TGF-β Effects on Prostate Cancer Cell Migration and Invasion Are Mediated by PGE₂ through Activation of PI3K/AKT/mTOR Pathway

Abstract

TGF-β plays an important role in the progression of prostate cancer. It exhibits both tumor suppressor and tumor-promoting activities. Correlations between cyclooxygenase (COX)-2 overexpression and enhanced production of prostaglandin (PG)E₂ have been implicated in cancer progression; however, there are no studies indicating that TGF-β effects in prostate cancer cells involve PGE₂ synthesis. In this study, we investigated TGF-β regulation of COX-1 and COX-2 expression in prostate cancer cells and whether the effects of TGF-β on cell proliferation and migration are mediated by PGE₂. COX-1 protein was ubiquitously expressed in prostate cells; however, COX-2 protein levels were detected only in prostate cancer cells. TGF-β treatment increased COX-2 protein levels and PGE₂ secretion in PC3 cells. Exogenous PGE₂ and PGF_2α had no effects on cell proliferation in LNCaP, DU145, and PC3 cells whereas PGE₂ and TGF-β induced migration and invasive behavior in PC3 cells. Only EP2 and EP4 receptors were detected at mRNA levels in prostate cells. The EP4-targeting small interfering RNA inhibited PGE₂ and TGF-β-induced migration of PC3 cells. TGF-β and PGE₂ induce activation of PI3K/AKT/mammalian target of rapamycin pathway as indicated by increased AKT, p70S6K, and S6 phosphorylation. Rapamycin completely blocked the effects of TGF-β and PGE₂ on phosphorylation of p70S6K and S6 but not on AKT phosphorylation. PGE₂ and TGF-β induced phosphorylation of AKT, which was blocked by antagonists of PGE₂ (EP4) receptors (L161982, AH23848) and PI3K inhibitor (LY294002) in PC3 cells. Pretreatment with L161982 or AH23848 blocked the stimulatory effects of PGE₂ and TGF-β on cell migration, whereas LY294002 or rapamycin completely eliminated PGE₂, TGF-β, and epidermal growth factor-induced migration in PC3 cells. We conclude that TGF-β increases COX-2 levels and PGE₂ secretion in prostate cancer cells which, in turn, mediate TGF-β effects on cell migration and invasion through the activation of PI3K/AKT/mammalian target of rapamycin pathway.

Prostaglandins (PGs) affect many mechanisms that have been shown to play a role in carcinogenesis such as cell proliferation, angiogenesis, apoptosis, and mutagenesis (1–3). PGs are derived from arachidonic acid released from plasma membrane by phospholipases, mainly phospholipase A2 (2, 3). Cyclooxygenase (COX), also known as prostaglandin-endoperoxidase synthase (PTGS), is a rate-limiting enzyme involved in the conversion of arachidonic acid to prostanoids (4). Two isoforms of COX have been identified: COX-1 or PTGS1 and COX-2 or PTGS2 (5). COX-1 is constitutively expressed and is considered as a housekeeping gene, whereas COX-2 is not detected in most normal tissues (4). COX-2 is an inducible enzyme that is rapidly up-regulated by mitogens, growth factors, and cytokines and thus is responsible for acute increases in PG synthesis (4). Five PGs have been identified: PGE₂, PGD₂, PGF_2α, PGI₂, and thromboxane (2, 3). PGE₂ is the most common and ubiquitously produced PG, which acts in autocrine and paracrine manners to elicit a wide range of physiologic functions (5). In addition to its normal function, PGE₂ has been implicated in a broad array of diseases including cancer. PGE₂ may contribute to tumorigenesis via induction of cell proliferation (6), angiogenesis (7, 8), invasion (9, 10), and metastasis (3, 11).

Multiple reports have shown that COX-2 expression in normal prostate tissue is weak or undetectable whereas prostate cancer tissues express high levels of COX-2 protein (12–16). Previous studies have also shown that the level of PGE₂ conversion from arachidonic acid is almost 10-fold higher in human malignant prostatic tissues than in benign prostatic tissues (17). PGE₂ also has been shown to stimulate cell growth in osteoblasts and prostate cancer cells (1, 6). PGE₂ interacts with four different E prostanoid (EP1–EP4) receptor subtypes, which belong to the superfamily of G protein-coupled receptors (18). Previous studies have shown that human prostate epithelial cells express EP2 and EP4 receptors whereas the expression of EP1 and EP3 receptors was not detected in these cells (7). Furthermore, protein kinase A-dependent pathways activated by EP2/EP4 receptors have been implicated in PGE₂ effects on secretion of vascular endothelial growth factor (7) and induction of c-Fos in prostate cancer cells (19).

