Introduction

Wilms tumor, also known as nephroblastoma, is the most common pediatric kidney cancer [1]. It accounts for over 90% of all the diagnosed kidney tumors in children [2]. The incidence rate of Wilms’ tumor varies geographically [3, 4]. The prevalence of Wilms tumor is about 7 cases per million children in the United States. Wilms tumor is also one of the most common renal tumors in children in China, with an incidence rate of ~ 3.3 per million. Wilms tumors are frequently diagnosed in young children with an average age of 2–3 years [5]. At present, long-term overall survival for the localized Wilms tumors exceeds 90% due to the improved risk stratification-adapted treatment [6]. However, nearly 20% of Wilms tumors are classified into high-risk subtype with frequent metastasis. Patients with high-risk tumors still subject to suboptimal outcomes [7,8,9]. Chronic health conditions secondary to intensified therapeutic regimens impact nearly 25% of Wilms tumor survivors [10].

The genetics of Wilms tumor tumorigenesis is complex, with multiple oncogenic drivers identified over the years. The currently known repertoire of oncogenic Wilms tumor driver alterations includes mutations in the WT1, CTNNB1, TP53, AMER1, as well as an abnormality of 11p15 methylation [11,12,13,14,15]. Apart from these, genetic association analyses in case-control studies also unveiled some Wilms tumor susceptibility loci [16,17,18,19]. Nevertheless, the well-established risk factors for Wilms tumor probably are only the tip of the iceberg. So far, all the known gene mutations can only explain less than 50% of Wilms tumor. Therefore, it is imperative to identify more causative variants to improve the understanding of the genetic susceptibility to Wilms tumor. In addition, detailed genetic information leads to new druggable targets, facilitating the development of more effective treatments for Wilms tumor.

N6-methyladenosine (m6A) is the most common internal chemical modification on eukaryotic mRNA [20]. m6A is mainly involved in the regulation of splicing, subcellular localization, translation, stability, and degradation of mRNA. m6A modulators are mainly classified into methyltransferase (writer), demethylase (eraser), and binding protein (reader). Methyltransferases include METTL3, METTL14, and WTAP, which mainly mediate m6A methylation of mRNA adenylate. Demethylases, consisting of FTO and ALKBH5, mainly remove m6A modification installed on RNA. Binding proteins include YTHDF1/2/3, YTHDC1/2, IGF2BP1/2/3, and eIF3, which are responsible for recognizing bases modified by m6A and regulating downstream pathways [21, 22]. The m6A modulator proteins play an important role in the occurrence and development of a variety of tumors [23,24,25]. However, research on the expression and function of m6A modulator genes in Wilms tumor has not yet been reported. The scarcity of investigation prompted us to contribute to our current report on associations between genetic variability of METTL14 and the risk of Wilms tumor. To this end, a total of five common SNPs in the METTL14 gene were genotyped and tested for their association with Wilms tumor susceptibility.

Methods

Sample selection

The study was carried out based on the principles of the Declaration of Helsinki. Approval of the study protocol was obtained from the institutional review board of Guangzhou Women and Children’s Medical Center (Ethics Approval No: 202016600). Eligible cases were all children newly diagnosed with a histologically confirmed Wilms tumor. Controls, recruited from the same hospital, were healthy volunteers of Chinese origin, without family history of Wilms tumor. Written informed consent was signed by all subjects’ guardians. All the subjects were enrolled from March 2001 to March 2018 and were genetically unrelated ethnic Han Chinese from China. A total of 414 cases diagnosed with Wilms tumor and 1199 hospital-based controls were included. They were enrolled from five hospitals (Guangzhou Women and Children’s Medical Center, The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University, The First Affiliated Hospital of Zhengzhou University, Second Affiliated Hospital of Xi’an Jiao Tong University, and Shanxi Provincial Children’s Hospital) in five different cities of China. Detailed information was previously reported [26, 27].

