Leptin rs2167270 G > A (G19A) polymorphism may decrease the risk of cancer: A case‐control study and meta‐analysis involving 19 989 subjects

Abstract Accumulating evidence has suggested that leptin (LEP) is very important for the development of cancer. Recently, a number of case‐control studies about the relationship of the rs2167270 G > A (G19A) variants in the LEP gene with the risk of cancer have yielded inconsistent results. In this study, we have carried out a case‐control study [1063 esophagogastric junction adenocarcinoma (EGJA) cases and 1677 controls] in a Chinese population. Furthermore, we carried out a pooled‐analysis of 13 studies involving 8059 cancer patients and 11 930 controls to assess whether the LEP G19A locus was associated with overall cancer susceptibility. Odds ratios (ORs) with the corresponding 95% confidence intervals (CIs) were harnessed to evaluate the potential association. In our case‐control study, we found an association between the carriers of LEP 19A allele and EGJA risk. In addition, the results of meta‐analysis also suggested significant associations with cancer risk (A vs G: OR = 0.92, 95% CI = 0.88–0.97, P = 0.001; AA vs GG: OR = 0.83, 95% CI = 0.74–0.93, P = 0.001, GA/AA vs GG: OR = 0.93, 95% CI = 0.88–0.99, P = 0.023 and AA vs GG/GA: OR = 0.83, 95% CI = 0.74–0.92, P < 0.001). Upon conducting a stratified analysis, we found that LEP 19A allele might decrease the susceptibility of non‐Hodgkin lymphoma (NHL) and colorectal cancer (CRC). In a stratified‐by‐ethnicity analysis, significant associations were also found in Asians, Caucasians, and mixed populations. We can conclude that the LEP G19A polymorphism constitutes a decreased risk of cancer.

and type 2 diabetes may contribute to an individual's cancer susceptibility. [5][6][7] Leptin (LEP), a hormone of energy expenditure, may contribute to control energy expenditure and balance by suppressing hunger. LEP, a 16 kDa glycol-protein, is predominantly made (>95%) by fat cells. 8 LEP interacts with LEP-receptor in the hypothalamus. A number of studies focused on the role of LEP in energy homeostasis and obesity. In addition, some investigations have demonstrated that LEP is associated with insulin signaling, inflammatory, and immune response. 9,10 Recently, several researchers reported that serum LEP levels might influence the development and progression of cancer. 11,12 It is found that the LEP rs2167270 G > A (G19A) locus is correlated with LEP levels and may also give a fascinating insight into the potential correlations with the development of cancer. 13,14 In a previous pooled study, it was found that individuals carrying a LEP 19A allele might have a lower tendency for cancer risk. 15 However, most of the eligible studies focused on Caucasians. The potential relationship of this singlenucleotide polymorphism (SNP) with cancer risk for Asians is unclear. Of late, several case-control studies investigating the association between LEP G19A polymorphism and cancer risk have been performed in Asians with relatively large samples. 16,17 Thus, it may be meaningful to obtain data from additional studies to get a more meaningful assessment of genetic effects.
In this study, to acquire an understanding of the relationship between LEP polymorphism and risk of cancer, we first studied LEP G19A polymorphism with the susceptibility of developing esophagogastric junction adenocarcinoma (EGJA). And then, we performed a meta-analysis to estimate the relationship of this polymorphism with overall cancer risk.  17 The disease stage was assessed according to AJCC criteria (version 7.0). The control group was composed of 1194 males and 483 females with the mean age of 63.91 ± 10.22 years. Information regarding smoking and drinking has been described in our previous study. 17,18 Each participant signed a written informed consent. This study was approved by the review boards of the Jiangsu University as well as the Fujian Medical University. The genomic DNA was carefully extracted from peripheral venous blood of participants by using DNA Kit (Promega, Madison, Wisconsin). The LEP G19A polymorphism was detected by SNPscan genotyping assay (Genesky Biotechnologies Inc., Shanghai, China) according to conditions described by Chen et al. 17

