c-Myc Acts as a Competing Endogenous RNA to Sponge miR-34a, in the Upregulation of CD44, in Urothelial Carcinoma

MicroRNAs (miRNAs) have been shown to play a crucial role in the progression of human cancers, including urothelial carcinoma (UC), the sixth-most common cancer in the world. Among them, miR-34a has been implicated in the regulation of cancer stem cells (CSCs); however, its role in UC has yet to be fully elucidated. In this study, bioinformatics and experimental analysis confirmed that miR-34a targets CD44 (a CSC surface marker) and c-Myc (a well-known cell cycle regulator) in UC. We found that, surprisingly, most UC cell lines and patient samples did express miR-34a, although epigenetic silencing by promoter hypermethylation of miR-34a expression was observed only in UMUC3 cells, and a subset of patient samples. Importantly, overexpression of c-Myc, a frequently amplified oncogene in UC, was shown to upregulate CD44 expression through a competing endogenous RNA (ceRNA) mechanism, such that overexpression of the c-Myc 3′UTR upregulated CD44, and vice versa. Importantly, we observed a positive correlation between the expression of c-Myc and CD44 in clinical samples obtained from UC patients. Moreover, overexpression of a dominant-negative p53 mutant downregulated miR-34a, but upregulated c-Myc and CD44, in UC cell lines. Functionally, the ectopic expression of miR-34a was shown to significantly suppress CD44 expression, and subsequently, suppression of cell growth and invasion capability, while also reducing chemoresistance. In conclusion, it appears that aberrant promoter methylation, and c-Myc-mediated ceRNA mechanisms, may attenuate the function of miR-34a, in UC. The tumor suppressive role of miR-34a in controlling CSC phenotypes in UC deserves further investigation.

Previous studies have demonstrated that dysregulation of miRNA, through promoter CpG island methylation, is an important mechanism in carcinogenesis [20]. Thus, we sought to identify an epigenetic mechanism downregulating miR-34a, in UMUC3 cells. Bisulphite pyrosequencing ( Figure 1B), and combined bisulfite restriction analysis (COBRA, Figure 1C) assays were used to determine the methylation status of the CpG island in which miR-34a resides. In agreement with the RT-PCR results ( Figure 1A), high miR-34a methylation levels were observed only in UMUC3 cells. In determining whether DNA methylation was responsible for epigenetic silencing of miR-34a, we found that epigenetic treatment with a DNA methyltransferase inhibitor, 5 aza-2 -deoxycytidine (5-AZA), and/or histone deacetylase (HDAC) inhibitor (trichostatin A, TSA), led to robust re-expression of miR-34a, in UMUC3 cells only ( Figure 1D), due to demethylation of the miR-34a promoter region ( Figure 1E). These results suggest that miR-34a is epigenetically silenced in UMUC3 cells.
Further clinical studies, using 55 UC patient samples, found the miR-34a promoter to be devoid of DNA hypermethylation, with only few samples having methylation levels > 50% ( Figure 1F). Interestingly, one group of samples (n = 35) had low miR-34a expression, regardless of promoter methylation level ( Figure 1F, red dot). After exclusion of those samples, the remaining samples (n = 20) showed an inverse correlation between miR-34a expression and promoter methylation ( Figure 1G, r = −0.31). Together, these studies of UC cell lines and patient samples showed that even lowly expressing miR-34a samples were devoid of high levels of DNA methylation. Thus, an alternative mechanism may be involved in the suppression of miR-34a function in UC.