TGF-β plays an important role in the progression of prostate cancer. It acts as tumor suppressor in the early stages of epithelial cancers by inhibiting proliferation and inducing apoptosis (20). However, in the later stages of the disease, TGF-β acts as a tumor promoter and is associated with aggressive form of cancers due to its effects on angiogenesis, immune suppression, and metastasis (20). Previous studies from several laboratories have investigated the role of TGF-β secreted by the epithelial and stromal cells in the development and progression of prostate cancer (21–23). Our laboratory has shown that TGF-β inhibited proliferation in WPE, RWPE1, and DU145 cells, but not in LNCaP or PC3 cells (24). Interestingly, TGF-β induced migration in PC3 cells, but not in DU145 cells (24). Previous studies have reported that TGF-β1 exerts stimulatory effects on COX-2 expression in various cell types such as colon cancer, intestinal epithelial, hen granulosa cells, and human mesangial cells (25–28). In addition, TGF-β effects on epithelial to mesenchymal transition in mammary epithelial cells involve its induction of COX-2 and PGE₂ secretion in these cells (29).

Several studies have demonstrated that COX-2 expression was found to increase PGE₂ production and metastatic potential of human colon and breast cancer cells (30, 31). However, a possible role of PGs in mediating differential effects of TGF-β in prostate cancer cells has not been investigated. In this study, we investigated TGF-β regulation of COX-1 and COX-2 expression in prostate cancer cells and whether the effects of TGF-β on cell proliferation and/or migration and invasion are mediated by PGE₂.

Materials and Methods

Chemicals and reagents

Recombinant human TGF-β1 was purchased from R&D Systems (Minneapolis, Minnesota). PGE₂ and PGF_2α were purchased from Sigma-Aldrich (St Louis, Missouri). The antibodies against phospho-AKT, AKT, phospho-p70 S6 kinase (Thr³⁸⁹), p70S6K, phospho-S6 ribosomal protein (Serine ^235/236), and S6 were purchased from Cell Signaling Technology, Inc. (Danvers, Massachusetts). Anti-COX-1, COX-2 antibodies, and EP2 and EP4 siRNAs (small interfering RNA) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, California). Enzyme immunoassay (EIA) kit for PGE₂ analysis was purchased from Assay Designs (Ann Arbor, Michigan). Anti-β-actin antibody and AH23848 (EP4 inhibitor) were purchased from Sigma-Aldrich. L161982 (EP4 inhibitor) and SB431542 (TGF-β receptor-I inhibitor) were obtained from Tocris Bioscience (Minneapolis, Minnesota). Smad3 inhibitor (SIS3) and rapamycin were obtained from Calbiochem (San Diego, California). The antirabbit and antimouse Igs coupled to horseradish peroxidase (IgG-HRP) were obtained from Promega Corp. (Madison, Wisconsin) and donkey antigoat IgG HRP was obtained from Santa Cruz Biotechnology, Inc.

Cell culture and cell treatments

Several cell lines derived from human prostate cells representing normal epithelial cells and different stages of prostate cancer progression were cultured using established procedures (24, 32). These cell lines included PZ-HPV7, RWPE1, RWPE2, LNCaP, LNCaP-C81, DU145, PC3, and PC3M cells.

To determine the effects of TGF-β1 on COX-1 and COX-2 expression and the effects of TGF-β1 and PGE₂ on phospho-AKT, p70S6K (Thr³⁸⁹), and S6 (Serine ^235/236), prostate cells were cultured in six-well plates at the density of 4 × 10⁵ cells per well. Before each experiment, the cells were incubated in serum-free or supplement-free media for 2 hours, followed by treatment with different doses of TGF-β1 or PGE₂ for specific time periods. The cells were washed with ice-cold PBS and lysed in lysis buffer (Cell Signaling Technology, Beverly, Massachusetts) containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO₄, 1 μg/mL leupeptin, and 1× protease inhibitor cocktail (Calbiochem, San Diego, California). Protein concentrations were determined by the Lowry HS assay using the Bio-Rad DC Protein Assay kit (Bio-Rad Laboratories, Inc., Hercules, California) according to the instructions provided by the manufacturer.

RNA isolation, reverse transcription (RT), and PCR

Total RNA was isolated from prostate cells using TRIzol (Invitrogen, Carlsbad, California) and the resulting RNA samples were quantified by optical density reading at 260 nm as described previously (24, 33). PCR was performed to detect mRNA levels of EP2, EP4, and L-19. The PCR mixture was composed of 0.1 mM deoxynucleotide triphosphates, 0.5 U Taq DNA polymerase, 10× PCR buffer with 3 mM MgCl₂ and 25 pM of the specific primers in a total volume of 15 μL. Primer information and the size of expected amplicons for individual genes are shown in Table 1. L-19 (a ribosomal protein) was used as a template control. RNA samples processed without RT and PCR amplified by the L-19 were used as negative controls. Amplification was performed at 94°C for 20 seconds, 60°C for 30 seconds, and 72°C for 45 seconds for 30 cycles for EP1-EP4, and L-19. For all PCRs, an initial step was at 95°C for 2 minutes, and a final extension was at 72°C for 10 minutes. The PCR products were separated on 1.5%–2% agarose gels and stained with ethidium bromide.

Table 1.

