Vascular endothelial growth factor gene polymorphisms and the risk of renal cell carcinoma: Evidence from eight case-control studies

Background Vascular endothelial growth factor (VEGF) protein plays important role in renal cell carcinoma (RCC) development and progression. VEGF gene polymorphisms can alter the protein concentrations and might be associated with renal cell carcinoma risk. However, the results of studies investigating the association between VEGF polymorphisms and renal cell carcinoma risk are inconsistent. Thus, a meta-analysis was performed. Methods We selected eligible studies via electronic searches. Only high-quality studies were included based on specific inclusion criteria and the Newcastle-Ottawa Scale (NOS). Results Eight studies primarily focusing on seven polymorphisms were included in our meta-analysis. Our results showed dramatically high risks for renal cell carcinoma were found regarding most genetic models and alleles of the +936C/T polymorphism (except CT vs. CC). In addition, significant increased renal cell carcinoma risks were found regarding all genetic models and alleles of the -2578C/A polymorphism. However, no significant associations were found between renal cell carcinoma risk and the +1612G/A, -460T/C, -634G/C, -405G/C or -1154G/A polymorphisms. Conclusions Our meta-analysis indicates that the +936C/T and -2578C/A polymorphisms of VEGF are associated with an increased risk for renal cell carcinoma. Additional rigorous analytical studies are needed to confirm our results.


INTRODUCTION
Approximately 337,860 cases of renal cell carcinoma (RCC) are diagnosed annually, and nearly 143,406 patients die from this cancer each year worldwide [1]. RCC is the third most common genitourinary malignancy. Moreover, both the incidence and mortality rates of RCC have steadily increased over the past several years [2]. The etiology of RCC is complex and multifactorial, and it involves multiple environmental and genetic factors [3,4]. Although an increasing number of studies have been performed on the etiology of RCC, the real causes of this cancer remain unclear. Previous studies have shown that many environmental factors such as cigarette smoking, alcohol drinking, occupational exposure to chemicals, hypertension and low frequencies of physical activity increase the risk of RCC [5][6][7]. Although many people are exposed to these risk factors during their lifetime, only a few people develop RCC. This finding suggests that genetic susceptibility plays a critical role in the etiology of this disease [8,9].
Vascular endothelial growth factor (VEGF) is an important pro-angiogenic growth factor, and it is one of the most potent endothelial cell mitogens [10,11]. VEGF Research Paper www.impactjournals.com/oncotarget plays a critical role in regulating the egress of the plasma proteins and cells that directly and indirectly stimulate angiogenesis [12]. Some research has indicated that the expression of VEGF affects tumor growth and metastasis, whereas the inhibition of VEGF signaling suppresses both tumor-induced angiogenesis and tumor growth [13]. The VEGF gene is located at chromosome 6p21.3 and consists of 8 exons. At least 30 single nucleotide polymorphisms (SNPs) exist in this gene [14] and some experimental studies have shown that certain SNPs can affect gene expression and change gene function [15].
Recently, numerous studies have been performed to evaluate the association between VEGF polymorphisms and RCC risk in diverse populations; however, the results of these studies conflict. To examine the association between VEGF polymorphisms and RCC risk, we performed a meta-analysis of all eligible published data up to June 5, 2016.

Study characteristics
We performed a literature search, and 286 potentially relevant publications were identified. After screening the title and abstract of each study, 277 studies were excluded because they did not involve both VEGF polymorphisms and RCC risk. After the subsequent data extraction, one study was excluded because it lacked controls [16].

-405G/C (rs2010963) and -1154G/A (rs1570360)
We also investigated the -405G/C (rs2010963) [18,19] and -1154G/A (rs1570360) [18,24] polymorphisms, both of which were examined in two studies including to the study-specific ORs and 95% CIs. The areas of the squares reflect the study-specific weights (which was the inverse of the variance). The diamonds represent the pooled ORs and 95% CIs. 262 and 373 cases and 477 and 516 controls, respectively. However, we did not identify any association between RCC risk and either the -405G/C (rs2010963) or -1154G/ A (rs1570360) polymorphism.

