storage Phosphor system-Packard Instr) and quantified using the Opti Quanti® software. GAGs extracted from each cell type were submitted to enzymatic degradation with heparitinases I and II from Flavobacterium heparinum for HS disaccharide analyses [36]. The degradation products were then analyzed in a PhenoSphere™ SAX 80 Å LC HPLC Column 150 × 4.6 mm. The Δ-disaccharides were eluted in a linear gradient of 0-1 M NaCl for 30 min at a flow rate of 1 ml/min. Individual fractions (0.5 ml) were collected and counted on a Micro-Beta counter. HS disaccharides were generated for three independent experiments and the products of digestion combined prior to analysis to allow detection. Hence, the results represent an overall trend but, cannot be further analyzed statistically.

Immunofluorescence
Transfected cells were seeded on coverslips at a concentration of 10 5 cells/ml. After 3 days, cells were fixed in methanol:acetone (1:1) for 2 min and incubated with primary antibody anti-SULF2 (H-80, Santa Cruz Biotechnology, CA, USA), polyclonal anti-human vimentin produced in goat (Santa Cruz Biotechnology, CA, USA), monoclonal anti-human-β-catenin produced in mouse (MAB13291-100, R&D Systems, MA, USA); Alexa 594 conjugated phalloidin (Invitrogen, Life Technologies Corporation, CA, USA) in PBS containing 5% FBS for 1 h. Subsequently, cells were incubated with secondary antibody conjugated with a fluorescent marker diluted 1:200 in PBS for 40 min in the dark. Cell nuclei were stained with DAPI 1:1000 in PBS with 0.01% saponin for 30 min. The controls were performed by omitting the primary antibody. The staining was observed and analyzed with a fluorescence microscope Nikon E-600 confocal microscope and LSM -510 NLO (Zeiss, Germany).

Viability assay
For the colorimetric proliferation assay, 10 4 cells/well were cultured in 96-well plates. After different times, cells were incubated with 20% of the dye bromide [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT, 5 mg/ml) (Sigma Chemical Co., MO, USA). For 2 hours at 37°C. The medium was carefully removed and formazan crystals produced were solubilized by addition of DMSO (MP Biomedicals, OH, USA). The plates were shaken for 10 min and the absorbance was measured in EXL800 ELISA plate reader, Universal MICROPLAT Reader (Bio-TEK Instruments, Inc.) at 540 nm. Cell viability was estimated by comparing the absorbance values with the controls at different times with the absorbance values of the controls.
Wound healing assay 2.10 5 cells/well were seeded in 24-well plates. After reaching confluence, a scratch was performed using a 200 μl pipette tip in the center of the plate. Closure of the wound was monitored using an inverted optical microscope (Zeiss, Germany) and images obtained by camera (Sony Cyber-shot) attached to the microscope.
Cell invasion assay 2.10 5 cells were seeded in Millicell® chambers (Millipore, MA, USA) containing polycarbonate membranes with pore diameter of 8 μm in medium without FBS. These chambers were placed in 24-well plates containing media with 10% FBS in the lower chamber. After 24 hours at 37°C and 5% CO 2 , the membranes were washed thoroughly with 10 mM PBS, fixed for 30 min in 4% paraformaldehyde, and stained with 0.2% crystal violet for 10 min. The remaining cells on the upper chamber were removed with a cotton swab. The cells were observed using an inverted optical microscope with photographic images obtained by camera (Sony Cyber-shot) attached to the microscope. To quantify cell migration, stained cells were solubilized in 10% acetic acid and absorbance was detected at 560 nm.

Colony formation assay (soft agar)
24-well plates were coated with 300 μl of 0.7% agarose and maintained at 4°C for 30 min. 6.10 3 cells were resuspended in medium containing 0.35% agarose and plated on plates previously covered with agarose. Cells were kept at 37°C with 5% CO 2 for 1 h, when it was added the respective culture medium of each cell. Formation of colonies was followed for 20 days. The colonies were counted and measured using an inverted optical microscope (Zeiss, Germany).