c-Myc Acts as a ceRNA of CD44 in Urothelial Carcinoma
One recent study revealed that mRNAs compete for binding to a shared set of miRNAs, thereby modulating miRNA-based regulation by acting as competing endogenous RNAs (ceRNAs) [21]. We therefore hypothesized that a ceRNA mechanism may be responsible for regulating miR-34a in UC. We then conducted a search of the intersection of three databases (TargetScan, miRanda, and PicTar) to identify any miR-34a targets overexpressed in UC. Interestingly, numerous studies have reported that c-Myc, a known miR-34a target [22], is amplified and upregulated in UC [23,24]. To determine whether miR-34a targets the c-Myc 3 UTR, in UC, we overexpressed miR-34a in UMUC3 cells with low miR-34a expression ( Figure 2A). As expected, transfection of miR-34a reduced mRNA and protein levels of c-Myc in UMUC3 cells ( Figure 2B). Studies have also shown that the stem cell marker, CD44 is a miR-34a target in UC [17]. Indeed, overexpression of miR-34a was shown to reduce both mRNA and protein of CD44 in UMUC3 cells ( Figure 2C). Taken together, these results suggest that miR-34a targets c-Myc and CD44 in UC.
The above results indicate that the ectopic expression of miR-34a repressed both c-Myc and CD44 in UC. Therefore, it was reasonable to assume a significantly positive correlation between CD44 and c-Myc (both targets of miR-34a and thus ceRNAs), in UC. To elucidate this ceRNA mechanism, we exogenously overexpressed the c-Myc and CD44 3 -UTRs, in TCCSUP and TSGH8301 UC cells having high endogenous miR-34a expression. Interestingly, overexpression of these 3 -UTRs significantly upregulated the expression of both CD44 and c-Myc mRNA and protein, in TCCSUP ( Figure 3A-C) and TSGH8301 ( Figure 3D,E) cells. Notably, these events were not due to a decrease, but slight increase, in miR-34a expression, both in pre-miR-34a and fully mature miR-34a ( Figure 4A,B). Taken together, these results demonstrate that overexpression of c-Myc 3 UTR leads to increased activation of CD44, and vice versa.

Dicer Knockdown Abolishes ceRNA Effect in UC
To determine whether c-Myc acts as a ceRNA of CD44 through miR-34a, we disrupted miRNA biogenesis via knockdown of the double-strand RNA endoribonuclease, Dicer [25], in TCCSUP and TSGH8301 cells. RT-PCR confirmed significant suppression of Dicer mRNA expression in both cell types transiently infected with anti-Dicer shRNAs ( Figure 5A,D). Compared to the negative-control (shGFP), overexpression of the c-Myc 3′UTR did not increase CD44 levels in Dicer-depleted TCCSUP ( Figure 5B) or TSGH8301 ( Figure 5E) cells. Reciprocally, overexpression of the CD44 3'UTR did not increase c-Myc levels in either Dicer-knocked down cell line ( Figure 5C,F). It is also noteworthy that TSGH8301 cells (with lower Dicer knockdown efficiency), had increased c-Myc and CD44 expression levels, following overexpression of either 3′UTR, in poorly knocked down Dicer cells ( Figure 5D-F). Taken together, these results confirm that c-Myc and CD44 act as ceRNAs, through miR-34a, in UC.

Dicer Knockdown Abolishes ceRNA Effect in UC
To determine whether c-Myc acts as a ceRNA of CD44 through miR-34a, we disrupted miRNA biogenesis via knockdown of the double-strand RNA endoribonuclease, Dicer [25], in TCCSUP and TSGH8301 cells. RT-PCR confirmed significant suppression of Dicer mRNA expression in both cell types transiently infected with anti-Dicer shRNAs ( Figure 5A,D). Compared to the negative-control (shGFP), overexpression of the c-Myc 3 UTR did not increase CD44 levels in Dicer-depleted TCCSUP ( Figure 5B) or TSGH8301 ( Figure 5E) cells. Reciprocally, overexpression of the CD44 3'UTR did not increase c-Myc levels in either Dicer-knocked down cell line ( Figure 5C,F). It is also noteworthy that TSGH8301 cells (with lower Dicer knockdown efficiency), had increased c-Myc and CD44 expression levels, following overexpression of either 3 UTR, in poorly knocked down Dicer cells ( Figure 5D-F). Taken together, these results confirm that c-Myc and CD44 act as ceRNAs, through miR-34a, in UC.
(shGFP), overexpression of the c-Myc 3′UTR did not increase CD44 levels in Dicer-depleted TCCSUP ( Figure 5B) or TSGH8301 ( Figure 5E) cells. Reciprocally, overexpression of the CD44 3'UTR did not increase c-Myc levels in either Dicer-knocked down cell line ( Figure 5C,F). It is also noteworthy that TSGH8301 cells (with lower Dicer knockdown efficiency), had increased c-Myc and CD44 expression levels, following overexpression of either 3′UTR, in poorly knocked down Dicer cells ( Figure 5D-F). Taken together, these results confirm that c-Myc and CD44 act as ceRNAs, through miR-34a, in UC.

