Introduction

Colorectal cancer (CRC) is the third most common cancer, the fourth leading cause of cancer-related death globally, and the second most common cancer in terms of the number of individuals living with cancer for five years or more, making up about 10% of all malignant diagnoses. It is been estimated that 1.4 million people are diagnosed with CRC per year, 65% of whom are found in developed countries; approximately 700,000 die of this disease per year; and 3.5 million individuals are living with CRC [1]. CRC is a complex disease that is influenced by multiple factors [2–4], however, numerous evidences have suggested that environmental factors and genetic variations are postulated to significantly affect the risk of CRC [5, 6]. In addition, single nucleotide polymorphisms (SNPs) also play an important role in the development and progression of CRC [7–9]. At present, sequencing of the human genome has revealed about 18.7 million SNPs that are thought to account for 90% of all sequence variation [10].

The KRAS proto-oncogene belongs to the RAS GTPase family which regulates cellular proliferation, apoptosis and other important biological processes [11]. KRAS protein plays a key role in the EGFR signaling pathway and oncogenic mutations in protein can drive downstream activation of this pathway even in the absence of upstream EGFR activation. The KRAS mutation is an essential step which is thought to contribute to cancer development by driving proliferation of cells and resisting to apoptosis with initiated mutations. KRAS 3’-untranslated regions (UTR) of human contains multiple putative tumor suppressor lethal-7 (let-7) complementary sites (LCS). The SNPs of KRAS 3’-UTR may prevent the let-7 miRNA from binding to KRAS and regulate the activity of KRAS so as to adjust the expression of its protein. Previous studies have investigated that the SNPs in the KRAS 3’-UTR might cause high levels of the KRAS oncogenic protein and lower levels of the let-7 miRNA, the overexpressed KRAS oncogenic protein can increase the activation of the RAF/MEK/MAPK pathway, which might promote the tumorigenesis of CRC [12, 13].

The human vascular endothelial growth factor (VEGF) gene is located on chromosome 6 at location 6p21.3 and consists of 8 exons [14]. VEGF gene is highly polymorphic, and its promoter, 5’-, and 3’-UTR has a variety of SNPs [15]. VEGF SNPs in the 3’-UTR have been found to be associated with variations in VEGF protein production [16]. These SNPs could result in high expression of the VEGF gene and increase tumor-related angiogenesis and metastasis, playing a key role in a series of pathologic processes involved in tumor growth and metastasis [17]. Moreover, VEGF-involved angiogenesis pathways are also believed to be an important target of chemotherapeutic treatment in CRC. VEGF mRNA expression influences tumor tissues by the combination of 3’-UTR polymorphic alleles present. Studies showed that polymorphic events affecting the mRNA expression of VEGF gene could affect the survival duration of cancer patients having received anti-angiogenic treatment [18]. Therefore, VEGF gene polymorphisms might also serve as potential molecular biomarkers to predict clinical outcomes [14].

Genetic variations in the KRAS gene and VEGF gene have been reported to influence protein translation efficiency and modulate gene expression. They have also been reported to associate with the clinical outcome of CRC [19–21]. However, although these studies have shown 90% of KRAS mutations occur in codon 12 or 13, there are still a lot of mutations that occur in the 3’-UTR. Moreover, no study has been reported on the association between the genetic variations of VEGF gene 3’-UTR and the risk of CRC in the Chinese population. In the present study, we analyzed the association of KRAS with VEGF gene 3’-UTR SNPs and genetic susceptibility to CRC among Chinese to determine the genetic effects of Chinese in relation to CRC.

Materials and methods

SNP selection and genotyping

The potentially functional SNPs of KRAS and VEGF gene in 3’-UTR were selected from NCBI dbSNP database (http://www.ncbi.nlm.nih.gov/projects/SNP/). Meanwhile, we selected SNPs with minor allele frequency (MAF) > 5% from Asian population in the HapMap database (http://hapmap.ncbi.nlm.nih.gov/). Eventually, one SNP in the KRAS 3’-UTR (rs1137188G > A) and four SNPs in the VEGF 3’-UTR (rs3025039C > T, rs3025040C > T, rs3025053G > A and rs10434A > G) were chosen and included in the analysis. All the five SNPs were genotyped by Shanghai Genesky Bio-Tech Co., Ltd. (http://biotech.geneskies.com/en/index.php/Index/fuwuer/id/28.html) using the improved multiplex ligase detection reaction (iMLDR) method, the primers are listed in Table 1. To verify the reproducibility, the assays were repeated for 5% of the samples at random as a quality control for genotyping, and the quality control samples confirmed 100% concordance.

