Association of KRAS and NRAS gene polymorphisms with Wilms tumor risk: a four-center case-control study

Wilms tumor is a type of pediatric solid tumor that arises partly due to somatic and germline mutations. Single-nucleotide polymorphisms (SNPs) in the RAS gene reportedly modify the risk for several types of human malignancies. We conducted a multicenter study to investigate whether RAS gene variants predispose individuals to Wilms tumor. Four SNPs in RAS were genotyped in 355 Wilms tumor cases and 1070 controls. The SNPs included rs12587 G>T, rs7973450 A>G and rs7312175 G>A in KRAS, and rs2273267 A>T in NRAS. Individuals harboring the rs12587 GT genotype were more likely to develop Wilms tumor than those carrying the GG genotype (adjusted odds ratio [OR]=1.30, 95% confidence interval [CI]=1.004-1.68, P=0.046). However, the other three SNPs seemed not to influence the risk for Wilms tumor. Compared to individuals without a risk genotype, those harboring one to three KRAS risk genotypes had an adjusted OR of 1.28 for developing Wilms tumor (95% CI=1.002-1.64, P=0.048). Stratification analysis revealed that rs12587 GT/TT was associated with Wilms tumor risk in children >18 months old (adjusted OR=1.39, 95% CI=1.02-1.89, P=0.037). Our findings indicate that the rs12587 G>T polymorphism in KRAS is associated with increased Wilms tumor susceptibility.


INTRODUCTION
Wilms tumor (nephroblastoma) is the most common pediatric renal malignancy [1]. It is normally derived from embryonal kidney precursor cells in which cell growth and/or differentiation are dysregulated during development [2,3]. The incidence rate of Wilms tumor is about 1 in 10,000 children in Western countries [4].

Association of KRAS and NRAS gene polymorphisms with Wilms tumor risk: a four-center case-control study
AGING The overall five-year survival rate exceeds 90% in developed countries [5][6][7]. Despite great achievements in the treatment of Wilms tumor, the outcomes for patients with high-risk disease (about 25%) remain disappointing [8]. Apart from this, high treatment costs and severe chronic health conditions that occur in nearly 25% of survivors are also challenging [9,10].
There is strong evidence that genetic factors contribute to Wilms tumor risk. To date, five Wilms tumor susceptibility loci have been well characterized, including Wilms tumor gene 1 (WT1), Wilms tumor gene on the X chromosome (WTX), catenin beta 1 (CTNNB1), tumor protein 53 (TP53) and the imprinted 11p15 region [11][12][13]. Although additional genetic variants continue to be identified, the carcinogenesis of Wilms tumor remains to be fully explained [14][15][16]. Therefore, it is indispensable to identify other genes that increase Wilms tumor susceptibility.
The RAS oncogene family has three members: KRAS, NRAS and HRAS. These genes encode a family of highly homologous GTPases that are involved in various cellular activities, such as growth, proliferation and differentiation [17,18]. RAS mutations have been detected in about 20% of human malignancies [19]. KRAS mutations are the most common, accounting for approximately 85% of all RAS mutations [20,21], followed by NRAS mutations (15%). HRAS mutations are very rare, constituting less than 1% of all RAS mutations [22].
The impact of RAS gene variants on the risk of cancer has been widely investigated, including in studies of colorectal cancer [23], lung cancer [24,25], breast cancer [26] and melanoma [27]. Clark et al. demonstrated that coordinated activation of RAS and βcatenin accelerated the growth and metastatic progression of Wilms tumor in a murine model [28]. They later reported that activating KRAS mutations were found in human Wilms tumor samples [29]. Recently, another team verified the importance of RAS mutations in the development and progression of Wilms tumor [30].
Despite these findings, the link between RAS gene polymorphisms and Wilms tumor risk remains obscure. To clarify the association of RAS with Wilms tumor risk, we selected single-nucleotide polymorphisms (SNPs) in the two most common diseased-related RAS genes, KRAS and NRAS, for analysis in a four-center hospital-based case-control study.

False-positive report probability (FPRP) analysis
In FPRP analysis (Table 4), only at a prior probability level of 0.25 and an FPRP threshold of 0.2 did the increased Wilms tumor risk remain noteworthy in carriers of rs12587 GT (FPRP=0.141), children >18 months old with rs12587 GT/TT (FPRP=0.131) and those with one to three risk genotypes (FPRP=0.139).

