UCCB tissue samples and normal tissue control samples were obtained from the UC Davis Cancer Center Biorepository. SV-HUC-1, T24, TCCSUP, J82 and UM-UC-3 cells were generous gifts from Dr Sweeney and Dr De Vere White (UC Davis). The cell lines were cultured in 10% fetal bovine serum (FBS) RPMI-1640 supplemented with glutamine, penicillin and streptomycin at 37°C and 5% CO₂. Transfections were carried out using Lipofectamine RNAiMAX (Invitrogen-Life Technologies, Carlsbad, CA, USA) at an oligonucleotide concentration of 50 nM. siRNAs used were: ON-TARGET plus control pool (non-targeting), and Dicer (custom sequence, AAGGCUUACCUUCUCC AGGCUUU) (both from Dharmacon, Lafayette, CO, USA).

T24, TCCSUP, J82 and UM-UC-3 cells were plated at 10,000–25,000 cells/well in 24-well plates. The cells were transfected with either control non-targeting scrambled oligonucleotides or Dicer targeting siRNA and viability was assessed at 48 and 96 h using Cell Counting Kit-8 (CCK-8) viability detection reagent (Dojindo). Conditions were plated in triplicate. Data were presented as mean ± standard deviation.

Whole cell lysates were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to 0.2-µm nitrocellulose membranes. The membranes were blocked in 5% milk in 0.1% Tween-phosphate-buffered saline (PBS-T) and were allowed to incubate with primary antibody overnight at 4°C. The following day, the membranes were incubated with secondary antibody conjugated to horseradish peroxidase (HRP), developed using the WesternBright Sirius kit (Advansta, Menlo Park, CA, USA) and visualized with a Konica Minolta SRX-101A developer or a Li-Cor C-DiGit Scanner. Tubulin, actin or Ponceau S stain served as loading controls. The antibodies used were: poly(ADP-ribose) polymerase (PARP), E-cadherin, N-cadherin, MMP-2 (Cell Signaling Technology, Inc., Danvers, MA, USA), Dicer (Abcam, Cambridge, UK), tubulin (Thermo Fisher Scientific, Rockford, IL, USA) and actin (Millipore, Billerica, MA, USA).

RNA was extracted using either the RNeasy kit (Qiagen, Hilden, Germany) or the mirVana miRNA Isolation kit (Ambion-Life Technologies, Carlsbad, CA, USA) and reverse transcribed into cDNA using either the QuantiTect reverse transcription kit or the miScript II RT kit (both from Qiagen). qPCR was performed using the KAPA SYBR FAST Universal qPCR kit (Kapa Biosystems) or the miScript PCR kit (Qiagen) on a ViiA 7 Real-Time PCR System (Applied Biosystems). Data were analyzed using the efficiency-corrected ΔCt method. The primers used were: miR-205, miR-31, miR-200a, miR-200b, miR-200c, miR-148a, miR-149 and miR-106b miScript primer assays (Qiagen); HPRT, F-GCCAGA CTTTGTTGGATTTG and R-CTCTCATCTTAGGCTTTG TATTTTG; Dicer, F-TCCACGAGTCACAATCAACACGG and R-GGGTTCTGCATTTAGGAGCTAGATGAG.

Invasion and migration were assessed using Transwell assays with or without Matrigel membrane coating, respectively (Cell Biolabs, Inc., San Diego, CA, USA; CBA-101-C). Cell seeding was normalized using a parallel viability assay as described below. Briefly, the cells were plated and then treated 24 h later with a control non-targeting oligonucleotide or Dicer targeting siRNA. After 48 h, 10% FBS containing RPMI-1640 was placed in the bottom chambers of the wells and the cells were seeded into the top chambers in media with no serum. Twenty-four hours later, invaded/migrated cells were removed, lysed and read using a fluorescent plate reader at 480/520 nm. Invasion/migration was normalized to cell viability from a parallel assay utilizing the cells from the same suspensions used to seed the invasion/migration assays. Viability was read using the CCK-8 reagent as previously described. The conditions were carried out in triplicate. Data were presented as mean ± standard deviation.

