Abstract
The present study analyzed single nucleotide polymorphisms (SNPs) located at putative microRNA(miRNA)-binding sites of the 3’-untranslated region (UTR) in different genes and investigated their impact on the susceptibility to colorectal cancer (CRC). Ninety-two SNPs were selected using an in silico analysis of 3’-UTR SNPs in an SNP database and their miRNA binding efficiency was calculated using several miRNA databases and the HapMap database. Two independent study sets were used: 380 healthy controls and 371 patients with colorectal adenocarcinoma for the discovery set, and 521 healthy controls and 524 patients with colorectal adenocarcinoma for the validation set. The SNP genotyping was performed using a Sequenom MassARRAY. In addition, a luciferase assay was used to investigate whether miR-370 modulated docking protein 3 (DOK3) gene expression when rs2279398G>A was included in the DOK3 3’-UTR region. For the discovery set, 16 out of 92 SNPs were significantly associated with the risk of CRC in at least one of the genetic models. The validation set showed that among these 16 SNPs, DOK3 rs2279398G>A was significantly associated with reduced risk of CRC in a recessive model [adjusted odds ratio (aOR)=0.65, 95% confidence interval (CI)=0.44-0.97, p=0.03]. In a combined analysis, DOK3 rs2279398G>A was associated with a significantly reduced risk of CRC in a co-dominant and recessive model (aOR=0.84, 95% CI=0.73-0.96, p=0.012; aOR=0.65, CI=0.49-0.88, p=0.004, respectively). Significantly lower Renilla activity was also observed with the rs2279398 AA construct when compared to the rs2279398 GG construct (p<0.001). DOK3 rs2279398G>A may affect the expression of DOK3 by altering the miRNA binding efficiency at the miRNA-binding sites of the 3’-UTR in DOK3, thereby impacting CRC tumorigenesis.
Colorectal cancer (CRC) is the third most common form of cancer and the fourth most common cause of cancer deaths globally (1). Since relapse following cancer surgery and resistance to chemotherapy persist as major obstacles to effective treatment, prognosis for CRC remains poor. Therefore, the rising incidence of CRC and its poor prognosis are pushing systematic approaches towards a more precise identification of patients with a higher susceptibility of CRC (2). While genetic factors play an important role and provide valuable information on the etiology of both sporadic and familial CRC, fewer than 6% of CRC cases can be explained by rare, high-penetrance variants of the CRC susceptibility genes identified to date (3, 4).
MicroRNAs (miRNAs) are a class of short, single-strand noncoding RNAs that play key roles in the regulation of gene expression by translational repression or mRNA degradation of the target, affecting critical functions in various physiological processes (5, 6). Alterations of miRNAs are involved in a wide range of human diseases, and can also affect tumor development and progression by regulating the expression of proto-oncogenes and tumor-suppressor genes (7, 8). Recent reports have shown that the expression of miRNAs is associated with the development or prognosis of CRC (9-11). Therefore, miRNAs are likely to be highly useful as biomarkers in CRC. Single nucleotide polymorphisms (SNPs) can occur in genes for the miRNA biogenesis pathway, primary miRNA, pre-miRNA, or mature miRNA sequences (12). Although the exact role of SNPs detected in miRNA has yet to be clarified, our previous studies demonstrated that specific SNPs are associated with susceptibility to CRC and its prognosis (13-15). Importantly, the predominant function of miRNAs is to regulate protein translation by binding to complementary sequences in the 3’-untranslated region (UTR) of target mRNAs, and thereby negatively regulate mRNA translation (16). Thus, polymorphisms of miRNA-binding sites in the 3’-UTRs of target genes could potentially affect the efficiency of miRNA binding to the target sites by altering the miRNA–mRNA interaction and changing the expression of the miRNA targets (12, 17, 18). However, these SNPs and their relationship to the risk of CRC have not yet been fully characterized as novel biomarkers for categorizing patients with different genetic risks (19). Accordingly, the present study analyzed 92 SNPs located at putative miRNA-binding sites in the 3’-UTR of different genes and investigated their impact on susceptibility to CRC in terms of altering the miRNA-binding efficiency.
