Genetic variation at the microRNA binding site of CAV1 gene is associated with lung cancer susceptibility

Single nucleotide polymorphism (SNP) may influence the genesis and development of cancer in a variety of ways depending on their location. Here we conducted a study in Chinese female non-smokers to investigate the relationship between rs1049337, rs926198 and the risk or survival of lung cancer. Further, we explored whether rs1049337 could alter the binding affinity between the mRNA of CAV1 and the corresponding microRNAs. Finally, we evaluated the relationship between expression level of CAV1 and prognosis of lung cancer. The results showed that the rs1049337-C allele and rs926198-C allele were the protective alleles of lung cancer risk. Haplotype analysis indicated that the C-C haplotype (constructed by rs1049337 and rs926198) was a protective haplotype for lung cancer risk. The result of luciferase reporter assay showed that rs1049337 can affect the binding affinity of CAV1 mRNA to the corresponding microRNAs both in A549 cell line and H1299 cell line. Compared with C allele, T allele had a relatively decreased luciferase activity. Compared with paired normal adjacent tissue or normal lung tissue, lung cancer tissue showed a relatively low level of CAV1. Refer to those patients at early stage of lung cancer, the expression level of CAV1 in patients at late stage of lung cancer was relatively low. In conclusion, the results indicated that rs1049337, it's a SNP located at 3′UTR region of CAV1 may affect lung cancer risk by altering the binding affinity between the mRNA of CAV1 and the corresponding microRNAs.


INTRODUCTION
Lung cancer is the most frequently diagnosed lethal malignancies in the world. As the poor 5-year survival rate, lung cancer has become the leading cause of death among females in more developed countries and males in the whole world. As is known to all, smoking is regarded as a common risk factor of lung cancer. Global cancer statistics state that 15% of men and 53% of women with lung cancer were not caused by smoking [1]. Smoking rates in Chinese women are lower than some European countries, but the incidence of lung cancer in China is relatively high [2]. This suggested us that some other factors may be involved in lung cancer in Chinese female population.
Caveolae is a plasma membrane invagination and juxtamembrane vesicles, which play a role in the macromolecular vesicle transportand can be regarded as a kind of lipid raft. Due to the ubiquitous tissue distribution of caveolae, it has been confirmed that caveolae involved a large number of cellular functions. Caveolin-1 (CAV1) protein is a principal structural component of caveolae and found in the cytosolic surface of caveolae, which called caveolar coat protein. In mammalian cells, CAV1 protein is quite small and highly conserved which can recruit many signaling molecules to caveloae and regulate their activity [3][4][5].
CAV1 has been found to be involved in a variety of cancer-associated biological processes, including cell migration and metastasis, cell transformation, angiogenesis, multidrug resistance [6][7][8][9]. CAV1 is an interesting gene for cancer as it may play different roles in different types of cancer at different stages. No consistent results regarding whether CAV1 is an oncogene or tumor suppressor gene. Most studies suggest that CAV1 may play a positive role in the early stages of cancer onset. In advanced tumors, CAV1 is frequently associated with poor prognosis and drug resistance [10].
Single nucleotide polymorphism (SNP) is a kind of single nucleotide change widely distributed in the human genome. There is increasing evidence suggesting that SNPs may play a critical role in the genesis and development of cancer [11][12][13][14]. The relationship between SNPs in CAV1 and cancer risk has not been extensively studied. In the present study, we performed a case-control study to assess the relationship between SNPs in CAV1 and the risk or survival of lung cancer in Chinese female non-smokers. Table 1 listed the basic information of the subjects. A total of 923 subjects were included in the present study, including 468 lung cancer cases and 455 age-matched controls. All subjects were Chinese female nonsmokers, and there was no significant difference between cases and controls with regard to age (P=0.861). Among cases, there were 402 non-small cell lung cancer (NSCLC) and 66 small cell lung cancer (SCLC). Among the cases with NSCLC, there were 322 cases with adenocarcinoma, 66 cases with squamous cell carcinoma and 14 cases with a variety of different rare pathologies.

