TMPRSS2-ERG and other gene fusions involving ETS factors and genes with strong promoter elements are common in prostate cancer. Although ERG activation has been linked to invasive properties of prostate cancers, the precise mechanisms and pathways of ERG-mediated oncogenesis remain poorly understood. Here, we show that ERG knockdown in VCaP prostate cancer cells causes an activation of cell adhesion, resulting in strongly induced active β1-integrin and E-cadherin expression as well as changes in WNT signaling. These observations were corroborated by data from ERG-overexpressing nontransformed prostate epithelial cells as well as gene expression data from clinical prostate cancer samples, which both indicated a link between ERG and epithelial-to-mesenchymal transition (EMT). Upregulation of several WNT pathway members was seen in ERG-positive prostate cancers, with frizzled-4 (FZD4) showing the strongest overexpression as verified by both reverse transcription-PCR and immunostaining. Both ERG knockin and knockdown modulated the levels of FZD4 expression. FZD4 silencing could mimic the ERG knockdown phenotype by inducing active β1-integrin and E-cadherin expression, whereas FZD4 overexpression reversed the phenotypic effects seen with ERG knockdown. Taken together, our results provide mechanistic insights to ERG oncogenesis in prostate cancer, involving activation of WNT signaling through FZD4, leading to cancer-promoting phenotypic effects, including EMT and loss of cell adhesion. Cancer Res; 70(17); 6735–45. ©2010 AACR.

Gene fusions between the prostate-specific and androgen-responsive TMPRSS2 gene and ETS transcription factors are found in roughly 50% of all prostate cancers. Most frequently, the fusion partner is ERG (ETS-related gene), followed by ETV1, ETV4, and ETV5 (14). As a result of these oncogenic fusions, androgens and the androgen receptor (AR) signaling drive ETS factor overexpression, resulting in prostate cancer development and progression (57). ERG overexpression has also been described in acute myeloid leukemia and acute T-lymphoblastic leukemia, whereas the EWS-ERG gene fusion is characteristic to a subset of Ewing's sarcomas (8). High ERG expression is an independent risk factor in acute T-lymphoblastic leukemia predicting poor relapse-free survival (9, 10). TMPRSS2-ERG gene fusions occur early during prostate carcinogenesis. The fusion may have prognostic significance depending on the cytogenetic nature of the rearrangement, overexpression of the ERG gene, and disease stage (1114).

The precise molecular events and signaling mechanisms contributing to ERG-initiated oncogenesis are incompletely understood. ERG has been shown to regulate invasion through the plasminogen activation pathway (15). The phenotypes in mouse models showed prostatic intraepithelial neoplasia formation, induction of β4-integrin expression, and loss of stromal cells in prostates from transgenic mice with probasin promoter–driven ERG overexpression (1517). Apparently, ERG overexpression in prostate epithelial cells is sufficient to induce neoplastic changes but not to produce carcinoma (15, 16). Coexpression of AR and ERG in mouse prostate epithelial cells induced invasive cancer with poorly formed, tightly packed prostate glands and loss of smooth muscle actin staining in stroma cells (17), suggesting that AR and ERG may act synergistically in the progression of PIN lesion to invasive adenocarcinoma. ERG is also associated with aberrant phosphoinositide 3-kinase pathway, which promotes prostate cancer progression (17). ERG overexpression in prostate tumor cells has been associated with induction of C-MYC and subsequent neoplastic changes (18). Finally, the results from our previous bioinformatics analyses of ERG coexpression in clinical prostate cancers suggest that ERG modulates epigenetic programming (19).

In this study, we examined the function of the ERG oncogene by knockdown and knockin experiments in prostate cell lines as well as carried out molecular pathologic analysis of ERG-positive clinical prostate cancers. We found that ERG silencing promotes cell adhesion and upregulates β1-integrin and E-cadherin expression as well as WNT signaling. We then observed that frizzled-4 (FZD4) was often co-overexpressed with ERG in clinical prostate cancers and was systematically modulated by ERG manipulation in vitro. Silencing of FZD4 mimicked the phenotypic effects seen with ERG knockdown, and FZD4 overexpression reverted the phenotypic effects seen with ERG knockdown. These observations link ERG activation in vitro and in clinical prostate cancers with activation of the WNT pathway through FZD4, resulting in epithelial-to-mesenchymal transition (EMT) and changes in cell adhesion.

Patient data and prostate cancer cell lines

Sixteen localized prostate cancer samples were taken freshly from total prostatectomy specimen. The patients were not treated by any other modalities (radiation, hormonal therapy) with the subsequent knowledge on cancer grade. The study protocol was approved by the local ethical committee, and informed consent of the patients was obtained. Frozen tissue blocks were sectioned, and 4 to 6 μm sections were collected for DNA and RNA extractions. Clinical data of the patients are summarized in Supplementary Table S1. The immortalized prostate cell line RWPE1 and the TMPRSS2-ERG gene fusion–positive tumor cell line VCaP were obtained from the American Type Culture Collection and cultured in RPMI medium.

Generation of GFP-ERG and GFP-FZD4 constructs

The full-length ERG cDNA was PCR amplified based on an open reading frame (ORF) clone purchased from Origene, Inc. (SC108516), by using ERG forward (5′-CAATCTCGAGCTATGGCCAGCACTATTAAGG AAGC-3′) and reverse (5′-CAATCCCGGGTTAGTAGTAAGTGCCCAGATGAGAAG-3′) primers with XhoI and XmaI restriction sites and Phusion High-Fidelity DNA polymerase (Finnzymes). The amplified cDNA was digested with XhoI and XmaI restriction enzymes, excised from agarose gels, and cloned into pEGFP-C1 (Invitrogen). Sequencing reactions using the same primers as in PCR were prepared by using the ABI BigDye Terminator V3.1 cycle sequencing kit and analyzed on the ABI 3100 genetic Analyzer (Applied Biosystems). The FZD4-GFP overexpression construct was ordered from Origene (RG217286). The pEGFP-ERG and pEGFP control vectors were transfected into RWPE1 cells by using the FuGENE-6 transfection reagent (Roche). RNA and protein samples were prepared 24, 48, and 72 hours after transfection. Harvested cells were used for RNA and protein isolation.