| Meta-analysis
We performed an extensive literature search in PubMed and EMBASE databases, covering all medical publications until 24 August 2018, with the following key words: LEP gene (e. g.: 'LEP' or 'leptin'), cancer (e.g. 'carcinoma,' 'cancer,' 'maligancy' or 'neoplasms'), and polymorphism (e.g.: 'polymorphism,' 'SNP' or 'variation'). In addition, we also carried out a manual search of the listed references of the included publications and related reviews.
The criteria of literature selection were as follows: (a) investigation designed as a case-control study; (b) focusing on the association of LEP G19A polymorphism with risk of cancer; (c) genotypes data listing in the publications. The major exclusion criteria of studies were as follows: (a) reviews; (b) duplicated reports; (c) not case-control study designs; (c) lack of data for genotype frequencies.
Two authors (J. Yang and Z. Zhong) extracted data from the included publications independently. The following information was collected: (a) first author; (b) publication year; (c) number of cases and controls; (d) country; (e) ethnicity; (f) source of controls; (g) cancer type; (h) genotyping method; and (I) genotype frequency. Ethnicities were defined as mixed, Asians, and Caucasians. For the source of controls, the publications were categorized as hospital-based and population-based studies.
In this study, we analyzed Hardy-Weinberg equilibrium (HWE) using a goodness-of-fit test using an online software (https://ihg.gsf.de/cgi-bin/hw/hwa1.pl). The strength of the correlation between LEP G19A locus and cancer risk was determined by calculating crude odds ratios (ORs) with their 95% confidence intervals (95% CIs). The following four models were calculated: homozygote comparison (AA vs GG), dominant model (AA/GA vs GG), recessive model (AA vs GG/GA), and allele model (A vs G). If I 2 > 50% or P < 0.1, it suggested that there was significant heterogeneity. Considering the heterogeneity among the included studies, a different model was used to pool the data. When no significant heterogeneity was identified, the Mantel-Haenszel method (fixed effects model) was used 19 ; otherwise, the Der Simonian and Laird method (random model) was utilized. 20,21 Sensitivity analysis was also carried out, which deletes an individual investigation and, in turn, recalculates the remainders. The source of heterogeneity among variables (e.g. cancer type, ethnicity) was explored by subgroup analysis. Begg's funnel plot and Egger's regression method were harnessed to examine the publication bias among the included studies. And P < 0.1 was defined as representative of significant bias. The Newcastle-Ottawa Quality Assessment Scale was used to assess the quality of the enrolled literatures. If scores ≥ 6 stars, the publication was considered as related high-quality. In this study, all P values for statistics were calculated with two-sided. STATA 12.0 software (Stata Corp, College Station, Texas) was used to analyze the data.

| Case-control study
A total of 2740 participants (involving 1063 EGJA patients and 1677 cancer-free controls) were included in this casecontrol study. Table 1 summarizes the primary information and our data for LEP G19A polymorphism. Table 2 shows the genotype distributions of LEP G19A polymorphism. In the analysis of LEP G19A polymorphism, differences in the distribution of LEP
When an analysis stratified by cancer type was conducted, we found that individuals carrying LEP

| 11001
We checked publication bias by using Begg's funnel plot and Egger's test. The statistical results showed that there was no significant bias in this meta-analysis (A vs G: Begg's test P = 1.00, Egger's test P = 0.825; AA vs GG: Begg's test P = 0.951, Egger's test P = 0.975; GA/AA vs GG: Begg's test P = 0.428, Egger's test P = 0.981; AA vs GA/GG: Begg's test P = 0.760, Egger's test P = 0.847; Figure 3). One-way sensitivity analysis was harnessed to confirm the stability of our findings. And we found that the corresponding results were not materially altered (Figure 4).
We assessed the quality score of the eligible studies by using the Newcastle-Ottawa Quality Assessment Scale. 32 The results are shown in Table 5. When the related lowquality studies (<6.0) were excluded, the findings were not substantially changed ( Table 6).