The effect of SNP on the risk of lung cancer
First we evaluated the relationship between the two SNPs and lung cancer risk. The results were summarized in Table 2

Haplotype analyses of the SNPs and lung cancer risk
We analyzed the relationship between haplotype and lung cancer risk. The haplotypes were constructed by rs1049337 and rs926198 (Table 4)

The effect of SNPs on the overall survival of lung cancer
The results of the relationship between genetic polymorphisms and survival time were summarized in Table  5, Supplementary Table 2 and Supplementary Figure 1. We have not observed any statistically significant association between rs1049337, rs926198 and lung cancer survival time.

Luciferase reporter assay
Supplementary Table 3 listed the predication results of microRNAs those may bind to CAV1 mRNA around rs1049337. Considering the predication score, finally hsa-miR-612 was selected to verify its relationship with rs1049337.
The result showed that rs1049337 can affect the binding affinity of CAV1 mRNA to some microRNAs ( Figure 2). In the A549 cell line and the H1299 cell line, we got essentially identical results. Compared with C allele, T allele had a relatively decreased relative luciferase activity. When miR-612 mimics existed in cell, T allele had a lower expression level of CAV1 compared with C allele. Compared with mimics negative control, the relative luciferase activity of vectors containing C allele and T allele was significantly decreased when miR-612 mimics existed. The results indicated that rs1049337 can affect the binding affinity of CAV1 mRNA to the corresponding microRNA, these microRNAs include miR-612.

The association of CAV1 expression level and lung cancer
Finally, we screened four datasets in GEO database which were suitable for our study. The results of the analysis of the relationship between CAV1 expression level and NSCLC were listed in Table 6. We found that compared with paired normal adjacent tissue or normal lung tissue, lung cancer tissue showed a relatively low level of CAV1 mRNA (P<0.001, P<0.001, respectively). In three datasets (GSE19804, GSE27716 and GSE40275), we found that referring to those patients at early stage of lung cancer, the expression level of CAV1 in patients at late stage was relatively low (P=0.007, P=0.005, P=0.001, respectively). In GSE27716, a borderline statistical difference was found in the comparison of CAV1 mRNA levels in noninvasive lung adenocarcinoma patients and invasive lung adenocarcinoma patients (P=0.08).