Sensitivity analyses
Hardy-Weinberg disequilibrium was observed in two studies (Shen et al. [20] and Lu et al. [21]). For +936C/T (rs3025039) polymorphism, our sensitivity analyses results indicated that exclusion of the aforementioned studies did not change the results for all the genetic models and allele (data not shown). In addition, for −2578C/A (rs699947) polymorphism, the sensitivity analyses results for all the genetic models and allele did not change either when excluding the study of Lu et al. [21] (data not shown). to the study-specific ORs and 95% CIs. The areas of the squares reflect the study-specific weights (which were the inverse of the variance). Diamonds represent the pooled ORs and 95% CIs.

Publication bias
Except for the -405G/C (rs2010963) and -1154G/A (rs1570360) polymorphisms, we used both funnel plots and Egger's test to assess the publication bias of each genetic model and allele. Our results did not show a publication bias for most of the genetic models and alleles (Supplementary Figure 1-2 showed the funnel plots of +936C/T and −2578C/A polymorphisms, respectively), except regarding CC vs. CT+TT of the -460T/C (rs833061) polymorphism (P=0.038).

DISCUSSION
VEGF, a growth factor that regulates angiogenesis and is involved in promoting endothelial cell proliferation [25]. VEGF protein likely plays an important role in the development and progression of cancer. Researchers have found that the expression of VEGF is significantly related to tumor stage, tumor size, and nuclear grade in patients with clear cell RCC [26]. In addition, the overexpression of VEGF has been detected in the vast majority of RCC tissues [27]. Currently, VEGF inhibition is a therapy for RCC [28]. However, the VEGF gene is highly polymorphic [29] and several functional SNPs in the VEGF gene alter the expression of the VEGF protein, thereby affecting tumor growth and progression. Recent studies have investigated the association between SNPs in the VEGF gene and the risk of RCC. However, these results are controversial. Thus, we conducted this meta-analysis to discuss the relationship between VEGF polymorphisms and RCC risk.
Zhang et al. [30] previously performed a metaanalysis that observed the association between VEGF polymorphisms and RCC risk. However, the author only reviewed 5 studies. In contrast, our meta-analysis included 8 relevant published studies. Moreover, our meta-analysis included many more cases and controls than the prior meta-analysis. In addition, we evaluated the quality of studies using the NOS. All of the included studies met high-quality standards, whereas the prior meta-analysis did not conduct any quality assessment. Thus, our meta-analysis is a more convincing and detailed evaluation compared with the prior study. Overall, we found that significant associations exist between VEGF polymorphisms and RCC risk (all of our results are summarized in Table 3 ). Specifically, most genetic models and alleles found high risks of RCC regarding the +936C/T (rs3025039) polymorphism. To the best of our knowledge, our study is the first meta-analysis to report that the +936C/T (rs3025039) polymorphism of VEGF can increase the risk of RCC. The +936C/T (rs3025039) polymorphism is located in the 3′-UTR and likely associated with obviously increased serum VEGF levels [31], which are related to tumor stage, tumor size, and nuclear grade. Interestingly, according to the results of Krippl P [32], the carriers of a +936 T allele had significant decreased risks of breast cancer and lower serum VEGF levels, which is opposite with our results. The reason of this discrepancy may be the tumor heterogeneity. Tumor heterogeneity is complex in many levels, including interdisease, intertumor, intratumor and tumormicroenvironment heterogeneity, etc. [33]. Furthermore, significant RCC risks were found in all genetic models and alleles of the -2578C/A (rs699947) polymorphism, whereas the prior meta-analysis only found increased RCC risks for the AA vs. CC genetic model and the A vs. C allele. Currently, several studies have reported that the -2578C/A (rs699947) polymorphism in the promoter region plays an influential role regarding plasma VEGF levels [34,35]. However, no significant associations were found between RCC risk and the +1612G/A (rs10434), -460T/C (rs833061), −634G/C (rs2010963), -405G/C (rs2010963) or -1154G/A (rs1570360) polymorphisms. All of the characteristics and results of the present study were compared with the former meta-analysis and summarized in Table 4 .
Certain limitations of this meta-analysis should be acknowledged. First, because our study only considered published articles, a publication bias might exist. However, the publication bias was only found for the CC vs. CT+TT of -460T/C (rs833061) polymorphism. The statistical results of the funnel plot and Egger's test support this finding. Second, the heterogeneities among certain genetic models and alleles were significant. The reasons underlying these heterogeneities included the source of the controls, the study design and differences in genetic backgrounds. Third, the control sample of two articles were in Hardy-Weinberg disequilibrium, however, all the results of +936C/T (rs3025039) and -2578C/A (rs699947) polymorphisms did not change significantly after sensitivity analyses. Fourth, as the most of the cases of +936C/T and -2578C/A polymorphisms were from Asians, so our results of these two SNPs may not represent Caucasians. Finally, because of the use of unadjusted data, potential confounds such as age, sex and residence might also have affected the effect estimates. Thus, a more precise and large scale evaluation based on adjusted data is needed.
In summary, our meta-analysis suggests that the +936C/T (rs3025039) and -2578C/A (rs699947) polymorphisms of VEGF are associated with increased risks for RCC. However, no significant RCC risks were obtained regarding the +1612G/A (rs10434), -460T/C (rs833061), -634G/C (rs2010963), -405G/C (rs2010963) or -1154G/A (rs1570360) polymorphisms. To the best of our knowledge, this meta-analysis is the first to report that the +936C/T (rs3025039) polymorphism can increase the risk of RCC. Larger and more rigorous analytical studies are required to confirm our results and evaluate the gene-environment interactions with regard to RCC risk.

