Cumulative evidence for relationships between multiple variants of HNF1B and the risk of prostate and endometrial cancers

Background To provide a synopsis of the current understanding of the association between variants of HNF1B and cancer susceptibility, we conducted a comprehensive research synopsis and meta-analysis to evaluate associations between HNF1B variants and prostate and endometrial cancers. Results Eighteen studies totaling 34,937 patients and 55,969 controls were eligible for this meta-analysis. Four variants showed a significant association with the risk of individual cancer. Strong significant associations were found between rs4430796 A and the risk of both prostate cancer (OR = 1.247, p = 2.21 × 10− 77) and endometrial cancer (OR = 1.217, p = 8.98 × 10− 16); the AA, AG genotypes also showed strong significant associations with the risk of prostate cancer (OR1 = 1.517, p = 4.46 × 10− 22; OR2 = 1.180, p = 0.002). There was a strong significant association between rs7501939 G and the risk of prostate cancer (OR = 1.201, p = 9.31 × 10− 31). Strong significant association was found between rs11649743 G (OR = 1.138, p = 1.08 × 10− 12), rs3760511 C (OR = 1.214, p = 1.57 × 10− 19) and the prostate cancer risk;the GG, AG genotypes of rs11649743 also showed strong significant associations with the risk of prostate cancer (OR1 = 1.496, p = 3.32 × 10− 6; OR2 = 1.276, p = 7.82 × 10− 6). All the cumulative epidemiological evidence of associations was graded as strong. Conclusions Our study summarizes the evidence and helps to reveal that common variants of HNF1B are associated with risk of prostate and endometrial cancer.


Background
Human cancers result in Considerable morbidity and mortality. Family history, ethnicity, lifestyle and region are potential risk factors for cancer development [1][2][3][4]. However, family-based and adoption studies have provided major evidence for the role of genes in the development of cancers [5][6][7].
Owing to advances in sequencing technologies and genome-wide association studies (GWAS), a large number of genetic variants correlated with various cancers have been identified [8,9]. Multiple studies have examined the relationship between the hepatocyte nuclear factor-1 beta (HNF1B, formerly known as TCF2) locus (on chromosome 17q12) and cancer risk [10][11][12][13]. HNF1B is a member of the homeodomain-containing superfamily of transcription factors and is involved in the tissue-specific regulation of many genes expressed in various organs [14] and during embryonic development [15]. Patients with a heterozygous HNF1B deletion exhibit renal disease, elevated liver enzymes, and diabetes [16]. HNF1B is strongly associated with the risks of many cancers, including prostate cancer [10,17],ovarian cancer [18][19][20],endometrial cancer [12,21,22] and lung cancer [13]. Recently, it has been reported that the rs7501939 single-nucleotide polymorphism (SNP) in HNF1B confers a poor overall survival in patients with multiple myeloma [23]. However, fine-mapping studies have revealed a complex genetic architecture of the HNF1B locus, demonstrating that variants of HNF1B and the direction of their effects differ between cancer types. SNPs rs4430796 and rs7501939, are both associated with the prostate cancer risk across many ethnic groups [24]. The same SNPs, are also associated with endometrial cancer risk in women of European background [12]. Yet, the SNP rs757210, in high linkage disequilibrium with rs4430796, is the most strongly associated with serous epithelial ovarian cancer [18].
Here, we collected data related to the associations between HNF1B variants and cancer phenotypes, and performed a comprehensive meta-analysis, involving a total of 34,937 patients and 55,969 controls, to derive more precise estimates of the associations between HNF1B variants and susceptibility to prostate and endometrial cancers.

Search strategy and inclusion criteria
The US National Library of Medicine's PubMed, Embase, OMIM, ISI Web of Science, and Chinese National Knowledge Infrastructure (CNKI) databases were searched in a systematic manner to retrieve all genetic association studies of HNF1B variants and cancers published before July 2017. The search strategy was based on a combination of the terms (Hepatocyte nuclear factor-1 beta or HNF1B) and (cancers or tumors). The references of all computer-identified publications were searched for additional studies, and the PubMed option "Related Articles" was also used to search for potentially relevant papers. Searches were performed by two independent reviewers (Yu Tong and Yibin Wang). The language of the publications did not influence article selections.
Studies were included if they met the following criteria. (1) the study reported original data from case-control or cohort studies, (2) the study reported alleles and genotypes for HNF1B variants, and (3) the numbers of subjects possessing each allele and genotype in the cancer and control groups were available. No restrictions were set for the source of controls (general population, clinic, or hospital). Studies were excluded when: (i) they lacked sufficient information; (ii) they were published as letters to editors or conference abstracts; (iii) they were studies about cancer mortality.

