Leptin rs7799039 (G2548A) polymorphism is associated with cancer risk: a meta-analysis involving 25,799 subjects

Introduction

Cancer is one of the major public health burden with over 18.1 million new cancer cases and 9.6 million cancer deaths in 2018.¹ According to assessments of World Health Organization in 2015, cancer is among the leading cause of death in most countries. The reasons may be very complex. Aging and growth of the population, as well as risk factors for cancer, might influence the development of cancer. Recently, accumulating evidence indicates that there is a connection between diabetes and obesity with cancer.² Thus, any variation in diabetes and obesity-related genes may influence the risk of cancer.

LEP, a 16-kDa adipocyte-derived peptide hormone, is a mediator of obesity and homeostasis. LEP interacts with the LEP receptor and its function is mediated through this receptor. Previous studies have demonstrated that the LEP signal may be transmitted through several signaling pathways (eg JAK/STAT, MAPK, PI3K, Wnt/β-catenin, and ERK).^3,4 It is also reported that LEP may affect angiogenesis, inflammation, thrombosis, and tumor growth, invasion, and metastasis.^4–13 Hardwick et al reported that LEP was very important for phosphorylation of the p42/44 mitogen-activated protein kinase and for enhancing proliferation of colonic epithelial cells.¹⁴ It is well known that single-nucleotide polymorphisms (SNPs) in genes may be implicated in the pathogenesis of a number of cancers and can be used as an indicator of early screening, diagnostics, and prevention measures.¹⁵ The human LEP gene maps to chromosome 7 (location: 128241278–128257629, NCBI Build 38). The LEP gene is polymorphic. And LEP SNPs may influence the risk of cancer.¹⁶ The rs7799039 G>A (G2548A) polymorphism in the LEP gene is the most widely studied for its relationship between this locus and the risk of human diseases. Terrasi et al suggested that the occurrence of LEP rs7799039 G>A variants could promote LEP protein expression in breast cancer cells through a Sp1- and nucleolin-dependent pathway, resulting in the LEP overexpression in tumor tissue.¹⁷

Recently, many molecular epidemiological studies have been carried out to identify the relationship between the LEP rs7799039 G>A polymorphism and cancer risk, but the findings have been conflicting. Three meta-analyses have been performed to explore the relationship between this SNP and cancer risk.^18–20 Results of these studies indicated that individuals carrying the LEP rs7799039 A allele might have an increased susceptibility of overall cancer. However, only a case–control study focusing on the association between the LEP rs7799039 G>A polymorphism and the risk of gastric cancer was included.²¹ The relationship of this polymorphism with cancer risk in Asians is unclear. Recently, several case–control studies conducted in Asians were carried out to explore the association between the LEP rs7799039 G>A polymorphism and cancer risk. To obtain a more precise assessment of the correlation of LEP rs7799039 G>A polymorphism with the risk of cancer, we performed an updated meta-analysis of all eligible studies focusing on the relationship of the rs7799039 G>A polymorphism to the susceptibility of developing cancer.

Materials and methods

Search strategy

In this meta-analysis, we carried out an electronic literature retrieval in PubMed, Embase, and the China Biology Medicine databases up to August 2018 using the following search strategies: (‘LEP” or “leptin”) and (“carcinoma” or “cancer” or ‘malignancy ‘ or “neoplasms”) and (“polymorphism” or “SNP” or “variation”). There was no restriction on language. The references in included studies and reviewers were carefully checked for other potential data. When a publication involved some subgroups, it was treated separately. This study was reported based on the Preferred Reporting Items for Meta-analyses (PRISMA) guideline (Table S1: PRISMA checklist).²²

Selection and exclusion criteria

The major selection criteria were as follows: (1) designed as case–control study that assessed the relationship between LEP rs7799039 G>A variants and cancer risk; (2) presented sufficient data (eg genotype number or other available data) to calculate the pooled-estimating; and (3) genotype distribution in controls did not violate Hardy–Weinberg equilibrium (HWE).

The exclusion criteria were defined as follows: (1) the publication was not designed as a case–control study; (2) the genotype data was not presented or could not be calculated; (3) genotype distribution in controls violated HWE; and (4) review articles and letters.

Data extraction

Two authors (W. Tang and C. Liu) independently extracted the information from each eligible study. If the extracted information was different, they would review the publication again and reached consensus. If they could not get a consistent assessment, another author (H. Qiu) would be invited to resolve the dispute and a final decision was made. The following data were extracted from each study: the surname of the first author, year of publication, country, ethnicity, numbers of participants, source of control, genotype frequencies, and genotyping method.