DISCUSSION
The role of CAV1 in cancer is oncogene or tumor suppressor gene hasn't been agreed yet, and the role of CAV1 in lung cancer has rarely been studied [3,4,7]. We conducted a study to explore the relationship between CAV1 gene polymorphism, expression level and lung cancer. We found that the CAV1-rs1049337, CAV1-rs926198 may affect the susceptibility of lung cancer in the Chinese female non-smokers. Further experiments revealed that compared with rs1049337-T allele, the increased relative luciferase activity was found in C allele. It was proved that the different allele of rs1049337 could affect the binding affinity of CAV1 to the corresponding microRNA. We also found that CAV1 expression may play a positive role in lung cancer.
MicroRNA is a class of single stranded noncoding RNA which is considered to regulate large number of genes at post transcriptional level. The regulatory effect of microRNA on gene expression is achieved by microRNA seed region complementary base pairing to the 3'UTR region of the target gene [15,16]. Any nucleotide change in 3'UTR region of target gene can affect the affinity of target gene and microRNA, and further influence the expression level of target gene. Rs1049337 is located at 3'UTR region of CAV1 gene. Our study found that compared with rs1049337-T allele, C allele was a protective allele for risk of lung cancer. Further investigation demonstrated that rs1049337 can affect the binding affinity of CAV1 mRNA to the corresponding microRNAs. Finally, we found that CAV1 expression may play a positive role in lung cancer. As far as we know, there are very few studies on the relationship between rs1049337 and disease. A study on the relationship between genetic variation in CAV1 and long-term pancreas transplant function found that rs1049337 is a tag SNP of CAV1, but there's no significant correlation between long-term graft function, long-term pancreas function and rs1049337 was observed in donors and recipients respectively [17]. The MuTHER study was found that rs1049337 is an expression Quantitative Trait Loci (eQTL), and another study based on the MuTHER study revealed that rs1049337 is an eQTL with an ageby-genotype interaction effect associated to GWAS hits [18,19]. Although we have studied the mechanism of how rs1049337 affects lung cancer, considering racial differences and the credibility of a single study, multicenter replication studies still need to be done. Rs926198 is located at intron region of CAV1. Although we haven't figured out how does intron SNPs play a role in disease, a large number of genome-wide association study (GWAS) have confirmed that intron SNPs are associated with a variety of diseases in a quite robust P value [20][21][22]. In our study, we found that compared with rs926198-T allele, C allele was a protective allele for lung cancer risk. To the best of our knowledge, it's the first study to explore the relationship between rs926198 and cancer. Previous studies indicated that rs926198 is associated with metabolic syndrome, fasting insulin levels, insulin resistance, hyperinsulinemic, hypertension and systemic sclerosis [23][24][25][26]. After further study, Rene Baudrand et al. found that in inadipose tissues, rs926198 minor allele is associated with a lower expression level of CAV1 [24]. So it is necessary to further explore the mechanism of how rs926198 plays a role in the disease.
CAV1 is localized to 7q31.2, closed to D7S522 locus which is a fragile site that is frequently absent in cancer and is thought to be a suspected tumor suppressor locus [27]. Indeed, CAV1 is considered to play a positive role in some cancer studies, but the results of some other studies were exactly the opposite. Whether in vitro or in animal or in cancer patients, CAV1 is a tumor promoter gene or tumor suppressor gene that has not been consistent. Breast cancer is the most widely studied cancer with CAV1. One study found that compared with normal mammary epithelial counterparts, CAV1 showed a significant decrease in breast cancer cell, and loss of CAV1 expression was found in P53 deficient cells [28]. CAV1 can inhibit the growth and metastasis of breast cancer [29]. The molecular mechanism study revealed that in human breast cancerassociated fibroblasts, down regulated CAV1 expression play a key role in maintaining the abnormal phenotype [30,31]. Breast cancer patients with loss CAV1 expression in stromal cells were associated with shorter survival time, increased risk of early tumor recurrence, higher recurrence rate and metastasis, higher CAV1 expression in tumor cells were associated with poor survival [32][33][34][35][36][37]. Other studies have suggested that CAV1 plays a negative role in breast cancer. In human breast cancer cell, CAV1 accumulates at invadopodia and its knockdown inhibits invadopodia formation [38]. CAV1 is thought to be a potential causative factor of trastuzumab resistance generation in breast cancer cells [39]. In prostate cancer cell, CAV1 is thought to promote metastatic activities, proangiogenic activities, promote tumor progression and promotes lymphangiogenesis [40][41][42][43]. In prostate cancer stroma, loss of CAV1 associated with reduced relapse-free survival [44]. The positive expression of CAV1 was more observed in high grade urothelial carcinoma and bladder cancer patients [45,46]. In human transitional bladder cancer cell with drug resistance, the expression level of CAV1 was found   to be elevated [47]. But one study has found that compared with normal tissue, CAV1 had a decreased expression in bladder cancer tissues [48]. The expression of CAV1 in colon carcinoma was up-regulated or down-regulated, the results of different studies have been quite different [49][50][51].
In Apc (min/+) mice with deficiency of CAV1, a promoted colorectal tumorigenesis was observed [52]. A meta analysis showed that overexpression of CAV1 was associated with a better overall survival in gastric cancer patients [53]. In CAV1 null (CAV-1 -/-) mouse model, hypercellularity with thickened alveolar septa was observed in lung parenchyma, disorganize and multilayer was observed in alveolar wall [54,55]. In lung cancer cell lines, there is controversy about whether CAV1 is expressed. Compared with lung epithelial cell, one study showed that the expression of CAV1 was reduced in lung adenocarcinoma cell lines [56]. In another study, CAV1 was found to be reduced in small cell lung cancer cell lines, but in 76% of NSCLC cell lines, it is still expressed [57]. In lung cancer patients, compared with the adjacent tissues or normal lung tissue, the expression of CAV1 was down-regulated in cancer tissues [58][59][60][61][62]. But CAV1 was more frequently expressed in NSCLC patients those with drug resistance and poor prognosis [60,[63][64][65]. In our study, CAV1 may be considered to play a positive role in

