Introduction

Bladder cancer is the 9th most common cancer in the world with 430,000 new cases diagnosed in 2012 and Belgium had the highest rate of bladder cancer in both sexes (17.5 per 100,000), followed by Lebanon (16.6 per 100,000), Malta (15.8 per 100,000) and Turkey (15.2 per 100,000) [1]. In Turkey 10,757 new cases were diagnosed and 4,690 patients died from bladder cancer complications in both sexes in 2012 [1]. Both environmental and genetic factors contribute to carcinogenesis of bladder cancer; and tobacco smoking, occupational exposure to aromatic hydrocarbons, family history of cancer, chemotherapy and radiotherapy are major etiological factors [2, 3].

Micro ribonucleic acids (miRNAs) are short (19–22 nucleotides) non-protein coding RNAs and are found in all eukaryotic cells. They regulate gene expression epigenetically through complementarily binding to the 3′ UTR region of a target mRNA transcript, and resulting in translational repression and gene silencing. More than 60 % of mammalian mRNAs contain conserved regions that serve as targets for miRNAs. In addition, one mRNA can be the target of multiple miRNAs, and each individual miRNA has the capacity to target hundreds of genes, with an average of 500 targets per miRNA [4]. As of its most recent release in June of 2013, over 2,578 mature miRNAs were identified in humans and cataloged in the miRBase 20.0 databases [5].

Accumulating evidence suggests aberrant miRNA expression patterns in most human malignancies, and some highly expressed miRNAs might function as oncogenes by repressing tumor suppressors; conversely, miRNAs expressed at low levels might function as tumor suppressors by negatively regulating oncogenes [6, 7]. Based on reports describing miRNA signatures, several down regulated and up regulated miRNAs have been discovered in bladder cancer [8], [9] and [10]. Some of those miRNAs are thought to be potential biomarkers for bladder cancer in diagnosis and prognosis prediction, as well as a treatment target [11, 12]. Especially current studies mainly focus on miRNAs for providing prognostic information of related cancer [12], [13], and [14].

However, the exact roles of these miRNAs and their target genes in oncogenic pathways have not been clearly elucidated in bladder cancer. Insights into these molecular mechanisms of bladder cancer initiation and progression can provide targets for preventative and therapeutic approaches while providing reliable biomarkers.

In this study, we aimed to predict potential mechanisms of bladder-cancer related miRNAs and target genes by miRNA profiling and bioinformatics tools. By combining miRNA profiling and bioinformatics tools in our study, we sought not only to provide insights into the pathogenesis of bladder cancer but also to discover potential biomarkers for diagnosis, prognosis and treatment.

Matherials and methods

Tissue preparation and RNA isolation

The collection of samples and research protocols were reviewed and approved by the Ethics Committee of Eskisehir Osmangazi University, and the investigation was performed with written informed consent from all the patients. Tissue samples were taken from 30 bladder cancer patients from the central Anatolian part of Turkey who had undergone prospective cystectomy or transurethral resection of BCs, and 9 normal urothelial tissue samples were derived from patients who underwent control cystoscopy. All patients were males ranging from 35 to 65 years old (50.63 + 9.87). Hematoxylin eosin stained sections were examined for tumor cell percentage, and tumor tissues containing more than 85 % tumor cells were selected for microarray analysis.

The tissue samples were deposited in RNAlater solution (Qiagen, USA) and subsequently stored at −20 °C until the RNA extraction. Total RNA including miRNAs was extracted from frozen bladder cancer and normal urothelial tissues using Trizol reagent (Invitrogen, USA) according to the manufacture’s protocol. Total RNA concentrations were measured using a spectrophotometer (NanoDrop Technologies, Wilmington, USA) and quantification analysis of total RNA was performed by a microfluidics-based platform (2100 Bioanalyzer, Agilent, Santa Clara, CA). Three tumor samples and one normal urothelial tissue sample were excluded from conducting miRNA array because of poor RNA quality.