Expression of c-Myc Correlates with Expression of CD44, in Samples from UC patients and Cell Lines
We also examined the ceRNA mechanism, in vivo, by assessing the expression of c-Myc and CD44 in samples obtained from 55 UC patients. Consistent with our cell line study, the expression of c-Myc in the clinical samples positively correlated with the expression of CD44 (R 2 = 0.1015, p < 0.05, Figure 6A). 3D scatter plot analysis of these samples also revealed a positive correlation between the expression of miR-34a and the expression of c-Myc and CD44 ( Figure 6B), although no correlation between expression of these genes with any of the clinical-pathological parameters was found (Table  1). Similar findings were obtained from a gene expression array (GDS1479, Figure 6C) and RNA-Seq dataset (from TCGA, Figure 6D). We also investigated the expression of c-Myc and CD44 in HUC1, normal urothelial cells and various UC cell lines ( Figure 6E). Except for the J82 cell line, which has surprisingly low levels of c-Myc expression ( Figure 6E, green dot), all cells demonstrated a positive correlation between c-Myc and CD44 expression, in UC cells (r = 0.672). In light of the strong association between c-Myc and the aggressiveness of cancer and eventual prognosis [26], we also examined the correlation between the expression of c-Myc and tumor recurrence in this sample cohort. Kaplan-Meier survival analysis revealed an association between higher c-Myc expression levels and poor recurrence-free survival (p < 0.05, Figure 6F).

Expression of c-Myc Correlates with Expression of CD44, in Samples from UC patients and Cell Lines
We also examined the ceRNA mechanism, in vivo, by assessing the expression of c-Myc and CD44 in samples obtained from 55 UC patients. Consistent with our cell line study, the expression of c-Myc in the clinical samples positively correlated with the expression of CD44 (R 2 = 0.1015, p < 0.05, Figure 6A). 3D scatter plot analysis of these samples also revealed a positive correlation between the expression of miR-34a and the expression of c-Myc and CD44 ( Figure 6B), although no correlation between expression of these genes with any of the clinical-pathological parameters was found (Table 1). Similar findings were obtained from a gene expression array (GDS1479, Figure 6C) and RNA-Seq dataset (from TCGA, Figure 6D). We also investigated the expression of c-Myc and CD44 in HUC1, normal urothelial cells and various UC cell lines ( Figure 6E). Except for the J82 cell line, which has surprisingly low levels of c-Myc expression ( Figure 6E, green dot), all cells demonstrated a positive correlation between c-Myc and CD44 expression, in UC cells (r = 0.672). In light of the strong association between c-Myc and the aggressiveness of cancer and eventual prognosis [26], we also examined the correlation between the expression of c-Myc and tumor recurrence in this sample cohort. Kaplan-Meier survival analysis revealed an association between higher c-Myc expression levels and poor recurrence-free survival (p < 0.05, Figure 6F). surprisingly low levels of c-Myc expression ( Figure 6E, green dot), all cells demonstrated a positive correlation between c-Myc and CD44 expression, in UC cells (r = 0.672). In light of the strong association between c-Myc and the aggressiveness of cancer and eventual prognosis [26], we also examined the correlation between the expression of c-Myc and tumor recurrence in this sample cohort. Kaplan-Meier survival analysis revealed an association between higher c-Myc expression levels and poor recurrence-free survival (p < 0.05, Figure 6F).

Dominant-Negative p53 Mutation Downregulates miR-34a in UC
Our previous experiment, using clinical samples, found a group of UC patient samples with low miR-34a expression, but also low miR-34a promoter methylation. This occurrence suggested an alternative mechanism(s) of controlling miR-34a expression in UC. In that regard, the miR-34a promoter was found to possess a perfect p53-binding site [27] and be directly regulated by the tumor suppressor p53 [28]. To determine whether p53 could regulate miR-34a expression, and the subsequent expression of c-Myc and CD44, we stably transfected p53 mutants (p53R175H and p53R273H) into TSGH8301 UC cells, with endogenous wild-type p53 [29], to create a dominant-negative effect. Compared to the control, overexpression of the p53 mutant downregulated miR-34a expression, while upregulating CD44 and c-Myc (Figure 7). These results confirmed that p53 regulates miR-34a expression directly, and subsequently, miR-34a's capability of downregulating CD44 and c-Myc expression, in UC cells.

