The study protocol was approved by the Medical Ethics Committee of Tongji Medical College and performed according to the declaration of Helsinki. All patients gave their written informed consent before participating in this study. The University of Padova and the Thomas Jefferson University Institutes’ ethics regulations on research conducted on human tissues were followed.

Cell lines T24, ScaBER and BIU-87 (bladder cancer); PC-3 and DU145 (prostate cancer); A549 (lung cancer); HeLa (cervical cancer); AGS (gastric cancer); A375 and A875 (melanoma); HepG2 and SMMC-7721 (hepatocarcinoma); MDA-MB-435, MDA-MB-231 and MCF-7 (breast cancer); Raji and K562 (leukemia); and L02 (embryo hepatocyte derived) were involved in this study. BIU-87, A875, MDA-MB-435, SMMC-7721, Raji, K562 and L02 were purchased from China Center for Type Culture Collection (CCTCC, Wuhan, China). T24, ScaBER, PC-3, DU145, A549, HeLa, AGS, A375, HepG2, MDA-MB-231 and MCF-7 were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco-BRL, Carlsbad, CA, USA) supplemented with 100 ml/l fetal bovine serum (HyClone, Logan, UT, USA) at 37°C in a humidified atmosphere containing 5% CO₂.

Plasmids used in this study included pIRES2-EGFP-TβRII-Fc (TβRII-Fc) and pIRES2-EGFP (control). Those interested in further details are kindly referred to the following citation (25). Plasmids were transfected into L02, T24 or ScaBER cell line using Lipofectamine Plus reagent, according to the manufacturer’s (Invitrogen, Carlsbad, CA, USA) instruction.

siRNAs, siRNA of Smad2 and Smad3 and their controls, on the other hand, were purchased from RiboBio Company (Guangzhou, China), and transfected into cells also using Lipofectamine Plus reagent, according to the manufacturer’s instruction. Forty-eight hours after the transfections, the cells were sorted for the detection of GFP expression or other experiments.

Total RNA was extracted from cells with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. For real-time RT-PCR assays, the cDNA sequences of genes were retrieved from NCBI database. The primers were designed with the Oligo Primer Analysis 4.0 software and the sequences were blasted (http://www.ncbi.nlm.nih.gov/BLAST/). Real-time RT-PCR was done as described previously (26). The mRNA level of the detected gene was expressed as the relative level to that of β-actin. The sequences of the primers were as follows: β-actin, 5′-CCTAGAAGCATTT GCGGTGG-3′ (sense) and 5′-GAGCTACGAGCTGCCT GACG-3′ (antisense); TGF-β1, 5′-ACTACTACGCCAAGGA GGTCAC-3′ (sense) and 5′-GAGCAACACGGGTTCAGGT-3′ (antisense); TGF-βRII, 5′-AGACGGCTCCCTAAACAC TAC-3′ (sense) and 5′-GAATGCTCTATGTCACCCACTC-3′ (antisense); TGF-βRI, 5′-GACAACGTCAGGTTCTGG CTCA-3′ (sense) and 5′-ATCGACCTTTGCCAATGCT TTC-3′ (antisense); p15, 5′-GGCAGACAGGTTTAGCTGTTT CATG-3′ (sense) and 5′-CCACAATGGAGCTAGAAGCA GGA-3′ (antisense); CUTL1, 5′-AAAGACCAGCCTGAAAGT CGG-3′ (sense) and 5′-CCAGGGATGAGCTGAAAAAGT-3′ (antisense); and Cdc25A, 5′-CTCCTCCGAGTCAACAGAT TCA-3′ (sense) and 5′-CAGCCACGAGATACAGGTCTTA-3′ (antisense).

A total of 46 clinical bladder cancer specimens were consecutively acquired during the period of October 2009 to October 2010, with approval by the Ethics Committee of the Medical Faculty of Tongji Medical College, by transurethral resection or radical cystectomy from untreated cancer patients (Table I). Informed consent was obtained in accordance with the Declaration of Helsinki from all subjects. Data on the patients’ clinical and pathologic states were collected, including sex, age, tumor size, pathologic T stage, tumor grade and multiplicity. The pathologic stage of bladder cancer was assessed according to the 2002 UICC TNM tumor stage classification by: i) the superficial bladder cancer (T1) includes pTa and pT1 tumors; and ii) the muscle invasive bladder cancer (T2) includes pT2, pT3 and pT4. Tumor grade was assessed according to the 2004 WHO/International Society of Urologic Pathology grading classification by: i) the well differentiated papillary urothelial neoplasm includes low malignant potential and low grade tumor (low); and ii) the poorly differentiated papillary urothelial neoplasm which includes high grade bladder cancer (high).