Statistical analysis

The strength of the correlation of the LEP rs7799039 G>A polymorphism with cancer susceptibility was determined by crude ORs with 95% CIs. The relationship between LEP rs7799039 G>A and cancer risk was evaluated using allele model (A vs G), homozygote model (AA vs GG), recessive model (AA vs GG/GA), and dominant model (AA/GA vs GG). We used the Q and I² test to check the heterogeneity among the included studies. A P>0.1 and I²<50% indicated that there was low heterogeneity, and then the Mantel–Haenszel method (fixed-effects model) was used to calculate the ORs and CIs;²³ otherwise, the DerSimonian and Laird method (random-effects model) was used to assess the association.^24,25 The sources of heterogeneity were analyzed by subgroup analyses. Sensitivity analysis was analyzed by omitting an individual study in turn and re-calculating the ORs and CIs. Publication bias was checked by using Bgger’s and Egger’s test. An internet chi-square test was used to determine whether the distribution of the genotypes in controls conformed to HWE (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). All data were calculated and analyzed by using Stata 12.0 software (Stata Corp., College Station, TX). A P<0.05 (two-sided) was considered as statistical significance.

Results

Study characteristics

Based on the selection criteria, 30 publications focusing on the association of the LEP rs7799039 G>A polymorphism with cancer risk were included.^21,26–53 One publication contained two independent case–control studies that we treated as two investigations.⁴⁹ The detail selecting process is shown in Figure 1. A total of 31 case–control studies involving 25,799 subjects were included in this meta-analysis. Among them, 19 were conducted in Caucasians,^26–44 eight performed in Asians,^{21,45–50,54} and four were in mixed populations.^51–53 Nine were population-based,^{27,28,30,33,37,38,43,44,50} and 22 case–control were hospital-based studies.^{21,26,29,31,32,34–36,39–42,45–49,51–54} Of all the eligible studies, 11 focused on breast cancer,^{33,37–40,44,48,49,51,53} four focused on colorectal cancer (CRC),^31,32,41,52 three focused on prostate cancer (PC),^26,30,34 and 13 focused on other cancers.^{21,27–29,35,36,42,43,45–47,50,54} Other information includingd case–control studies in the pooled analysis is summarized in Table 1. The genotypes and alleles of LEP rs7799039 G>A polymorphism are shown in Table 2.

Table 1 Characteristics of the studies in meta-analysis

Table 2 Distribution of LEP rs7799039 G>A polymorphism genotype and allele

Figure 1 Flow diagram of the meta-analysis of the association between LEP rs7799039 G>A polymorphism and overall cancer risk.

Meta-analysis results

Table 3 summarizes the results of this meta-analysis. We found that the LEP rs7799039 G>A polymorphism was associated with overall cancer risk (Figure 2). The A vs G genetic model has an OR = 1.10 with a 95% CI = 1.00–1.21 and a P = 0.051. The AA vs GG genetic model has an OR = 1.22 with a 95% CI = 1.01–1.48, and a P = 0.042. When we compared AA/GA vs GG model, we found an OR = 1.16 with a 95% CI = 1.02–1.33 and P = 0.026. Comparing the AA genotype with GA/GG, we calculated an OR =1.12 with a 95% CI = 1.00–1.26 and a P = 0.059.

Table 3 Results of the meta-analysis from different genetic models

Figure 2 Meta-analysis of the association between LEP rs7799039 G>A polymorphism and overall cancer risk (dominant model, random-effects model).

In a subgroup analysis by ethnicity, we found an association in Asian populations with AA/GA vs GG having an OR = 1.23, a 95% CI = 1.01–1.49 and a P = 0.044, Table 3.

In a subgroup analysis by cancer type, we found that the LEP rs7799039 G>A polymorphism moderately increased the risk of PC; AA vs GA/GG: OR =1.24, 95%CI = 1.04–1.470, P = 0.014. However, we found that this G>A polymorphism might actually confer a decreased the risk to CRC, AA vs GA/GG: OR =0.88, 95%CI = 0.77–1.00, P = 0.046. When we conducted a subgroup analysis by cancer system, we found that this G>A polymorphism might increase the susceptibility of hematopoietic cancer; A vs G: OR = 1.13, 95% CI = 1.01–1.26, P = 0.038; AA vs GG: OR = 1.25, 95% CI = 1.00–1.55, P = 0.049) and other system cancers (AA/GA vs GG: OR = 2.06, 95% CI = 1.05–4.02, P = 0.035.

Heterogeneity analysis

For this meta-analysis, we found that there was significant heterogeneity among the included case–control studies (Table 3). To identify the major sources of heterogeneity, we carried out subgroup analyses. The results indicated that Asians, small sample size studies (<1000), and hospital-based studies might lead to the major heterogeneity in this meta-analysis.

