Study subjects

In the present study, we included 685 histopathologically confirmed bladder transitional cell carcinoma and 730 cancer-free controls. Included study subjects were recruited from the First Affiliated Hospital and Huai-An Affiliated Hospital of Nanjing Medical University, and Jiangsu Province Hospital of Traditional Chinese Medicine (TCM) between January 2003 and January 2010. The detailed method of recruiting study subjects for the study had been described previously [29]. Pathological diagnosis for bladder tumor stage was according to the 2002 International Union Against Cancer tumor–nodes–metastasis classification and the World Health Organization 1973 grading of urothelial papilloma was used to define the bladder cancer grade: well differentiated (grade 1, G1), moderately differentiated (grade 2, G2) or poorly differentiated (grade 3, G3). Bladder cancer patients were excluded, if which had previous cancer, metastasized cancer from other origin, previous radiotherapy or chemotherapy. The cancer-free subjects were recruited from those who were seeking health care in the outpatient departments at the hospital. The cancer-free controls were matched by age (± 5 years) and sex to the cases, which were genetically unrelated to the cases and had no individual history of cancer including melanoma skin cancer. The cancer-free subjects who had symptoms suggestive of bladder cancer, such as hematuria, were excluded. We used a short questionnaire to obtain the demographic and risk factor information from the included subjects. In this study, we defined ever smokers (former and current smokers) based on smoking condition. Subjects who smoked daily for >1 year were defined as ever smokers. Ever smokers who had quit smoking for >1 year were defined as former smokers and the others as current smokers. This case-control study was approved by the institutional review board of Nanjing Medical University. All individuals signed informed consents, and each subject donated 5 ml blood sample for genomic DNA extraction.

SNPs selecting and genotyping

In this study, we studied the genetic variants of DICER and DROSHA genes, which play the important roles in the miRNA biogenesis. Here, we focused on studying the single nucleotide polymorphisms (SNPs) spanning these two genes (HapMap Data Release 27), including 2 kb upstream and 2 kb downstream using the Haploview software [30]. The following criteria should be included: (i) SNPs should be located in the 5’ flanking region, 5’ UTR, 3’UTR, and coding region with amino acid changes, (ii) minor allele frequency (MAF) > 5% in Han Chinese in Beijing (CHB). According to the criteria, four SNPs were identified in DICER (rs12323635, rs13078, rs1057035, and rs3742330) and three SNPs in DROSHA (rs2291109, rs10719, and rs642321).

Genomic DNA was extracted from peripheral blood lymphocytes of the subject. The included seven SNPs were genotyped in all 1415 subjects using the MGB TaqMan probe Assay (7900HT Real Time PCR System, Applied Biosystems, Foster City, USA). About 10% of the samples were randomly selected for repeated genotyping for validation, and the results were 100% concordant. Genotype analysis was performed by two persons independently in a blinded fashion and controls were included in each plate to ensure accuracy of the genotyping. However, several samples failed in genotyping were due to DNA quality, and we would exclude them in the further analyses. Table 1 provided the primary information of the selected seven SNPs.

Bioinformatics analysis of DROSHA 3’UTR

Based on bioinformatics analysis, we predicted that hsa-miR-27a/b can bind with the 3’UTR region of DROSHA by using four common websites (Target Scan: http://www.targetscan.org/, miRanda: http://www.microrna.org/, Microcosm: http://www.ebi.ac.uk/enright-srv/microcosm/cgi-bin/targets/v5/genome.pl, and PITA: http://genie.weizmann.ac.il/) (Figure 1A). We considered that the combination of these approaches would greatly reduce the possibility of false positive.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Characterization and functional analysis of the DROSHA 3’UTR.

(A) DROSHA 3’UTR was predicted a binding site for hsa-miR-27a/b. The sequence of hsa-miR-27a and hsa-miR-27b only had one base difference (underlined). Scanning about ±100bp regions of the binding site, we only found that rs10719T>C was located in this region. Predict effect of allelic variation at rs10719 on hsa-miR-27a/b recognition and the construct of pGL3-DROSHA 3’UTR-T/C containing renilla luciferase gene and full-length 3’UTR of DROSHA gene with different alleles of rs10719 (arrow: T>C substitution). (B, C) Luciferase reporter assays to measure rs10719T or C allele difference with the presence or interference of hsa-miR-27a/b. In (B), T24 cells seeded on 24-well plates were transiently co-transfected with constructs and hsa-miR-27a/b mimics or stable negative control (NC). In (C), J82 cells seeded on 24-well plates were transiently co-transfected with constructs and hsa-miR-27a/b inhibitors or inhibitor NC. Results are shown as relative luciferase activity versus NC. Data were from three independent transfection experiments. _**, P < 0.05.