Fresh bladder cancer specimens were used for tumor cell isolation as described previously (27). Briefly, tumor tissue was washed three times in cold DMEM medium containing 1% FBS and digested with hyaluronidase (Sigma-Aldrich, St. Louis, MO, USA), collagenase and DNase for 1 h at 37°C. After grinding with semifrosted slides and lysis of RBC, the dissociated cells were incubated on ice for 20 min and then spun down at 500 rpm for 1 min. This process was repeated twice and the cells were first incubated for 2 h to get rid of adhesive cells. Tumor cells were then cultured in DMEM supplemented with 10% FBS, 2 mmol/l L-glutamine, 1.0 mmol/l sodium pyruvate, 100 U/ml penicillin G sodium, and 100 μg/ml streptomycin sulfate in 6-well plate in a humidified incubator at 37°C with 5% CO₂.

Total RNA, extracted from cell lines or primary bladder cancer cells, was reversely transcribed to cDNA. The latter was used to amplify the whole coding sequence of TGF-βRII by PCR with primers (5′-CTGGAAGATGCTGCTT CTC-3′ and 5′-ACTGCTCTGAAGTGTTCTGC-3′). PCR products were sequenced by Sangon Biotech (Shanghai, China).

Cells were incubated with 1 μg of human IgG/10⁵ cells for 15 min at room temperature prior to staining, and then stained with phycoerythrin-conjugated anti-human TGF-β1 or TGF-βRII antibodies and their isotypes (eBioscience, San Diego, CA, USA). The stained cells were used for flow cytometric analysis (BD Biosciences, LSR II).

To detect TGF-βRII protein, cells were grown in glass slides overnight. After washing twice, slides were fixed in dry acetone for 10 min at room temperature and air-dried for another 10 min. Rabbit anti-human TGF-βRII primary antibody (Millipore), biotinylated goat anti-rabbit IgG, and streptavidin-conjugated horseradish peroxidase (eBioscience) were used for immunohistochemical staining.

Cells were incubated with PKH-26 dye for membrane staining, and then cultured in the presence or absence of TGF-β1 (2 ng/ml, PeproTech) in 24-well plate, and 24 to 48 h later, the proliferation was analyzed by flow cytometry and expressed with stimulation index (SI).

Bladder cancer cell motility and invasion were evaluated by Transwell assay as described previously (28). In migration or invasion experiment, cells were allowed to reach confluence in serum-containing complete growth medium and then incubated for 16 h in serum-free medium before treatment of TGF-β1 or TβRII-Fc and smad2+3 siRNA transfection. Matrigel invasion assay was performed in Transwell plates with polycarbonate membrane filters (Corning, Corning, NY, USA). Precoated filters (6.5 mm in diameter, 8 μm pore size, Matrigel 100 μg/cm²) were rehydrated with 0.1 ml medium. Then, 2×10⁵ pretreated cells in 0.2 ml DMEM were added to the top chamber. Medium (0.6 ml) supplemented with 20% fetal bovine serum was added to each well of the plate to act as a chemoattractant in the lower chamber. Following incubation for 18 h at 37°C, non-invading cells at the upper surface of the filter were wiped off with a cotton swab, and the invading cells at the lower surface of the filter were fixed for 5 min in 100% methanol and stained with hematoxylin and eosin. Cells that moved through the insert were counted in five random fields and expressed as the average number of cells per field. Experiments were repeated in triplicate. Transwell migration assays were done using the same procedure but without coating filters with Matrigel.

Co-immunoprecipitation was performed as described previously (29). Briefly, cells were harvested at 0, 5, 10 and 20 min following TGF-β1 (2 ng/ml) treatment. Cell extracts were first precleared with 25 μl of protein A-agarose (Sigma-Aldrich). The supernatants were immunoprecipitated with anti-TGF-βRII antibody for 1 h at 4°C, followed by incubation with protein A-agarose overnight at 4°C. The complexes were collected by centrifugation for western blot analysis.

Cell lysates and prestained m.w. markers were separated by SDS-PAGE followed by transfer onto nitrocellulose membranes. The membranes were blocked in TBST (Tris-buffered saline with 0.5% of Triton X-100) containing 5% non-fat milk or BSA, and probed with primary antibodies (R&D Systems). After incubation with the secondary Ab conjugated with HRP, membranes were extensively washed, and the immunoreactivity was visualized by ECL according to the manufacturer’s protocol (ECL kit, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Antibodies smad2/3, psmad2/3, β-actin, p15, CUTL1 and Cdc25A were purchased from Santa Cruz Biotechnology.