Dominant-Negative p53 Mutation Downregulates miR-34a in UC
Our previous experiment, using clinical samples, found a group of UC patient samples with low miR-34a expression, but also low miR-34a promoter methylation. This occurrence suggested an alternative mechanism(s) of controlling miR-34a expression in UC. In that regard, the miR-34a promoter was found to possess a perfect p53-binding site [27] and be directly regulated by the tumor suppressor p53 [28]. To determine whether p53 could regulate miR-34a expression, and the subsequent expression of c-Myc and CD44, we stably transfected p53 mutants (p53R175H and p53R273H) into TSGH8301 UC cells, with endogenous wild-type p53 [29], to create a dominantnegative effect. Compared to the control, overexpression of the p53 mutant downregulated miR-34a expression, while upregulating CD44 and c-Myc (Figure 7). These results confirmed that p53 regulates miR-34a expression directly, and subsequently, miR-34a's capability of downregulating CD44 and c-Myc expression, in UC cells.

miR-34a Suppresses Tumorigenicity and Drug Resistance in UC
miR-34a has been reported to be a key regulator of tumor suppression, and its downregulation has been linked to the occurrence and development of several cancers [30]. Here, we sought to elucidate the functional role of miR-34a, in UC, by stably overexpressing miR-34a in UMUC3 cells. Such overexpression significantly inhibited cell growth ( Figure 8A), and anchorage-independent growth ( Figure 8B). It also significantly inhibited the invasive ability of UMUC3 cells ( Figure 8C) and resensitized them to the effects of cisplatin ( Figure 8D). Finally, a nude mouse model demonstrated

miR-34a Suppresses Tumorigenicity and Drug Resistance in UC
miR-34a has been reported to be a key regulator of tumor suppression, and its downregulation has been linked to the occurrence and development of several cancers [30]. Here, we sought to elucidate the functional role of miR-34a, in UC, by stably overexpressing miR-34a in UMUC3 cells. Such overexpression significantly inhibited cell growth ( Figure 8A), and anchorage-independent growth ( Figure 8B). It also significantly inhibited the invasive ability of UMUC3 cells ( Figure 8C) and resensitized them to the effects of cisplatin ( Figure 8D). Finally, a nude mouse model demonstrated that miR-34a overexpression inhibited the growth of UMUC3 cell xenografts in vivo ( Figure 8E). Overall, these findings revealed that miR-34a overexpression inhibited both the growth and invasive ability of cancer cells, while also reducing chemoresistance in UC cells. Figure 7. p53 controls miR-34a expression in TSGH8301 urothelial carcinoma cells. Co-expression of p53 mutants p53R175H or p53R273H, with wild-type p53, in TSGH8301 cells, downregulated miR-34a and upregulated CD44 and c-Myc, as determined by RT-PCR. * p < 0.05; ** p < 0.01.

miR-34a Suppresses Tumorigenicity and Drug Resistance in UC
miR-34a has been reported to be a key regulator of tumor suppression, and its downregulation has been linked to the occurrence and development of several cancers [30]. Here, we sought to elucidate the functional role of miR-34a, in UC, by stably overexpressing miR-34a in UMUC3 cells. Such overexpression significantly inhibited cell growth ( Figure 8A), and anchorage-independent growth ( Figure 8B). It also significantly inhibited the invasive ability of UMUC3 cells ( Figure 8C) and resensitized them to the effects of cisplatin ( Figure 8D). Finally, a nude mouse model demonstrated that miR-34a overexpression inhibited the growth of UMUC3 cell xenografts in vivo ( Figure 8E). Overall, these findings revealed that miR-34a overexpression inhibited both the growth and invasive ability of cancer cells, while also reducing chemoresistance in UC cells.

Discussion
miR-34a plays an important role in cancer progression and drug resistance. Previous studies have reported miR-34a downregulation in many cancer types [31], including urothelial cancer (UC) [32]. Epigenetic silencing of miR-34a has also been addressed in several previous studies of human cancer [19,33,34]; however, the mechanism underlying miR-34a downregulation, in UC, has yet to be elucidated. In this study, when evaluating miR-34a expression in seven UC cell lines, we were