Patients and Methods
Study populations. A total of 371 patients who were histologically confirmed to have sporadic CRC at the Kyungpook National University Hospital (KNUH) in Daegu, Korea, were included in the discovery set. The controls (n=380) were randomly selected from a pool of healthy volunteers who visited the general health check-up center at the hospital and were matched with the patients as regards gender. For an independent validation set, a total of 524 patients and 521 controls were collected. This study was approved by the Institutional Review Board of KNUH (KNUMCBIO_11-1002) and written informed consent was obtained from all participants. The genomic DNA samples from the patients with CRC and healthy controls were provided by the National Biobank of Korea-KNUH, which is supported by the Korean Ministry of Health, Welfare and Family Affairs. All patients and controls were ethnic Koreans residing in Daegu or the surrounding regions. The CRC diagnosis and staging were determined according to the WHO classifications and TMN classifications included in the 6th edition of the American Joint Committee on Cancer (20).
SNP selection and genotyping. A total of 12,877 polymorphisms in the 3’-UTR of genes that may be affected by miRNAs were selected from a public SNP database (http://www.ncbi.nlm.nih.gov/SNP) and Ensemble data. To select all the potentially functional miRNA-binding polymorphisms, this study used the miRNA target prediction program miRanda algorithm version 3.1 (http://www.microrna.org/microrna), where SNPs with minimum free energy (MFE) difference <6 between two alleles were excluded when calculating the MFE for all the possible miRNA and SNP complex sequences. As a result, a total of 391 SNPs were collected with a minor allele frequency (MAF) ≥0.1 from the HapMap JPT data in the public SNP database and a Gibbs binding free energy (ΔG, kJ/mol) ≥5.7 kJ/mol in a miRNA and RNA hybrid model. The selected 391 SNPs were analyzed with a healthy control population using SEQUENOM's MassARRAY® iPLEX assay (SEQUENOM Inc., San Diego, CA, USA), which eliminated 54 SNPs with a MAF <0.1. Among the remaining 337 SNPs, 92 SNPs applicable to the SEQUENOM's MassARRAY were genotyped.
The genomic DNA was extracted from normal tissues using as QIAamp® genomic DNA kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. The genotype analysis was performed using SEQUENOM's MassARRAY® iPLEX assay according to the instructions of the manufacturer. To validate the genotyping, approximately 5% of the cohort samples were randomly selected for re-genotyping by another researcher using a sequencing method or a restriction fragment length polymorphism assay, and the results were 100% concordant.
Cloning of luciferase reporter gene and luciferase assay. In the in silico analysis, since docking protein 3 (DOK3) gene expression was affected by miR-370-binding efficiency, miR-370 was selected using the MicroSNIPer program (http://epicenter.ie-freiburg.mpg.de/services/microsniper/index.php). A luciferase assay was used to investigate whether miR-370 modulated DOK3 gene expression when rs2279398 was included in the DOK3 3’-UTR region. A 142-bp fragment of the DOK3 3’-UTR region was synthesized using polymerase chain reaction and cloned into the dual luciferase vector psiCHECK2 (Promega, Madison, WI, USA). A forward primer with an XhoI restriction site (5’- CCC CTC GAG ACA CAC GCC TGT GTC CAC C -3’) and reverse primer with a NotI restriction site (5’-CCG CGG CCG CTG CAG TGG GTT TGA GC -3’) were then used to amplify the DOK3 3’-UTR region. DNA sequencing was also used to verify that the clone sequences were all correct. Using a 12-well plate, 293T cells (ATCC, Manassas, VA, USA) were seeded in a Dulbeco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum. The cells were then transfected with psiCHECK2-DOK3 constructs containing the 3’-UTR of DOK3 that included rs2279398G>A, in the presence of a miR-370 mimic (Qiagen) using Effectene (Qiagen) transfection reagent. After 48 hours, the cells were collected and cell lysates prepared according to Promega's instruction manual. The Renilla luciferase activity was measured using a Lumat LB953 luminometer (EG & G Berthhold, Bad Wildbad, Germany), and the results normalized using the luciferase activity. All the experiments were performed in triplicate.
Statistical analysis. The results for the patients and controls were compared using Student's t-test and the chi-square test for the continuous variables and categorical variables, respectively. The p-value for the luciferase assay was determined using Student's t-test. The Hardy–Weinberg equilibrium was tested by comparing the observed and expected genotype frequencies using a χ2 test with 1 degree of freedom. An unconditional logistic regression analysis was used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs), with adjustment for age and gender. All the analyses were performed using Statistical Analysis Software for Windows, version 9.2 (SAS Institute, Cary, NC, USA). The statistical significance of difference between the groups was determined using Student's t-test with the treatment and experiment as factors. Differences were considered statistically significant at p<0.05.
Characteristics of the study population.