Results were expressed as mean values ± SD and interpreted by ANOVA or χ² test (in statistics of patient sample, we used χ² continuity correction when 1≤T<5). SPSS version 12.0 (Chicago, IL, USA) was used for statistical analysis. Differences were considered to be statistically significant when P<0.05.

This study first examined TGF-β1 expression and its receptors TGF-βRI and TGF-βRII by real-time RT-PCR in human tumor cell lines derived from a variety of tissues, including breast, liver, lung, gastric, prostate, skin, bladder and bone marrow. As expected, all the tumor cell lines were shown to highly express TGF-β1 but to have low expression of TGF-βRI (Fig. 1A), which was consistent with previous reports (14–16,20–23). Unexpectedly however, we found that although TGF-βRII was weakly expressed in most tumor cell lines but strongly expressed in T24 bladder cancer cells (Fig. 1A). Such abnormal high expression of TGF-βRII in T24 cells was further confirmed by flow cytomety and cellular immunohistochemical staining (Fig. 1B and C). Thus, bladder cancer cell line T24 was identified as highly expressing TGF-βRII.

Next, this study turned to the question of whether TGF-βRII in T24 bladder cancer cells was functional. TGF-β1 impacts on cell growth arrest are well known in the literature (12–14). First we performed the in vitro proliferation assay, and found that the addition of TGF-β1 effectively inhibited the growth of normal liver cell line L02 and the bladder cancer cell line ScaBER, but not T24 cell growth (Fig. 2A). On the other hand, an in vitro Transwell assay found that the addition of TGF-β1 markedly promoted the migration and invasion of T24 cells (Fig. 2B, C and D). To test and confirm these initial results, a comparable approach was used to transfect TGF-βRII-Fc vector or smad2/3 siRNA, resulting in the blocking of TGF-βRII or silencing smad2/3, the downstream molecules of TGF-βR signaling (Fig. 2E). Under such conditions, we found that the above effects of TGF-β on T24 cells were relieved (Fig. 2A, B, C and D). These findings suggest that TGF-βRII does not mediate T24 cell growth arrest but is capable of promoting T24 cell migration and invasion through the TGF-β signal pathway.

Having identified a vital role for TGF-βRII in cell migration and invasiveness, this study turned next to explore the possible molecular mechanism through which TGF-β1 binding TGF-βRII resulted in the failure of growth arrest but the enhancement of motility in T24 bladder cancer cells. The literature has already identified the mechanisms of TGF-β1-mediated growth arrest which involve upregulating cyclin-dependent kinase (CDK) inhibitor p15 and downregulating CDK4/6-activating phosphatase Cdc25A (30–32). Looking further, this study found that TGF-β1 stimulation was capable of upregulating p15 and downregulating Cdc25A in both ScaBER and L02 cell lines without any undue effect on the expression of p15 and Cdc25A in T24 cell line, evaluated by both RT-PCR and real-time PCR (Fig. 3A and B). Consistently, p15 and Cdc25A proteins in T24 cells were not affected by TGF-β stimulation, evaluated by a western blot analysis (Fig. 3C). On the other hand, homeobox transcription factor CUTL1 has been well demonstrated as a critical target of TGF-β1 signaling to mediate the promotion of cancer cell motility and invasiveness (33–35). Therefore, besides p15 and Cdc25A, this study was able to also determine the expression of CUTL1. Interestingly, the expression of CUTL1 was significantly upregulated by the addition of TGF-β1 in T24 cells but not in L02 or ScaBER cells (Fig. 3A, B and C). Therefore, in T24 bladder cancer cells, the TGF-βRII signaling pathway is ineffective in the regulation of p15 and Cdc25A expression but highly efficacious in the regulation of CUTL1 expression, leading to the evasion of growth arrest and the enhancement of motility and invasion.