Discussion
miR-34a plays an important role in cancer progression and drug resistance. Previous studies have reported miR-34a downregulation in many cancer types [31], including urothelial cancer (UC) [32]. Epigenetic silencing of miR-34a has also been addressed in several previous studies of human cancer [19,33,34]; however, the mechanism underlying miR-34a downregulation, in UC, has yet to be elucidated. In this study, when evaluating miR-34a expression in seven UC cell lines, we were surprised to observe epigenetic silencing of miR-34a, via promoter hypermethylation, only in UMUC3 UC cells. Clinical analyses also found that most UC patient samples were devoid of promoter miR-34a hypermethylation. For UC cells (BFTC905 and J82) and patient samples showing low expression, but also low miR-34a promoter methylation, the possibility of other epigenetic mechanisms (e.g., histone modifications), in regulating miR-34a, cannot be excluded. Since miR-34a is expressed in most UC cell lines and patient samples, alternative mechanisms likely exist for controlling miR-34a function in UC.
A new type of miRNA regulation, referred to as a competitive endogenous RNA (ceRNA) mechanism, involves crosstalk between miRNA targets and the regulation of gene expression, through competition in binding to shared miRNAs has recently been described [35]. Several research groups, including ours, have reported this phenomenon in human cancers [36][37][38][39]. In particular, we recently demonstrated that estrogen receptor-driven upregulation of E2F6 can 'sponge' miR-193a, to facilitate c-Kit overexpression, in ovarian cancer [36]. In the current study, we demonstrated that overexpression of the c-Myc 3 UTR could upregulate CD44 through miR-34a. Importantly, we also observed a positive correlation between the expression of c-Myc and CD44 in samples from UC patients and cell lines. Taken together, our results indicate that c-Myc may act as a ceRNA of CD44, in UC.
Several miR-34a targets have been shown to control the function of miR-34a, through ceRNA mechanisms. For example, one oncogenic, long noncoding (lnc)RNA (lncRNA-SNHG7) has been shown to sponge miR-34a, upregulating the expression of the oncogene GALNT7 in colorectal cancer [40]. In another study, lnc015192 was shown to sponge miR-34a to upregulate Adam12 expression in breast cancer [41]. However, the current study was the first to demonstrate that c-Myc mRNA can act as a ceRNA, to upregulate CD44 expression, by targeting miR-34a, in UC. Our results indicate that c-Myc participates in cancer stemness, not only via protein regulation [42][43][44], but also via mRNA. Note that the genomic amplification of chromosome 8q (where c-Myc resides), commonly associated with UC, provides an additional driving force for the overexpression of CD44 [23,24]. This may also explain the significantly shorter recurrence-free survival of patients presenting high c-Myc expression.
As in previous studies, we observed that miR-34a overexpression in UC cells inhibited tumor invasion and reduced drug resistance, probably through the inhibition of CD44 [15][16][17]. In the future, it may be possible to implement epigenetic interventions to restore miR-34a expression, and inhibit cancer stemness, in a subset of UC patients with epigenetic silencing of this micro-RNA [16].
On the other hand, the tumor suppressor gene p53, the most frequently mutated tumor suppressor in human cancer, including urothelial carcinoma [45,46], also directly upregulates miR-34a [47][48][49], thereby explaining the fact that p53 signaling can control miR-34a expression, in UC samples, without miR-34a promoter methylation. In this regard, overexpression of a dominant-negative p53 mutant resulted in miR-34a downregulation, and the subsequent upregulation of CD44 and c-Myc, in UC cells with wild-type p53.
It is also noteworthy to point out that the ceRNA effect of c-Myc and CD44 overexpression, by the c-Myc 3 UTR, in UC cell lines ( Figure 4A,B), was not due to a decrease-but surprisingly, a slight increase-of mR-34a expression in those cells. This phenomenon may be due to the presence of a feedback loop that increases miR-34a expression upon the increase of c-Myc/CD44 [50]. However, further experiments need to be performed to validate such feedback loop.

Cell Cultures and Patient Samples
Human UC cell lines (SV-HUC, BFTC905, HT1376, TSGH8301, TCCSUP, J82, and UMUC3) were used in this study. SVHUC, BFTC905, HT1376, and TSGH8301 cell lines were maintained in RPMI 1640 medium (Gibco, Grand Island, NY), TCCSUP was maintained in DMEM medium, and J82 and UMUC3 were maintained in MEM medium (Gibco). Cultured cells were supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin, in a humidified atmosphere of 5% CO 2 , at 37°C . Fifty-five urothelial cancer samples from bladders or upper urinary tracts (ureter or renal pelvis) were procured from Chia-Yi Christian Hospital, Chia Yi, Taiwan. The clinical-pathological data of all samples is summarized in Table 2. All tissue samples were acquired from either transurethral resection (TURBT) or radical surgery. Patients were then followed up by either cystoscopy or radiographic detection (CT or MRI) for recurrence. All patients provided informed consent. This human sample study was approved by the Institutional Review Board of Chia-Yi Christian hospital, Taiwan.

