Association between polymorphisms in the integrin gene predicted microRNA binding sites and bladder cancer risk

Abstract

Bladder cancer (BC) is the ninth most frequent malignancies in the world and the occurrence of this disease has dramatically increased in recent years. Integrins have been demonstrated to play an important role in the development and progression of BC. However, the association between polymorphisms in integrin genes and BC susceptibility was still unclear. A number of studies mainly focused on polymorphisms in the coding regions of integrin genes previously, while in this study, polymorphisms in the 3’ untranslated regions (3’UTR) were investigated in Chinese Han population. According to previous study, seven single-nucleotide polymorphisms (SNPs) in predicted microRNA (miRNA) target sites were chosen as potential targets. And four SNPs including rs11902171, rs2675, rs17664, and rs1062484, were finally examined for their effect on BC risk and clinical prognosis. These four polymorphisms were genotyped by using the high-resolution melting method (HRM) in 317 BC patients with long-time follow-up together with 317 age-matched healthy controls. AC carriers of rs2675 in ITGB5 were associated with an increased risk of BC (OR 1.44, 95% CI 1.02-2.03). No significant relationship was detected between these SNPs and the recurrence-free survival time of overall study population or non-muscle invasive BC subgroups in univariable analysis. In conclusion, rs2675 in miRNA binding sites of ITGB5 might be a potential target for BC susceptibility prediction.

Introduction

Bladder cancer (BC) has been prevalent in the United States with an estimated 72,570 new cases and 15,210 deaths in 2013 [1]. The incidence rate of BC in Chinese population was rising with the average growth rate 4.60% per year during 1998-2008 [2]. Various risk factors have been demonstrated to be associated with the development of BC, including cigarette smoking, occupational and environmental exposure [3]. Recently, polymorphisms in individuals’ genetic susceptibility have been revealed to be correlated with the risk of BC [4]. However, mechanisms explaining gene interactions and signal pathways in the development of BC are still unclear.

Integrins, a family of cell surface receptors, are composed of non-covalently linked type 1 transmembrane glycoprotein subunits α and β [5-8]. They mediate the connection between extracellular matrix (ECM) and intracellular actin cytoskeleton, which may lead to the activation of related signaling pathways and cause changes of cell function [9,10]. A number of studies have confirmed the critical role of integrins in tumor development and progression [11,12]. Integrins are highly over-expressed on different types of cancer cells. They participate in the regulation of cell cycle following specific ECM binding, and induce cellular quiescence or malignant transformation by controlling cyclin D1 and cyclin-dependent kinase inhibitor family [13]. Besides, integrins could cooperate with oncogenes or receptor tyrosine kinases to influence tumorigenesis through cross-talking various signaling pathways, including PI3K-AKT, NF-κB, and p53 pathways [14]. Therefore, integrins have profound contribution on the alteration of biological behavior of tumor cells, including migration, invasion, and metastasis [15]. Recently, the possible association between integrins and BC has been identified by studies. The expression of integrin α₂ subunit was proved to be associated with the reduction of adherence between BC cells and promotion of invasive potential [16]. Abnormal expression of integrin α₂β₁ and/or α₆β₄ were reported to be an early event in the initiation of BC [17,18]. Integrin α₃β₁ increased a lot in BC with the progressive loss of integrin α₂β₁, which was verified to contribute to migration and invasion of BC. Moreover, loss of co-expression and co-localization of integrin α₆β₄ and collagen VII were identified to be related to more aggressive behavior of BC cells.

MicroRNAs (miRNAs) belong to the family of non-coding small RNA molecules, which contain 21-25 nucleotides. Gene expression was controlled by post-transcriptional regulation through strict matching between the seeding region, 2-7 nucleotides in the 5’ miRNA sequence and the non-coding region, 3’ untranslated region (3’UTR) of the target genes. Gene expression was down-regulated following specific mRNAs inhibition after miRNA-target interaction [19,20]. Thus, same gene with different genotypes might result in different levels of gene expression, which was caused by miRNAs suppression. By regulating the expression of tumor suppressor genes or proto-oncogenes, miRNAs were confirmed to contribute to changes in the individual susceptibility with multiple human malignancies [21].

