Background
Worldwide, breast cancer (BC) represents the leading cause of female gynecological cancer death. Its estimated deaths (189,000) were reported almost equal to the estimated number of deaths from lung cancer (188,000 deaths) [
1], indicating that BC has become a global burden. Known risks factors contribute to BC included reproductive events, hormonal level, and family histories [
2], but they account for less than 47% cases [
3]. Other etiology, though remains unknown, is believed causing by an integrated function of carcinogen exposure and polymorphisms in genes, especially genes involved in carcinogen metabolism [
4].
Cytochrome P4502E1 (CYP2El), a member of Cytochrome P450 (CYP) super-family which involved in the metabolism of many endogenous and exogenous substance, is of pivotal importance in metabolizing ethanol and low-molecular-weight carcinogens such as
N-nitrosamines in cigarettes [
5,
6]. Alcohol intake, a risk factor for cancer of various organs, has been proved associated with BC [
7]. The relationship between tobacco smoke and BC development, also continuously, been emphasized that smoking could increased breast cancer risk, no matter passively or actively [
8]. Furthermore, their association with BC according to carcinogen-metabolizing genotype was also investigated by more than 50 epidemiologic studies. Results indicating that gene polymorphisms, including CYP450s, glutathione
S-transferases,
N-acetyltransferases, and sulfotransferases, interacting with carcinogen exposure, may modified one’s susceptibility to cancer [
9]. Among the various CYPs, CYP2E1 is an important Phase I enzyme involved in the metabolism of alcohol and tobacco-generated
N-nitrosamines, altering its activity has been suggested might link to the development of BC [
4].
CYP2E1 is located at chromosome 10. So far, more than 100 single nucleotide polymorphisms (SNPs) have been found (
http://www.ncbi.nlm.nih.gov/SNP). However, only several common mutations were extensively investigated as they might alter the activity of CYP2E1 [
10,
11]. Rs3813867
G>
C and rs2031920
C>
T were two key SNPs among them, with the former one associated with Pst I restriction enzyme site and the later one with Rsa I restriction enzyme site, their linkage disequilibrium also lead to the CYP2E1*5 haplotype and form three types of distinct genotypes: (1)
c1/c1(Rsa I+
Pst I−
), homozygous of normal alleles; (2)
c1/c2, heterozygous; (3)
c2/c2 (Rsa I−
Pst I+
), homozygous alleles after nucleotide exchanged [
12]. Another polymorphism, recognized by Dra I restriction enzyme in intron 6 (rs6413432), form CYP2E1*6 polymorphism and also result in three genotypes:
C/C, C/D and D/D [
13].
The relationship between above CYP2E1 polymorphisms and BC has been investigated by various studies, however, presenting conflict results. One studies conducted on patients suffering from primary unilateral BC demonstrated the absence of any association between CYP2E1*5 polymorphisms with BC, no matter in premenopausal or postmenopausal women [
14]. However, Wu et al. [
15], who carried out a studies on non smoker and non drinker women, reported that individuals with the
c2/c2 genotype of CYP2E1*5 had a lower BC risk than that of
c1/c1 (OR = 0.24, 95% CI = 0.08–0.74). While the most recently study by Chong et al. [
16]. indicated that the c1/c2 genotype or c2 allele carriers with CYP2E1*5 variation have an approximately 1.8-fold higher risk of BC. Such controversy results may due to the relatively low mutation frequency of CYP2E1 and small epidemiologic studies with low statistic power; we therefore systematically reviewed and performed a meta-analysis to quantitatively evaluate the role of CYP2E1 polymorphisms in BC development.
Methods
Search strategy
A comprehensive search of literature listed in various databases (PubMed, Web of Science and Google scholar), published before August 2016, was performed using the following key words ‘breast cancer’ or’ ‘breast carcinoma’, ‘polymorphism’ or ‘variant’ and ‘mutation’, all combined with Medical Subject Heading (MeSH) term ‘CYP2E1’. The eligible studies were retrieved, and their reference lists were screen by hand to find every relevant paper. No any restriction such as time and language was made during the searches, as well as attempts to obtain unpublished studies.