Sensitivity analysis

Sensitivity analysis was conducted by deleting an individual study in turn and calculating the pooled ORs and CIs of the remaining studies. For this SNP, the results under all genetic comparisons were not influenced by removing any eligible study (Figure 4).

Figure 3 Begg’s funnel plot of meta-analysis of the association between LEP rs7799039 G>A polymorphism and cancer risk (dominant genetic model, random-effects model).

Publication bias

Begg’s test and Egger’s test were used to determine whether there was publication bias in genetic comparisons. The shapes of the Begg’s funnel plot revealed that they were symmetrical; A vs G had a P _Begg’s= 0.588, AA vs GG had a P _Begg’s= 0.802; AA/GA vs GG had a P _Begg’s = 0.953; and AA vs GA/GG had a P _Begg’s= 0.887 (Figure 3). The results of Egger’s test also highlighted that there was no evidence of publication bias (A vs G: P_Egger’s = 0.559; AA vs GG: P_Egger’s = 0.579; AA/GA vs GG: P_Egger’s = 0.639 and AA vs GA/GG: P_Egger’s = 0.660).

Figure 4 Sensitivity analysis of the influence of dominant model (random-effects estimates).

Discussion

The adipocyte-derived peptide hormone LEP has a well-known influence on inflammation, tumor growth, and metastasis. Rs7799039 G>A is a common promoter SNP in the LEP gene, that may affect the transcriptional level and LEP expression.⁵⁵ We therefore hypothesized that the LEP rs7799039 G>A polymorphism might be closely related to the susceptibility of cancer. Although a number of studies have focused on the relationship between the LEP rs7799039 G>A polymorphism and cancer risk, the observed results have been inconsistent. Three meta-analyses carried out by Liu et al¹⁸, He et al¹⁹, and Yang et al²⁰, including 12, 15, and 15 eligible case–control studies, respectively, yielded conflicting results in some subgroups. Of late, some new data regarding the relationship of the LEP rs7799039 G>A polymorphism and cancer risk have been reported.^{36–50,53,54} Therefore, an updated meta-analysis is needed to address this issue. In our meta-analysis, data of 31 independent case–control studies including 11,276 cancer cases and 14,523 controls were pooled, which is more participants than were in the meta-analyses mentioned above. Thus, this updated analysis should be more comprehensive. To the best of our knowledge, the present study is the most convincing pooled analysis to explore the association between the LEP rs7799039 G>A polymorphism and cancer risk. Results of our meta-analysis did indicate that the LEP rs7799039 G>A polymorphism was associated with an increased risk of overall cancer, especially in Asians, PC, hematopoietic malignancy, and other system cancer subgroups.

With the accumulating evidence of genetic association investigations, it is urgent to synthesize all available data to obtain a robust result. According to the findings, the association of increased cancer risk with the LEP rs7799039 G>A polymorphism was found in overall populations. Race also could be a critical biological factor for the genetic comparison. In previous meta-analyses,^18–20 most of the eligible studies contained only Caucasians. In the current study, more case–control studies included Asians.^{21,45–50,54} The results suggest that the LEP rs7799039 G>A polymorphism might increase the risk of cancer in Asians. We are the first to report the relationship between this SNP and cancer risk in this ethnicity.

An interesting phenomenon observed during stratified analysis was that the LEP rs7799039 G>A polymorphism decreased the risk of CRC, while this SNP increased the risk of PC, other cancers, and hematopoietic malignancy. One possible explanation is that there were insufficient sample sizes for subgroup analysis. Although our findings were stable by one-way sensitivity analysis, publication bias was not found.

Among the included studies, significant heterogeneity was found in four genetic models for overall analysis. Stratified analyses indicated that heterogeneity was significant in some subgroups (eg Asians, small sample sizes, and hospital-based studies). These factors may contribute to the major heterogeneity in this study.

Several limitations, in this meta-analysis, should be acknowledged. First, although the Begg’s funnel plot and Egger’s test suggested no significant publication bias, it is possible that certain unpublished data are yet to be included. Selection bias for this study might have existed. Second, for lack of detailed information in the included studies, only crude ORs and CIs were calculated. We did not carry out the analysis adjusted for other potential risk factors (eg smoking, alcohol consumption, body mass index, and vegetable intake). Finally, heterogeneity among the eligible case–control studies was statistically significant in multiple genetic models. These findings should be considered with caution.

In conclusion, this study performed an extensive assessment based on a larger sample size than the previous pooled analysis. Our study indicates that the LEP rs7799039 G>A polymorphism may contribute to the development of cancer. In order to further verify or refute our findings, large well-designed epidemiological studies are warranted. As investigations among Asian populations are limited, further well-designed epidemiological studies involving a wider spectrum of subjects to explore the potential role of this SNP in Asians are needed.