Polymorphism selection and genotyping

The selection of the five potentially functional METTL14 gene SNPs (rs1064034 T > A, rs298982 G > A, rs62328061 A > G, rs9884978 G > A, and rs4834698 T > C) was described in detail in our previous studies [28,29,30]. Genomic DNA from each sample was extracted from peripheral blood. Genotypes were determined using the TaqMan method. Replicate samples (10% of the samples) were picked out of all genotyping batches, and the concordance levels for blind duplicate samples were 100% for all SNPs assayed.

Statistical analysis

SNP genotypes were tested for consistency with Hardy-Weinberg equilibrium (HWE) within the control sample using a Goodness-of-fit χ2 test. Differences between cases and controls in the distribution of demographic and clinical variables were checked using a two-sided χ2 test. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) and two-sided P-values were calculated using unconditional logistic regression to estimate the relative risk associated with each genotype. Associations were further estimated in the groups stratified by age, gender, and clinical stages. Haplotype frequency distributions were deduced from observed genotypes using logistic regression analyses [31, 32]. False-positive report probability (FPRP) analysis was applied to assess noteworthy associations with detailed methods presented elsewhere [33, 34]. We performed expression quantitative trait loci (eQTL) and splicing quantitative trait loci (sQTLs) analyses through the Genotype-Tissue Expression (GTEx) project (http://www.gtexportal.org/) to evaluate the correlations between genotypes of candidate SNPs and genes expression as well as alternative splicing (AS) events of genes [35]. A probability value (P value) less than 0.05 was considered significant. All statistical analyses were performed using SAS version 9.1 software (SAS Institute, Inc., Cary, North Carolina).

Results

Effect of METTL14 gene SNPs on Wilms tumor risk

Clinical characteristics of the participants were depicted in our previous study (Table S1) [27]. Here, we successfully genotyped the five METTL14 gene SNPs (rs1064034 T > A, rs298982 G > A, rs62328061 A > G, rs9884978 G > A, and rs4834698 T > C) in 403 cases and 1198 controls, out of 414 cases and 1199 controls samples. The correlation between these SNPs and Wilms tumor risk is shown in Table 1. All these SNPs followed Hardy-Weinberg equilibrium (HWE) in controls (HWE P > 0.05). The rs1064034 variant alleles were remarkably associated with reduced risk of Wilms tumor (TA vs. TT: adjusted OR = 0.78, 95% CI = 0.61–0.99, P = 0.041; TA/AA vs. TT: adjusted OR = 0.83, 95% CI = 0.70–0.995, P = 0.044). Similar association was found for the rs298982 (GA/AA vs. GG: adjusted OR = 0.69, 95% CI = 0.53–0.91, P = 0.009). We then defined rs1064034 TA/AA, rs298982 GA/AA, rs62328061 AG/AA, rs9884978 GA/GG, and rs4834698 TT/TC as protective genotypes based on their ORs. Participants with 5 protective genotypes showed a 0.69-fold decrease in the risk of developing Wilms tumor when compared with those with 0–4 protective genotypes (95% CI = 0.52–0.91, P = 0.008).

Table 1 Association between METTL14 gene polymorphisms and Wilms tumor susceptibility
Full size table

Stratification analysis of significant SNPs

We analyzed the association between the METTL14 gene polymorphisms and susceptibility to Wilms tumor in subgroups separated by age, gender, and clinical stages (Table 2). Further stratification study revealed that the rs1064034 was associated with reduced Wilms tumor risk in groups with age > 18 months, female, and clinical stage IV diseases. Moreover, stronger protective effects was found for the GA/AA genotypes of rs298982 and combined five protective genotypes among children age > 18 months, females, clinical stage I + II tumors, and clinical stage III + IV tumors.

Table 2 Stratification analysis of protective genotypes with Wilms tumor susceptibility
Full size table

METTL14 haplotype analysis

We next evaluated whether the haplotypes of the five METTL14 gene SNPs are linked with Wilms tumor risk (Table 3). When compared to reference haplotype TGAAC, haplotypes AGAGT (P = 0.016), AAGGT (P = 0.010), and AAAGC (P = 0.002) were linked with significantly decreased Wilms tumor risk.