Selection criteria
In this study, we performed the meta-analysis according to the proposal of Meta-analysis of Observational Studies in Epidemiology group (MOOSE) [
17]. The eligible studies were requested to meet the following inclusion criteria: (1) any type of comparative study; (2) evaluated the association between CYP2E1 gene polymorphisms and breast cancer risk; (3) in cases and controls, provided sufficient data to estimate the odds ratio (OR) with their 95% confidence intervals (95% CIs). Studies were excluded if one of them existed: (1) insufficient data to extract; (2) without control population; and (3) some CYP2E1 polymorphisms that rarely reported. If overlapping data was found, either the study with lower quality or the earlier published one would be excluded in the following analysis.
Data extraction from each eligible study was conducted by two independent investigators, which included: (1) the first author’s name; (2) year of publication; (3) study region or country; (4) ethnicity; (5) cancer confirmation; (6) sample size (both cases and controls); (7) source of control (together with matching criteria); (8) polymorphisms of CYP2E1; (9) genotyping method; (10) genotype distribution in cases and controls and whether P value of the control population consistent with the Hardy–Weinberg equilibrium (HWE). In the event of different results, discussion was conducted to solve the discrepancies. When a study reported the results on different CYP2E1 polymorphisms, we treated them as separate studies in our meta-analysis.
Quality assessment
To evaluate the quality of the included studies, a set of predefined criteria originally proposed by Thakkinstian et al. [
18]. was used. The predefined criteria, which cover the credibility of controls, the representativeness of cases, specimens of cases when determining genotypes, Hardy–Weinberg equilibrium in controls, and total sample size, was structured as a 16-item list with scores ranging from 0 to 15 by Qin et al. [
19]. and has been quoted by several meta-analyses [
20,
21] (see Additional file
1: Table S1). As done previously, the studies with scores ≥10 were defined as high-quality studies, while the rest were low-quality studies.
Statistical analysis
The association of each CYP2E1 polymorphisms with breast cancer risk was estimated by calculating pooled ORs and 95% CIs under different comparison models, including additive models, recessive model, and dominant model. Firstly, the heterogeneity between studies would be assessed by the Q test and I2 statistics. According to the presence (PQ < 0.1 or I2 ≥ 50%) or absence (PQ ≥ 0.1 and I2 < 50%) of heterogeneity, different models would be used to calculate the pooled ORs, with the former situation using DerSimonian–Laird random-effects model while the later using Mantel–Haenszel fixed-effects model. If heterogeneity existed, Galbraith plot analyses would be carried out to investigate the sources of heterogeneity among studies. Then, subgroup analysis by ethnicity would be performed to address possible effects of these polymorphisms on different population. To assess the stability of the results, sensitivity analysis was performed by sequential omission of individual studies, especially studies whose genotype frequencies in the control populations were deviated from HWE, as they may generate bias. HWE in the control group population, if not reported in the original article, would be tested via a goodness-of-fit Chi square test. Finally, for each polymorphism, the Begg’s funnel plots and Egger’s linear regression test was used to test the publication bias (P < 0.05 indicated a significant publication bias). All analyses were performed with Stata software (Stata/SE version 12.0, Stata Corp, College Station, TX) and all P values were two-sided.
Discussion
Breast cancer, of which heredity explains approximately 10–15% of the cases, with only 5% can be clarified by known genetic polymorphisms such as BRCA1 and BRCA2 [
28]. Such fact suggesting that other potential, common, but low-penetrance genetic variants may contribute to individual’s susceptibility to breast cancer. CYP2E1, a Phase I enzyme responsible for the metabolic activation of various carcinogens such as
N-nitrosaminesan and alcohol, was in different activity among individuals [
12]. It has been assumed that polymorphisms of CYP2E1*5 and CYP2E1*6 may lead to a decreased activity in CYP2E1 enzyme, thus linked to a lower risk of cancer. Nevertheless, the power of a single study was too small to draw a precise conclusion, we therefore investigated breast cancer and CYP2E1 polymorphisms in these common mutations using a meta-analysis.