Polymorphisms in the miRNA binding sites of specific genes have been proved by recent studies to be associated with susceptibility of various tumors. MiRNA-target interaction between miRNAs Let-7e and Let-7f and human IL23R gene was disrupted by single-nucleotide polymorphism (SNP) rs10889677 in the 3’UTR in patients with breast cancer [22]. Such SNPs have also been identified to have association with other types of cancers, including colorectal cancer [23] and non-small cell lung cancer [24]. Furthermore, rs10719T > C polymorphism located in 3’UTR of DROSHA gene disrupted the binding site for hsa-miR-27b, which was demonstrated to be associated with the increased risk of BC [25]. Therefore, genetic variation in the miRNA binding sites of integrin genes may also contribute to the risk of BC, including the occurrence and progression. In the present hospital-based case-control study, four SNPs in the 3’ UTR of four integrin genes were investigated in Chinese Han population in order to provide data for screening high-risk individuals.

Materials and methods

Study population

A total of 317 patients recruited in this study were identified from the Department of Urology, West China Hospital, Sichuan University from 2000 to 2012. All these patients were sporadic cases with the pathological diagnosis of BC (mean age 63.35 years; 242 males and 75 females). Patients with immune disease or any histopathologic diagnosis other than BC were excluded. 317 healthy individuals with no evidence of cancer or immune disease were selected in a routine check-up or health awareness campaigns as controls (mean age 64.13 years; 236 males and 81 females) from West China Hospital, Sichuan University. All controls were matched to cases by gender and age (± 5 years) at baseline. Clinical and epidemiology information were collected from all subjects, including smoking status and history, tumor grade, tumor stage, and event (recurrence/non recurrence, and death/live). 1 ml peripheral blood samples were collected from all subjects and stored at -80°C for further detection. Written informed consent was obtained from each subject.

The follow-up process was conducted in 125 patients to evaluate the potential effect of polymorphisms in the microRNA binding sites of integrin genes on the recurrence of BC. Face to face interview and telephone calls were made to collect data of these patients. Malignant urothelial tumors that have not invaded the detrusor were defined as non-muscle invasive BC, while the others were muscle invasive BC. The recurrence of BC was defined as newly growing focus in cystoscopy or imaging findings. The recurrence-free survival time was the specific period from the date of surgery to the date of recurrence or death. Follow-up lasted until May 2013.

SNP selection

SNPs were selected with detailed selection method for polymorphisms in the miRNA binding sites of integrin genes, which was provided by Brendle et al. [26]. Polymorphisms in the 3’UTRs of 10 integrin genes (ITGA3, ITGA5, ITGA6, ITGAv, ITGB1, ITGB3, ITGB4, ITGB5, ITGB6, and ITGB8) and predicted putative miRNA binding sites were reported to be screened and selected through online available tools, including microInspector (http://mirna.imbb.forth.gr/microinspector/), PicTar (http://pictar.bio.nyu.edu/), and TargetScan (http://www.targetscan.org/). Seven SNPs were chosen in the previous study; however, ITGB3 rs3809865, ITGB4 rs743554 and ITGB1 rs17468 were given up in this study as a result of no appropriate PCR condition. Finally, four SNPs were selected, including rs11902171, rs2675, rs17664, and rs1062484. The variations of miRNA targets are listed in Table 1.

Table 1.

Information about predicted miRNA binding sites SNPs

gene	SNP	Variation	chromosome	MAFa	MiRNA/wild type allele	MiRNA/variant allele
ITGAv	Rs11902171	[G/C]	2	0.166	mir-382, mir-30a-3p, mir-30e-3p,
ITGB5	Rs2675	[A/C]	3	0.160	mir-192, mir-215, mir-449, mir-504	mir-192, mir-215, mir-449, Mir-30a-3p, mir-30e-3p.
ITGA6	Rs17664	[A/G]	2	0.404	mir-195, mir-214, mir-223	mir-195, mir-214, mir-223, mir-152.
ITGA3	Rs1062484	[C/T]	17	0.137	mir-150a, mir-187, mir-191, mir-211, mir-326, mir-339, mir-362, mir-370, mir-486, mir-501, mir-517	mir-150, mir-191, mir-145, mir-218, mir-296, mir-326, mir-339, mir-362, mir-370, mir-432, mir-433, mir-478, mir-486, mir-501, mir-517, mir-518c, mir-517a,

^aMAF (Minor Allele Frequency) was cited from http://www.ncbi.nlm.nih.gov/SNP.