MiRNA expression profiling

After quality assessment remaining 27 tumor samples (12 high grade and 15 low grade; 6 muscle invasive and 21 non-invasive) and 8 normal urothelial tissue samples were loaded onto Agilent’s miRNA arrays with 723 human and 76 viral miRNA represented probes based on Sanger 10.1 (Agilent V 2.0). The input for the miRNA labeling system was 100 ng of total RNA. Following dephosphorylation and denaturation, total RNA was labeled with cyanine 3-pCp (Cy-3) and then hybridized on Human miRNA Microarray (Agilent V 2.0) using the miRNA Complete Labeling and Hyb Kit (Agilent). Each sample was hybridized at 55 °C for 20 h in an agitated hybridization oven at 20 rpm. Then, the slides were washed using the Gene Expression Wash Buffer Kit (Agilent) and scanned by an Agilent Scanner (G2505C). The images were processed and analyzed with Feature Extraction Software (Agilent Technology Ver. 10.1). For genomic profiling, spots were background-corrected using the median foreground minus the median background signal intensities for dye, and log2 ratios were calculated. Following the microarray signal data transformation (linear to log values), a filtration process based on the flag value of spots was performed. Raw microarray data were normalized and analyzed using GeneSpring GX 11.5 (Agilent Technologies, San Francisco, USA). We performed quantile normalization for signals and used the median (50th percentile) method,which normalizes each chip on its median for “per chip” normalization.

Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) analysis

Independent RNA samples prepared in parallel with the samples for the miRNA array were used to validate the miRNA array data. In the validation of microarray data, quantitative RT-PCR reactions were performed on randomly selected deregulated miRNAs including hsa-miR-143 and hsa-miR-145. The reverse-transcription reactions were carried out according to the manufacturer’s protocol by using a MiRCURY LNA Universal Reverse Transcriptase kit (Exiqon, Denmark). Initially, reverse transcription was performed with 5 ng/µl total RNA in a 20-µl volume. The assays were heated to 42 °C for 60 min and then 95 °C for 5 min.Before real-time assays, total RNAs were diluted 80 fold in nuclease-free water. Real-time PCR was carried out on a Roche Lifecycler 480 by using SYBR Green master mix and LNA primers of hsa-miR-143, hsa-miR-145, and housekeeping 5S RNA according to the manufacturer’s protocol. The transcription level of 5SRNA was used as an internal control. Then, melting curve analysis was performed to evaluate the specificity of assays. Each sample was analyzed in triplicate.

Bioinformatic analysis

Mature sequences of miRNAs were loaded into the DIANA-Tarbase v 6.0 databases for target analysis [15]. We checked the mature sequences of our deregulated miRNAs from the Mirbase 20.0 databasebefore target analysis. Target genes of deregulated miRNAs were listed using DIANA-TarBase database v 6.0, which include experimentally validated miRNA targets in the literature. To explore the potential function of the whole miRNA and target gene signature, DIANA-mirPath, a web-based computational tool that was developed to identify molecular pathways that are potentially altered by the expression of multiple miRNAs, was used to incorporate miRNAs into KEGG molecular pathways [16]. The pathways were obtained with a p value < 0.05 and gene count >2. The overlapping parts of the “Pathways in cancer (hsa05200)”, “Bladder cancer (hsa05219)”, and “ErbB signaling pathway (hsa04012)” categories of enriched KEGG pathways and related miRNAs were visualized by Cytoscape software [17].

Statistical analysis

All miRNAs in the bladder cancer and normal samples were compared using a t-test to define differentially expressed miRNAs. Multiple testing corrections were done using the Benjamini-Hochberg method [18]. A false discovery rate (FDR) less than 0.05 and absolute log fold change (logFC) greater than 2 were set as the significant cut-offs. The expression levels of the transcripts were evaluated using the comparative CT method (2−∆∆Ct) in qRT-PCR.

Results

Identification of differentially expressed genes

According to the cut-off criteria of |logFC| > 2 and p value < 0.05, we finally gained 19 deregulated miRNAs, including 5 up- and 14 downregulated (Table 1).

Table 1 Differentially expressed miRNAs between bladder cancer and normal bladder urotelium. A value of p < 0.05 was considered as statistically significant and an absolute log fold change greater than 2 were set as significant cut-off
Full size table
Table 2 Number of known targets of deregulated miRNAs. Listed from Tarbase 6.0 database
Full size table
Table 3 The Enriched KEGG Pathways of miRNA Targets
Full size table

Validation of differentially expressed miRNAs using qRT-PCR

Independent RNA samples that were prepared in parallel with the samples for the miRNA array were used to validate the miRNA array data. The expression of randomly selected deregulated miRNAs including hsa-miR-143 and hsa-miR-145 was validated using qRT-PCR (Table 4).