Data were presented as mean ± standard deviation. A two-tailed, two-sampled equal variance Student's t-test was used to assess the differences between the samples. P≤0.05 was considered to indicate a statistically significant result.

Previous studies have demonstrated that Dicer expression is altered in the context of UCCB, albeit the nature of this expression alteration is unclear (10–12). Using qPCR, we examined the expression of Dicer in UCCB tumor samples. We found a statistically significant decrease in Dicer expression in tumor tissues as compared to normal tissues (Fig. 1A). Thus, we hypothesized that Dicer may have an unidentified tumor-suppressor role in this disease. Additionally, we performed western blot analysis to define the Dicer protein levels in the T24, TCCSUP, J82 and UM-UC-3 UCCB cell lines and we found that Dicer expression was variable (Fig. 1B).

We aimed to better understand how Dicer functions in UCCB by knocking down Dicer in the four cell lines mentioned above and assessing viability (Fig. 2A). We found that a decreased expression of Dicer led to the attenuation of cell growth in all the cell lines except for J82 (Fig. 2B). The results are consistent with previous studies that show that decreased Dicer impairs proliferation (12,13). The reason for J82 cells being refractory to this treatment is unclear. However, this result highlights known heterogeneity found in different tumors. These data suggest that in certain contexts, Dicer increases cell viability and plays an oncogenic role.

We assessed whether the decrease in cell viability in response to Dicer knockdown was due to an increase in apoptosis (Fig. 2C). Using western blot analysis to detect cleaved PARP, we demonstrated that a decreased Dicer expression induced apoptosis in T24 and UM-UC-3 cells. As with the viability assay above, we observed no change in J82. Notably, no change was observed in TCCSUP, suggesting that decreased viability previously noted in this cell line may not be due to apoptosis in these cells. Alternatively, apoptosis in this cell line may not result in PARP cleavage.

Recent evidence has shown that lower levels of Dicer in breast and ovarian cancers are associated with an epithelial-to-mesenchymal transition (EMT), a mesenchymal phenotype, and leads to increased motility (9,15,16). Since we and other investigators have shown that Dicer can be downregulated in UCCB, we hypothesized that Dicer may function similarly in this context and thus is also able to serve a tumor-suppressor role (10,11). To assess a possible role in EMT for Dicer, we examined the expression of the mesenchymal marker N-cadherin and the epithelial marker E-cadherin (Fig. 3). We found that Dicer knockdown led to an increase in N-cadherin expression in T24 and TCCSUP cells. Depletion of Dicer did not cause an alteration of N-cadherin expression in J82 cells. In UM-UC-3 cells, we detected a very low expression of N-cadherin which appeared to decrease with Dicer knockdown. E-cadherin was detected in TCCSUP cells but not in T24, J82 and UM-UC-3 cells. However, E-cadherin did not change in response to Dicer knockdown in TCCSUP cells. Additionally, Dicer depletion did not induce the expression of E-cadherin in T24, J82 and UM-UC-3 cells (data not shown). These data indicate that decreased Dicer expression promotes a more mesenchymal phenotype in specific cell contexts via the upregulation of N-cadherin. However, Dicer downregulation failed to promote complete EMT.

Additionally, we determined whether Dicer knockdown promotes a mesenchymal phenotype in an immortalized non-transformed cell line. Using SV-HUC-1 cells, we again assessed EMT marker expression and found that the attenuation of Dicer did not reduce E-cadherin expression (Fig. 3). N-cadherin was not detected in these cells and its expression was not induced by Dicer knockdown (data not shown).