Western blot analysis

Cells were plated at 70% confluence and left to attach overnight. Protein content was measured using the Bio-Rad protein assay kit. Twenty micrograms of total protein were taken up in Laemmli buffer containing 3% of β-mercaptoethanol, denatured for 5 minutes at 95°C, separated on a 7% SDS-polyacrylamide gel, and transferred to a Protran nitrocellulose transfer membrane (Schleicher & Schuell). Green fluorescent protein (GFP; 1:1,000, rabbit polyclonal, A11122, Molecular probes), ERG (1:1,000, rabbit polyclonal, SC-353, Santa Cruz), and FZD4 (1:1,000, goat polyclonal, SC-66450, Santa Cruz) antibodies were used for Western blotting. Antibody against HSPA8 (heat shock 70 kDa protein 8, 1:1,000, rat monoclonal, SPA-815, Stressgen) was used as the loading control. Proteins from the cytosolic, membranous, and nuclear fractions were isolated using a Compartment Protein Extraction Kit (Chemicon) according to the manufacturer's instructions. The signal was detected with a 1:5,000 dilution of horseradish peroxidase (HRP)–conjugated anti-goat, anti-rat, or anti-rabbit secondary antibodies (Amersham ECL-HRP Linked Secondary Antibodies). ECL (GE Healthcare) Western Blotting Detection Systems was used for Western blot analysis.

siRNA and lentiviral shRNA knockdown of ERG in VCaP cells

The most effective siRNA against ERG (SI03089443, Qiagen; target all the isoforms of ERG gene), FZD4 (SC-39983, Santa Cruz), or siCONTROL Non-Targeting siRNA (1027310, Qiagen) was transfected using the Lipofectamine 2000 reagent (Invitrogen) and tested for knockdown by quantitative reverse transcription-PCR (qRT-PCR) using the universal probe library (Roche). Samples were prepared after 24, 48, and 72 hours. To generate stable VCaP cells with reduced ERG mRNA expression, Mission ERG shRNA lentiviruses (NM_004449.3-2067s1c1, NM_004449.3-214s1c1 or 214, NM_004449.3-643s1c1 or 643, NM_004449.3-1404s1c1, and NM_004449.3-823s1c1) and Mission nontarget shRNA Lentiviral Particles (Mission/Sigma) with a titer of >106 TU/mL were used to infect VCaP cells. Subsequently, stable clones were selected using puromycin (2 μg/mL).

Fluorescence-activated cell sorting analysis

Twenty thousand cells per well were used for fluorescence-activated cell sorting (FACS) analysis of stable ERG 214 shRNA– and FZD4 siRNA–treated VCaP cells as well as for the respective controls. A similar protocol was used for the analysis of GFP-ERG–, GFP-FZD4–, and GFP-transfected RWPE1 cells. Cells were trypsinized and fixed with 2% paraformaldehyde (PFA) with 0.5% Triton X-100 and resuspended in PBS. Cells were washed and stained with active β1-integrin and E-cadherin (the same antibodies as those used in immunofluorescence staining; 1:1,500 dilutions) overnight at 4°C. After washings, cells were stained with Alexa 647–labeled secondary antibodies (1:2,000) for 1 hour at room temperature. Cells were washed and the mean fluorescence intensity was measured using Accuri C6 Flow Cytometer.

Cell adhesion assay

Plates (96 wells) were coated with laminin or fibronectin (0.50 μg/mL) overnight and blocked with 0.1% bovine serum albumin (BSA; 1 hour, 37°C). ERG siRNA and scrambled siRNA–transfected VCaP cells were harvested 72 hours after transfection. Trypsin was inactivated with 0.2% (wt/vol) soybean trypsin inhibitor (Sigma). Cells were suspended in 0.5% BSA in serum-free RPMI, seeded (5,000 cells per well) on the plates, and allowed to adhere for 30 minutes at 37°C. After washing with PBS, cells were fixed (4% PFA, 10 minutes) and stained with propidium iodide (PI). Adhesion was measured by counting the number of PI-stained cells using Acumen Assay Explorer (absorbance at 488 nm).

Real-time quantitative PCR analysis

Total cellular RNAs were isolated using the Trizol reagent. For cDNA synthesis, 100 ng of total RNA were reverse transcribed with the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). qRT-PCR analysis was done on an Applied Biosystems 7900HT instrument, using specific primers designed by the Universal Human Probe Library Assay Design Center. Results were analyzed using SDS 2.3 and RQ manager software (Applied Biosystems), and the relative expression of mRNA was determined using β-actin as an endogenous control. Data are from two separate biological experiments, with triplicate samples. Primers are listed in Supplementary Table S2.

Illumina bead array expression profiling and bioinformatics analysis

Total RNA was isolated from GFP and GFP-ERG–transfected RWPE1 cells as well as from transfected ERG siRNA–, ERG 214 shRNA–, and ERG 643 shRNA–treated VCaP cells in duplicate. The integrity of the RNA was monitored before hybridization, using the Bioanalyzer 2100 (Agilent). Purified total RNA (500 ng) was amplified with the Total Prep Kit (Ambion), and the biotin-labeled cRNA was hybridized to Sentrix HumanRef-8 Expression Bead Chips (Illumina). The arrays were scanned with the Bead Array Reader. Raw Illumina data were processed using the R/Bioconductor software and the lumi package. Variance-stabilizing transformation and quantile normalization were applied as preprocessing steps and for background correction. Differential expression of genes from two independent biological replicates of each condition was identified by fitting a linear model for each gene using the limma package. For network analyses, data were uploaded into the MetaCore Analytical Suite version 2.0 (GeneGo). The raw data of the microarray analysis have been deposited in the GEO database (GSE16671).