Co-cultures of prostate cancer and fibroblasts
Sterile glass cloning rings (O-rings) were placed on top of glass coverslips in 24-well culture dishes. Human fibroblasts isolated from amniotic fluid were kindly donated by Dr. Walter Pinto Júnior. Fibroblasts were seeded around the O-ring at a density of 1.5.10 4 cells per well in DMEM containing 10% FBS. Prostate cancer cells were seeded inside the O-ring at a cell density of 0.5x10 4 cells per well in RPMI containing 10% FBS. The cells were maintained in culture for 48 h (37°C, 5% CO 2 ). The O-rings were then removed, and the cells were maintained in culture until the cells spread out into the O-ring area (2 days). The cells were sequentially fixed using 4% paraformaldehyde and submitted to immunofluorescence, as previously described.

Statistical analyses
Statistical analyzes were performed using Student's T test in Microsoft Excel software (Microsoft, WA, USA). The results were presented as the mean ± standard deviation of triplicates of each experiment and were considered statistically significant if p ≤ 0.05. All experiments were performed three times, unless stated otherwise.

SULF2 expression in normal and prostate cancer cells
SULF2 gene expression was analyzed in RWPE-1 normal prostate epithelial cells and in LNCap, PC3 and DU-145 prostate cancer cells. Total RNA was extracted from the cells, and gene expression was analyzed by semi-quantitative electrophoresis after PCR. Our result showed that SULF2 is similarly expressed in normal and prostate cancer cells ( Figure 1A). Subsequently, we transfected RWPE-1 normal cell, PC-3 and DU-145 cancer cells with the expression plasmid pcDNA 3.1 containing the SULF 2 gene. As control, we transfected the same cells with the empty vector. The cells were clonally selected and the transfection efficiency confirmed by quantitative RT-PCR, western blotting and immunofluorescence. We observed a significant increase (50-fold) of SULF2 gene expression in the transfected cells ( Figure 1B). The overexpression of SULF2 did not affect the mRNA levels of SULF1 in all the transfected cells (data not shown). Moreover, western blotting analyzes demonstrated that SULF2 was increased in cells extracts, but mainly in the medium ( Figure 1C). The increased expression of SULF2 was also observed by immunofluorescence ( Figure 1D).

SULF2 enzymatic activity in prostate cancer cells
In order to analyze whether the forced overexpression of SULF2 gene resulted in up-regulation of the active enzyme, we verified the content of sulfated HS in PC3 and DU-145 transfected cells. Indeed, we observed a decrease of approximately 50% of sulfated HS in all of the compartments studied, medium, cell, and ECM, in both cells (Figure 2A). We also performed the analyses of HS disaccharides from PC3 and DU-145 transfected cells, using a strong anion-exchange (SAX) column. It was possible to observe an expressive decrease of the trisulfated disaccharide UA(2S)-GlcNS(6S) ( Figure 2B). Our result is consistent with previous data from the literature, which describes that the trisulfated disaccharide from HS is the main substrate for both SULF1 and SULF2 [24,26,37,38].

Consequences of SULF2 overexpression in cell viability and migration
After analyzing the effects of SULF2 forced expression on the structure of HS from prostate cancer cells, we studied the differences on cell viability and migration. Initially, cell viability was measured by MTT colorimetric assay. The overexpression of SULF2 had no effect on the viability of RWPE-1 cells ( Figure 3A). However, both prostate cancer cells, PC3 and DU-145 presented increase on cell viability. The migration was also analyzed by wound healing assay. A scratch was performed on confluent cell cultures and the cells were allowed to migrate for 24 h. Interestingly, the normal prostate epithelial cell line RWPE-1 transfected with SULF2, did not present any increase on cell migration ( Figure 3B). However, prostate cancer cells showed a robust migratory phenotype. These results indicate that forced expression of SULF2 increased cell viability and migration solely on prostate cancer cells, but did not enhance these characteristics on normal cells.