However, in the present study, no significant association was found between SNP rs2031920
C>
T polymorphism and BC. The haplotype CYP2E1*5, which consist of SNP rs2031920
C>
T and 3813867
G>
C, also failed to identify any significant association with BC risk. Single SNP rs3813867
G>
C was also taken into consideration, but further analysis was not carried out due to the suboptimal study numbers (n = 1) [
27]. Such insignificant results may be partially attributed to the different distribution of the CYP2E1*5 polymorphism between varies races, with the rarest 0.05 in Caucasians and the highest 0.23 in oriental populations [
12]. Nevertheless, after stratified the study populations into different races, where we mainly focus on Asian populations, the results still failed to indicate any association between CYP2E1*5 polymorphism and BC development. The pooled results of rs2031920
C>
T SNP and BC risk were consistent with those studies included in the present meta-analysis, all indicating an insignificant relationship between them; and the overall results of CYP2E1*5 polymorphism and BC were also accordance with half of those included, though one of the rest observed a decreased risk of BC while another revealed an increased risk. Taken together, it may be concluded that CYP2E1*5 polymorphisms are not associated with BC risk in the overall population.
Interestingly, when considering the CYP2E1*6 polymorphism, our study found that individuals with the D/C and C/C + D/C genotype had a significantly higher risk of BC compared to those with the D/D genotype, similar increased result was also found in the C allele carriers when compared with the D allele carriers, especially in Caucasian population. These results suggested that polymorphism in CYP2E1*6 could be a risk factor for BC development. But such result was inconsistent with those of the original studies, of which all suggested no significant relationship between any comparison model of CYP2E1*6 variation and BC development.
Actually, our results should explain with caution as there is increasing evidence that metabolizing enzymes do not act alone. In the study carried out by Choi et al. [
24]. that explored the role of alcohol and genetic polymorphisms of CYP2E1*5 in BC development, no significant overall differences were found in the
c1/c2 genotype frequencies between BC cases and controls. However, after taking the drinking situation into consideration, a 1.9-fold increasing risk for developing BC was found when comparing the ‘ever’-drinking women with the c2 mutation to the non-drinkers with the
c1/c1 mutation. Another study, investigated lifetime passive cigarette smoke exposures together with genetic variants and BC risk in women who had never smoked, found that interaction between passive smoke exposure and CYP2E1*6
AA/AT (namely
CC/CD) polymorphism could significant increased breast cancer risk among premenopausal women [
27]. In sum, such gene-environment interaction should be taken into consideration when investigating CYP2E1 polymorphism in the development of BC, however, due to the limited studies included, our study could not conduct further analysis with these factors taken into consideration.
To our knowledge, this is the first meta-analysis carried out to date to evaluate the role of CYP2E1 polymorphisms in breast cancer susceptibility. Despite the findings mentioned above, this study had several limitations. First, we haven’t taken the gene-environment interaction into consideration. As is known to all, apart from genetic factors, smoking status and alcohol consumption are important risk factors for BC; however, we could not conduct subgroup analyses stratified by environmental exposure due to the limited information on our included studies. Second, the overall results of our study were based on crude ORs, but a more precise evaluation should be adjusted for the know risk factors such as age and menopause status. Third, the number of studies included in this study is relatively small, with three or four studies for each polymorphism, which may lead to low statistical power and prevent us from exploring a real association of the CYP2E1 polymorphism and BC risk. Fourth, because no attempts were made to access unpublished studies and studies in languages other than English, publication bias may exist, though results of our Begg’s funnel plot and Egger’s test did not reveal any publication bias. Fifth, as most studies were conducted in Asian and Caucasian population, the relative lack of ethnic diversity demands for further studies.