Gene tissue index method

The gene tissue index (GTI) algorithm is comparable with the cancer outlier profile analysis (COPA) algorithm in detecting outliers (2). It is designed to lower the bias created by other outlier statistics that depend on the mean, mean absolute deviation, and outlier sums by normalizing the score based on the number of samples. A positive GTI score means that there are more outliers in the target group (e.g., cancer group) than in the reference group (e.g., normal group). A GTI score close to zero indicates no difference in the expression levels of that gene in the two groups.

Immunofluorescence and confocal microscopy

siRNA- and shRNA-treated cells were grown on coverslips in 24-well plates and fixed with 2% PFA in PBS, permeabilized with 0.5% Triton X-100 in fixative solution for 15 minutes at room temperature, rinsed in PBST (PBS, 0.01% Tween 20), and blocked with 20% horse serum for 60 minutes in a humidified chamber (1× PBS/0.05% NaN3, 25 μL/coverslip). After fixing and blocking, coverslips were rinsed again in PBST and labeled with 25 μL primary antibody overnight. FZD4, active β1-integrin (mouse, monoclonal, 12G10, Abcam), activation state–specific phosphor-epitope (20), and E-cadherin (mouse, monoclonal, ab1416, Abcam) primary antibodies (1:100 dilution) were used for immunostaining. The coverslips were washed and labeled with 25 μL of secondary antibodies conjugated with Alexa 488, Alexa 555, or Alexa 647 (1:500 dilution in PBST, Invitrogen) for 60 minutes. After PBST washes, coverslips were mounted with 3 μL Vectashield with 4′,6-diamidino-2-phenylindole (DAPI; H-1200, Vector Laboratories). A confocal laser scanning microscope (Axioplan 2 with LSM510) equipped with Plan-Apochromat 63×/1.4 oil objectives was used in imaging. Confocal images represent a single Z-section of 1 μm. Image J software was used to analyze the confocal images (21).

Immunohistochemical staining

Based on the qRT-PCR results, five ERG-positive and five ERG-negative prostate cancer samples as well as two normal prostate samples were selected for FZD4 staining by immunohistochemistry. Frozen sections (6 μm thick) were cut to silan-coated glasses and dried at room temperature for 30 minutes, fixed at −20°C temperature with acetone for 10 minutes, and again dried for 10 minutes at room temperature. Slides were washed in Tris-buffered saline (TBS; pH 7.4–7.6) for 3 × 5 minutes. Endogenous peroxidase was blocked with 0.3% hydrogen peroxide in TBS for 20 minutes. Slides were washed and incubated in goat normal serum (15 μL goat normal serum in 1 mL TBS) for an hour in a moist chamber. Affinity-purified IgG to human ERG (1:100, rabbit polyclonal, Santa Cruz, SC-353) and FZD4 (1:100, goat polyclonal, Santa Cruz, SC-66450) was diluted with 1% BSA in TBS and the slides were incubated with the primary antibody at 4°C overnight in a moist chamber. For the negative control, one slide in each staining procedure was incubated with rabbit normal serum (primary antibody omitted). After washing with TBS (3 × 5 min), the slides were incubated with biotinylated anti-goat secondary antibodies (in TBS-buffered 1% BSA) at room temperature, in a moist chamber, for 30 minutes. The slides were washed and incubated with Vectastain ABC reagent (Vector Laboratories) at room temperature in a moist chamber for 30 minutes. Slides were washed and stained with diaminobenzidine, counterstained with Mayer's hematoxylin, dehydrated, treated with xylene, and mounted.

T-cell factor/lymphoid enhancer factor reporter assay

The Cignal T-cell factor/lymphoid enhancer factor (TCF/LEF) reporter (GFP) assay kit (CCS-018G) was used to measure the activity of WNT signaling in prostate epithelial cells in response to ERG modulation by using inducible TCF/LEF-responsive GFP reporter construct. Control GFP reporter was used to confirm the transfection efficiency. RWPE1 cells (20,000) were cultured in 96-well cluster plates and cotransfected with GFP or GFP-ERG overexpression construct (200 ng/well) together with TCF/LEF reporter construct (200 ng/well) for 48 and 72 hours. Similar experiments were done for ERG knockdown using both siRNA transfection– and stable shRNA expression–based approaches in VCaP cells. Also, FZD4 siRNA–transfected VCaP cells were studied with appropriate scrambled controls at 48 and 72 hours. All transfections were done at least in triplicate, and GFP intensity was measured by Acumen Explorer. Graph Pad Prism 4 was used to plot the results and calculate the significance.

Silencing and overexpression of ERG in prostate cells

Five different shRNA lentiviral constructs were used to induce permanent knockdown of the ERG gene in VCaP cells that contain the TMPRSS2-ERG fusion gene (2, 22, 23). Selection with puromycin resulted in two stable cell lines with different ERG shRNAs (nos. 214 and 643). qRT-PCR and Western blot analyses (Fig. 1A) indicated 40% to 60% knockdown compared with a scrambled shRNA control. The partial knockdown efficacy may indicate that a more complete knockdown was not compatible with long-term cell survival. The ERG 214 shRNA cell line showed the most efficient reduction in ERG mRNA and was therefore subsequently used in functional experiments. Also, 10 different siRNAs targeting ERG were tested in transient transfection studies in VCaP cells, with those providing >50% knockdown chosen for further experiments (Fig. 1B). For overexpression studies, the full-length ORF for ERG was PCR amplified, cloned in-frame to enhanced GFP vector, and sequence verified. This construct and a control vector expressing only GFP were transfected into the immortalized, nonmalignant RWPE1 cells. The results indicate that the GFP-ERG fusion protein was almost exclusively nuclear whereas GFP localized to the cytoplasm (Supplementary Fig. S1A). Similar results were obtained by Western blot analysis of different subcellular protein fractions, indicating that the GFP-ERG fusion protein is effectively translocated into the nucleus (Supplementary Fig. S1B).