DISCUSSION
Thus far, only a small portion of genetic loci have been found to increase the risk of Wilms tumor. This underscores the need to reveal more genetic loci that could predispose individuals to this disease. Herein, we evaluated the impact of KRAS and NRAS gene SNPs on AGING    cancer [32] and triple-negative breast cancer [33]. In a population-based case-control study conducted in the US by Christensen et al., the KRAS-LCS6 variant genotype (rs61764370) was not associated with the overall risk of head and neck squamous cell carcinoma, but was associated with a significantly reduced survival time [34].
Wang et al. [7] reported that the rs712 polymorphism in the KRAS 3' UTR was associated with a reduced risk for oral squamous cell carcinoma, while rs1137282 in KRAS exon 6 was not [35]. In contrast, in a study of 181 gastric cancer patients and 674 cancer-free controls, Li et al. found that the T allele of rs712 significantly enhanced the susceptibility to gastric cancer [36]. As different types of tissues and cells have different miRNA profiles, the effects of SNPs in specific 3' UTRs may vary accordingly. Moreover, differences in the population sources, environmental exposures, sample sizes and selection criteria of subjects may also have influenced the contribution of RAS SNPs to cancer susceptibility of different types. Therefore, it is necessary to define the impact of RAS polymorphisms on the risk of a certain cancer type in a certain population.
Our findings indicated that carriers of the KRAS rs12587 GT genotype had a genetic predisposition to Wilms tumor risk. Unexpectedly, rs7973450 A>G, rs7312175 G>A and rs2273267 A>T were not significantly associated with Wilms tumor risk. The rs12587 G>T, rs7973450 A>G and rs2273267 A>T polymorphisms reside in different complementary miRNA sites. The different locations of these SNPs may be one reason for their different effects on cancer risk. Other plausible interpretations of the null association include the relatively small sample size and the low-penetrance susceptibility of single polymorphisms.
One limitation of this study was the relatively small sample size, which may have impaired the statistical power, especially for the stratification analysis. Another limitation was the restriction of the included population to a single ethnicity (Chinese Han), which may render the findings inapplicable to other populations. Further, though we analyzed four SNPs in the current study, additional SNPs should be considered in future studies. Lastly, the current study only focused on genetic factors, and gene-environment interaction analysis was not performed due to the lack of relevant information. Wilms tumor is a heterogeneous disease, and both genetic and environmental factors contribute to its tumorigenesis. Thus, more comprehensive studies are warranted.
In conclusion, this was the first multi-center evaluation of the association of KRAS and NRAS gene SNPs with Wilms tumor susceptibility. Our study has provided the first evidence that KRAS gene SNPs may increase Wilms tumor susceptibility. Ongoing epidemiological studies in other independent populations are warranted prior to extrapolation of the current conclusions.

Study subjects
In total, 355 cases and 1070 healthy controls were included in this study (Supplemental Table 1). The subject selection criteria were described in detail in our previous study [37][38][39][40][41][42][43]. In brief, cases with newly diagnosed and histologically confirmed Wilms tumor were recruited from four centers in China (Guangzhou Women and Children's Medical Center [37][38][39][40][41][42][43], The First Affiliated Hospital of Zhengzhou University, The Second Affiliated Hospital and Yuying Children's Hospital of Wenzhou Medical University, and The Second Affiliated Hospital of Xi'an Jiao Tong University). All the included cases were sporadic cases. The controls were healthy volunteers without a history of Wilms tumor, matched to the cases by age, gender and city of residency. All the subjects or their guardians provided written informed consent before participating. Approval of the study protocol was obtained from the Institutional Review Board of each center prior to the study.

Polymorphism selection and genotyping
We analyzed three potential functional SNPs in the KRAS gene and one potential functional SNP in the NRAS gene. SNPs were selected from the NCBI dbSNP database (http://www.ncbi.nlm.nih.gov/projects/SNP) and SNPinfo (http://snpinfo.niehs.nih.gov/snpfunc.htm). These four SNPs could capture an additional 89 SNPs with R 2 ≥ 0.8 (Supplemental Table 2). The selection criteria were set as previously described [42,44]. Genomic DNA was extracted from venous blood with a TIANamp Blood DNA Kit (TianGen Biotech Co. Ltd., Beijing, China). SNP genotyping was performed with a TaqMan SNP Genotyping Assay (Applied Biosystems, Foster City, CA, USA). Negative controls with water and 10% replicates were also genotyped to ensure genotyping accuracy. No discordant genotypes were found in the replicates.

Statistical analysis
Statistical analysis was performed in SAS release 9.1 (SAS Institute, Cary, NC, USA). The genotype frequency distributions of the polymorphisms were first AGING evaluated among the controls, and Hardy-Weinberg equilibrium was assessed with the χ 2 test. The distribution of subject characteristics between cases and controls was examined with a two-sided χ 2 test. Unadjusted and adjusted (for age and gender) ORs and 95% CIs were generated for both single and combined SNPs. We then determined the association of the SNPs with Wilms tumor risk using the OR and 95% CI calculated from multivariable logistic regression analysis. FPRP analysis was performed as described previously [45]. All results were considered statistically significant if P<0.05.

CONFLICTS OF INTEREST
There are no competing interests to declare.