LIN28A gene polymorphisms confer Wilms tumour susceptibility: A four‐centre case‐control study

Abstract Wilms tumour is a renal malignancy that commonly occurs in children. LIN28A gene overexpression has been reported to be involved in various human malignancies, while its roles in Wilms tumour risk are still under investigation. Here, we genotyped four LIN28A polymorphisms in 355 Wilms tumour patients and 1070 healthy controls from four hospitals in China. The genotyped single nucleotide polymorphisms (SNPs) include the following: rs3811464 G>A, rs3811463 T>C, rs34787247 G>A and rs11247957 G>A. Overall, we found that rs3811463 T>C and rs34787247 G>A were associated with increased risk of Wilms tumour. Combination analysis of risk genotypes showed that, compared to non‐carriers, subjects with 1 risk genotype and 1‐3 risk genotypes were more likely to develop Wilms tumour, with an adjusted odds ratio (OR) of 1.58 and 1.56, respectively. Stratified analysis further demonstrated that the risk effect remained prominent in some subgroups. We also found that presence of 1‐3 risk genotypes was associated with Wilms tumour risk in subgroups > 18 months of age, females, males and those with clinical stage I + II diseases. Furthermore, expression quantitative trait locus (eQTL) analysis indicated that rs3811463 C allele was significantly associated with increased transcripts of LIN28A gene. These findings suggest that LIN28A gene polymorphisms may be associated with increased predisposition to Wilms tumour.

trols from four hospitals in China. The genotyped single nucleotide polymorphisms (SNPs) include the following: rs3811464 G>A, rs3811463 T>C, rs34787247 G>A and rs11247957 G>A. Overall, we found that rs3811463 T>C and rs34787247 G>A were associated with increased risk of Wilms tumour. Combination analysis of risk genotypes showed that, compared to non-carriers, subjects with 1 risk genotype and 1-3 risk genotypes were more likely to develop Wilms tumour, with an adjusted odds ratio (OR) of 1.58 and 1.56, respectively. Stratified analysis further demonstrated that the risk effect remained prominent in some subgroups. We also found that presence of 1-3 risk genotypes was associated with Wilms tumour risk in subgroups > 18 months of age, females, males and those with clinical stage I + II diseases. Furthermore, expression quantitative trait locus (eQTL) analysis indicated that rs3811463 C allele was significantly associated with increased transcripts of LIN28A gene. These findings suggest that LIN28A gene polymorphisms may be associated with increased predisposition to Wilms tumour.

K E Y W O R D S
case-control study, LIN28A, polymorphism, susceptibility, Wilms tumour

| INTRODUC TI ON
Wilms tumour is a common paediatric renal cancer worldwide, which mostly affects children. 1 The prevalence of Wilms tumour is nearly 1 out of 10 000 in North America. 2,3 Despite substantial advances in the managements of Wilms tumour, prognosis of relapsed patients remains poor even after re-treated with chemotherapy and surgery alone; a 4-year overall survival rate is less than 50% for this subgroup of patients. 4 To be note, previous study has been shown that VHL, PBRM1, SETD2 and BAP1 are most frequently mutated in regular kidney cancers. 5 However, because of the different genome mutation landscape from the adult counterparts, there is a critical need to identify more causal genetic variations in Wilms tumour.
Lin28 is an RNA-binding protein. It is implicated in cell growth, glucose metabolism and pluripotency, through regulating the biogenesis of miRNAs. In mammals, LIN28 gene encodes two RNA-binding paralogs, Lin28A and Lin28B. Lin28A/B inhibits let-7 microRNAs maturation and then promotes translation of certain target mRNAs of let-7. 6 The key mechanism by which Lin28A represses let-7 biogenesis is that cytoplasmic Lin28A targets precursor form of let-7 (pre-let-7) and then recruits TUTase 4 to induce oligo-uridylation of pre-let-7. Polyuridylation then facilitates pre-let-7 destabilization and ultimately decreases the level of mature let-7. 7 The molecular mechanism underlying the Lin28A-mediated carcinogenesis has been elucidated, to some extent. However, the contribution of LIN28A gene single nucleotide polymorphisms (SNPs) to Wilms tumour risk has not been investigated previously. Using clinical samples from a large multi-centre study, we attempted to determine the associations between LIN28A SNPs and Wilms tumour risk. We identified rs3811463 T>C and rs34787247 G>A polymorphisms in LIN28A significantly associated with Wilms tumour risk.
These data implicate polymorphisms in LIN28A gene as an important mechanism of Wilms tumour tumorigenesis.