Figure 1.

Validation of reduced ERG expression in ERG shRNA or siRNA-treated VCaP cells. A, qRT-PCR analysis, and Western blot analysis and its quantification to measure ERG expression in response to ERG silencing by stable expression of shRNA molecules (ERG 214 or 643) in VCaP cells. B, qRT-PCR analysis, and Western blot analysis and its quantification to measure ERG expression in response to ERG siRNA transfection at the 48 h time point in VCaP cells. To obtain the relative expression, ERG mRNA was normalized to β-actin mRNA (*, P < 0.02) and ERG protein was normalized to HSPA8. C, morphologic phenotype of scrambled shRNA–, ERG 214 shRNA–, and ERG 643 shRNA–expressing VCaP cells after puromycin selection, and scrambled siRNA– and ERG siRNA–treated VCaP cells 72 h after transfection. D, cell-cell adhesion analysis of ERG siRNA versus scrambled siRNA control in VCaP cells. siRNA-treated cells were allowed to attach on the laminin- or fibronectin-coated plates at different time points, and the PI-labeled cells were counted by using Acumen Explorer (**, P < 0.005; ***, P < 0.001).

Figure 1.

Validation of reduced ERG expression in ERG shRNA or siRNA-treated VCaP cells. A, qRT-PCR analysis, and Western blot analysis and its quantification to measure ERG expression in response to ERG silencing by stable expression of shRNA molecules (ERG 214 or 643) in VCaP cells. B, qRT-PCR analysis, and Western blot analysis and its quantification to measure ERG expression in response to ERG siRNA transfection at the 48 h time point in VCaP cells. To obtain the relative expression, ERG mRNA was normalized to β-actin mRNA (*, P < 0.02) and ERG protein was normalized to HSPA8. C, morphologic phenotype of scrambled shRNA–, ERG 214 shRNA–, and ERG 643 shRNA–expressing VCaP cells after puromycin selection, and scrambled siRNA– and ERG siRNA–treated VCaP cells 72 h after transfection. D, cell-cell adhesion analysis of ERG siRNA versus scrambled siRNA control in VCaP cells. siRNA-treated cells were allowed to attach on the laminin- or fibronectin-coated plates at different time points, and the PI-labeled cells were counted by using Acumen Explorer (**, P < 0.005; ***, P < 0.001).

Close modal

ERG silencing promotes cell adhesion by induction of active β1-integrin and E-cadherin

Next, we explored phenotypic changes in VCaP prostate cancer cells in response to ERG knockdown either by stable shRNA or transient siRNA transfection. ERG shRNA– and ERG siRNA–transfected cells formed round cell clusters that were strongly adherent to culture plates (Fig. 1C). To confirm and quantitate the increased adhesion in response to ERG silencing, we performed cell adhesion assays using laminin and fibronectin matrix–coated plates at 5, 10, and 30 minutes time points. The results indicate that ERG siRNA silencing increased cell-cell adhesion compared with the scrambled control (Fig. 1D). To study if this change was due to the differential expression of known regulators of cell attachment, VCaP cells with a stable ERG shRNA knockdown or transient ERG silencing were stained with antibodies against E-cadherin (CDH1) and active β1-integrin and compared with the respective controls. ERG silencing induced strong active β1-integrin (Fig. 2A) and E-cadherin (Fig. 2B) expression, which may explain the increased cell adhesion of VCaP cells in response to ERG knockdown. FACS analysis was performed to quantify the increased expression of active β1-integrin and E-cadherin in response to ERG 214 shRNA–treated VCaP cells compared with the scrambled shRNA control cells. The results shown in Supplementary Fig. S2 indicate that ERG 214 shRNA significantly increased the expression of active β1-integrin (>2-fold) and E-cadherin (>3.5-fold). These results support the earlier findings indicating increased invasion in response to TMPRSS2-ERG overexpression in prostate epithelial cells.

Figure 2.

ERG silencing induces active β1-integrin and E-cadherin expression. Immunostaining of stable scrambled shRNA and ERG 214 shRNA as well as scrambled siRNA and ERG siRNA exposed in VCaP cells was performed with antibodies against (A) active β1-integrin and (B) E-cadherin. DAPI staining (blue) was used to visualize the nuclei. Taqman quantitative real-time PCR and Western blot analysis of (C) FZD4 expression in RWPE1 cells transiently transfected with GFP-ERG vector and (D) stable ERG 214 shRNA–expressing VCaP cells with their respective control (*, P < 0.05).

Figure 2.

ERG silencing induces active β1-integrin and E-cadherin expression. Immunostaining of stable scrambled shRNA and ERG 214 shRNA as well as scrambled siRNA and ERG siRNA exposed in VCaP cells was performed with antibodies against (A) active β1-integrin and (B) E-cadherin. DAPI staining (blue) was used to visualize the nuclei. Taqman quantitative real-time PCR and Western blot analysis of (C) FZD4 expression in RWPE1 cells transiently transfected with GFP-ERG vector and (D) stable ERG 214 shRNA–expressing VCaP cells with their respective control (*, P < 0.05).

