Human 769-P renal carcinoma cells, A549 human lung epithelial cells, HeLa cervix adenocarcinoma cells, and PC3 prostate adenocarcinoma cells were obtained from the American Type Culture Collection (ATCC, USA) and cultivated in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS), 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin in a 37°C CO₂ incubator.

LY294002 (LY), SB203580 (SB), PD98059 (PD), and MG132 were obtained from Sigma (USA). SP600125 was obtained from Cayman. β-catenin inhibitor BH535 was obtained from Calbiochem. TGF-β was obtained from Peprotech (USA). Antibodies against HIF-1α, E-cadherin, N-cadherin, and β-catenin were obtained from BD Biosciences. Antibodies against fibronectin, vimentin, tubulin, and actin were obtained from Sigma. Antibodies against HA were obtained from Covance. Antibodies against CA9, S6K, S6, cyclin D1, and GAPDH were obtained from GeneTex. Antibodies against mTOR, mTOR-Ser2448, Akt-Ser473, Akt, S6K-Thr389, S6-Ser235/236, GSK-3β, GSK3β-Ser9, Smad2-Ser465/467, active β-catenin, and Snail were obtained from Cell Signaling Technology.

The sequences targeted by the anti-Snail siRNA were 5′-GGUGUGACUAAUGCAATT-3′ and 5′-UUGCAUAGUUAGUCACACCTT-3′. siRNAs were synthesized by and purchased from Biotools (USA). Transfection of siRNA was conducted using Lipofectamine RNAiMAX (Invitrogen, USA) at a final concentration of 0.1–0.3 μM. ON-TARGETplus SMARTpool siRNAs (Dharmacon, Chicago, IL, USA) targeting human Akt or mTOR were transfected as siRNA duplexes for 48 h. Knockdown of HIF-1α achieved by transfection with the pSUPER-based plasmid expressing sh-HIF-1α, which was a generous gift from Dr K.J. Wu (Taiwan). Transfection of sh- HIF-1α siRNA was performed according to the manufacturer's protocol. Briefly, cells in the exponential phase of growth were plated in 6-well tissue culture plates at a density of 1.5×10⁵ cells/well, grown for 24 h, and transfected with siRNA using Lipofectamine 2000 and OPTI-MEM serum free medium. After 48 h, cells were harvested and lysed for western blotting.

Cells were lysed in TEGN buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 420 mM NaCl, 10% glycerol, and 0.5% Nonidet P-40) containing 1 mM Na₃VO₄ and 1 mM protease inhibitor cocktail (Roche). For immunoprecipitation, lysates were diluted 1:1 with TEG buffer (10 mM Tris, pH 7.5, 1 mM EDTA, and 20% glycerol) and rocked with antibodies and 20 μl of 50% protein G beads (Upstate) for 3 h at 4°C. The immunoprecipitates were washed four times in TEG:TEGN buffer (1:1), boiled in protein sample dye (2 M β-mercaptoethanol, 12% SDS, 0.5 M Tris, pH 6.8, 0.5 mg/ml bromophenol blue, and 30% glycerol), analyzed by SDS-polyacrylamide gel electrophoresis (PAGE), and subjected to western blotting.

Transient transfections of cancer cell lines were performed using Lipofectamine 2000 (Invitrogen). Whole cell extracts were prepared at specific time-points after transfection and subjected to SDS-PAGE/western blotting or immunoprecipitation assays as described previously (12).

Cells were plated on cover slips 1 day before drug exposure. After treating the cells with TGF-β with or without Rha or LY for the specified times, the cells were fixed with 3% paraformaldehyde and 2% sucrose in PBS for 10 min. Coverslips were permeabilized with Triton X-100 buffer (0.05% Triton X-100, 20 mM HEPES-KOH, pH 7.9, 50 mM NaCl, 3 mM MgCl₂, 300 mM sucrose) for 10 min and blocked in PBS containing 1% BSA for 1 h at room temperature, after which the cells were washed three times with PBS and incubated with Alexa Fluor 488- or rhodamine-conjugated secondary antibodies (Molecular Probes) in Hoechst 33342 stain containing PBS for 1 h. Finally, the treated cells were washed with PBS three times and mounted in Fluoromount™ Aqueous Mounting Medium (Sigma).