Plasmid Constructs and shRNA Knockdown
The miR-34a full-length DNA sequence was amplified via PCR with specific primers (Table 3) inserted into pSilencer4.1 as miRNA expression vectors. The CD44 3 UTR and c-Myc 3 UTR regions (each of which contains two miR-34a binding sites) were ligated into EGFP-C1. All shRNA plasmids in the pLKO.1 plasmid were purchased from the National RNAi Core Facility at Academia Sinica, Taiwan (rnai.genmed.sinica.edu.tw) and the clone IDs were Dicer-1 TRCN0000290426 and Dicer-2 TRCN0000290486).

RNA Expression
All RNA samples were extracted with TRIzol Reagent (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. The expression of CD44 and c-Myc, in UC cell lines, was analyzed by RT-PCR, using the StepOne Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) with specific primers (Table 3). Relative gene expression levels were determined by comparing the threshold cycle (i.e., Ct value) of the test gene, normalized to the Ct of GAPDH, in each sample.

DNA Methylation Analysis
All DNA was extracted using Genomic DNA Mini Kits (Geneaid, Taipei, Taiwan) according to the manufacturer's instructions, and underwent bisulfite modification, using EZ DNA Methylation-Gold™ kits (Zymo Research, Orange, CA, USA), according to the manufacturer's protocol. Combined Bisulfite Restriction Analysis (COBRA) was used to determine miR-34a promoter methylation status, through digestion using BstUI (New England BioLabs, Beverly, MA, USA). Digested products were separated on 1.5% agarose gels for visualization. Pyrosequencing was performed using the PyroMark Q24 (Qiagen, GmbH, Hilden, Germany), with Pyro Gold Reagents (Qiagen), according to the manufacturer's instructions. The methylation status of nine CpG sites, proximal to miR-34a, was measured. The methylation percentage of each cytosine was determined by the fluorescence intensity of a cytosine, divided by the sum of fluorescence intensity of cytosines and thymines, at each CpG site (multiplied by 100%). Primers for methylation analysis are shown in Table 3.

Cell Viability Assay
Cells were seeded in 96-well plates for 4 d without or with cisplatin in various concentrations. Cell growth was determined using a cell counting kit (CCK-8, Sigma), according to the manufacturer's instructions. Cell absorbances were then determined at 450 nm, using a microplate reader.

Soft Agar Assay
For soft agar assays, a 3-mL base layer of agar (0.5% agar in culture medium) was allowed to solidify in a six-well flat-bottomed plate, before the addition of 2.0-mL cell suspensions (10,000 cells) in 0.3% agar in culture medium. The cell-containing layer was then solidified at room temperature for 20 min. Colonies were allowed to grow for 14-21 days at 37 • C with 5% CO 2 before imaging. Plates were then stained with 1 mg/mL iodonitrotetrazolium chloride (Sigma) overnight at 37 • C.

In Vivo Tumorigenicity Assay
Eight-week-old, athymic nude mice (BALB/cByJNarl) were obtained from the Taiwan National Laboratory Animal Center or BioLASCO Taiwan Co., Ltd., Taipei, Taiwan. All mice were kept under specific pathogen-free conditions, using a laminar airflow rack, with free access to sterilized food and autoclaved water. All animal experiments were approved by the Animal Experimentation Ethics Committee of National Chung Cheng University (Taiwan). 1 × 10 6 cells, in a 1:1 mixture of 0.1 mL medium and Matrigel (BD Biosciences), were injected subcutaneously into each flank of each mouse (day 0). Tumor size was measured daily, using calipers, in length (L) and width (W). Tumor volumes were calculated using the formula (L × W2/2). At the end of the experiments, all mice were sacrificed by cervical dislocation, as previously approved.

Statistical Calculation
All statistical analysis was performed using GraphPad (San Diego, CA, USA) Prism 5 software (version 5.01). Non-parametric variables were assessed using the Mann-Whitney test, when appropriate. Recurrence-free survival was assessed using Kaplan-Meier analysis with log-rank test. A p value of less than 0.05 was considered significant.

Conclusions
In conclusion, miR-34a was epigenetically silenced only in a subset of UC cell lines and patient samples. Moreover, overexpression of c-Myc was shown to upregulate CD44, via a miR-34a-mediated ceRNA mechanism, in UC cell lines and clinical samples. The overexpression of miR-34a was also shown to suppress tumor growth and invasive capability, while reducing drug resistance. Thus, amplification of c-Myc is an important mechanism in controlling cancer stemness, through ceRNA mechanisms, in UC. These novel modes of endogenous competition could explain the regulation of any number of genes, both in healthy and diseased tissues.