| Polymorphism selection and genotyping
A total of four LIN28A SNPs with potential function were included in the analysis, following standard selection criteria. The locations of these SNPs in the LIN28A are as below: rs3811464 G>A in the upstream, rs3811463 T>C, rs34787247 G>A and rs11247957 G>A are all in the 3'-UTR. The details of these four SNPs are shown in Table S1 in Appendix S1. DNA was mainly extracted from study participants using a TIANamp Blood DNA Kit (TianGen Biotech Co Ltd).
Genotyping was carried out following the manufacturer's instructions by TaqMan methodology. We used water as negative controls to ensure genotyping accuracy. 10% random selected samples were re-genotyped with 100% concordance rates.

| Statistical analysis
The chi-square test was firstly applied to assess each SNP for deviation from the Hardy-Weinberg equilibrium (HWE) among the controls. The differences in the distribution of categorical variables between the cases and controls were analysed using the twosided chi-square test. We evaluated the association between the four selected SNPs and Wilms tumour risk by using multivariable logistic regression analysis. The association results are presented as odds ratios (ORs) with 95% confidence intervals (CIs), with adjustment for gender and age. We calculated the false-positive report probability (FPRP) to test for false-positive associations.
A probability of <.2 was considered a noteworthy association.
We also conducted the expression quantitative trait loci (eQTL) analysis using GTEx portal website (http://www.gtexp ortal.org/ home/), to predict the influence of SNPs on expression level of LIN28A. 10 Statistical significance was defined when P < .05. The SAS release 9.1 (SAS Institute) was adopted to conduct all statistical analyses. Table S2 in Appendix S1 gives information on the baseline characteristics of the included cases and controls. Similar distributions of age (P = .131) and gender (P = .182) were observed between cases and controls. As to clinical stages of cases, 119 (33.52%), 92 (25.92%), 79 (22.25%) and 47 (13.24%) patients were classified into clinical stage I, II, III and IV, respectively, according to the NWTS-5 criteria. 11 Clinical stage was unable to determine for a small subset of 18 (5.07%) patients.

| Risks associated with the LIN28A polymorphisms
The associations of all the four LIN28A SNPs (rs3811464 G>A, rs3811463 T>C, rs34787247 G>A and rs11247957 G>A) with Wilms tumour risk were listed in

| False-positive report probability results
We further calculated false-positive report probability (FPRP) values for the significant findings, with results shown in Table S3 in Appendix S1. At the prior probability level of .1, significance of the statistically significant findings for rs3811463 T>C (TC/ CC vs GG) and rs34787247 G>A disappears. As to the stratification analyses, association with rs3811463 T>C in clinical stages I + II patients and association with rs34787247 G>A in subjects older than 18 months old remain noteworthy. As for combined risk genotype analysis, we found that significant findings for the 1 vs 0 and 1-3 vs 0 genotypes remained noteworthy. FPRP analysis for stratification analyses of 1-3 vs 0 polymorphism also revealed a noteworthy result in subgroup of age > 18, males and stages I + II.

| Expression quantitative trait locus (eQTL) analysis
To further assess the putative functional relevance of LIN28A rs3811463 T>C, we evaluated the correlation between rs3811463 T>C polymorphism and LIN28A expression in transformed fibroblasts tissues. We found that the risk allele (C) of rs3811463 could enhance LIN28A expression significantly ( Figure S1 in Appendix S1). This study leads a vital part as a pioneer in exploring the association between LIN28A gene SNPs and Wilms tumour risk. Merit of this study is its origination on the first large-scale multi-centrebased analysis. The study also accompanies some minor limitations.

| D ISCUSS I ON
First, even though our sample size is considerable given the low incidence of Wilms tumour, the sample size is still not enough to provide a strong statistical power. Some negative results obtained here might need to be further validated, especially for stratification analysis. Second, only four SNPs were analysed in the current study.
There might be more SNPs in LIN28A gene associated with Wilms tumour risk. Third, all the included subjects were Chinese Han population although selected from four different hospitals. The findings might not be applicable to other populations. Last, we only took into consideration of the genetic factors because the information of environmental factors was not accessible. Thus, we are unable to determine whether the analysed LIN28A polymorphism is modified by certain environmental factors.
In all, we performed a multi-centre study to investigate the association between LIN28A polymorphisms and Wilms tumour risk in Chinese population. Our findings for the first time provided the insight into the potential role of LIN28A gene polymorphisms in Wilms tumour risk. Nevertheless, our conclusion based on genetic analysis is far from enough to fully elucidate the aetiology of Wilms tumour.
Comprehensive analysis, considering both genetic factors, environmental factors and genetic-environmental interactions, will help to reveal the pathogenesis of Wilms tumour.