Primers used for the Detection of Target mRNAs

Gene	Primer	Sequence (5′ → 3′)	Product size (bp)
EP1	Forward	CCGCCACCTTCCTGCTGTTCG	1037
	Reverse	GGTGGGCTGGCTTAGTCGTTG
EP2	Forward	ACCTACTTCGCTTTCGCCAT	208
	Reverse	CGTACTGCCCATAGTCCAGC
EP3	Forward	CGCCTCAACCACTCCTACACA	837
	Reverse	GCAGACCGACAGCACGCACAT
EP4	Forward	CGAATTGCTTCTGTGAACCCCAT	344
	Reverse	GAGGTGGTGTCTTCCTGGGCCAG
l-19	Forward	GAAATCGCCAATGCCAACTC	405
	Reverse	TCTTAGACCTGCGAGCCTCA

Western blot analysis

Western blot analyses were performed as described previously (33). Briefly, cell lysates were mixed with Laemmeli's buffer (62.5 mM Tris, pH 6.8, 2% sodium dodecyl sulfate, 5% β–mercaptoethanol, and 10% glycerol). Individual samples (30–35 μg proteins) were subjected to SDS-PAGE in 8% or 10% gels and transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, Massachusetts). After blocking the membranes with 5% fat-free milk in TBST (50 mM Tris, pH 7.5, containing 0.15 M NaCl, and 0.05% Tween 20) for 1 hour at room temperature, the membranes were incubated with appropriate dilutions of specific primary antibodies (1:1000 for p-AKT, AKT, p-p70S6K [Thr³⁸⁹], p70S6K, p-S6 [Serine ^235/236], and S6; 1:500 for COX-1 and COX-2; 1:20 000 dilution was used for β-actin) overnight at 4°C. After washing, the blots were incubated with antirabbit or antimouse IgG HRPs for 1 hour. The blots were developed in ECL mixture (Thermo Fisher Scientific Inc, Rockford, Illinois), exposed to an X-ray film and visualized by autoradiography. The density of specific protein bands was determined by QuantityOne image analysis software.

EIA of PGE₂

Prostate cancer (PC3) cells were cultured in 10-cm² dishes to approximately 80% confluence. The cells were placed in MEM containing 0.2% BSA (Sigma-Aldrich) the next day. The cells were treated with TGF-β1 (1 and 5 ng/mL) for different time periods to detect changes in the amount of PGE₂ production. The conditioned media were collected and centrifuged to remove cellular debris. Attached cells were lysed as described earlier (33). Two independent sets of samples were prepared for estimation of PGE₂ secretion. Concentrations of PGE₂ were measured in the conditioned media using an EIA kit (Assay Designs, Ann Arbor, Michigan) according to the instructions provided by the manufacturer. Total protein concentrations in total lysates were measured as described previously (33), and PGE₂ concentrations were normalized with total protein concentrations.

Cell proliferation assays

The effects of PGE₂, PGF_2α, and TGF-β on cell proliferation were determined using ³H-Thymidine incorporation assays as described previously (24, 33). The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was also performed to confirm the effects of PGE₂, PGF_2α, and TGF-β on cell proliferation using assay kits purchased from Promega Corp. as described previously (34).

Cell migration and invasion assays

In vitro cell migration assay was performed using 24-well transwell inserts (8 μm) as described previously (24, 35). Briefly, the outside of the transwell insert membrane was coated with 50 μL rat tail collagen (50 μg/mL) overnight at 4°C. The next day, aliquots of rat tail collagen (50 μL) were added into the transwell inserts to coat the inside of the membranes for 1.5 hours at room temperature. The cells were harvested from cell culture dishes by EDTA-trypsin into 50-mL conical tubes and centrifuged at 500 × g for 10 minutes at room temperature. The pellets were resuspended in MEM supplemented with 0.2% BSA at a cell density of 3 × 10⁵ cells/mL. Chemoattractant solutions were made by diluting PGE₂ and PGF_2α (0.1, 1, 10, 100 ng/mL), TGFβ1 (5 ng/mL), or epidermal growth factor (EGF) (3 ng/mL) into MEM supplemented with 0.2% BSA. MEM containing 0.2% BSA served as a control. EGF was used as a positive control (36). Control and chemoattractant solutions (400 μL) were added into different wells of a 24-well plate. Aliquots of 100-μL cell suspension were loaded into transwell inserts that were subsequently placed into the 24-well plate. The transwell insert-loaded plate was placed in a cell culture incubator for 5 hours. For invasion assay, inserts (BD Biosciences, Palo Alto, California) were coated with 50 μL of a 1:4 Matrigel/Medium dilution (BD Biosciences) and allowed to solidify at 37°C for 1 hour. Cells were resuspended (3 × 10⁴ cells/mL) in MEM with 0.1% fetal bovine serum (FBS), and 500 μL of cell suspension were added to each insert. Cells were treated with PGE₂ and PGF_2α (0.1, 1, 10, 100 ng/mL), TGFβ1 (5 ng/mL), or EGF (3 ng/mL) and allowed to invade through a porous membrane coated with Matrigel at 37°C for 48 hours. For migration and invasion assays, the cells inside transwell inserts were removed by cotton swabs. The cleaned inserts were fixed in 300 μL of 4% paraformaldehyde (pH 7.5) for 20 minutes at room temperature. Cells that had migrated to the outside of the transwell insert membrane were stained using HEMA 3 staining kit (Fisher Scientific, Inc., Pittsburgh, Pennsylvania). The number of stained cells was counted as described previously (24, 34). The results were expressed as migration/invasion index defined as: the average number of cells per field for test substance/the average number of cells per field for the medium control. The experiments were conducted at least three times using independent cell preparations.