Effects of SULF2 knockdown on prostate cells
In order to confirm that the previous cell behaviors were indeed acquired due to the overexpression of SULF2, we studied the consequences of SULF2 knockdown on the same prostate cells. For this purpose, prostate cells were transfected with siRNAs targeting SULF2 mRNA. Gene silencing was confirmed by quantitative RT-PCR ( Figure 4A), and the levels of SULF2 mRNA were reduced in at least 95%. After that, the cells were submitted to viability and migration assays as previously described. Interestingly, the knockdown of SULF2 reduced the cell viability of RWPE-1 normal prostate cells, as well as reduced the cell viability of DU-145 and PC3 prostate cancer cells ( Figure 4B). In addition, SULF2 silencing impaired cell migration ( Figure 4C). Apparently, the overexpression of SULF2 was not sufficient to increase normal epithelial prostate cells growth and migration. However, the enzyme must be important for these cell properties, since its knockdown also decreased normal prostate cells migration and viability.

SULF2 overexpression increases colony formation and invasion of prostate cancer cells
In order to determine whether SULF2 increase was able to exacerbate the tumorigenic phenotype of prostate cancer cells in vitro, PC3 and DU-145 cells were submitted to colony formation and transmigration assays. For transmigration assay, cancer cells were plated on the top of membranes with pore diameter of 8 μm and allowed to migrate for 24 h. DU-145 cancer cells transfected with SULF2 presented a 3-fold increase on migration through the membrane, and PC3 cancer cells transfected with SULF2 presented a 2-fold increase on migration ( Figure 5A). For colony formation assay, the cancer cells were embedded in soft agar and the colonies growth was followed for 20 days. DU-145 and PC3 prostate cancer cells overexpressing SULF2 presented an increase of 3fold on the size of the colonies formed, compared to cells transfected with empty vector ( Figure 5B). Equally important, the SULF2 overexpressing cells also formed more colonies on soft agar ( Figure 5B).

SULF2 overexpression increases the expression of epithelial-mesenchymal transition markers
The epithelial-mesenchymal transition (EMT) is a key developmental program that is often activated during cancer invasion and metastasis [39]. Recently, EMT markers have been found to confer malignant traits, such as motility, invasiveness and resistance to apoptosis [39]. Since we have observed an increase of these characteristics on prostate cancer cells with forced expression of SULF2, we decided to analyze some EMT markers in these cells. Thus, prostate cancer cells were immunostained for CD44, vimentin, and N-cadherin and the presence as well as their quantity analyzed by flow cytometry. Indeed, PC3 and DU-145 prostate cancer cells overexpressing SULF2 exhibited increased levels of CD44, vimentin, and N-cadherin ( Figure 6). Hence, our results indicate that prostate cancer cells overexpressing SULF2 become more undifferentiated, which is in agreement with the increased cell growth and migration presented by them.
Previous studies have shown that SULFs regulate WNT signaling pathway [25,37]. Therefore, we analyzed the presence of WNT 3A and β-catenin in DU-145 and PC3 cells overexpressing SULF2. By flow cytomery, we observed an increase of active unphosphorylated β-catenin in both SULF2 transfected cells ( Figure 7A). Moreover, the proportion of cells presenting both WNT 3A and β-catenin ( Figure 7B) also increased. DU-145 cancer cells overexpressing The content of GAGs from these compartments was analyzed through agarose gel electrophoresis. The gel was exposed to a radiosensitive screen and quantification was performed by densitometry with Opti-Quanti Software (A). Each sample was digested using a mixture of heparin lyases and analyzed on a PhenoSphere™ 5 μm SAX 80 Å LC Column 150 × 4.6 mm. Δ-disaccharide were eluted with a linear gradient of NaCl 0-1 M over a 30-min period at a flow rate of 1 ml.min -1 . Individual fractions (0.5 ml) were collected and counted using a micro-beta counter (B). The bars indicate the average of three independent experiments. The arrows show the trisulfated disaccharide UA(2S)-GlcNS(6S). *P ≤ 0.05. SULF2 showed 33.7% of cells double stained for WNT 3A and β-catenin respectively, in comparison to 20.7% in control cells transfected with empty vector. PC3 cells overexpressing SULF2 showed 40.2% of cells double stained for WNT 3A and β-catenin respectively, compared to 26.0% in control cells transfected with empty vector. Finally, by confocal microscopy, we could detect β-catenin located close to the cells membrane in control cells, while cells with forced expression of SULF2, presented a nuclear staining for β-catenin (arrows, Figure 7C).