Close modal

ERG modulation affects WNT signaling in vitro

To study the genome-wide consequences of ERG modulation, gene expression changes were measured by Illumina bead arrays in RWPE1 cells 48 to 72 hours after transfection of GFP-ERG compared with GFP only. Furthermore, these results were compared with data from transient ERG silencing in VCaP cells and with the stable VCaP cell lines expressing ERG shRNAs 214 or 643, all with the respective controls. We listed the top differentially expressed genes in response to ERG modulation as Supplementary Tables S3 and S4. The WNT pathway was one of the most systematically altered signaling pathways in the ERG modulation experiments according to bioinformatics analysis by gene set enrichment analysis and MetaCore and GeneGo pathway analyses (Supplementary Table S5). For example, in the ERG-GFP–transfected cells, the genes implicated in WNT signaling were FZD4 and disabled homologue 2 (DAB2; both upregulated) and dickkopf homologue 1 (DKK1; downregulated). In contrast, axin-related protein (AXIN2); lymphoid enhancer binding factor-1 (LEF1); and protein phosphatase 1, catalytic subunit, β isoform (PPP1CB) were all upregulated in the permanent ERG-shRNA knockdown cell line variants. On the other hand, FZD4 and plasminogen activator and urokinase receptor (PLAUR) were downregulated.

FZD4, a WNT receptor, was one of the most consistently differentially expressed genes in response to both ERG knockdown and upregulation in human prostate cells. We confirmed a relationship between FZD4 and ERG by qRT-PCR in ERG-overexpressing RWPE1 cells and in the stable ERG 214 shRNA–expressing VCaP cells. The results indicate that ERG overexpression increased FZD4 expression whereas ERG silencing resulted in decreased FZD4 at both the mRNA and protein levels (Fig. 2C and D).

Coexpression of WNT signaling genes in ERG-positive prostate cancers

To explore the relationship between ERG and the WNT pathway in vivo in clinical prostate cancer specimens, a systematic bioinformatics analysis of genes involved in the WNT pathway was performed. We identified 226 WNT pathway genes from the Biocompare database (http://gspd.biocompare.com/). The GeneSapiens gene expression database (http://www.genesapiens.org) was used to identify those that had both an outlier expression in prostate cancer and were positively correlated with the ERG oncogene expression (24) in 70 ERG-positive and 267 ERG-negative prostate cancer samples compared with samples from 159 nonmalignant prostates. A correlation value >0.3 (P < 0.05) with ERG and a positive GTI outlier score in ERG-positive samples were used for selecting genes to explore in a coexpression network.5

5Mpindi et al., submitted for publication.

The results are shown in Supplementary Table S6. In accordance with our previous study (19), histone deacetylase inhibitor 1 (HDAC1) showed the highest association with ERG-positive prostate cancers. FZD4, identified as an ERG-modulated gene in functional experiments with prostate epithelial cells, was among the top WNT pathway outlier genes correlating with ERG.

We then studied the coexpression patterns of ERG, WNT-pathway genes, and markers for epithelial and mesenchymal differentiation in the 496 prostate tissue samples. The expression data presented as a correlation map (Fig. 3A) indicate that ERG mRNA correlated with WNT pathway activation, increased expression of mesenchymal markers such as cadherins CDH2 and CDH11, and reduced expression of epithelial markers such as the luminal keratins KRT7 and KRT18. To find out if EMT-regulating transcription factors are differentially expressed in response to ERG modulation, SNAI1 and SNAI2 mRNA expression were studied. The results are shown in Supplementary Fig. S3 and indicate that ERG silencing reduced SNAI1 and SNAI2 expression in VCaP cells, whereas increased SNAI1 mRNA was seen in response to ERG overexpression in RWPE1 cells. These results are in accordance with the profiles we observed in the in vitro functional experiments and support the role of ERG in regulating WNT signaling and activation of EMT.

Figure 3.

Coexpression analysis of WNT signaling genes in prostate tissues. A, coexpression analysis of WNT signaling genes in ERG-positive and ERG-negative prostate cancers as well as in normal prostate. Red areas indicate strong, positive correlation between genes; blue areas indicate negative correlation or anticorrelation. The map indicates increased correlation between the WNT pathway genes in ERG-positive tumors, less correlation in ERG-negative tumors, and no correlation in normal prostate tissues. The expression of epithelial and mesenchymal marker genes is shown for reference to visualize the association of induced WNT signaling with EMT. B, validation of increased expression of WNT genes in ERG-positive prostate cancers: The top five genes, HDAC1 (**, P < 0.001), GNAI1 (*, P < 0.03), FZD4 (*, P < 0.01), and PRKD1, were significantly upregulated in ERG-positive (n = 9) compared with ERG-negative (n = 7) prostate cancers. All experiments were done in triplicates.

Figure 3.

Coexpression analysis of WNT signaling genes in prostate tissues. A, coexpression analysis of WNT signaling genes in ERG-positive and ERG-negative prostate cancers as well as in normal prostate. Red areas indicate strong, positive correlation between genes; blue areas indicate negative correlation or anticorrelation. The map indicates increased correlation between the WNT pathway genes in ERG-positive tumors, less correlation in ERG-negative tumors, and no correlation in normal prostate tissues. The expression of epithelial and mesenchymal marker genes is shown for reference to visualize the association of induced WNT signaling with EMT. B, validation of increased expression of WNT genes in ERG-positive prostate cancers: The top five genes, HDAC1 (**, P < 0.001), GNAI1 (*, P < 0.03), FZD4 (*, P < 0.01), and PRKD1, were significantly upregulated in ERG-positive (n = 9) compared with ERG-negative (n = 7) prostate cancers. All experiments were done in triplicates.