Transfection with EP2/EP4 siRNA

To knock down endogenous EP2/EP4 expression, prostate cells were cultured in six-well plates at the density of 2 × 10⁵ cells per well in 2 mL antibiotic-free normal growth medium supplemented with 5% FBS. The cells were cultured at 37°C until the cells were 60%–80% confluent. Control and EP2/EP4 siRNAs were transfected into PC3 cells according to the manufacturer's instructions (Santa Cruz Biotechnology, Inc.). Briefly, the cells were washed once with 2 mL of siRNA transfection medium, after which the siRNA duplex was overlaid onto the washed cells, incubated for 6 hours, followed by addition of 10% FBS to the cells without removing the transfection mixture, and the cells were incubated for 24 hours. The medium was changed with fresh medium and the cells were incubated for an additional 24 hours. The cells were harvested at specific times for determination of the changes in EP2/EP4 mRNA levels and cell migration assay.

Statistical analysis

All experiments were repeated at least three times using a different cell preparation. The results are presented as mean ± SEM of three independent experiments, and images from a single representative experiment are presented. Student's t test, ANOVA, and Duncan's modified multiple-range tests were employed to assess the significance of differences among various treatment groups.

Results

The effects of TGF-β on COX-2 levels and PGE₂ production in prostate cancer cells

Total proteins were extracted from immortalized prostate epithelial cell line (PZ-HPV7), immortalized prostate luminal epithelial cell line (RWPE1), k-ras transformed RWPE1 (RWPE2) cells, and prostate cancer cell lines (LNCaP, LNCaP-C81, DU145, PC3, and PC3M). To examine the presence of COX-1 and COX-2 proteins in these prostate cell lines, total cellular proteins were analyzed by Western blots using antibodies specific for COX-1 and COX-2. As shown in Figure 1A, COX-1 protein was ubiquitously expressed in all prostate cell lines. COX-2 protein was highly expressed in all prostate cancer cells (LNCaP, LNCaP-C81, DU145, PC3, and PC3M) but was undetectable in normal prostate epithelial cell lines (PZ-HPV7, RWPE1, and RWPE2). Therefore, we chose advanced stages of prostate cancer (PC3) cells as the model for further studies.

Figure 1.

The Effects of TGF-β on COX-1 and COX-2 Expression A, Western blot analysis of COX-1 and COX-2 protein levels in prostate cell lines. Total cellular proteins were separated by SDS-PAGE and blotted using COX-1 and COX-2 antibodies. β-Actin antibody was used as internal control. B, Western blot analysis of COX-1 and COX-2 protein levels in PC3 cells after time-dependent (0–8 h) exposure to exogenous TGF-β (5 ng/mL). C, Western blot analyses of COX-2 and β-actin in PC3 cells after pretreatment with TGF-β receptor-I inhibitor (SB431542; 10 μM) or Smad3 inhibitor (SIS3; 1 μM). D, PC3 cells were serum starved for 24 hours and treated with different doses of TGF-β1 (1 and 5 ng/mL) for specific time periods (8H and 24H). Analysis of PGE₂ in the culture medium was performed by EIA. PGE₂ levels were normalized to the total cellular proteins in each sample. *, Significantly different (P < .05) compared with untreated controls.

Next, we determined the effects of TGF-β1 on COX-1 and COX-2 protein expressions in prostate cancer (PC3) cells. Cells were cultured in the presence or absence of TGF-β1 for specific times, and the expression of COX-1 and COX-2 was determined by Western blotting. TGF-β1 treatment had very little, if any, effect on COX-1 protein levels (Figure 1B). On the other hand, TGF-β1 treatment increased COX-2 protein levels in PC3 cells in a dose-dependent (data not shown) and time-dependent manner. The increase in COX-2 protein after treatment with TGF-β1 (5 ng/mL) was readily apparent after 30 minutes and remained high up to 8 hours (1.6 ± 0.02 fold; P < .001) (Figure 1B).

To determine the specificity of TGF-β1-dependent induction of COX-2, PC3 cells were pretreated with a specific inhibitor of TGF-β receptor-I (SB431542; 10 μM) or Smad3 inhibitor (SIS3; 1 μM) for 1 hour. As shown in Figure 1C, pretreatment with both inhibitors suppressed basal and TGF-β1-induced increase in COX-2 protein levels.

Previous studies have shown that the level of PGE₂ conversion from arachidonic acid is almost 10-fold higher in human malignant prostatic tissues than in benign prostatic tissues (17). We determined the secretion of PGE₂ by PC3 cells after treatment with TGF-β1 (1 and 5 ng/mL) for 8 and 24 hours. The conditioned media were collected and analyzed for PGE₂ levels by EIA. As shown in Figure 1D, PC3 cells secrete significant amounts of PGE₂ in culture. The secretion of PGE₂ by PC3 cells was increased by treatment with TGF-β1. The highest levels of PGE₂ concentration were observed after 24 hours in response to TGF-β1 treatment at 1 ng/mL (2.0 ± 11.86 fold; P < 0.001).