https://doi.org/10.1371/journal.pone.0081524.g001

Real-time quantitative reverse transcription-PCR (RT-PCR) assay

In order to evaluate the endogenous expression level of hsa-miR-27a/b, four bladder cancer cell lines (EJ, T24, J82, and 5637) seeded into 20 cm² plates were subjected to extraction of the total RNA isolated from cells using Trizol Reagent (Invitrogen, CA, USA). Table S1 in File S1 showed the primers information of hsa-miR-27a/b and U6. The reverse transcriptase reactions (10 μl) contained 2 μl total RNA (500 ng/μl), 1 μl 10×AMV RT buffer, 10 pmol each of dNTPs (Toyobo, Tsuruga, Japan), 0.75 μl antisense looped primer mix, 0.25 U/μl RNase Inhibitor (Toyobo, Tsuruga, Japan), 1U/μl AMV reverse transcriptase. The mixture was incubated at 16 °C for 15 min, 42 °C for 60 min, and 85 °C for 5 min. Next, Applied Biosystems 7900HT Real Time PCR System was used to perform real-time quantification PCR (ABI, CA, USA) based on the SYBR-Green method (Toyobo, Tsuruga, Japan). All reactions were conducted in triplicate. Fold changes were normalized to the expression levels of U6.

For detection of the correlation between the DROSHA mRNA levels and rs10719 T>C polymorphism in vivo, a total of 61 bladder tumor tissues with different genotypes (32 for TT, 24 for TC, and 5 for CC genotypes) were subjected to extraction of the total RNA using Trizol Reagent (Invitrogen, CA, USA). Bladder tumor tissues were preserved in liquid nitrogen after being removed from the body. The total RNA was evaluated by both reverse transcriptase reaction and real-time quantitative PCR based on the SYBR-Green method (Toyobo, Tsuruga, Japan). The cDNA was used for the amplification of DROSHA gene and an endogenous control gene GAPDH. The primers information of DROSHA and GAPDH genes were included in the Table 1. Fold changes were normalized by the expression levels of GAPDH and each assay was performed in triplicate.

DROSHA 3’UTR containing rs10719TC reporter gene constructs and Luciferase reporter assays

To construct the luciferase reporter plasmids of DROSHA 3’UTR, DROSHA 3’UTR fragments (937bp) carrying the major rs10719T allele were amplified by PCR. The primers were 5’-ACCTTGGTACCCCAGATGAGACTGAAGACATC-3’ (forward) and 5’-ACCTTCTCGAGGCACTCACTATATATTTGCTG-3’ (reverse). The PCR products were extracted and separated by agarose gel, which were cloned with TA cloning Kit (Invitrogen, CA, USA). In addition, the fragment containing minor rs10719C allele was conducted using the following primers: 5’-TAGTTTTCCTGCAGACAATGAACGAAGTGTGC-3’ (forward) and 5’-TTTATTTCAATGAGCACACTTCGTTCATTGTC-3’ (reverse). Finally, the amplified fragment carrying T or C allele was inserted downstream of the luciferase gene in a pGL3-promoter plasmid and then the plasmid containing T or C allele was conducted, which were confirmed by sequencing.

For luciferase reporter assay, T24 and J82 cells were placed in 24-well plates (1×10⁵ cells per well) and then cotransfected with pGL3-DROSHA 3’UTR-T or pGL3-DROSHA 3’UTR-C and pRL-SV40 (50:1). The mimics and inhibitors of hsa-miR-27a/b and their negative controls (GenePharma, Shanghai, China) were cotransfected with the reporter plasmids at a final concentration of 20nmol/μl. Forty-eight hours after transfection in T24 and J82 cells, luciferase activity in lysates was measured with a Dual-Luciferase Reporter Assay System (Promega, WI, USA) and normalized against the activity of the pRL-SV40. Assays were followed by the manufacture’s suggestions. Independent triplicate experiments were performed for each plasmid construct.

Statistical analysis

The Hardy-Weinberg equilibrium of the genotype distribution among the controls was applied by using a goodness-of –fit χ² test. The frequency distributions of selected demographic variables between the cases and controls were tested using χ² test. Genotype-specific odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated by unconditional univariate and multivariate logistic regression analyses. The multivariate adjustment included the age, sex and smoking status (never and ever smoking). In addition, Kruskal-Wallis one-way ANOVA tests were used for analyzing the results of DROSHA mRNA expression in vivo. In this study, the relative luciferase reporter gene expression for T or C allele was calculated separately. Student’s t test was used to evaluate the differences in the expression levels of luciferase reporter gene among subgroups. All tests were two-sided using the SAS software (version 9.1; SAS Institute, Inc, Cary, NC, USA) and P < 0.05 was considered statistically significant.