Study subject
This study is approved by the ethics committee of China Medical University. Each participant signed an informed consent before participating in the study. All of the participants were unrelated Chinese Han population and non-smoking women. Patients with lung cancer were recruited from the First Affiliated Hospital of China Medical University and Liaoning Cancer Hospital& Institute. All of the patients were newly detected and had not received any radiotherapy or chemotherapy. Over the same period, controls were recruited from the health examination center of the same hospital and frequency-matched to cases in age.

Data collection and SNP genotyping
Clinical information of patients was collected from patients' medical records. By face-to-face interviews, we obtained the personal information from each subject. The data was recorded age, status of smoking, family history of cancer, etc. Subjects who had smoked less than 100 cigarettes in their lifetime were defined as non-smokers. Each subject was collected 10ml of venous blood. DNA was extracted from whole blood using standard Phenol-chloroform Method and was stored at -20°C for subsequent experiments. We carried out the follow up by telephone among the patients once every three months. SNPs were genotyped by using the Illumina 660W SNP microarray (Illumina Inc., San Diego, CA).

Cell culture and transfection
We purchased A549 and H1299 cell line from Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. A549 and H1299 cell lines were cultured in RPMI-1640 medium (Biological Industries, Israel) supplemented with 10% fetal bovine serum (Biological Industries, Israel), 80U/ml penicilin and 0.08mg/ml streptomycin, followed by incubation with 5% CO 2 at 37°C. A549 and H1299 cell lines transiently transfected using Opti-MEM (Gibco, USA) and Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer's instructions.
MicroRNA mimics were purchased from GenePharma (GenePharma, China). A 602-bp fragment of CAV1 3'UTR containing rs1049337-C or rs1049337-T were amplified and cloned in GV272 vector (GeneChem, China). All reporter plasmid were verified by direct sequencing. Cell were cultured in 24-well plates, transient transfection was performed at the logarithmic phase. 48 hours after transfection, the firefly luciferase and renilla luciferase were quantified by Dual-Luciferase Reporter Assay System (Promega, USA). We performed three independent experiments and each was done in triplicate.

The association of CAV1 expression level and lung cancer
In order to evaluate whether the expression level of CAV1 can affect the progression and prognosis of lung cancer patients, we systematically searched the GEO database to obtain the NSCLC gene expression profile datasets with prognostic information [70,71]. Data extraction were using GEO2R web tool in GEO database. If there are multiple probes corresponding to the same gene, the average value of these probes was calculated as the expression level of the gene.

Statistical analysis
Student's t-test (for continuous variable) and chi-squared test (for categorical variables) were used to measure the difference between two groups in the distribution of demographic characteristics. Comparison of continuous variables between three or more groups were used one-way analysis of variance (ANOVA). In control group, the Pearson chi-squared test was adopted to determine whether the genotype frequencies follow Hardy-Weinberg equilibrium (HWE). Unconditional logistic regression model was applied to calculate the odds ratios (ORs) and their 95% confidence intervals (CIs) to evaluate the relationship between SNPs and lung cancer risk with adjustment for age. The haplotypes frequency estimation and the analysis of association between haplotypes and lung cancer risk were using SHEsis online web-server [72,73]. Statistical analyses were carried out using SPSS 22.0 software (IBM, USA). All p value were two sided and p<0.05 was considered statistically significant.

Author contributions
Conceived and designed the experiment: BZ ZY XF. Performed the experiments: XF XL XQ. Analyzed the data: XF XL. Contributed reagents/materials/analysis tools: BZ ZY YZ. Wrote the paper: XF BZ. Statistical analysis and interpretation: XF ZY XL.

ACKNOWLEDGMENTS
We thank the GEO database for making the data public. Heartfelt thanks to each author's and subjects's contribution.

CONFLICTS OF INTEREST
There is no potential conflict of interest in this study.

FUNDING
This study was supported by grants No. 81272293 from the National Natural Science Foundation of China.