Cells (5×10⁵) in 6-well plates were pre-treated with or without LY294002 for 30 min, followed by exposure to TGF-β for 24 h. Later, the cells were trypsinized and counted, after which 1.25×10⁵ cells were seeded in Matrigel-coated Transwell inserts (24-well, 8-μm pore size; BD Biosciences, Canada) with serum-free medium with inhibitors or inhibitors/TGF-β. DMEM with 10% FBS was added to the lower chamber to act as a chemo-attractant. After incubation for 48 h, the inner cells that did not invade the lower chamber via the pores were removed with a cotton swab. The cells that adhered to the underside of the membrane were fixed in methanol and stained with 0.5% crystal violet. For each sample, six random fields were counted and examined under a microscope with a x100 objective to determine the number of cells that had invaded across the membrane. Three independent experiments were conducted.

Quantitative measurements of the obtained results were performed using ImageQuant software following the manufacturer's protocol (Molecular Dynamics) and presented as mean + standard error of the mean (SEM) for three independent experiments (n=3). Statistical analyses were performed using Student's t-test. Results were considered significant at p-value <0.05 (P<0.05).

We investigated the mechanism underlying TGF-β-induced EMT in A549 (lung), HeLa (cervical), and 769-P (renal) cancer cells, all of which have been established as suitable models for studying EMT in vitro (12–15). As expected, TGF-β treatment resulted in the acquisition of spindle-fibroblastic morphology, downregulation of epithelial marker E-cadherin, and upregulation of mesenchymal markers fibronectin (FN), vimentin and N-cadherin (Fig. 1A and C). FN induction was observed at day 1 (Fig. 1A, upper panel). TGF-β dose-dependently induced expression of FN and N-cadherin, as well as phosphorylation of Smad2 (Fig. 1B). E-cadherin reduction by TGF-β occurred on day 3 in in A549 cells (Fig. 1A, bottom panel). Additionally, TGF-β treatment for 2 days led to upregulation of HIF-1α expression in 769-P cells (Fig. 1E). Changes in protein expression were confirmed by immunofluorescence (IFA) experiments. When 769-P cells were stimulated with TGF-β for 2 days, FN expression was slightly increased. IFA experiments also confirmed upregulation of FN in 769-P cells (Fig. 1F). Similar protein alterations were observed in HeLa cells (Fig. 1D).

Several studies suggest a role for PI3K in TGF-β signaling (16–18). The involvement of various signaling pathways in regulating TGF-β-induced EMT was assessed. A series of pharmacological inhibitors, including LY (PI3K inhibitor), PD (MAPK inhibitor), SB (p38 inhibitor), and SP (JNK inhibitor), were used to determine the involvement of various signaling pathways in regulating TGF-β-mediated responses. Serine/threonine kinase Akt, a well-known downstream effector of PI3K, is involved in cell survival signaling. Akt phosphorylation at serine 473 was strongly activated by TGF-β. When cells were treated with LY, PD, SB, or SP, AKT phosphorylation at serine 473 was downregulated in HeLa cells (Fig. 2A). Relative to SB and SP, LY more effectively abolished phosphorylation of Akt, indicating that TGF-β activation may occur via the PI3K/Akt signaling.