The effects of PGE₂ and PGF_2α on cell proliferation, migration, and invasion of prostate cancer cells

TGF-β exerts effects on proliferation, migration, and invasion of specific prostate cancer cell lines. To determine whether these effects are mediated by PGs, we investigated the effects of different concentrations of PGE₂ and PGF_2α on cellular proliferation, migration, and invasion under identical experimental conditions in selected prostate cell lines (LNCaP, DU145, and PC3). As shown in Figure 2A, treatment with different doses of PGE₂ and PGF_2α had no effect on proliferation of three prostate cancer cell lines. On the other hand, exogenous TGF-β1 inhibited proliferation in DU145 prostate cancer cells but had no effect on proliferation in PC3 and LNCaP cells. These data suggested that PGE₂ and PGF_2α do not mediate TGF-β effects on proliferation of prostate cancer cells. On the other hand, treatment with PGE₂ (10 ng/mL) stimulated cell migration (2.6 ± 0.21 fold; P < .05) in PC3 cells but not in DU145 cells, similar to the effects of TGF-β under identical conditions (Figure 2B). PGF_2α treatment at the same concentrations had no significant effects on cell migration in PC3 cells (Figure 2B). EGF used as a positive control induced cell migration in both DU145 (2.7 ± 0.05 fold; P < .05) and PC3 (2.7 ± 0.23 fold; P < .05) cells (Figure 2B).

Figure 2.

Effects of PGE₂ and PGF_2α on Cell Proliferation, Migration, and Invasion of Prostate Cancer Cells A, The effects of PGE₂, PGF_2α, and TGF-β on DNA synthesis in LNCaP, DU145, and PC3 cells as determined by [³H]rsqb]thymidine incorporation assay. The cells were serum starved for 24 hours and treated with different concentrations of PGE₂ and PGF_2α for 18 hours in the presence of 5% FBS. TGF-β1 (5 ng/mL) was used as a positive control. The cells were then pulse labeled for 4 hours with 1 μCi/ml [³H]thymidine, and radioactivity was determined by liquid scintillation counting. Each bar represents mean ± SEM (n = 3). *, Significantly different (P < .05) when compared with untreated controls. B, TGF-β and PGE₂ induced migration in PC3 cells in a dose-dependent manner but not in DU145 cells in a transwell migration assay. Representative images of DU145 and PC3 cell lines after different treatments. Cells were visualized under ×10 objectives. EGF (3 ng/mL) was used as a positive control. Each bar represents mean ± SEM (n = 3). *, Significantly different (P < .05) compared with untreated controls. C, PC3 cells were treated with different concentrations of PGE₂ (0.1, 1, 10, 100 ng/mL). TGF-β1 and EGF were used as positive controls. The cells were allowed to invade through a porous membrane coated with Matrigel for 48 hours. Each bar represents mean ± SEM (n = 3). *, Significantly different (P < .05) compared with untreated controls.

Previous studies have shown that TGF-β induced motility and invasive behavior in prostate cancer cells (34). We determined the effects of various concentrations of PGE₂ on cell invasion in PC3 cells using the BD BioCoat Matrigel Invasion inserts. As shown in Figure 2C, treatment with different concentrations of PGE₂ significantly increased invasiveness of PC3 cells (2.1 ± 0.10 fold; P < .001). TGF-β1 (1.3 ± 0.06 fold; P < .001) and EGF (1.7 ± 0.06 fold; P < .001) also had a significant effect on cell invasion in PC3 cells (Figure 2C). These results indicate that PGE₂ secreted in response to TGF-β may mediate its effects on cell migration and invasion but not on cell proliferation.

The effects of PGE₂, TGF-β, and EGF on migration in PC3 cells with or without EP2 or EP4 knockdown

The biological actions of PGE₂ are mediated by four different G protein-coupled receptors, namely EP1, EP2, EP3, and EP4. Therefore, we determined the steady-state mRNA levels of EP1,-2, -3, and -4 receptors in prostate cell lines. Total RNAs were extracted from immortalized normal prostate epithelial cells (PZ-HPV7 and RWPE1), k-ras transformed RWPE1 (RWPE2) cells, and prostate cancer cell lines (LNCaP, LNCaP-C81, DU145, PC3, and PC3M). As shown in Figure 3A, RT-PCR detected only EP2 and EP4 mRNAs but not EP1 or EP3 in all prostate cell lines.

Figure 3.