DNA isolation and SNP genotyping

DNA samples were collected from peripheral blood samples of patients with EDTA as the anticoagulant. Genomic DNA was extracted with QIAamp DNA extraction kit (Qiagen) following the manufacturer’s protocol. After extraction, DNA purity and concentration were determined by spectrophotometer. Genotyping of the selected microRNA binding site SNPs were done with Sequenom MassARRAY & iPLEX assay of Capitalbio Company. In brief, primers for PCR and iPLEX reaction were synthesized with the tool designed by Genotyping Tools & MassARRAY Assay Design software. PCR amplification was performed in 384-well plate and shrimp alkaline phosphatase was used to deal with the product for the purpose of dephosphorylating unincorporated dNTPs in PCR system. Then the Mass ARRAY iPLEX reaction was performed. After the iPLEX reaction, iPLEX products with different molecular weight were created by each base of SNP sites. Finally, different iPLEX product could be identified by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF), and then genotyping information of each SNP site for every subject could be obtained in the result of MALDI-TOF with TYPER 4.0 software.

Statistical analysis

Hardy-Weinberg equilibrium was detected in each SNP in controls. Pearson χ² or Fisher’s exact tests was carried out to analysis the potential difference in the distribution of SNP genotypes between patients and controls. Student’s t test was used to explore the difference in the distribution of age. The odds ratios (ORs) and 95% condense intervals (CIs) for association between gene miRNA binding site SNP genotype and BC risk were calculated with the risk option of crosstabs. Kaplan-Meier and Cox proportional hazards models were used to detect the association between genotypes and recurrence-free survival (RFS) time. Univariable analyses were applied throughout the whole study. All the analyses in this study were performed with the Statistical Package for Social Sciences software, v.13.0 (SPSS, Chicago, IL) and p < 0.05 in two sides was considered statistically significant.

Results

Characteristics of the study objects

Basic information of the study subjects and clinical characteristics of each patient are summarized in Table 2. No significant difference in the distribution of age were found between the cases (63.35 ± 12.95) and the controls (64.13 ± 12.38) (p = 0.443). Among recruited BC patients, the majority was in advanced clinical grade, including clinical grade II and III (34.1% (n = 108) and 48.9% (n = 155), respectively). The percentage of patients in clinical grade I was 13.6% (n = 43), and the rest 3.4% (n = 11) were mixed grade tumor. In addition, patients with superficial BC accounted for 43.5% (n = 138), while the remaining 56.5% (n = 179) were invasive BC patients.

Table 2.

Demographic characteristics in the cases and controls

	Variables	Cases (n = 317) (%)	Controls (n = 317)	P
Sex	Male	242	236
	Female	75	81
	Age (Mean ± SD)	63.35 ± 12.95	64.13 ± 12.38	0.443
Smoking	Non smokers	168
	Smokers	149
Grade	I	43
	II	108
	III	155
	Mixed	11
Stage	Superficial	138
	Invasive	179
Event	Recurrence/non recurrence	71/54
	Death/live	38/87

Association of SNPs with BC susceptibility

The genotype and allele frequencies of 4 selected SNPs in all subjects are shown in Table 3. All the SNPs were in Hardy-Weinberg equilibrium (P > 0.05). Genotype and allele distributions of ITGB5 gene polymorphism rs2675 exhibited significant difference between cases and controls. Compared with the AA genotype, AC genotype was strongly associated with an increased risk of BC (OR = 1.44; 95% CI = 1.02, 2.03; p = 0.04). The frequency of C allele is significantly higher than the A allele in the cancer subjects than the controls (OR = 1.34; 95% CI = 1.01, 1.77; p = 0.04). However, for the other three SNPs, the genotype and allele frequencies did not differ significantly between the patients and the controls (P > 0.05).

Table 3.