Table 4 Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) validation of microarray data
Full size table

Target genes of miRNAs and their roles in the KEGG pathways

The experimentally validated target genes of deregulated miRNAs were obtained from DIANA-Tarbase v 6.0 databases. The results showthat hsa-miR-30a-5p had the highest number of targets (411), whereas hsa-miR-370 had only 4 targets. Hsa-miR-195-3p, hsa-miR-574-3p, hsa-miR-638, and hsa-miR-193b-3p had no validated targets in the database (Table 2). To gain further insights into the function of deregulated miRNAs and their targets, DIANA-mirPath was applied to identify the significant KEGG pathways. Enrichment results showed that most pathways were related to cancer or cell signaling pathways (Table 3). Most target genes of miRNAs (35) were enriched in the “Pathways in cancer (hsa05200)” category (Table 5). Significantly,13 cancer-associated target genes and 7 related miRNAs were enriched in the “ErbB signaling pathway (hsa04012)” category (Table 5). E2F1, ERBB2, NRAS, RAF1, CDKN2A, KRAS, TP53, CCND1, MMP1, E2F3, MYC, RB1, HRAS, CDKN1A, and VEGFA target genes (Fig. 1) as well as the related hsa-let-7c, hsa-miR-125b-5p, hsa-miR-143-3p, hsa-miR-145-5p, hsa-miR-195-5p, hsa-miR-572, hsa-miR-26a-5p, hsa-miR-23b-3p,and hsa-miR-210 were enriched in the “Bladder cancer (hsa05219)” category (Table 5).

Table 5 Kinds of enriched target genes and realted miRNAs in three selected KEGG pathways
Full size table
Fig. 1
figure 1

Bladder cancer (hsa05219) pathway in KEGG pathways. Yellow rectangles represent genes targeted by only one miRNA, whereas orange rectangles represent genes targeted by one more miRNA in the bladder cancer pathway. Green rectangles represent genes which are not targeted by any miRNA

Full size image

MiRNA regulated gene networks associated with bladder cancer

The “Pathways in cancer (hsa05200)”, “Bladder cancer (hsa05219)”, and “ErbB signaling pathway (hsa04012)” categories from enriched KEGG pathways were selected for further investigation of the miRNA regulatory networks in bladder cancer. All of the interactions of enriched target genes and miRNAs of three pathways are listed in Table 6. Based on these data, the overlapping parts of three pathways were visualized by Cytoscape software (Fig. 2). We determined hub CDK6, BCL2, E2F3, PTEN, MYC, RB, and ERBB3 target genes along with hub hsa-let-7c, hsa-miR-195-5p, hsa-miR-141-3p, hsa-miR-26a-5p, hsa-miR-23b-3p, and hsa-miR-125b-5p miRNAs for these interaction networks.

Table 6 Presentation of miRNA based target genes in three selected KEGG pathways
Full size table
Fig. 2
figure 2

Cytoscape representation of overlapping parts of three selected KEGG pathways and related miRNAs. Pink circles represent target genes of miRNAs, and green circles represent hub genes. Pink diamonds represent miRNAs, and yellow diamonds represent hub miRNAs. Colored lines represent target genes and miRNA interactions. We used “C”, “B”, and “E” abbreviations for “Pathways in cancer (hsa05200)”, “Bladder cancer (hsa05219)”, “ErbB signaling pathway (hsa04012)”, respectively. Blue lines represent interactions in only C. Black lines represent interactions in only B. Green lines represent interactions in only E. Pink lines represent interactions of overlapping parts of C + B + E. Yellow lines represent interactions of overlapping parts of C + B. Light blue lines represent interactions of overlapping parts of C + E. Red lines represent interactions of overlapping parts of B + E

Full size image

Discussion

In this study, we analyzed the expression profile in 27 tissue samples of histologically confirmed bladder cancer samples (12 high grade and 15 low grade; 6 muscle invasive and 21 non-invasive) and 8 non-malignant tissue samples of the bladder. A total of 19 miRNAs, including 5 ups- and 14 down regulated, displayed significant differential expression in cancerous tissue compared to non-cancerous tissue.To date, differential expressions of miRNAs in bladder cancer have been published in several studies [19, 20]. Although the results of these studies are not the same, the down regulation of hsa-miR-143, hsa-miR-145, hsa-miR-125b, and hsa-miR-26a and the up regulation of hsa-miR-141 are in agreement with most studies, including our study [19, 20].