Effects of PGE₂, TGF-β, and EGF on Migration in PC3 Cells with or without EP2 or EP4 Knockdown A, Steady-state mRNA levels of EP1,-2, -3, and -4 receptors in prostate cells. Total RNAs were isolated and semiquantitative RT-PCR was performed to determine the mRNA levels of EP1, -2, -3, and -4 in prostate cell lines. L-19 was used as an internal control. No reverse transcriptase (RT) samples derived from the same RNAs were included. B, Knockdown of endogenous EP2 and EP4 mRNA levels in PC3 cells by specific EP2 or EP4 siRNA. RNA samples were harvested 72 hours after transfection. EP2 and EP4 were detected with RT-PCR. No RT samples derived from the same RNAs were included. L-19 was used as an internal control. Control siRNA did not affect mRNA levels of EP2 and EP4. The EP2 or EP4-targeting siRNA specifically reduced EP2 or EP4 mRNAs, respectively. The bar diagram shows the statistical and quantitative analysis of EP2 or EP4 mRNA in PC3 cells after incubation in the presence of absence of EP2/EP4 siRNA or control siRNA. Different letters denote significant differences among various groups (n = 3; ANOVA; P < .05). C, PGE₂-, TGF-β-, and EGF-induced cell migration in the PC3 cells that were treated with no siRNA (no), control siRNA (control), EP2, or EP4 siRNA. The effects of PGE₂ and TGF-β on cell migration were completely eliminated in the PC3 cells that were treated with the EP4-targeting siRNA but not EP2 siRNA. Different letters represent significant differences among various groups (ANOVA; P < .05).

These results confirmed previous studies showing that prostate cancer cells primarily express EP2 and EP4 PG receptors (7). To determine whether EP2 or EP4 receptor is responsible for PGE₂-induced cell migration in prostate cancer cells, we performed transient transfection in PC3 cells to knock down endogenous EP2 or EP4 receptor using siRNA specific for human EP2 and EP4. Control siRNA did not affect mRNA levels of EP2 and EP4 in PC3 cells. The EP2- and EP4-targeting siRNAs significantly reduced EP2 and EP4 mRNAs by more than 90% in PC3 cells (Figure 3B). PGE₂-, TGF-β-, and EGF-induced cell migration was not affected in the PC3 cells that were treated with control siRNA. Similarly, EP2-targeting siRNA had no effect on cell migration (Figure 3C). On the other hand, the effects of PGE₂ and TGF-β on cell migration were completely eliminated in the PC3 cells that were treated with the EP4-targeting siRNA. EP4 siRNA also caused a significant reduction in basal migration. EP4 knockdown caused only a slight but statistically insignificant decrease in EGF-induced migration in PC3 cells.

Next, we investigated the effects of TGF-β and PGE₂ on migration of PC3 cells in the presence of pharmacologic inhibitors of EP4. As shown in Figure 4, A and B, antagonists of EP4 receptor (L161982 and AH23848) suppressed both TGF-β- and PGE₂-induced cell migration in PC3 cells (P < .05). Pretreatment with AH23848 also caused a significant reduction in basal migration.

Figure 4.

Pretreatment with EP4 Receptor Inhibitors Blocked the Migration of PC3 Cells Induced by PGE₂ and TGF-β Treatment A and B, PC3 cells were pretreated with EP4 antagonist (L161982; 24 nM) or (AH23848; 20 μM) for 1 hour and then treated with PGE₂ (10 ng/mL) and TGF-β (5 ng/mL). The data are presented as mean ± SEM (n = 3). Different letters represent significant differences among various groups (ANOVA; P < .05).

TGF-β and PGE₂ activate phosphatidylinositol phosphate 3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway in prostate cancer cells

Previous studies have shown that TGF-β effects on cell migration and invasion are mediated via the activation PI3K/AKT pathway (37, 38). Therefore, we determined the effects of PGE₂ and TGF-β on AKT phosphorylation. PC3 cells were serum starved and treated with TGF-β (0–10 ng/mL) and PGE₂ (0–100 ng/mL) for specific time periods. The levels of phosphorylation of AKT were analyzed by Western blot analysis using a specific antibody for phospho-AKT^ser473. As shown in Figure 5A, treatment with TGF-β (5 ng/mL) induced dose-dependent AKT phosphorylation, and maximum levels of p-AKT were observed 30 minutes after TGF-β treatment (1.6 ± 0.01 fold; P < .001). In parallel experiments, PGE₂ (10 ng/mL) also caused a dose-dependent increase in AKT phosphorylation in a time-dependent manner (2.0 ± 0.04 fold; P < .001) (Figure 5A).

Figure 5.

TGF-β and PGE₂-Induced AKT Phosphorylation Is Mediated Primarily by the EP4 Receptor A, Western blot analyses of phosphorylated AKT and total AKT in PC3 cells after treatment with different doses of exogenous TGF-β (0.1, 1, 5, and 10 ng/mL) and PGE₂ (0.1, 1, 10, and 100 ng/mL), and after treatment for specific time periods (0–2 h) with TGF-β (5 ng/mL) and PGE₂ (10 ng/mL). B and C, PC3 cells were pretreated with a specific inhibitor of PI3K (LY29400; 10 μM) or EP4 receptor inhibitors (L161982; 24 nM) or (AH23848; 20 μM) for 1 hour and then treated with TGF-β, PGE₂, or EGF. The levels of phosphorylated AKT and total AKT were analyzed by Western blot.