Close modal

Validation of WNT pathway activation and FZD4 in ERG-positive clinical samples

We validated the expression levels of 15 ERG coexpressed genes, including several genes involved in WNT signaling, by qRT-PCR in nine ERG-positive and seven ERG-negative prostate cancer samples. The results confirmed that HDAC1, guanine nucleotide binding protein (G protein), α inhibiting activity polypeptide 1 (GNAI1), FZD4, and protein kinase D1 (PRKD1) were highly correlated with ERG in clinical prostate samples (Fig. 3B). For some genes, associations may link these to other central signaling molecules important for prostate cancer progression. For example, protein kinase D1 (PRKD1), which is upregulated in ERG-positive prostate cancer, interacts with E-cadherin, which is dysregulated in prostate cancer and associated with altered cell aggregation and motility (25). FZD4 was also analyzed immunohistochemically in frozen sections derived from the same clinical prostate cancer samples used in qRT-PCR experiments. There was a trend toward higher FZD4 protein expression in ERG-positive tumors compared with ERG-negative samples (Fig. 4). FZD4 protein was predominantly membrane bound and associated with cancer cells.

Figure 4.

Immunohistochemical staining of FZD4 in clinical prostate tumors. ERG-positive cancer samples (tumors 217O and 241V with Gleason grades 4 and 3) showed intense membrane-bound FZD4 staining in carcinoma cells, whereas ERG-negative cancers (tumors 350O and 224V with Gleason grades 4 and 5) showed weak FZD4 expression in epithelial cell membranes.

Figure 4.

Immunohistochemical staining of FZD4 in clinical prostate tumors. ERG-positive cancer samples (tumors 217O and 241V with Gleason grades 4 and 3) showed intense membrane-bound FZD4 staining in carcinoma cells, whereas ERG-negative cancers (tumors 350O and 224V with Gleason grades 4 and 5) showed weak FZD4 expression in epithelial cell membranes.

Close modal

FZD4 silencing induces active β1-integrin and E-cadherin expression

To explore the role of FZD4 in cell-cell contacts and in cell adhesion, we knocked down FZD4 by siRNA in VCaP cells and confirmed by qRT-PCR, Western blot, and immunostaining (Supplementary Fig. S4). FZD4 silencing resulted in upregulation of active β1-integrin (Fig. 5A) and E-cadherin (Fig. 5B) expression in VCaP cells in an identical way as observed for ERG knockdown and also confirmed by FACS by using labeled active β1-integrin and E-cadherin antibodies (Supplementary Fig. S5A and B). Similar FACS analysis was performed with GFP-ERG– and GFP-FZD4–transfected VCaP cells at the 72-hour time point. The results indicate that GFP-ERG and GFP-FZD4 overexpression reduced active β1-integrin and E-cadherin expression (Supplementary Fig. S5C and D). This supports the idea that FZD4 is sufficient to mediate the effects of ERG on EMT and cell adhesion in prostate cancer cells.

Figure 5.

Immunostaining of active β1-integrin and E-cadherin in response to FZD4 silencing in VCaP cells. Immunostaining of scrambled siRNA– and FZD4 siRNA–treated VCaP cells with antibodies against (A) active β1-integrin and (B) E-cadherin (green) after 48 hours transfection.

Figure 5.

Immunostaining of active β1-integrin and E-cadherin in response to FZD4 silencing in VCaP cells. Immunostaining of scrambled siRNA– and FZD4 siRNA–treated VCaP cells with antibodies against (A) active β1-integrin and (B) E-cadherin (green) after 48 hours transfection.

Close modal

FZD4 as a critical regulator of WNT signaling in ERG-positive prostate cancer

To assess the activation of WNT signaling in response to ERG overexpression, TCF/LEF GFP Reporter Assay was performed in GFP-ERG–transfected RWPE1 cells at 48 and 72 hours. The results are shown in Fig. 6A and indicate a 2.4-fold higher WNT signaling in response to ERG overexpression. Moreover, the activity of the WNT pathway was 3-fold reduced by ERG knockdown, as evaluated by comparison of TCF/LEF reporter activity in stable ERG 214 shRNA–expressing VCaP cells with scrambled shRNA control cells (Fig. 6B). Transient silencing of ERG by siRNA transfection in VCaP cells also resulted in reduced TCF/LEF reporter activity in accordance with ERG shRNA knockdown results. As expected based on the role of the FZD4 gene in the WNT pathway, FZD4 siRNA knockdown reduced the TCF/LEF reporter activity in VCaP cells (Fig. 6C). To validate our hypothesis that FZD4 mediates the effect of ERG, we overexpressed FZD4 in VCaP cells and simultaneously performed ERG silencing using siRNAs. The results indicate that FZD4 overexpression reversed the effect of ERG knockdown on WNT activity in VCaP cells (Fig. 6D).

Figure 6.

TCF/LEF reporter assay: ERG and FZD4 regulates WNT signaling. WNT activity was measured by Acumen Explorer by using inducible TCF/LEF reporter (GFP) at different time points. A, RWPE1 cells were cotransfected with TCF/LEF reporter and GFP-ERG or GFP vector followed by WNT activity measurement at the 48 and 72 h time points. B, WNT activity in ERG-silenced VCaP cells by shRNA and siRNA compared with the respective controls. C, WNT activity in FZD4 siRNA–treated VCaP cells. D, WNT activity was rescued by FZD4 overexpression in ERG siRNA–treated VCaP cells (*, P < 0.05; **, P < 0.001).

Figure 6.

TCF/LEF reporter assay: ERG and FZD4 regulates WNT signaling. WNT activity was measured by Acumen Explorer by using inducible TCF/LEF reporter (GFP) at different time points. A, RWPE1 cells were cotransfected with TCF/LEF reporter and GFP-ERG or GFP vector followed by WNT activity measurement at the 48 and 72 h time points. B, WNT activity in ERG-silenced VCaP cells by shRNA and siRNA compared with the respective controls. C, WNT activity in FZD4 siRNA–treated VCaP cells. D, WNT activity was rescued by FZD4 overexpression in ERG siRNA–treated VCaP cells (*, P < 0.05; **, P < 0.001).