A study by Nieman et al demonstrated that N-cadherin expression is more indicative of a motile and invasive phenotype rather than a loss of E-cadherin (17). They also showed that a forced expression of N-cadherin in E-cadherin-expressing cells promoted motility and invasion despite the presence of E-cadherin (17). Thus, we conclude from our EMT data that decreased Dicer expression can promote mesenchymal properties and therefore may promote a more motile and invasive phenotype via upregulation of N-cadherin in T24 and TCCSUP cells.

Using transwell migration assays, we determined whether decreased Dicer expression promotes migration and invasion (Fig. 4A and B). Because we hoped to augment function, we wanted to choose the line that would allow us to detect the most change. Based on previous studies of the invasive propensities of UCCB cell lines, we decided that TCCSUP cells would be best suited for these studies since they exhibit a low invasive capability compared to T24 (18). Furthermore, TCCSUP cells exhibited the most robust induction of N-cadherin expression in response to Dicer knockdown. We also used SV-HUC-1 in the present study (also shown to have low invasiveness) to further assess whether Dicer inhibition regulates motility and whether it may function to promote malignancy in a non-transformed context (19). Our data demonstrated that Dicer knockdown increased the migration of TCCSUP cells but only potentiated a trend towards increased motility in the SV-HUC-1 cells (p=0.06) (Fig. 4A). We conclude from these studies that Dicer can play a role in regulating motility of UCCB cells.

Studies have also shown that decreased levels of Dicer led to increased cell invasion (20). Thus, in addition to assessing migration, we examined whether Dicer regulates the invasive ability of UCCB cells (Fig. 4B). Using a matrigel-coated transwell assay, we demonstrated that Dicer knockdown increases invasion almost 2-fold in the TCCSUP and SV-HUC-1 cells suggesting that lower levels of Dicer may potentiate a more metastatic phenotype. Western blot analyses demonstrated that Dicer knockdown also led to increased levels of MMP-2 in the two cell lines (Fig. 4C). Additionally, we showed that Dicer knockdown increased the expression of MMP-2 in T24 and UM-UC-3 cells, but not in J82 cells Fig. 4C). It has been shown that MMP-2 expression and activity are associated with increased stage and invasiveness of bladder cancer, respectively (21,22). These data suggest that decreased Dicer may increase the invasive abilities of UCCB cells and demonstrate a mechanism by which Dicer plays a tumor-suppressor role.

We used qPCR to assess the expression of miRNAs associated with a motile/invasive phenotype (Fig. 5). Based on Dicer's known functions, we hypothesized that miRNAs may be ultimately responsible for the effects observed when Dicer is knocked down. In a study by Wszolek et al, a panel of invasion-associated miRNAs was determined using qPCR comparing 31 invasive UCCB lesions to 26 non-invasive UCCB lesions (23). From this panel, 5 miRNAs were selected to investigate in response to Dicer knockdown. We selected the well-characterized miR-200a/b/c in addition to miR-31, which has been shown to negatively correlate with UCCB patient progression and mortality, and miR-205 which was demonstrated to have strong discriminatory power in distinguishing invasive UCCB tumors from non-invasive UCCB tumors (24–26). Wzsolek et al showed that the overexpression of these miRNAs in UM-UC-3 cells attenuated invasion using a transwell matrigel-coated membrane assay (23). In response to Dicer knockdown, we found small decreases in the levels of miR-200a/b/c (Fig. 5). However, we found a >5-fold reduction in miR-205 levels and a ~2-fold decrease in miR-31 expression (Fig. 5). These data along with the study by Wzsolek suggested that decreased Dicer expression may potentiate cell invasion via decreased expression of miRNAs (23). However, we hypothesized that decreased Dicer could lead to global downregulation of miRNAs. Using qPCR, we assessed the expression of additional miRNAs (miR-148a, miR-149 and miR-106b) associated with invasion in other contexts and found these were also downregulated in response to attenuated Dicer levels (Fig. 5) (27–29). Our data suggest that pan miRNA expression may be decreased along with decreased Dicer levels and that the global downregulation of miRNAs promotes an invasive phenotype.