Furthermore, pretreatment of PC3 cells with a specific inhibitor of PI3K (LY29400; 10 μM) resulted in significant inhibition (> 95%; P < .001) of both TGF-β- and PGE₂-induced phosphorylation of AKT (Figure 5B). EGF was used as a positive control that induced AKT phosphorylation, which was blocked by LY29400 inhibitor. Next, we determined the effects of TGF-β and PGE₂ on activation of PI3K/AKT signaling in the presence of specific EP4 inhibitors. As shown in Figure 5C, two EP4 inhibitors blocked phosphorylation of AKT in PC3 cells in response to both PGE₂ (L161928, 0.54 ± 0.03; AH23848, 0.33 ± 0.01; P < .001) and TGF-β (L161928, 0.11 ± 0.006; AH23848, 0.12 ± 0.01; P < .001). These inhibitors did not influence the EGF-induced phosphorylation of AKT in PC3 cells.

In the next set of experiments, we investigated whether activation of PI3K/AKT by TGF-β, PGE₂, or EGF leads to activation of mTOR signaling. The cells were treated with TGF-β, PGE₂, or EGF for 30 minutes in the presence of rapamycin to inhibit mTOR signaling. As shown in Figure 6A, the activation of mTOR signaling was demonstrated by the phosphorylation of its downstream targets. There was significant increase in the phosphorylation of p70S6K and S6 protein in response to TGF-β (1.7 ± 0.22 fold and 1.3 ± 0.15 fold; P ≤ .001), PGE₂ (1.7 ± 0.34 fold and 1.4 ± 0.14 fold; P ≤ .05), or EGF (2.4 ± 0.68 fold and 1.8 ± 0.32 fold; P ≤ .05) treatment, respectively (Figure 6A). Treatment with rapamycin blocked the effects of TGF-β, PGE₂, or EGF on phosphorylation of p70S6K and S6 but not on AKT phosphorylation (Figure 6A).

Figure 6.

TGF-β and PGE₂ Activate PI3K/AKT/mTOR Pathway in Prostate Cancer Cells A, PC3 cells were pretreated with rapamycin (10 μM) for 30 minutes and then treated with PGE₂, TGF-β, or EGF. The levels of phosphorylated AKT, p70S6K (Thr³⁸⁹), S6 (Serine ^235/236) and total AKT, p70S6K, S6 were analyzed by Western blot. Total AKT, p70S6K, and S6 antibodies were used as loading controls. B and C, Pretreatment with PI3K inhibitor (LY294002; 10 μM) or rapamycin (10 μM) for 1 hour blocked the migration of PC3 cells induced by PGE₂ (10 ng/mL), TGF-β (5 ng/mL), and EGF (3 ng/mL) in treated cells. The data were presented as mean ± SEM (n = 3). Different letters represent significant differences among various groups (ANOVA; P < .05).

Furthermore, we showed that both LY29400 and rapamycin completely blocked the stimulatory effects of both TGF-β and PGE₂ on migration of PC3 cells (Figure 6, B and C). Interestingly, both inhibitors also blocked the effects of EGF on cell migration. These results suggest that both TGF-β and PGE₂ activate PI3K/AKT/mTOR signaling pathway that is mediated by EP4 receptors. The activation of this pathway is essential for stimulatory effects of both TGF-β and PGE₂ on migration and invasion of prostate cancer cells.

Discussion

In this study, we report that TGF-β increases COX-2 levels and PGE₂ secretion in prostate cancer cells. We also show that TGF-β and PGE₂ activate PI3K/AKT/mTOR signaling pathway that is mediated by prostaglandin (EP4) receptors and that activation of this pathway is essential for stimulatory effects of both TGF-β and PGE₂ on migration and invasion of prostate cancer cells.

TGF-β exerts differential effects on cell proliferation and migration in prostate cell lines. These differential effects of TGF-β during different stages of cancer progression presumably depend on selective loss or acquisition of specific intracellular signals that are required to elicit different biological responses to TGF-β. A loss of TGF-β receptors and/or Smad proteins has been shown to result in TGF-β resistance in cancer cells (20, 23). However, most cancer cells retain classical TGF-β signaling components throughout cancer progression but modify or recruit additional signaling pathways to exert novel or different biological effects. Our results show that TGF-β increases PGE₂ secretion in prostate cancer cells, which, in turn, mediates its effects on cell migration but does not play a role in TGF-β effects on cell proliferation.

COX is the main regulated enzyme for the biosynthesis of PGs (4). We found that COX-1 protein was ubiquitously expressed in all prostate cell lines whereas COX-2 protein levels were highly expressed in all prostate cancer cells but were undetectable in normal prostate epithelial cells. Several studies have demonstrated that TGF-β1 has stimulatory effects on COX-2 expression in various cell types, including colon, mesangial, hen granulosa, and intestinal epithelial cells (25–28). Treatment with TGF-β in rat intestinal epithelial cells causes transient induction of COX-2 and increased PG production (40). In addition, COX-2 expression was also found to increase PGE₂ production and the metastatic potential of human colon and breast cancer cells (30, 31). In the present study, we showed that prostate cancer (PC3) cells expressed higher levels of COX-2 protein, and these cells secrete significant amounts of PGE₂ in culture. Treatment with TGF-β1 further increased COX-2 protein levels and PGE₂ secretion in these cells. Our results confirm previous studies in many cell types that TGF-β1 is a potent inducer of COX-2 expression and PGE₂ production. In addition, Shao et al. (25) have shown that the level of COX-2 protein in rat intestinal epithelial cells was significantly reduced when TGF-β1 was removed from the culture medium, suggesting that the increased expression of COX-2 in these cells depends upon the presence of exogenous TGF-β. Also, higher basal levels of COX-2 protein and PGE₂ secretion by PC3 cells may also be maintained through autocrine effects of TGF-β isoforms secreted by these cells (34) and would suggest that development and progression of prostate cancer may be dependent on autocrine effects of TGF-β on COX-2 levels and PGE₂ secretion.