Known targets of 19 deregulated miRNAs and their roles in biological processes were demonstrated via KEGG pathway analysis. The KEGG enrichment analysis results in this study confirmed the reliability of our findings, and many of them have been implicated in various cancers and related cell signaling pathways. We considered that the “Pathways in cancer (hsa05200)”, “Bladder cancer (hsa05219)”, and “ErbB signaling pathway (hsa04012)” categories from KEGG pathways were highly related to bladder cancer initiation and progression [21].Studies suggest that more centralized genes in the network are more prone to be key deliverers to proper cellular function than peripheral genes [22, 23]. Therefore; we constructed miRNA regulatory networks based on overlapping parts of these three pathways. Our results showed that CDK6, BCL2, E2F3, PTEN, MYC, RB, and ERBB3 hub genes along with hsa-let-7c, hsa-miR-195-5p, hsa-miR-141-3p, hsa-miR-26a-5p, hsa-miR-23b-3p, and hsa-miR-125b-5p hub miRNAs might play key roles in miRNA regulatory networks in bladder cancer.

We also reviewed some of the literature and reconfirmed the roles of some hub genes in bladder cancer.Bladder cancer has been reported to be associated with loss of function, affecting tumor-suppressor genes such as Tp53, RB, and PTEN [24, 25]. Oncogenes, especially E2F genes (e.g.E2F1and E2F3), which affect the RB pathways, have also been reported to be involved in the progression of this neoplasm [26]. Huang et al. [27] previously reported on the association between hsa-miR-125b and E2F3. Similar to previous studies, we demonstrated that PTEN, RB1, and E2F3 were hub genes in our network, and their regulatory hub miRNAs, including hsa-miR-141-3p, hsa-miR-26a-5p, hsa-miR-23b-3p, and hsa-miR-125b-5p, might play a central role in bladder cancer development in our regulatory networks (Fig. 2). Accumulating evidence suggests that aggressive bladder tumors are associated with the overexpression of some oncogenes [28]. The overexpression of MYC, BCL2, andERBB3 hub oncogenes might be related to the downregulation of hsa-let-7c, hsa-miR-195-5p, hsa-miR-26a-5p, hsa-miR-23b-3p, and hsa-miR-125b-5p in our study. These gene network associations have to be confirmed with further comprehensive functional studies.

The CDK6 gene is a member of the cyclin-dependent protein kinase (CDK) family.This kinase is a catalytic subunit ofthe protein kinase complex that is important for cell cycle G1 phase progression and G1/S transition.This kinase, along with CDK4, has been shown tophosphorylate and thus regulate the activity of tumor-suppressor protein RB. Although increased expression of this oncogene has rarely been reported in bladder cancer, we speculate that CDK6 might have a significant effect on bladder cancer development [29].

Lin et al. [30] reported that CDK4 is a novel target in has-miR-195-mediated cell cycle arrest in bladder cancer cells. As mentioned, CDK4 and CDK6 had similar functions in the cell cycle. In our networks, CDK6 was related to hsa-let-7c, hsa-miR-195-5p, and hsa-miR-125b-5p (Fig. 2). From this point, we speculated that these miRNAs, CDK4, and CDK6 genes might exist in the same bladder cancer pathway. However, the exact regulatory networks remain elusive, and it is hard to estimate the actual false-positive rate of bioinformatic tools. Therefore, further experimental studies should be carried out in order to confirm our results.

Conclusion

Overall, deregulated miRNAs and known target genes might play key roles in the pathogenesis of bladder cancer according to the pathway enrichment analysis. Particularly, hub genes and miRNAs of our constructed network might be central actors of molecular alterations in bladder cancer and candidate biomarkers for diagnostic, prognostic and therapeutic purposes. Of course, further research is needed to confirm their exact roles.