Further investigation was necessary into the molecular basis by which TGF-β1 signaling had no effect on p15 and Cdc25A expression but was efficacious in the upregulation of CUTL1 in T24 cells. In bulk tumor cell population the growth of TGF-βR^high tumor cells is declined due to TGF-β1 signaling-mediated growth arrest, leading, in turn, to the domination of TGF-βR^low/neg tumor cells, depending on the type of cultured cells involved. Paradoxically, T24 bladder cancer cell line maintains the high expression of TGF-βRII. To explain this, it was reasonable to speculate that T24 cells employed a mutation strategy to encrouch TGF-β signaling. In this regard, a pair of primers was designed to amplify 1707 bp cDNA fragment covering the coding region of TGF-βRII. By 25 cycles, the visible PCR products from T24 and ScaBER cell lines were confirmed in an agarose gel (Fig. 4A). Further sequencing analysis indicated that TGF-βRII was not mutated in either ScaBER or BIU-87, but mutated with several single nucleotides in T24 (Fig. 4B and C). Notably, one point mutation was the GAG to AAG transition, leading to corresponding amino acid transition of Glu²⁶⁹→Lys in the cytoplasmic domain of TGF-βRII. Glu and Lys belong to acidic and basic amino acids, respectively. Thus, this Glu²⁶⁹→Lys mutation might profoundly change the electric property of TGF-βRII, leading to altered conformations and abnormal signaling pathways.

The fact, that Smad2/3 are the classical downstream molecules for TGF-β signaling transduction (23,24), begged the question of whether Smad2/3 were also required for the mutated TGF-βRII signaling pathway in T24 cells. This, in turn, first called for the measurement of the active form of Smad2/3 by measuring its phosphorylation state by western blot analysis. It was found that the phosphorylation was induced 30 min after the stimulation of 2 ng/ml TGF-β1 in L02 and ScaBER cells as well as T24 cells (Fig. 5A), suggesting that the mutated TGF-βRII signaling transduction might be through the classical Smad2/3 pathway. Moreover, by pulling down the TGF-βRII complex with anti-TGF-βRII antibody, it was observed that Smad2/3 proteins were bound to the complex (Fig. 5B), suggesting the interaction of Smad2/3 and TGF-βRII in T24 cells. When silencing Smad2 and Smad3 in T24, the binding complex can not be detected with additional TGF-β1, expression of CUTL1 then having almost no change (Fig. 5C), indicating that elevated level of CUTL1 was Smad2/3 dependent in T24 cells. Taken together, these findings suggested that the transduction of TGF-β signaling by mutated TGF-βRII is Smad2/3-dependent in bladder cancer T24 cell line.

To translate the implications of the above in vitro data in vivo, the sequence of TGF-βRII cDNA in bladder cancer patients was further analyzed. Primary bladder cancer cells were isolated from fresh specimens of bladder cancer patients (n=46, Table I), and then used for RT-PCR to amplify the TGF-βRII cDNA (25 cycles) for sequencing. As expected, similar mutation patterns were found in primary bladder cancer cells as in T24 cells. A total of 18 specimens showed GAG→AAG mutation (39.1%), which was also companied by other mutations observed in T24 cells (Fig. 4C). Therefore, it is safe to say that GAG→AAG mutation of TGF-βRII also occurs in bladder cancer patients. To explore the possible clinical significance of such point mutations, patients were subdivided into 2 groups: non-mutated TGF-βRII (group 1, n=28) and point-mutated TGF-βRII (group 2, n=18). In cases of bladder cancer, tumor grade is used to reflect the relapse and metastasis, and pathologic T stage, however, may reflect the degree of infiltration and invasion by tumor cells. By comparing the clinical and pathologic features of 2 groups, significant differences were found in both pathologic T stage and tumor grade stages (P<0.01; Table II). Together, these findings suggested that GAG→AAG mutation of TGF-βRII may also occur in patients with bladder cancer and such TGF-βRII mutations seem to be correlated with worse malignant phenotypes and poor prognosis.

The mutation-induced relief from growth arrest and strengthened motility in T24 bladder cancer cell line TGF-βRII were found to be similarly at work in primary bladder cancer cells. On the basis of the sequencing results, the primary bladder cancer cells from 12 patients were classified into two groups: no TGF-βRII mutation (n=6) and mutation with GAG→AAG (n=6). The expression of TGF-βRII between these two groups was then compared. The immunohistochemical staining showed that TGF-βRII protein was strongly expressed in the mutation with GAG→AAG group but not in the no mutation group (Fig. 6A). As a parallel observation, the primary bladder cancer cell growth was not affected by TGF-β treatments in the mutation group (Fig. 6B). On the other hand, TGF-β treatment enhanced the invasion of these cells from GAG→AAG group, which were impaired by TGF-βRII blockade (Fig. 6B). Measurements, furthermore, of the expression of p15, Cdc25A and CUTL1 in the cultured primary bladder cancer cells in the presence or absence of TGF-β, as expected, showed that p15 and Cdc25A were not affected, whilst CUTL1 was upregulated by TGF-β in GAG→AAG group (Fig. 6C and D). Therefore, TGF-βRII GAG→AAG mutation in bladder cancer patients might favor cancer cell migration and survival.