It has been reported that nonsteroidal anti-inflammatory drugs and COX-2 inhibitors suppress invasiveness of human prostate cancer cell lines; however, PGE₂ treatment can reverse this effect (41). PGE₂ also has been shown to stimulate cell growth in osteoblasts and prostate cancer cells (1, 6). In the present study, TGF-β1 and PGE₂ exerted similar effects on cell motility: both induced cell migration and invasion in PC3 cells, but not in DU145 cells. TGF-β inhibits proliferation in normal prostate cells and prostate epithelial cancer cells; however, PGE₂ had no effects on cell proliferation in LNCaP, DU145, and PC3 cells. These results indicate that PGE₂ secreted in response to TGF-β may mediate its effects on cell migration and invasion but not on cell proliferation. Hence, TGF-β-induced PGE₂ secretion may represent a shift in intracellular signaling involved in the escape from inhibition of proliferation to the stimulation of more invasive and metastatic behavior by TGF-β in advanced stages of prostate cancer. Our results also showed that these effects of PGE₂ on migration in PC3 cells are mediated by the EP4 receptors. These results are in line with the previous studies and the current study showing that prostate cancer cells express primarily EP2 and EP4 receptors (7). Kim et al. (42) reported that PGE₂ activates its cognate EP4 receptor subtype to induce the migration of lung cancer cells by activating c-Src through a mechanism that involves the adaptor protein β-arrestin 1. We show that both PGE₂- and TGF-β-induced cell migration require EP4-dependent signaling pathway. The selective EP4 knockdown by siRNA or pharmacologic inhibition of EP4 completely blocked the effects of PGE₂ and TGF-β on cell migration in PC3 cells. On the other hand, EP4 inhibition did not affects EGF effects on activation of PI3K/AKT pathway and stimulation of migration in PC3 cells. The possibility that PGE₂ may exert other biologic effects on prostate cancer cells, which may involve other PG (EP1–3) receptors, has not been ruled out.

PI3K/AKT pathway plays an important component in the migratory and prosurvival signaling pathway (38, 43). TGF-β can activate or inhibit PI3K/AKT signaling pathway depending on the cell type and the stimulus (37). Previous studies have also shown that PGE₂ induces the phosphorylation of AKT and nuclear factor-κB-p65 subunit in endothelial cells and prostate cancer cells (7, 8). A study in human renal mesangial cells suggested that PI3K/AKT activation participates in TGF-β-mediated induction of both COX-2 protein expression and PGE₂ production (28). Additionally, EP receptors have been shown to activate PI3K/AKT pathway, which, in turn, regulates Smad2/3 signaling in response to TGF-β (29). We have recently shown that TGF-β1 and TGF-β3 induce migration and invasive behavior in prostate cancer cells and that this effect is mediated by PI3K/AKT pathway (34). The present study shows that both PGE₂ and TGF-β induce activation of PI3K/AKT pathway, which is blocked by the specific antagonists of EP4 receptors, indicating that both TGF-β and PGE₂ require EP4 for activation of PI3K pathway. The activation of mTOR pathway by TGF-β and PGE₂ was confirmed by phosphorylation of downstream target of mTOR cascades (p70S6K and S6). In addition, specific inhibitors of PI3K/AKT (LY29400) and mTOR (rapamycin) significantly blocked PGE₂- and TGF-β-induced cell migration in PC3 cells. These results suggested that PGE₂ mediates TGF-β effects via EP4 receptor leading to cell migration and that PI3K/AKT/mTOR signaling pathway is critical for these effects. Interestingly, EGF effects on migration of PC3 cells were also blocked by these inhibitors, suggesting that PI3K/AKT/mTOR pathway is essential for migration of prostate cancer cells in response to multiple upstream signals.

Based on the results presented in this study, we propose (Figure 7) that TGF-β increases the levels of COX-2 and PGE₂ secretion by prostate cancer cells in the advanced stages of the disease. PGE₂, in turn, acting through EP4 receptors, activates PI3K/AKT/mTOR pathway, which is required for TGF-β and PGE₂ effects on cell migration and invasive behavior. Therefore, targeting of the downstream signaling components regulated by PGE₂ might be a better therapeutic approach for overcoming severe side effects and health risks of COX-2 inhibitors in prostate cancer patients.

Figure 7.

A proposed model suggesting that TGF-β increases COX-2 levels and PGE₂ secretion in prostate cancer cells and both TGF-β and PGE₂ activate PI3K/AKT/mTOR signaling pathway.