Download:

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

Table 1. SNPs and PCR primer for KRAS and VEGF allele genotyping.

https://doi.org/10.1371/journal.pone.0174140.t001

Results

Clinical characteristics of the subjects

The demographic characteristics of the population were shown in Table 2. The total number of CRC cases and healthy controls included in the study was 617. Among the subjects were 371 patients (238 males and 133 females, mean ± SD age 57.3 ± 12 years, range 20 to 83) and 246 healthy controls (139 males and 107 females, mean ± SD age 55.9 ± 13 years, range 20 to 82). The range of BMI was 15.4–32.7 (mean: 23.5) and 15.8–36.1 (mean: 23.6) in the CRC cases and healthy controls, respectively. Because the controls were frequency matched with the cases, there were no significant difference in the distribution of age, gender and BMI between the CRC cases and healthy controls (P > 0.05). Most of the patients had healthy lifestyle, and non-smoking and non-drinking patients accounted for 59.3% and 56.3%, respectively. There were no statistically significant difference between the groups with respect to smoking status (P = 0.305) and alcohol consumption (P = 0.588). Patients with colon and rectum cancer accounted for 32.9% and 67.1% of CRC respectively, and 43.4% of the patients had lymph node metastasis. Most of the patients were later Dukes stages.

Download:

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

Table 2. Demographic data of CRC cancer cases and healthy controls.

https://doi.org/10.1371/journal.pone.0174140.t002

Association between KRAS and VEGF 3’-UTR SNPs and susceptibility to CRC

Genotype distribution of one KRAS SNP and three VEGF SNPs were summarized in Table 3. Genotyping call rate of the candidate SNPs were 98.91–100%. The genotype distribution among the healthy controls were in HWE (P = 0.203 for rs1137188, P = 0.275 for rs10434, P = 0.743 for rs3025039 and P = 0.474 for rs3025053). At the VEGF rs3025039 locus, significant difference was found in genotype distribution of CRC cases and healthy controls (χ² = 4.039, P = 0.043). When the rs3025039TT genotype was used as the reference, the C variant genotype was associated with an increased protection of CRC (AOR = 0.693, 95% CI = 0.485–0.989 for CC and CT+TT). Furthermore, the rs3025040 was in high LD with rs3025039 and had similar results, the CC variant genotype was also associated with an increased protection of CRC (AOR = 0.709, 95% CI = 0.524–0.963 for C/T and AOR = 0.693, 95% CI = 0.485–0.989 for CC and CT+TT) (data were not shown). However, these were no significant difference after Bonferroni correction. Other three SNPs (rs1137188GG, rs3025053AA and rs10434AA) were all not associated with CRC susceptibility when compared with their common genotypes.

Download:

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

Table 3. Statistical analysis of associations between the genotypes of SNPs and CRC risk.

https://doi.org/10.1371/journal.pone.0174140.t003

The association of KRAS and VEGF 3’-UTR SNPs and clinicopathological parameters of CRC

In genetic polymorphisms analysis, the association between KRAS and VEGF 3’-UTR SNPs and CRC risk was further analyzed by tumor site, tumor size, differentiation degree, invasive depth and TNM staging. As shown in Table 4, we found that the KRAS rs1137188 variant AA genotype was highly associated with tumor size (≥ 5 cm) (P = 0.01; Bonferroni-adjusted P = 0.04). The result revealed a significantly increased risk in CRC patients in the rs1137188 variant AA genotype when compared with that of the wild-type homozygous GG and heterozygous GA genotype. Meanwhile, the VEGF rs3025039 variant CC genotype was significant difference in tumor site (adjusted P = 0.024), tumor size (adjusted P = 0.019) and colorectal TNM staging (adjusted P = 0.032). The KRAS rs1137188 variant AA genotype was more evident in colon (adjusted P = 0.040). The VEGF rs3025053 variant GG genotype was more evident in colon (adjusted P = 0.030). However, these were no significant different after Bonferroni correction. At rs10434 locus, there were no difference in the risk estimates between the variant genotypes and CRC patients.

Download:

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

Table 4. The associations of SNPs polymorphisms and clinicopathological parameters of CRC.

https://doi.org/10.1371/journal.pone.0174140.t004

Discussion

We have investigated the important role of rs1137188 polymorphism of KRAS gene and rs3025039, rs3025040, rs3025053 and rs10434 polymorphisms of VEGF gene on the risk of the pathogensis, genetic sensitivity, and progression of CRC patients in a hospital-based case-control study. The study results had shown that CRC patients with KRAS rs1137188 AA genotype were more likely to have higher portion of tumor size (≥ 5 cm). The finding suggested that the rs1137188 may play an important role in the etiology of CRC. There were no significant association between the rest of polymorphisms and clinical outcome of CRC patients after Bonferroni correction.