Next, in order to verify the importance of the PI3K pathway in TGF-β regulation, 769-P cells were pre-treated with or without PI3K inhibitor LY for 30 min, followed by co-treatment with TGF-β for 20 h. As shown in Fig. 2B and C, LY dramatically blocked TGF-β-mediated expression of EMT markers in 769-P cells. Additionally, TGF-β-induced expression of HIF-1α and EMT-related genes Snail, FN, and N-cadherin were abrogated by LY in HeLa and A549 cells, implicating the PI3K signaling pathway in the regulation of TGF-β-mediated responses (Fig. 2D and E). To assess the role of Snail in TGF-β-induced EMT (16,18), Snail protein was knocked-down via small interfering RNA (siRNA) in 769-P cells. In 769-P cells lacking Snail protein, TGF-β-induced EMT was abolished, demonstrating the crucial role of Snail in the EMT (Fig. 2F). These results suggest that Snail may play a crucial role in TGF-β-induced EMT.

To verify our hypothesis regarding Rha, we assessed the impact of Rha on TGF-β-mediated Snail expression, which is a transcriptional regulator active during EMT. Therefore, various cancer cell types were exposed to Rha in the presence or absence of TGF-β. Together with previous studies, the results described above show that TGF-β can alter expression of EMT-related genes, including Snail (10,11,19,20). As shown in Fig. 3A–C, TGF-β triggered increased Snail expression in diverse cancer cells. However, when cancer cells were pre-treated with Rha, TGF-β-induced Snail expression was blocked by Rha in all tested types of cancer cells. To examine whether the proteasome degradation pathway is involved in Rha-induced HIF-1α protein degradation, 769-P cells were stimulated with TGF-β or co-treated with Rha for 20 h, followed by exposure to 26S proteasome inhibitor MG132 for 6 h. As shown in Fig. 3D, MG132 completely blocked the reduction in TGF-β-induced Snail expression induced by Rha, suggesting that Snail was recruited to the 26S proteasome for degradation (Fig. 3D, upper). Similar results were observed in PC3 cells (Fig. 3D, bottom), indicating that the ability of MG132 to reverse Snail expression by Rha was ubiquitous.

In previous studies, Rha was recognized as an inhibitor of HIF-1α accumulation in PC3 cells. Therefore, TGF-β mediated downregulation of HIF-1α expression by Rha was investigated. Rha reduced protein levels of HIF-1α in a concentration-dependent manner (Fig. 4A). In addition, protein levels of EMT-related markers N-cadherin, vimentin, and the noted HIF-1α target gene carbonic anhydrous IX (CA9) were reduced by Rha in a concentration-dependent manner (Fig. 4B). The mechanism underlying the effect of Rha on HIF-1α was assessed. As shown in Fig. 4C, MG132 completely abolished Rha-induced degradation of HIF-1α, indicating that the reduction in HIF-1α expression induced by Rha was partially due to degradation by the 26S proteasome. Next, to determine whether HIF-1α was recruited to the 26S proteasome as a result of ubiquitination, an immunoprecipitation (IP) experiment was performed, in which PC3 cells were transfected with a plasmid containing HA-tagged ubiquitin for 24 h, followed by exposure to TGF-β in the presence or absence of Rha for 24 h. Finally, the cells were treated with MG132 for 6 h. As shown in Fig. 4D, PC3 cells co-treated with TGF-β and Rha had approximately 3-fold the level of HA-tagged ubiquitin as those treated only with TGF-β. Because of the changes in EMT markers that accompanied HIF-1α expression, we assessed the involvement of HIF-1α in TGF-β-mediated changes in EMT markers. Therefore, 769-P cells were transfected with pSuper-HIF-1α or the pSuper control plasmid for 24 h, followed by exposure to TGF-β with or without Rha for 24 h. As shown in Fig. 4E, HIF-1α silencing had little impact on alterations in EMT markers, indicating that HIF-1α was not a critical factor in TGF-β-mediated EMT. Collectively, these results indicate that Rha treatment promotes TGF-β-induced HIF-1α ubiquitination, subsequently leading to HIF-1α degradation by the 26S proteasome. However, HIF-1α itself is not required for TGF-β-induced EMT.