Association study between the polymorphisms of DICER and DROSHA and bladder cancer risk

The characteristics of the 685 bladder transitional cell carcinoma patients and 730 controls are summarized in Table 2. Herein, we did not observe statistical difference in the distribution of age (P = 0.144) and sex (P = 0.825) between patients and controls. However, there were more ever smokers (55.6%) among the patients more than among the controls (38.4%), and this difference was statistically significant (P < 0.001). These variables were adjusted for the subsequent multivariate logistic regression analysis. Of the 685 cases, there were 315 tumor grade 1 (46.0%), 261 tumor grade 2 (38.1%) and 109 tumor grade 3 (15.9%) patients. In addition, 435 (63.5%) had superficial tumors and remaining 250 (36.5%) had invasive tumors.

The seven SNPs genotype frequencies among the control were in agreement with the Hardy-Weinberg equilibriums (P > 0.05; Table 1). As shown in Table 3, we observed that subjects with the DROSHA 3’UTR rs10719C allele (TC and CC genotypes) had a 1.24-fold increased risk of bladder cancer (Adjusted OR = 1.25, 95% CI = 1.01-1.55, P = 0.041) compared with the rs10719TT genotype. Meanwhile, we observed that the distribution of rs10719TC genotypes between the cases and controls showed significant difference (P = 0.017). However, we did not observe any significant differences in genotype distribution of the DICER rs12323635CT, rs13078TA, rs1057035TC, rs3742330AG and DROSHA rs2291109AT, rs642321CT polymorphisms between the cases and controls (all P >0.05, Table 3). We further evaluated the effect of included seven polymorphisms on bladder cancer risk by stratifying by sex and smoking status. As shown in Table S2 in File S1, we found that rs10719TC polymorphism can increase risk of bladder cancer among male patients (Adjusted OR = 1.34, 95% CI = 1.05-1.70, P = 0.018), and ever smokers (Adjusted OR = 1.56, 95% CI = 1.14-2.14, P = 0.006).

DROSHA 3’UTR rs10719TC affects DROSHA expression by regulating the hsa-miR-27b binding

In order to explore the possible mechanism of the DROSHA 3’UTR in the bladder cancer risk, we performed the functional assays. Based on bioinformatics analysis, the DROSHA 3’UTR was predicted a binding site for hsa-miR-27a/b (Figure 1A). Here, we demonstrated that rs10719TC was located 46bp downstream of the hsa-miR-27a/b binding site in the DROSHA 3’UTR.

As shown in Figure S1 in File S1, real-time quantitative RT-PCR assay suggested that endogenous expression levels of hsa-miR-27a/b in J82 cell line were more significantly higher than other cells (T24, EJ, and 5637) (P < 0.001). Here, we chose J82 and T24 cell lines in the further luciferase assay. Then, we used the plasmids for transient co-transfection with the T24 cells [stable negative control (NC) miRNA: stable NC or hsa-miR-27a/b mimics] and J82 cells (inhibitor NC or hsa-miR-27a/b inhibitor). As shown in Figure 1B (T24 cell), hsa-miR-27a/b suppressed luciferase expression in the presence of rs10719T allele, comparing with NC (P < 0.05), but not the rs10719C allele (P > 0.05). Additionally, we also found that the inhibition of hsa-miR-27a/b expression can increase luciferase expression efficiently for the rs10719T-containing plasmid rather than the rs10719C-containing plasmid in J82 cells (P < 0.05; Figure 1C). However, we also found that a significant decrease in luciferase expression, when we added hsa-miR-27a inhibitor to allele C in J82 cells. The similar result did not be observed in adding hsa-miR-27b inhibitor to allele C in J82 cells. Maybe, hsa-miR-27a did not directly affect DROSHA luciferase expression. Our data suggested that rs10719 T to C substitution acts as a loss-of-function mutation and would affect DROSHA luciferase expression by hsa-miR-27b target.

In the present study, we also performed RT-PCR assay to explore whether hsa-miR-27b affected DROSHA expression by degrading mRNA or suppressing mRNA post-translational translation (Figure S1 in File S1). A total of 61 bladder tumor tissues with different genotypes of the DROSHA rs10719TC polymorphism were used to assess the expression of DROSHA mRNA. No significant difference levels of DROSHA mRNA among individuals with the TT, TC and CC genotypes was observed (P > 0.05). Taken together, hsa-miR-27b may affect DROSHA expression by regulating protein translation.