Close modal

Knockdown of ERG in the TMPRSS2-ERG fusion–positive VCaP prostate cancer cells resulted in clusters of strongly adherent cells with induction of active β1-integrin and E-cadherin expression. Increased ERG activity in turn was associated with EMT, evidenced as both repression of epithelial-specific genes, such as E-cadherin and cytokeratins, and increased expression of mesenchymal-specific genes, such as CDH2 and CDH11. Loss of E-cadherin is a hallmark of EMT, a process implicated in aggressive tumor cell behavior, invasion, and metastasis (2629). Our results indicate that ERG represses E-cadherin expression, by inducing Snail transcription factors, central players in EMT transitions. In addition, increased cell adhesion accompanied with induced active β1-integrin expression may indicate reduced motile/invasive behavior of the tumor cells in response to ERG silencing because decrease in β1-integrin expression has been linked to increased invasion in prostate cancer cells (30). Interestingly, ERG overexpression in prostate cells has been shown to increase invasion in RWPE 1 and silencing of activated ERG to reduce invasion in VCaP cells (15). Taken together, our data suggest that this effect is primarily mediated by EMT accompanied with profound changes in cell adhesion. The integrated evidence from both ERG manipulation in vitro and gene coexpression studies in clinical tumors indicated activation of the WNT pathway, and in particular overexpression of the FZD4 gene, in ERG-positive tumors. FZD4 silencing mimicked the ERG knockdown phenotype by also inducing active β1-integrin and E-cadherin expression. Furthermore, FZD4 overexpression reversed the impact of ERG knockdown in prostate cancer cells. These results suggest that in VCaP cells, FZD4 is both necessary and sufficient as a mediator of the oncogenic effects of ERG overexpression. The association between ERG and FZD4 gene expression in clinical specimens suggests that such a mechanism of oncogenesis may be operational in vivo as well. The detected mRNA and protein expression changes in clinical prostate cancer samples indicate that EMT is a relatively early change in carcinogenesis, and FZD4 overexpression (31) could be driven by the downstream effects of the ERG-fusion gene.

Our study therefore provides novel insights into the critical pathways by which ectopic ERG oncogene expression may initiate and promote prostate cancer progression. In particular, we provide evidence to support the hypothesis that ERG-mediated oncogenesis in prostate cancer involves activation of WNT signaling through FZD4, leading to key cancer-promoting phenotypic effects, such as EMT and loss of cell adhesion. Our preliminary results also indicate that an experimental WNT pathway inhibitor, 4-(chloromethyl)benzoyl chloride, reduced the growth of the ERG-positive VCaP prostate cancer cells more efficiently than that of ERG-negative RWPE1 prostate cells.6

6S. Gupta et al., unpublished data.

This suggests that in exploring specific treatment strategies for ERG-positive prostate cancer patients, it may be possible to target downstream signaling events, such as WNT signaling (32).

No potential conflicts of interest were disclosed.

We thank the Finnish DNA Microarray Centre for performing TaqMan qRT-PCR analysis.