Data extraction
Data were extracted independently by two investigators (Yu Tong and Yibin Wang), who used recommended guidelines for reporting on meta-analyses of observational studies. The following data were extracted from the eligible studies: authors, journal title, year of publication, country of origin, selection and characteristics of cases and controls, demographic data, ethnicity of the study population, numbers of eligible and genotyped cases and controls, and genotype distributions in cases, controls, and available subgroups. Furthermore, we examined whether genotype frequencies in control groups conformed to the Hardy-Weinberg equilibrium (HWE) was determined. Any disagreement was adjudicated by a third author (Yi Qu).

Statistical analysis
The odds ratio was used as the metric of choice for each study. To detect overall genetic associations, allele frequencies were computed for studies reporting allele and genotype data. Pooled odds ratios were computed by the fixed effects model and the random effects model based on heterogeneity estimates. Once an overall gene effect was confirmed, the genetic effects and mode of inheritance were estimated using the genetic model-free approach suggested by Minelli et al. We performed Cochran's Q test and calculated І 2 statistic to evaluate heterogeneity between studies. Harbord's test was performed to evaluate publication bias. Potential small-study bias was evaluated by Egger's test [25]. Sensitivity analyses were conducted to examine if the significant association would be lost when the first published report was excluded, or studies deviated from HWE in controls were excluded. All analyses were conducted using Stata, version 14.0 (StataCorp, 2017), with the metan, metabias, metacum, and metareg commands.
Venice criteria [26] were applied to evaluate the epidemiological credibility of significant associations identified by meta-analysis. Credibility was defined in three categories: amount of evidence (graded by the sum of test alleles or genotypes among cases and controls: A for > 1000, B for 100-1000, and C for < 100), replication of the association (graded by the heterogeneity statistic: A for I 2 < 25%, B for I 2 between 25 and 50%, and C for I 2 > 50%), and protection from bias (graded as A: there was no observable bias, and bias was unlikely to explain the presence of the association, B: bias could be present, C: bias was evident or was likely to explain the presence of the association, association. C was also assigned to an association with a summary OR less than 1.15, unless the association had been replicated by GWAS or GWAS meta-analysis from collaborative studies. With no evidence of publication bias). Cumulative epidemiological evidence for significant associations was thought to be strong if all three grades were A, moderate if all three grades were A or B, and weak if any grade was C.
To determine whether a significant association could be excluded as a false positive finding, FPRP (false positive report probability) was calculated using the method described by Wacholder et al. [27]. FPRP < 0.05, 0.05 ≤ FPRP ≤0.20, and FPRP > 0.20 were considered strong, moderate, and weak evidence of true association, respectively.

Eligible studies
Our initial database search identified 113 potentially relevant studies. Based on a review of titles and abstracts, 55 articles were retained. The full text of these 55 articles was reviewed in detail, and 18 studies containing 36 datasets were eligible for inclusion in the meta-analysis. The specific process for identifying eligible studies and inclusion and exclusion criteria are summarized in Fig. 1a.
rs7501939 A > G and the risk of prostate cancer Six publications were included in the evaluation of the association between the HNF1B rs7501939 and prostate cancer ( Table 1) rs11649743 A > G and the risk of prostate cancer Two publications included data regarding the association between the HNF1B rs11649743 and prostate cancer ( Table 1). There was a significant difference in the between-study heterogeneity among the eligible studies (Q = 15.1, p = 0.035, I 2 = 53.6%). Strong significant association was observed with the prostate cancer risk (p = 1.08 × 10 − 12 , random effect OR = 1.138, 95% CI: 1.062,

rs7501939 G > A and the risk of endometrial cancer
Two publications were included in the analysis of the association between the HNF1B rs7501939 and endometrial cancer (

Cumulative evidence of association Epidemiological credibility of significant associations
Venice criteria were applied to evaluate these significant associations. Details of protection from bias for genetic variants significantly associated with prostate and endometrial cancer risk in meta-analyses are shown in Table 4. Grades of A were given to all these meta-analyses for amount of evidence, replication of association, and protection from bias. Therefore, strong evidence of true association with cancer risk is assigned to rs4430796, rs7501939, rs11649743, and rs3760511 for prostate cancer and rs4430796 for endometrial cancer.

Probability of true association with cancer risk
To evaluate the probability of true association with cancer risk for the nominally significant variants, FPRP value was calculated. All associations with cancer risk had a FPRP value < 0.001. Thus, all the cumulative epidemiological evidence of associations was graded as strong.