Previous studies demonstrated the importance of glycogen synthase kinase-3β (GSK-3β) and mTOR as downstream effectors of Akt following activation by TGF-β in diverse cell types (21,22). However, the functional relevance of Akt and its downstream effector pathways in TGF-β signaling remains largely unknown. We tested the involvement of pathways downstream of Akt in the regulation of TGF-β-mediated HIF-1α expression. The role of GSK-3β in the regulation of TGF-β-mediated HIF-1α expression was investigated. When cells are exposed to Wnt, GSK-3β is phosphorylated and becomes inactive. When Wnt signaling is blocked, active GSK-3β leads to phosphorylation, ubiquitination, and degradation of β-catenin. In brief, β-catenin retains the GSK-3β consensus motif for ubiquitination (23,24). As shown in Fig. 5A, Wnt/β-catenin inhibitor FH535 inhibited TGF-β-induced expression of HIF-1α and phosphorylation of GSK-3β at serine 9 in a dose-dependent manner. As the concentration of FH535 was increased, the decreased protein level of HIF-1α was concomitant with the decrease in phosphorylation of GSK-3β at serine 9 and active (non-phosphorylated) β-catenin in PC3 cells. In addition, FH535 treatment also significantly decreased the protein level of cyclin D1 in a dose-dependent manner in PC3 cells, demonstrating the association between β-catenin and cyclin D1 (25,26). We next examined the inhibitory effect of Rha on the mTOR and Akt signaling pathways. As shown in Fig. 5B, Rha abolished mTOR signaling by reducing phosphorylation of mTOR at Ser2448, S6K at Thr389, and S6 at Ser235/236. In addition, phosphorylation of molecules involved in Akt signaling, such as Akt (Ser473), GSK-3β (Ser9), and β-catenin, was also attenuated by Rha treatment. Furthermore, expression of HIF-1α downstream gene CA9 was also reduced by Rha treatment. These results suggest that Akt and mTOR may be important for TGF-β-induced expression of HIF-1α; therefore, to confirm this hypothesis, 769-P cells were transfected with siRNA against Akt and mTOR for 24 h, followed by exposure to TGF-β in the presence or absence of Rha for 24 h, after which the cells were collected for western blotting. As shown in Fig. 5C, knockdown of Akt and mTOR by siRNA markedly attenuated TGF-β mediated HIF-1α expression in 769-P cells, indicating that Akt and mTOR are required for TGF-β mediated expression of HIF-1α. However, the effect of Akt inhibition was more significant than that of mTOR inhibition. Similar effects were observed in A549 cells (Fig. 5D).

Taken together with previous studies, the results described above prompted us to investigate whether the alterations in EMT induced by TGF-β enhanced invasiveness and metastasis (5,15,16,27–30). Therefore, IFA was performed. As shown in Fig. 6B, cells exposed to TGF-β for 48 h had higher expression of mesenchymal marker FN. In contrast, cells treated with LY or Rha in the presence of TGF-β lacked FN protein. Because of the connection between increased FN expression and cell invasiveness, cell invasion assays were employed to confirm the inhibitory effect of Rha. 769-P cells were exposed to TGF-β in the presence or absence of Rha for 20 h, followed by re-trypsinization and re-seeding of the treated cells into an invasion chamber harboring Matrigel (1.25×10⁵ cells). After 48 h of incubation, the invasive ability of the cells that had penetrated through the inserts was assessed by crystal violent staining microscopic examination. Cells stimulated with TGF-β had an enhanced invasive ability in comparison with that of the control cells. Moreover, cells exposed to TGF-β and Rha had less invasive ability than cells treated with TGF-β only (Fig. 6A). Taken together, our results demonstrate that TGF-β is a contributing factor in cancer progression and metastasis. Moreover, our results suggest that Rha may suppress TGF-β-mediated invasion via inhibition of the PI3K/Akt/mTOR/GSK-3β/β-catenin signaling pathway (Fig. 6C).