The KRAS mutations, resulting in activation of the mitogen-activated protein kinase (MAPK) signaling pathway that mediates extracellular signals involved in cell proliferation, apoptosis and differentiation [22, 23]. The KRAS polymorphisms have been reported to influence the sensitivity and prognosis of various cancers of anti-targeted drug therapy, including lung cancer [24], gastric cancer [25], triple-negative breast cancer [26]. However, recent reports have mainly focused on the common genetic variations in the KRAS gene, published studies on the association between KRAS 3’-UTR SNPs and the susceptibility to various cancers are limited, Kim et al. [27] investigated KRAS rs9266 polymorphism was associated with risk of non-small cell lung and ovarian cancer. Kazmi et al [28] implicated that KRAS rs61764370 polymorphism was significantly associated with increased risk and prognosis of gallbladder cancer in North Indian population. Meanwhile, Dobre et al. [29] found that KRAS polymorphism was associated with statistically significant reduced survival of CRC. However, our study indicated that rs1137188 of KRAS gene in 3’-UTR was not associated with CRC susceptibility. Similar results have been observed in previous studies. Zhang et al. [30] investigated rs61764370 GT/GG polymorphism in KRAS 3’-UTR was not a genetic susceptible risk factor for CRC and could not be used as a biomarker for estimating cancer risk in Caucasian population. Langevin et al. [31] performed a review and meta-analysis to clarify the relationship between Let-7 microRNA-binding-site polymorphism in the 3’-UTR of KRAS and CRC outcome. Their results implicated that KRAS-LCS6 genotype was not associated with overall or progression-free survival (PFS) in the CRC patients. The reason of these results may be that the occurrence and development of CRC was related to a variety of factors, including gene-environmental interactions and ethnic ethics.

Another gene, VEGF, is a key player in the process of tumor angiogenesis and stimulates endothelial cell proliferation, survival and vascular maturation [32]. VEGF-related angiogenesis pathway is an important target of anti-cancer drug development [33]. There are a number of studies showing the influence of the VEGF SNPs on the risk and prognosis of various cancers, such as breast cancer [34] and non-small cell lung cancer [35]. However, few have reported the association of VEGF 3’-UTR SNPs with the susceptibility to CRC patients among Chinese. The association of polymorphisms in the VEGF pathway with CRC remains inconclusive. Rezaei et al. [36] found VEGF rs699947 polymorphism was correlated with development of breast cancer. Chen et al. [37] also reported that VEGF +936C/T gene polymorphism influenced the response to chemotherapy and overall survival of non-small cell lung cancer patients. Zhang et al. [38] found that VEGF rs3025039 was significantly associated with glioma susceptibility and might serve as genetic markers. Janardhan et al. [39] found that VEGF rs3025039 showed immense promise as a marker for disease aggression and recurrence and a factor for poor prognosis of epithelial ovarian cancer. Jeon et al. [40] carried out a case-control study to assess the relationship between VEGF 3’-UTR polymorphisms and CRC susceptibility in Koreans, and found that VEGF 1725G > A and 1451C > T may contribute to CRC susceptibility. In some solid tumors, expression of VEGF is significantly associated with angiogenesis and metastasis. The genetic polymorphisms in VEGF gene 3’-UTR can modify the potential binding sites of transcription factors, which lead to protein dysfunction and malformation. Lower VEGF expression can decrease tumor-related angiogenesis, growth and metastasis. However, our study pointed out that the SNPs in 3’-UTR of VEGF gene were not associated with CRC susceptibility in Han Chinese in Sichuan province after Bonferroni correction. Of course, the limited sample size and the differences of ethnicity could be the potential sources for inconsistent results.

There are several limitations in this study including the generalizability issue, since our study was restricted to the single Han Chinese patients. However, frequency of genetic polymorphisms often varies between ethnic groups. So further study is needed to clarify the generalizability of these findings to other ethnic populations. Moreover, the inadequate study design, such as a limited sample size, should be considered. The possible selection bias and information bias might have been present. The present study lacks information regarding lifestyle risk factors and clinical characteristics in the CRC patients, such as dietary intake, physical activity, aspirin use, PFS, 10-year PFS, overall survival and chemotherapy. These factors may also be associated with CRC risk. The mechanism by which polymorphisms in the KRAS gene and VEGF gene 3’-UTR affect development of CRC is still unclear. So, these results need to be confirmed further.

In summary, in this case-control study, our results indicated that KRAS rs1137188 AA may be a significantly increased risk for the Chinese CRC patients. However, the SNPs selected in our study were not adequate to support a role for CRC susceptibility, since after Bonferroni correction, the difference turns out to be insignificant. The study with a larger sample size are needed to draw firmer conclusions as to whether KRAS and VEGF SNPs in 3’-UTR are related to the risk of CRC. Further study of the relationship between KRAS and VEGF SNPs and survival analysis and the efficacy of chemotherapy-based drugs in CRC is needed. Additional functional studies will be done in order to better demonstrate the susceptibility and molecular mechanisms of CRC affected by KRAS and VEGF SNPs. However, our study identified clinical value of KRAS and VEGF 3’-UTR SNPs in Han Chinese in Sichuan province. It might provide a new measure for the individualized treatment of CRC patients in future.

Acknowledgments

We gratefully thank Dr. Mei Feng from Southwest Medical University for her assistance in improving this manuscript. We also thank Dr. Xing-Wang Sun in the Affiliated Hospital of Southwest Medical University (China) for her critical reading of this manuscript.