Discussion
To our knowledge, this is the first general overview of the association between HNF1B variants and susceptibility to prostate and endometrial cancers. Our primary analysis revealed that, rs4430796 A, showed strong   significant associations with risk of both prostate cancer (OR = 1.247, p = 2.21 × 10 − 77 ,) and endometrial cancer (OR = 1.217, p = 8.98 × 10 − 16 ); the AA, AG genotypes also showed strong significant associations with risk of prostate cancer (OR1 = 1.517, p = 4.46 × 10 − 22 ; OR2 = 1.180, p = 0.002). Sensitivity analyses in Caucasians demonstrated patterns similar to that of the full population. However, lack of significant association was found in Africans, which is likely due to the considerably smaller sample size. There was a strong significant association between rs7501939 A and the risk of prostate cancer (OR = 1.201, p = 9.31 × 10 − 31 ); however, lack of significant association with endometrial cancer risk was observed (OR = 1.104, p = 0.751. For rs11649743 G, strong significant association was found with the prostate cancer risk (OR = 1.138, p = 1.08 × 10 − 12 ), and the GG, AG genotypes also showed strong significant associations with the risk of prostate cancer (OR1 = 1.496, p = 3.32 × 10 − 6 ; OR2 = 1.276, p = 7.82 × 10 − 6 ). Strong significant association was also found between rs3760511 C and the risk of prostate cancer (OR = 1.214, p = 1.57 × 10 − 19 ). Using the Venice criteria and false-positive report probability tests, we graded all the cumulative evidence of significant associations with prostate and endometrial cancers risk as strong.
Our findings were based on several gene-association studies, including several thousand participants, and were robust in terms of study design and sensitivity analyses. We found no evidence of publication bias or small study bias based on funnel plots. Between-study heterogeneity was found in allelic association studies (G versus A) of rs7501939, and in allelic (G versus A) of rs11649743 for prostate cancer. When HWE was examined, one study showed deviation. Our results were robust to the removal of this study.
HNF1B encodes three isoforms: isoforms (A, B and C); isoform A and B act as transcriptional activators and isoform C acts as a transcriptional repressor [44]. HNF1B is involved in the regulation of cell proliferation, and genetic variation in HNF1B might modulate the risk of cancer [45]. However, the precise pathomechanism by which the genetic variation affects susceptibility to cancers is still unclear. In a recent GWAS, rs4430796 and rs7501939 in HNF1B were associated with the risks of both endometrial cancer in women of European background [43] and prostate cancer [28] . Several studies examined the associations between HNF1B and prostate cancer and endometrial cancer across various populations [12,46,47]. According to these studies, the two variants are associated with the risks of prostate cancer and endometrial cancer. Moreover, the rs4430796 G allele is significantly associated with an increased risk of lung cancer [13] . In 2013, Pharoah et al. identified that the HNF1B rs757210 is specific to serous epithelial ovarian cancer by pooling data from GWAS and follow-up genotyping; the analysis included 43 studies from the Ovarian Cancer Association Consortium [18]. At the same time, Shen et al. found evidence for a differential effect of HNF1B on the serious and clear cell subtypes of ovarian cancer. They found that HNF1B loss-of-function role and gain-of-function are related to serous and clear cell ovarian cancers, respectively [20]. Another research discovered HNF1B rs7501939 was a susceptibility locus for testicular germ cell tumor [48]. Taken together, these studies suggest that specific HNF1B variants predispose individuals to clear cell ovarian, endometrial, lung and prostate cancers, et al.
There are several limitations of the study. First, it is likely that some publications were overlooked although we conducted an exhaustive literature search, some relevant published studies with null results were not identified. Second, due to insufficient data, we were unable to evaluate publication bias for associations between several variants in 8q24 region and prostate and endometrial cancer. Third, a unified analysis standard across studies could not be defined for lack of raw data from the original publications. Therefore, future studies with larger sample size are warranted to confirm these associations.

Conclusions
Given the relevance of HNF1B variants to cancer biology, we attempted to estimate the strength of the genetic associations between these variants and prostate and endometrial cancers. This Human Genome Epidemiology (HuGE) systematic review presents strong evidence for an association between HNF1B variants and prostate and endometrial cancers, both overall and in Caucasians, Asians, Africans, and Indians, suggesting a multiplicative genetic model for variants of HNF1B among different ethnic populations. Our study results also suggest that HNF1B plays an important role in prostate and endometrial cancers, and these variations may serve as efficient and economical biomarkers for the diagnosis of prostate and endometrial cancers.