The polymorphisms of the integrin gene and the risk of BC

Polymorphism		Patients (n = 317)		Controls (n = 317)a		OR (95%)	P value b
		n	%	n	%
rs11902171	GG (540)	268	84.54	272	85.80	1	-
	GC (91)	48	15.14	43	13.56	0.88 (0.57-1.38)	0.58
	CC (3)	1	0.32	2	0.64	1.97 (0.18-21.86)	1.00
	G (1171)	584	92.11	587	92.59	1	-
	C (97)	50	7.89	47	7.41	0.94 (0.62-1.42)	0.75
rs2675b	AA (412)	219	69.08	193	60.88	1	-
	AC (195)	86	27.13	109	34.38	1.44 (1.02-2.03)	0.04
	CC (27)	12	3.79	15	4.74	1.42 (0.65-3.11)	0.38
	A (1019)	524	82.65	495	78.08	1	-
	C (249)	110	17.35	139	21.92	1.34 (1.01-1.77)	0.04
rs17664	GG (453)	222	70.03	231	72.87	1	-
	GA (168)	91	28.71	77	24.29	0.81 (0.57-1.16)	0.25
	AA (13)	4	1.26	9	2.84	2.16 (0.66-7.12)	0.31
	G (1074)	535	84.28	539	85.02
	A (195)	99	15.62	95	14.98	0.95 (0.70-1.29)	0.76
rs1062484	CC (623)	309	97.48	314	99.05	1	-
	CT (11)	8	2.52	3	0.95	0.37 (0.10-1.40)	0.22
	C (1257)	626	98.74	631	99.53
	T (11)	8	1.26	3	0.47	0.37 (0.10-1.41)	0.23

^aHardy-Weinberg equilibrium calculation for control subjects: p = 0.83 for rs11902171, p = 0.94 for rs2675, p = 0.41 for rs17664, p = 0.93 for rs1062484.

^bThe bold numbers mean the p value is < 0.05.

No association between SNPs and BC recurrence-free survival

Among 317 BC patients, 125 patients completed the follow-up and 71 patients suffered from BC recurrence. According to the collected information, the mean and median follow-up time of this study was 38.25 ± 3.62 months and 24 months, and the mean and median recurrence-free survival time for the 125 patients was 38.87 ± 2.79 months and 25 months, respectively. No association was detected between the four polymorphisms and BC recurrence-free survival time among the 125 patients (Table 4). In the subgroup of non-muscle invasive BC cases, no relationship was found either (Table 5).

Table 4.

Recurrence-free survival analysis of 125 bladder cancer patients

polymorphisms	genotype	N (all, n = 125)	N (reoccurence, n = 71)	HR (95% CI)
Rs2675	AA	83 (66.4%)	48 (57.8%)	1
	AC	37 (29.6%)	20 (54.1%)	0.858 (0.393, 1.871)
	CC	5(4.0%)	3 (60.0%)	1.094 (0.173, 6.897)
Rs1062484	CC	120 (96.0%)	70 (58.3%)	1
	CT	5 (4.0%)	1 (20.0%)	0.179 (0.019, 1.646)
Rs11902171	GG	99 (79.2%)	56 (56.6%)	1
	GC	25 (20.0%)	14 (56.0%)	0.977 (0.404, 2.366)
	CC	1 (0.8%)	1 (100.0%)	—
Rs17664	GG	91 (72.8%)	53 (58.2%)	1
	GA	32 (25.6%)	17 (53.1%)	0.813 (0.362, 1.826)
	AA	2 (1.6%)	1 (50.0%)	0.717 (0.043, 11.825)

Table 5.

Recurrence-free survival analysis of 50 non-muscle invasive bladder cancer patients

non-muscle invasive bladder cancer
polymorphisms	genotype	N (all, n = 50)	N (reoccurence, n = 29)	HR (95% CI)
RS2675	AA	32 (64.0%)	21 (65.6%)	1
	AC	15 (30.0%)	7 (46.7%)	0.458 (0.131, 1.599)
	CC	3 (6.0%)	1 (33.3%)	0.262 (0.021, 3.219)
RS1062484	CC	49 (98.0%)	29 (59.25%)	-
	CT	1 (2.0%)	0 (0.0%)	-
RS11902171	GG	39 (78.0%)	20 (51.3%)	1
	GC	10 (10.0%)	8 (80.0%)	3.800 (0.714, 20.224)
	CC	1 (2.0%)	1 (100.0%)	-
RS17664	GG	36 (72.0%)	20 (55.6%)	1
	GA	14 (28.0%)	9 (64.3%)	1.440 (0.402, 5.157)

Discussion

Previous studies mainly focused on the relationship between polymorphisms in the coding region of integrin genes and BC susceptibility. In the present study, polymorphism rs2675 in the predicted microRNA binding site of ITGB5 gene was identified to be associated with the occurrence and progress of BC, which might provide a potential predicted target for the evaluation of BC susceptibility.