Search strategy and selection criteria
According to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), we performed an electronic systematic search of PubMed, the Cochrane Library database, EMBASE, Google Scholar and the China National Knowledge Infrastructure (CNKI) without any restriction on language up to June 5, 2016. The combinations of keywords used were "renal cancer" or "renal carcinoma"; "polymorphism" or "variant"; and "vascular endothelial growth factor" or "VEGF." In addition, the reference lists of the papers retrieved and recent reviews were also examined. We included all studies that (1) evaluated the association between VEGF polymorphisms and the risk of RCC in humans; (2) used a case control design; (3) confirmed RCC using the accepted diagnostic criteria; (4) reported sufficient published data, including ORs and their 95% CIs, or the number of events for the purposes of calculation. The exclusion criteria were (1) a lack of sufficient data to calculate ORs with corresponding 95% CIs; and (2) overlapping cases or controls. Only the most recent or the largest research study was included in the case of overlap.

Data extraction
Two investigators (GMC and DWJ) extracted the raw data independently based on the inclusion and exclusion criteria. The following information was extracted from all of the enrolled studies (see Table  1): the surname of the first author, date of publication, participant ethnicity, quality scores, sources of controls, number of cases and controls and the HWE P-value. All disagreements were resolved via discussion.

Quality assessment
Two authors (GMC and SZR) assessed the study quality using the NOS [36] which evaluates methodological quality using a star rating system. Nine stars was defined as a full score; 5 to 9 stars was considered as being of high methodological quality; and 0 to 4 stars was considered as being of poor quality [37]. The quality of all the included studies is listed in Table 2. For conflicting NOS scores, an agreement was reached via a comprehensive reassessment, and only high-quality studies were included in our meta-analysis.

Statistical analysis
The relationship between VEGF polymorphisms and the risk of RCC was evaluated via pooled ORs with 95% CIs. The significance of the pooled ORs was tested using the Z-test, and a (two-tailed) P-value of <0.05 was regarded as significant. The HWE was calculated in the control groups using the chi-square test, and P<0.05 signified a departure from HWE. Betweenstudy heterogeneity was calculated using the I 2 test. If the heterogeneity was significant (I 2 >50%) [38], then a random-effects model was used (the DerSimonian and Laird method) [39]; otherwise, the fixed-effect model (the Mantel-Haenszel method) [40] was applied. To assess the stability of the results, sensitivity analyses were conducted to evaluate the impact of the studies, especially which not in HWE. Because publication bias is always a concern for meta-analyses, funnel plots and Egger's test were both used to examine publication bias (P<0.05 was considered as significant publication bias) [41]. All statistical analyses were performed using STATA statistical software (Version 12.0; Stata Corporation, College Station, TX, USA).