Table 3 The frequency of inferred haplotypes of METTL14 gene based on observed genotypes and their association with the risk of Wilms tumor
Full size table

False-positive report probability (FPRP) analysis

The obtained significant findings above were further assessed using false-positive report probability (FPRP) analysis (Table 4). At the prior probability of 0.1 and FPRP threshold value of 0.2, the associations between rs1064034 and Wilms tumor risk remained noteworthy in models TA/AA vs. TT and subgroup of children > 18 months in TA/AA vs. TT. Noteworthy results were also found for the GA vs. GG, GA/AA vs. GG, and subgroup of children > 18 months in GA/AA vs. GG. In addition, a significant decrease of Wilms tumor risk was detected in the carrier of 5 vs. 0–4 protective genotypes and subgroup of children > 18 months in 5 vs. 0–4 protective genotypes. Significant findings remained noteworthy in the haplotype TGGGC when compared to reference haplotype TGAAC.

Table 4 False-positive report probability analysis for significant findings
Full size table

Effect of SNPs on gene expression (eQTLs) and splicing (sQTLs)

We further used GTEx to analyze the expression quantitative trait loci (eQTLs) and splicing quantitative trait loci (sQTLs) of rs1064034 and rs298982. Interestingly, rs1064034 was significantly associated with mRNA expression of RP11-384 K6.6 in the whole blood (Fig. 1A) and cells-cultured fibroblasts (Fig. 1B), as well as SNHG8 in cells-cultured fibroblasts (Fig. 1C). We found that the rs1064034 could affect the splicing events of RP11-384 K6.6 (Fig. 1D) and SNHG8 (Fig. 1E) genes in cells-cultured fibroblasts. Similarly, rs298982 was significantly associated with mRNA expression of RP11-384 K6.6 in the whole blood (Fig. 2A) and cells-cultured fibroblasts (Fig. 2B), as well as SNHG8 in cells-cultured fibroblasts (Fig. 2C). SNP rs298982 could also affect the splicing events of RP11-384 K6.6 (Fig. 2D) and SNHG8 (Fig. 2E) genes in cells-cultured fibroblasts.

Fig. 1
figure 1

Functional relevance of rs1064034 on gene expression and splicing events in GTEx database. rs1064034 was significantly associated with RP11-384 K6.6 level in the A whole blood (P = 9.9*10−14) and B cells-cultured fibroblasts (P = 3.5*10−12) as well as CSNHG8 mRNA level in the cells-cultured fibroblasts (P = 1.8*10−5). rs1064034 can affect the splicing events of DRP11-384 K6.6 (P = 2.3*10−7) and ESNHG8 (P = 4.1*10−5) genes in cells-cultured fibroblasts

Full size image
Fig. 2
figure 2

Functional relevance of rs298982 on gene expression and splicing events in GTEx database. rs298982 was significantly associated with RP11-384 K6.6 level in the A whole blood (P = 3.9*10−9) and B cells-cultured fibroblasts (P = 9.4*10−9) as well as CSNHG8 mRNA level in the cells-cultured fibroblasts (P = 1.8*10−6). rs1064034 can affect the splicing events of DRP11-384 K6.6 (P = 8.7*10− 7) and ESNHG8 (P = 4.3*10− 6) genes in cells-cultured fibroblasts

Full size image

Discussion

This is the first genetic epidemiological study on the association of genetic variants in the METTL14 gene and Wilms tumor risk. We found that common variants in the METTL14 gene were significantly associated with susceptibility to this malignancy. This study may contribute to uncovering the underlying biology and genetics of Wilms tumor.