Effects of SULF2 overexpression in stroma-cancer co-cultures
The O-ring co-culture system is an attempt to mimic a tumor microenvironment. The stromal cells are seeded and cultured immediately around the tumor cell line, allowing cell-cell contact besides establishing a gradient of soluble factors throughout the stromal cells, similar to that found in tissues [35]. As expected, after 2 days of culture, PC3 and DU-145 prostate cancer cells overexpressing SULF2 had already connected to fibroblasts, whereas the control cells transfected with empty vectors  MTT viability assay (B) and wound healing assay (C) were performed as described previously. (CTRL NEG: scramble siRNA sequence). *P ≤ 0.05. remained distant from fibroblasts ( Figure 8). This probably occurred due to the increased migration presented by PC3 and DU-145 with forced expression of SULF2. Moreover, by using this system coupled to immunocytochemistry, we analyze the region of intersection between stromal cells and tumor cells (Figure 8). We observed an apparent increase in vimentin in the intersection area between cancer cells and fibroblasts for both transfected cells.

Discussion
Recent progress in cancer biology suggests that a limited number of pathways are critical for initiating and maintaining deregulated cell proliferation, and migration, which are the major cellular alteration responsible for cancer advance and metastasis [32]. New agents in development, target several of these critical pathways and many of them have ligands to which cell-surface or ECM PGs act as co-receptors [16]. Studies of the newly discovered family of HS 6-O-endosulfatases, SULF1 and SULF2, suggest that HSPGs in the ECM or on the cell surface can sequester growth factor ligands and cytokines in a sulfationdependent manner and release them when desulfated by heparan-degrading endosulfatases [25].
The SULFs are a family of enzymes that are secreted via the Golgi and are located on the cell surface or released into the ECM. These enzymes selectively remove the 6-Osulfate groups from HS, with preference for those present in trisulfated disaccharides [24,25]. Importantly, such partial and oriented desulfatation can differentially modify the interaction of protein ligands to HS. When SULFs remove the 6-sulfate, they trigger the release of HSPGs ligands, allowing them to act in cells.
A limited number of studies reported the involvement of SULFs in prostate cancer. SULF1 is present in prostatic stromal cells in the transition regions between cancer and stroma and SULF2 chromosome locus is associated to prostate cancer susceptibility regions [33,34].
In the present study, we found that SULF2 acts as an oncogenic protein in prostate cancer cells once cells overexpressing SULF2 presented increased cell viability and migration, which has already been observed in different tumor cells previous studied, where the overexpression of SULF2 had also been performed [25,31,32]. These effects were reverted when SULF2 mRNA was silenced using siR-NAs. Interestingly, SULF2 knockdown on normal prostate epithelial cells, RWPE-1, has also decreased cell growth and migration.
Moreover, DU-145 and PC3 prostate cancer cells with forced expression of SULF2 presented an augmentation of invasiveness and tumor colony formation in vitro. Therefore, SULF2 appears to act as a proto-oncogene in prostate cancer cells, increasing their ability to growth and migrate. However as cancer is a multifactorial disease, the augmentation of SULF2 alone was not sufficient to produce these effects in normal prostate epithelial cells.
In order to determine the mechanisms involved in the increase of cell growth and migration, we investigated the effects of SULF2 overexpression on EMT markers. In recent years, EMT has been found to confer malignant characteristics to cells, such as motility, invasiveness, and resistance to apoptosis, on neoplastic cells [40][41][42]. During the process of tumor metastasis, which is often enabled by EMTs [43], disseminated cancer cells would seem to require self-renewal capability, similar to that exhibited by stem cells, in order to establish new focus of metastases. This raises the possibility that the EMT process, which enables cancer cell dissemination, may also provide self-renewal capability to the disseminating cancer cells.