Integrin β₅, encoded by the ITGB5 gene [27], was proved to play a critical role in the process of tumor angiogenesis, proliferation, invasion and migration. Blockage of integrin αvβ5 was identified to result in the reduction in tumor angiogenesis, which was caused by the suppression of phosphoinositide-3-kinase-dependent pathway [28]. In addition, integrin β₅ participated in promoting tumor cell invasion through cooperation with protease activated receptor-1 [29]. And tumor cell migration could be facilitated by integrin β₅ via reactive oxygen species-induced mitochondrial dysfunction [30]. Integrin β₅ was also reported to have the potential to promote tumorigenesis by contributing to the TGF-β-induced epithelial-mesenchymal transition, which was mediated by the Src-focal adhesion kinase and MEK-extracellular signal-regulated kinase induced signaling pathways independently [31]. Moreover, a function research showed that depletion of integrin β₅ in tumor cells resulted in decreased tumor intake and angiogenesis, which led to the reduction in proliferative capacities and migration of tumor. However, this phenotype could be recovered after re-expression of integrin β₅ [32]. So integrin β₅ might be functionally involved in the BC development and progression. In this study, our results indicated that rs2675 in ITGB5 might be correlate with BC risk in Chinese Han population. Due to the lack of validated studies investigating the relationship between various genotypes of this SNP and diverse ITGB5 mRNA/protein expression, it is too early to draw a conclusion about the application of this SNP to predict BC risk. In addition, the recruitment of large samples in different population is required in the future.

Furthermore, the accessibility of miRNA-target combination might be changed as a result of the base transition from A to C in rs2675 (ITGB5), which led to the loss of miRNA targets for mir-504. MiR-504 was demonstrated to promote the tumorigenicity of cells in vivo [33]. As reported by Ma X et al. [34], up-regulation of miR-504 was associated with primary glioblastoma, while ectopic miR-504 increased invasion and migration in oral squamous cell carcinoma [35]. New miRNA targets in ITGB5 gene appeared for mir-30a-3p and mir-30e-3p at the same time. The emergence of mir-30a-3p and mir-30e-3p might have the opposite effect against miR-504. These two miRNAs inhibited tumor proliferation, invasiveness and metastasis, which were identified to be downregulated in hepatocellular carcinoma and clear renal cell carcinoma [36,37]. But it is still unclear about the accurate regulation mechanisms of these miRNAs on rs2675 in BC susceptibility. Moreover, whether this predicted miRNA binding site is the exact miRNA binding site needs further analysis.

No association was detected between the other three polymorphisms and BC risk in this study. For rs17664 in ITGA6 gene and rs1062484 in ITGA3 gene, no relationship between these polymorphisms and the risk of other tumor types was demonstrated by existing studies as well. However, polymorphism rs11902171 in ITGAv gene was reported to be associated with the occurrence and progression of prostate cancer [38] and gastric cancer [39], which might be relevant with the characteristics of tumor itself or the limitation caused by the sample size.

Integrins were proved to play an important role in the process of tumor recurrence. Previous studies identified that integrins could predict the intravesical recurrence of non-muscle invasive bladder cancer (NMIBC) [40]. However, none of these four SNPs in the integrin gene predicted miRNA binding sites mediates BC recurrence-free survival of 125 patients who completed the follow-up. Treatment and prognosis for muscle invasive bladder cancer (MIBC) and NMIBC are quite different. The majority of patients with MIBC received radical cystectomy or radiotherapy, and distant metastases were the most common cause of death [41]. For the treatment of NMIBC, although a combination of transurethral resection (TURBT) and intravesical therapy was widely applied clinically, the recurrence rate was rather high in both low-and high-grade disease, which fluctuated between 50% and 70% [42,43]. In addition, the degree of malignancy increases with the raised frequency of recurrence. No exact indicators have been found for the prediction of NMIBC recurrence. Therefore, subgroup analysis in NMIBC patients is necessary to verify the prognostic role of these polymorphisms. However, no significant association was detected between these polymorphisms and the recurrence-free survival of NMIBC, which might be caused by survival index selection or limited sample size.

In conclusion, our data proved the association between polymorphism rs2675 in ITGB5 gene and the risk of BC. This finding will be of benefit to the screening of high BC risk population and the prediction of prognosis in BC patients. However, regional population recruitment, SNPs selection limitation, and lack of functional analysis contributed to the defect of the study. Therefore, further well-designed studies in a large scale are needed to confirm the results in our study.

Disclosure of conflict of interest

None.