| DISCUSSION
In this case-control study, we found that LEP G19A polymorphism decreased the risk of EGJA. To the best of our knowledge, the first pooled-analysis that carried out an extensive evaluation of the G19A polymorphism in the LEP 5′-UTR region with the risk of overall cancer was conducted in 2014. 15 In our meta-analysis, 13 publications involving 8059 cases and 11 930 controls were included. Compared with the previous study, more new studies performed in Asian population were recruited. 16,17 Although some studies suggested that LEP G19A polymorphism could increase the risk of cancer, 17 the pooled ORs of our study confirmed that G19A polymorphism in the LEP gene was correlated with a decreased risk of overall cancer. It is worth noting that this potential association was also observed in Caucasians, Asians, mixed populations, and NHL and CRC subgroups.
In the past few decades, some case-control studies have been designed to explore the potential relationship between G19A polymorphism in the LEP gene and the risk of cancer. 16,17,[22][23][24][25][26][27][28][29][30][31] Skibola et al 24 found that LEP G19A polymorphism decreased the risk of NHL in Caucasians. Another study also identified similar findings regarding CRC in mixed populations. 30 A previous meta-analysis indicated that a tendency to decrease risk was noted between LEP G19A polymorphism and cancer. 15 However, for Asian population, only two casecontrol studies with small sample sizes were included in this pooled analysis. 22,23 The association of LEP G19A polymorphism with cancer risk in Asians was unclear.
Recently, several studies investigated the relationship between LEP G19A polymorphism and cancer risk in Asians. 16,17 And they found no association between this SNP and cancer risk. Recently, Zhang et al 17   observed results were more controversial. In the current study, we conducted a case-control study to identify the correlation between LEP G19A variants and the development of EGJA. We first found that LEP G19A polymorphism decreased the risk of EGJA in Asians. To estimate the relationship of LEP G19A polymorphism with cancer risk more extensively, we conducted an updated meta-analysis. It was found that LEP G19A polymorphism may have a lower risk of overall cancer. LEP G19A polymorphism, a SNP in the 5′-untranslated region, could not be translated into amino acid peptides. However, this SNP may influence the RNA translation, stability, and transcription, and then alter the expression of LEP protein. A recent study indicated that LEP 19A allele is correlated with lower levels of LEP. 14 A meta-analysis found that the decreased serum LEP levels were a protective factor to breast cancer risk. 11 It is conceivable that the reduced levels of serum LEP associated with LEP 19A allele may attenuate the risk of cancer. In this meta-analysis, we confirmed this phenomenon. Furthermore, we identified a significant association in Caucasians and Asians for the first time.
The results of the heterogeneity analysis are shown in Table 6. We found that there was no evident heterogeneity across studies. Publication bias was evaluated by Begg's funnel plots and Egger's linear regression test. The results showed that no significant bias was observed. In this metaanalysis, we assessed quality of the included studies. We found that the related low-quality studies did not influence the findings of overall evaluation. These findings indicated that our conclusions were credible and stable.
Some limitations of the present pooled-analysis should be acknowledged, even though it was powered by involving the latest literatures as well as our case-control study. First, when the data were extracted and pooled, it was found that significantly heterogeneities existed among certain subgroups. Thus, these observed results should be explained with caution in these subgroups. Second, for the lack of critical data (eg such as age, sex, BMI, and environmental factors), gene-environment interaction could not be carried out. Third, in this study, only LEP G19A polymorphism was studied; the interaction of gene-gene was not evaluated. Fourth, in this study, a functional study focusing on the LEP G19A polymorphism was not conducted. Finally, because the eligible studies were limited, our analysis may be underpowered in some subgroups.
In conclusion, it is highlighted that the G19A polymorphism in the LEP 5′-UTR region is associated with a decreased risk of EGJA. In addition, the subsequent meta-analysis also indicates that this SNP decreases the risk of overall cancer. To confirm or refute our findings, large scale case-control studies are needed.