Results
Characteristics of study populations. The descriptive characteristics of the study participants are provided in Table I. No significant differences were found between the patients and the controls as regards the mean age and sex in the discovery set and validation set, yet the mean age of the patients in the discovery set was significantly higher than that of the controls (p=0.006). The median age of patients with CRC was 65 years, and 505 (56.4%) had colonic cancer and 385 (43.0%) had rectal cancer. The pathological stages after surgical resection were as follows: stage I: n=148, 16.5%; stage II: n=327, 36.5%; stage III: n=329, 36.8%; and stage IV: n=91, 10.2%. Twenty percent of the patients had elevated carcinoembryonic antigen levels.
Genotypic frequencies and their association with CRC risk. For the discovery set, 16 of the 92 SNP genotypes were significantly associated with the risk of CRC in at least one of the genetic models (Table II). Meanwhile, for the validation set, among these 16 SNPs, DOK3 rs2279398G>A was significantly associated with reduced CRC risk in a recessive model [adjusted odds ratio (aOR)=0.65, 95% CI=0.44-0.97, p=0.033]. In the combined analysis, the A allele of DOK3 rs2279398G>A was associated with a significantly reduced risk of CRC co-dominant and recessive models (aOR=0.84, 95% CI=0.73-0.96, p=0.012, aOR=0.65, CI=0.49-0.88, p=0.004, respectively) (Table III). Finally, the subgroup analysis according to the primary tumor site also revealed a strong association of the A allele of DOK3 rs2279398G>A with reduced risk of colonic cancer (Table IV).
Effect of DOK3 rs2279398G>A polymorphism on miR-370 binding. To verify whether the rs2279398G>A polymorphism directly affected the miR-370-binding efficiency in DOK3 3’-UTR, a Renilla luciferase reporter plasmid cloned downstream to a segment of the DOK3 3’-UTR containing rs2279398G>A was generated. Each construct, psiCHECK2: rs2279398GG and psiCHECK2: rs2279398AA, was then co-transfected into 293T cells with miR-370, and the Renilla luciferase activity measured after 48 hours. As shown in Figure 1, significantly lower Renilla activity was observed using the rs2279398AA construct when compared to the rs2279398GG construct (p<0.001), suggesting that the rs2279398G>A polymorphism may affect DOK3 expression by the direct binding of miR-370 to its 3’-UTR.
Discussion
A relatively large population-based study was used to investigate the influence of 92 SNPs located at putative miRNA-binding sites of the 3’-UTR of different genes on the risk of CRC. As a result, the DOK3 rs2279398G>A polymorphism was found to be associated with a significantly lower risk of CRC under a co-dominant and recessive model for the minor A allele. Therefore, these findings suggests that the polymorphism at the miRNA-binding site of DOK3 (rs2279398) may be involved in CRC carcinogenesis and could be useful as a marker for genetic susceptibility to CRC. This result is worthy of note, since as far as we are aware, there are no previous data on the association of a DOK family polymorphism with CRC risk.
Genotypes in the discovery and validation studies, and their association with the risk of colorectal cancer.
The 3’-UTR contains miRNA response elements and plays a crucial role in gene expression by influencing the localization, stability, export, and translation efficiency of mRNA, contributing to gene regulation (5, 6). Therefore, a polymorphism in the 3’-UTR of the gene can create as well as destroy the miRNA-binding site and have a similar influence on an SNP located within the miRNA seed region (21). In particular, when considering that miRNAs play a pivotal role in CRC initiation and development, it is important to evaluate the CRC risk associated with such SNPs. Several studies have already investigated the functional effect of various polymorphisms on the miRNA recognition sites on target genes and their association with CRC risk (22). Landi et al. studied selected SNPs within putative miRNA sites of genes known to be involved in CRC, and found that two polymorphisms, namely rs17281995 in [cluster of differentiation 86 (CD86)] and rs1051690 in [insulin receptor (INSR)], were associated with an increased risk of CRC (23, 24). Moreover, another recent study found an association between the SNPs in [replication protein A2 (RPA2)] and [general transcription factor IIH subunit (GTF2H1)] of a 3’-UTR involved in a DNA-repair pathway and the risk of rectal cancer (25). However, such biological functions still need to be experimentally verified through a functional study.
Interestingly, there are several important differences between the data in the present study and that in previous CRC association studies. Firstly, given the homogeneous ethnic background of Korean patients, any potential confounding effect due to ethnicity is likely to be small in the current study. Moreover, the luciferase reporter assay revealed that the DOK3 (rs2279398) polymorphism directly affected gene expression and the risk of cancer. In addition, this study included a relatively large number of patients.