METTL14 is a key component of the m6A methyltransferase complex. METTL14 has different roles in different tumors and can be either a cancer promoter or suppressor. Chen et al. [36] identified METTL14 as a tumor suppressor in colorectal cancer. The low METTL14 was significantly associated with poor overall survival. Further functional experiments demonstrated that METTL14 inhibited the progression of colorectal cancer by regulating the production process of m6A-dependent precursor miR-375. Ma et al. [37] found that METTL14 was remarkedly downregulated in hepatocellular carcinoma. The reduced METTL14 expression was significantly associated with unfavorable recurrence-free survival and overall survival. The inhibitory role of METTL14 on hepatocellular carcinoma may be partly attributed to its facilitation of the primary miR-126 maturation in a m6A-dependent manner. METTL14 exerted an oncogenic role in acute myeloid leukemia via mRNA m6A modification [38]. Lang et al. [39] observed that METTL14 was an important driver in EBV-induced oncogenesis. They found that knockdown of METTL14 caused a decreased tumorigenic activity of EBV-transformed cells in the xenograft animal model systems. METTL14 could promote the growth and metastasis of pancreatic cancer by up regulating the m6A level of PERP mRNA [40].

Since the function and mechanism of m6A modification in mammals have not been studied for a long time, the effect of SNPs of m6A modification genes on genetic susceptibility to tumors has been hardly understood. Through adopting a two-stage case-control study, Meng et al. [41] conducted the first study to explore whether m6A gene SNPs could predispose to colorectal cancer in the Chinese population. All the five METTL14 gene SNPs (rs115267066, rs167246, rs2029399, rs298981, and rs441216) failed to show impacts on colorectal cancer risk. By enrolling 898 patients with neuroblastoma and 1734 controls, our group found that the METTL14 gene rs298982 G > A and rs62328061 A > G could significantly reduce the risk of neuroblastoma in children, while rs9884978 G > A and rs4834698 T > C could significantly increase the risk of neuroblastoma [28]. Regarding Wilms tumor, no studies investigating the role of METTL14 gene SNPs were available by far.

In the current study, rs1064034 and rs298982 variant alleles were found to protect from developing Wilms tumor. The combination of five protective genotypes led to a 0.69-fold decrease in the risk of developing Wilms tumor in comparison to 0–4 protective genotypes, indicating the stronger effect of the combined SNPs. It is believed that association studies based on haplotypes of multiple SNPs instead of individual SNP remarkedly strengthen the power for mapping and characterizing disease-causing genes [42, 43]. Thus, we examined whether haplotypes of METTL14 gene are associated with Wilms tumor risk. Expectedly, METTL14 gene haplotypes showed a significantly increased protection against Wilms tumor, indicating the synergistic effects of these SNPs. Genetic variation can modulate gene expression, thereby affecting phenotypes and susceptibility to complex diseases such as Wilms tumor. Here we harnessed the GTEx database to evaluate the effect of SNPs rs1064034 and rs298982 on expression and alternative splicing events of genes. We found that rs1064034 and rs298982 were significantly correlated with the expression and splicing of its nearby genes SNHG8 and RP11-384 K6.6. LncRNA SNHG8 acts as a vital role in tumorigenesis [44,45,46,47,48]. Thus, it is biologically possible that changes of the expression and splicing of SNHG8 and RP11-384 K6.6 caused by SNP rs1064034 and rs298982 may influence Wilms tumor risk (Fig. 3). Our results bring new insights into genetic mechanisms of how METTL14 affects Wilms tumor risk. Our findings identify METTL14 gene SNPs as risk markers in pediatric Wilms tumor. These findings not only show the relationship between some METTL14 gene SNPs and Wilms tumor risk but also can help to improve risk stratification strategies for Wilms tumor patients. In all, in-depth mechanism of how METTL14 SNPs affects Wilms tumor risk by regulating the gene expression and splicing pattern awaits to be elucidated. Potential limitations of our study include relatively small sample size, a lack of independent validation, and failure to incorporate other confounders. We also acknowledged that the conclusion obtained here was limited to Chinese. Cautions should be taken when interpreting this conclusion in other populations.

Fig. 3
figure 3

Possible mechanism of how SNPs rs1064034 and rs298982 confer to Wilms tumor risk

Full size image

Conclusion

In summary, we demonstrated the significant effects of METTL14 gene SNPs on the risk of Wilms tumor. However, further validation studies with larger sample size and involving different populations are required to strengthen this association.