We found that the up-regulated SULF2 cells increased EMT markers, including CD44, vimentin and N-cadherin. CD44 is a multifunctional class I transmembrane glycoprotein [44,45] that generally acts as a specific receptor for hyaluronic acid, promoting migration in normal cells. Also, CD44 presents cytokines and chemokines to their complimentary receptors on the cellular membrane [41]. It is mainly associated with proteins that monitor the extracellular changes and is critical in regulating cell adhesion, proliferation, growth, survival, motility, migration, angiogenesis, and differentiation [39], and is highly  expressed in almost every cancer cell in its standard or variant form [45,46].
Loss of E-cadherin expression in cancer cells may be associated with gain of N-cadherin expression, leading to a fibroblastic phenotype with increased motility and invasive potential in vitro and in vivo [47][48][49]. Moreover, the reduced expression of E-cadherin, abnormal expression of N-cadherin, transformation from E-cadherin to N-cadherin and the increased expression of TGF-β 1 and Twist play an important role in the occurrence and development of prostate cancer [50,51].
Vimentin is an intermediate filament which supports cellular mechanostructural integrity participating in cell adhesion, migration, survival, and signaling [52,53]. High vimentin expression has been reported in bone metastasis of prostate cancer and has been implicated in prostate cancer cell invasion [54,55]. Consistent with this result, we observed an up-regulation of vimentin expression in cocultures of stromal cells and metastatic prostate cancer cells. Accordingly, our previous work had already demonstrated an increased expression of vimentin when stromal cells were exposed to prostate tumor cell lines, besides changes in its cellular arrangement from punctate to a fibrilar distribution [53].
A series of studies have demonstrated that the WNT/ β-catenin signaling pathway is one of the major pathways involved in EMT regulation in different types of tumor, including prostate cancer [56][57][58][59]. Interestingly, one of the known consequences of SULF overexpression is the promotion of WNT signaling pathway. According to the model proposed by Ai et al. [32], the action of SULFs weakens the association of WNT to HSPGs at the cell surface, which allows the WNT to activate its Frizzled signal transducing receptors. β-catenin is stabilized by WNT and translocated into the nucleus, where it binds to the T cell factor and lymphoid enhancer factor (TCF/LEF) family of transcriptional cofactors. Successively, β-catenin-TCF/LEF complexes activate transcriptional cascades that induce EMT programs.
Our previous study, with human colorectal cancer cell lines, confirmed that the forced expression of SULFs results in increased WNT 3A signaling pathway, evinced by the accumulation of active unphosphorylated β-catenin [60]. Consistent with this, prostate cancer cells overexpressing SULF2 presented an increase of WNT 3A and β-catenin double-stained cells, in addition to a nuclear location of β-catenin. Therefore, our results indicate that the WNT/β-catenin signaling pathway could be regulating the EMT in those cells.
In summary, SULF2 overexpression increases metastatic prostate cancer cells growth and migration, leading to an augmentation of tumor colony formation and invasiveness. In addition, forced expression of SULF2 resulted in an increment of EMT markers and in a stronger contact between prostate cancer cells and stromal cells. Therefore, SULF2 may contribute to the metastatic process in prostate cancer.

Conclusions
Our results demonstrated a possible pro-tumorigenic role of SULF2 in prostate cancer. However, due to the limitations of in vitro experiments, further in vivo studies are necessary to better understand the complex function of SULF2 in prostate cancer. As previous studies have already indicated an involvement of SULFs in different types of tumors [22][23][24][25][26][27], and as there are some evidences of the involvement of SULFs in prostate cancer [28,29], we believe that the study of this enzyme will contribute to a better understanding of this disease, as well as emerge with new therapeutic opportunities.