The significant finding from the present study is an association between a DOK3 (rs2279398) polymorphism and the risk of CRC. DOK3 is an adaptor protein that functions in feedback loops to modulate tyrosine kinase signaling. The activation of protein-tyrosine kinases induces tyrosine phosphorylation of the target proteins, triggering molecular interactions of the proteins. Therefore, the alteration of DOK3 can negatively regulate the rat sarcoma virus-extracellular signal-regulated kinase (RAS-ERK) pathway, which is a positive signaling cascade critical for cell activation downstream of protein-tyrosine kinases (26). In a recent in vitro study of the response of CRC cell lines to a combination of 5-fluorouracil and radiotherapy, the authors reported that DOK3 blocked the complex formation of downstream molecules, thereby behaving as an inhibitor of the RAS signaling pathway (27). Therefore, given these results, the rs2279398 polymorphism located in the 3’-UTR of the DOK3 gene may influence miRNAbinding efficiency and thus potentially modulate tumor suppression. In fact, a recent phase III trial demonstrated that the addition of cetuximab, a monoclonal antibody that targets the epidermal growth factor receptor (EGFR), to chemotherapy significantly increased survival in patients with metastatic CRC (28). However, an anti-EGFR treatment may not be optimal for CRC, and interpreting the exact therapeutic effect of such a treatment is apparently difficult in patients with a V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation. Therefore, the current results may also provide essential SNP data to explain the precise biological significance of cancer development, as well as treatment resistance in patients with CRC. Similarly, Zhang et al. have reported that the KRAS 3’-UTR polymorphism can predict the cetuximab responsiveness in patients with wild-type KRAS with CRC treated with cetuximab monotherapy (29). Yet despite increasing evidence that DOK3 plays important roles in several aspects of tumorigenesis, it is still unclear whether the DOK3 (rs2279398) polymorphism itself alters the protein expression. Notwithstanding, the current study provides explicit evidence that the rs2279398 polymorphism is associated with a significantly lower risk of CRC under a co-dominant and recessive model for the minor A allele.
Renilla/luciferase assay of miRNA suppression with docking protein 3 (DOK3) rs2279398G>A polymorphism. 293T cells were co-transfected with miR-370a and a psiCheck2 plasmid containing DOK3 3’-UTR with the rs279398GG or rs2279398AA polymorphism. Each bar represents the mean±s.e.m. renilla luciferase activity normalized to firefly luciferase activity. p-Value was determined using Student's t-test.
To date, several risk factors have been associated with a higher incidence of CRC (30). The modifiable risk factors of CRC include smoking, physical inactivity, being overweight/obese, eating processed meat, and drinking alcohol excessively. The presence of polyps and hereditary diseases also considerably increases the risk of CRC. More recently, based on the extensive study of genetic alterations in CRC, tumorigenesis has been identified as a multistep process that involves the accumulation of mutations in tumor-suppressor genes and oncogenes such as adenomatous polyposis coli (APC), p53, and KRAS (31).
Docking protein 3 (DOK3) genotypes of cases and controls, and their associations with the risk of colorectal cancer.
Docking protein 3 (DOK3) genotypes of cases and controls, and their associations with the risk of colorectal cancer according to the primary site.
The present study also showed that patients at high risk for developing CRC might be identified by genetic testing for specific polymorphisms. This finding can potentially benefit patients through the identification of the DOK3 polymorphism, enabling the early detection of disease and providing information for appropriate treatment strategies. However, even though the present data identified certain gene variants as statistically significant risk factors, these results should be interpreted cautiously. First of all, when considering the multiple comparison issue, the possibility of a type I error cannot be ruled-out in the analysis of individual SNPs. Therefore, additional studies with larger sample sizes are required. Secondly, the present finding that the DOK3 polymorphism itself alters the protein expression still needs to be confirmed. In addition, gene interactions with environmental and lifestyle factors can also affect the relevance of different variants in cancer susceptibility. And last but not least, the current results do not conclusively reveal the exact role of the DOK3 polymorphism in CRC tumorigenesis or its relationship with the RAS-ERK pathway.
The current findings indicate that genetic variations of DOK3 may influence the risk of CRC. However, since the exact mechanism and function of these gene variants have not yet been fully defined, the present findings need to be confirmed in further studies with other populations in order to clarify the association between these polymorphisms and the risk of CRC.
Acknowledgements
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Korean Ministry of Education, Science and Technology (NRF-2011-0015862 and NRF- 2014M3A9E2064642).
Footnotes
↵* These Authors contributed equally to this work.
Conflicts of Interest
The Authors declare no conflicts of interest with regard to this study
- Received April 12, 2016.
- Revision received May 17, 2016.
- Accepted May 24, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.