Grant Support: Academy of Finland Center of Excellence program, Sigrid Juselius Foundation, Finnish Cancer Society, and EU-FP7 Epitron.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Helgeson
BE
,
Tomlins
SA
,
Shah
N
, et al
. 
Characterization of TMPRSS2:ETV5 and SLC45A3:ETV5 gene fusions in prostate cancer
.
Cancer Res
2008
;
68
:
73
80
.
2
Tomlins
SA
,
Rhodes
DR
,
Perner
S
, et al
. 
Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer
.
Science
2005
;
310
:
644
8
.
3
Tomlins
SA
,
Mehra
R
,
Rhodes
DR
, et al
. 
TMPRSS2:ETV4 gene fusions define a third molecular subtype of prostate cancer
.
Cancer Res
2006
;
66
:
3396
400
.
4
Tomlins
SA
,
Mehra
R
,
Rhodes
DR
, et al
. 
Integrative molecular concept modeling of prostate cancer progression
.
Nat Genet
2007
;
39
:
41
51
.
5
Hermans
KG
,
van Marion
R
,
van Dekken
H
,
Jenster
G
,
van Weerden
WM
,
Trapman
J
. 
TMPRSS2:ERG fusion by translocation or interstitial deletion is highly relevant in androgen-dependent prostate cancer, but is bypassed in late-stage androgen receptor-negative prostate cancer
.
Cancer Res
2006
;
66
:
10658
63
.
6
Mehra
R
,
Tomlins
SA
,
Shen
R
, et al
. 
Comprehensive assessment of TMPRSS2 and ETS family gene aberrations in clinically localized prostate cancer
.
Mod Pathol
2007
;
20
:
538
44
.
7
Perner
S
,
Mosquera
JM
,
Demichelis
F
, et al
. 
TMPRSS2-ERG fusion prostate cancer: an early molecular event associated with invasion
.
Am J Surg Pathol
2007
;
31
:
882
8
.
8
Giovannini
M
,
Biegel
JA
,
Serra
M
, et al
. 
EWS-erg and EWS-Fli1 fusion transcripts in Ewing's sarcoma and primitive neuroectodermal tumors with variant translocations
.
J Clin Invest
1994
;
94
:
489
96
.
9
Marcucci
G
,
Mrozek
K
,
Bloomfield
CD
. 
Molecular heterogeneity and prognostic biomarkers in adults with acute myeloid leukemia and normal cytogenetics
.
Curr Opin Hematol
2005
;
12
:
68
75
.
10
Baldus
CD
,
Burmeister
T
,
Martus
P
, et al
. 
High expression of the ETS transcription factor ERG predicts adverse outcome in acute T-lymphoblastic leukemia in adults
.
J Clin Oncol
2006
;
24
:
4714
20
.
11
Attard
G
,
Clark
J
,
Ambroisine
L
, et al
. 
Duplication of the fusion of TMPRSS2 to ERG sequences identifies fatal human prostate cancer
.
Oncogene
2008
;
27
:
253
63
.
12
Hermans
KG
,
Boormans
JL
,
Gasi
D
, et al
. 
Overexpression of prostate-specific TMPRSS2(exon 0)-ERG fusion transcripts corresponds with favorable prognosis of prostate cancer
.
Clin Cancer Res
2009
;
15
:
6398
403
.
13
Tu
JJ
,
Rohan
S
,
Kao
J
,
Kitabayashi
N
,
Mathew
S
,
Chen
YT
. 
Gene fusions between TMPRSS2 and ETS family genes in prostate cancer: Frequency and transcript variant analysis by RT-PCR and FISH on paraffin-embedded tissues
.
Mod Pathol
2007
;
20
:
921
8
.
14
Wang
J
,
Cai
Y
,
Ren
C
,
Ittmann
M
. 
Expression of variant TMPRSS2/ERG fusion messenger RNAs is associated with aggressive prostate cancer
.
Cancer Res
2006
;
66
:
8347
51
.
15
Tomlins
SA
,
Laxman
B
,
Varambally
S
, et al
. 
Role of the TMPRSS2-ERG gene fusion in prostate cancer
.
Neoplasia
2008
;
10
:
177
88
.
16
Klezovitch
O
,
Risk
M
,
Coleman
I
, et al
. 
A causal role for ERG in neoplastic transformation of prostate epithelium
.
Proc Natl Acad Sci U S A
2008
;
105
:
2105
10
.
17
Zong
Y
,
Xin
L
,
Goldstein
AS
,
Lawson
DA
,
Teitell
MA
,
Witte
ON
. 
ETS family transcription factors collaborate with alternative signaling pathways to induce carcinoma from adult murine prostate cells
.
Proc Natl Acad Sci U S A
2009
;
106
:
12465
70
.
18
Sun
C
,
Dobi
A
,
Mohamed
A
, et al
. 
TMPRSS2-ERG fusion, a common genomic alteration in prostate cancer activates C-MYC and abrogates prostate epithelial differentiation
.
Oncogene
2008
;
27
:
5348
53
.
19
Iljin
K
,
Wolf
M
,
Edgren
H
, et al
. 
TMPRSS2 fusions with oncogenic ETS factors in prostate cancer involve unbalanced genomic rearrangements and are associated with HDAC1 and epigenetic reprogramming
.
Cancer Res
2006
;
66
:
10242
6
.
20
Mould
AP
,
Askari
JA
,
Barton
S
, et al
. 
Integrin activation involves a conformational change in the α1 helix of the β subunit A-domain
.
J Biol Chem
2002
;
277
:
19800
5
.
21
Collins
TJ
. 
ImageJ for microscopy
.
BioTechniques
2007
;
43
:
25
30
.
22
Perner
S
,
Demichelis
F
,
Beroukhim
R
, et al
. 
TMPRSS2:ERG fusion-associated deletions provide insight into the heterogeneity of prostate cancer
.
Cancer Res
2006
;
66
:
8337
41
.
23
Demichelis
F
,
Fall
K
,
Perner
S
, et al
. 
TMPRSS2:ERG gene fusion associated with lethal prostate cancer in a watchful waiting cohort
.
Oncogene
2007
;
26
:
4596
9
.
24
Kilpinen
S
,
Autio
R
,
Ojala
K
, et al
. 
Systematic bioinformatic analysis of expression levels of 17,330 human genes across 9,783 samples from 175 types of healthy and pathological tissues
.
Genome Biol
2008
;
9
:
R139
.
25
Syed
V
,
Mak
P
,
Du
C
,
Balaji
KC
. 
β-Catenin mediates alteration in cell proliferation, motility and invasion of prostate cancer cells by differential expression of E-cadherin and protein kinase D1
.
J Cell Biochem
2008
;
104
:
82
95
.
26
Margineanu
E
,
Cotrutz
CE
,
Cotrutz
C
. 
Correlation between E-cadherin abnormal expressions in different types of cancer and the process of metastasis
.
Rev Med Chir Soc Med Nat Iasi
2008
;
112
:
432
6
.
27
Schmalhofer
O
,
Brabletz
S
,
Brabletz
T
. 
E-cadherin, β-catenin, and ZEB1 in malignant progression of cancer
.
Cancer Metastasis Rev
2009
;
28
:
151
66
.
28
Peinado
H
,
Olmeda
D
,
Cano
A
. 
Snail, zeb and bHLH factors in tumour progression: an alliance against the epithelial phenotype?
Nat Rev Cancer
2007
;
7
:
415
28
.
29
Brabletz
T
,
Hlubek
F
,
Spaderna
S
, et al
. 
Invasion and metastasis in colorectal cancer: Epithelial-mesenchymal transition, mesenchymal-epithelial transition, stem cells and β-catenin
.
Cells Tissues Organs
2005
;
179
:
56
65
.
30
Dedhar
S
,
Saulnier
R
,
Nagle
R
,
Overall
CM
. 
Specific alterations in the expression of α3β1 and α6β4 integrins in highly invasive and metastatic variants of human prostate carcinoma cells selected by in vitro invasion through reconstituted basement membrane
.
Clin Exp Metastasis
1993
;
11
:
391
400
.
31
Acevedo
VD
,
Gangula
RD
,
Freeman
KW
, et al
. 
Inducible FGFR-1 activation leads to irreversible prostate adenocarcinoma and an epithelial-to-mesenchymal transition
.
Cancer Cell
2007
;
12
:
559
71
.
32
Barker
N
,
Clevers
H
. 
Mining the wnt pathway for cancer therapeutics
.
Nat Rev Drug Discov
2006
;
5
:
997
1014
.