FASEB J. : Fast, Stable Macromolecular Detection
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published as doi: 10.1096/fj.08-124073.
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.08-124073v1
23/7/2274    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yu, K.-D.
Right arrow Articles by Shao, Z.-M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yu, K.-D.
Right arrow Articles by Shao, Z.-M.
(The FASEB Journal. 2009;23:2274-2287.)
© 2009 FASEB

A functional polymorphism in the promoter region of GSTM1 implies a complex role for GSTM1 in breast cancer

Ke-Da Yu*,1, Gen-Hong Di*,1, Lei Fan*,1, Jiong Wu*, Zhen Hu*, Zhen-Zhou Shen*, Wei Huang{dagger} and Zhi-Ming Shao*,2

* Breast Cancer Institute, Cancer Hospital, Department of Oncology, Shanghai Medical College, Institutes of Biomedical Science, Fudan University, Shanghai, China; and

{dagger} Chinese National Human Genome Center at Shanghai, Shanghai, China

2 Correspondence: Department of Breast Surgery, Cancer Hospital/Cancer Institute, Breast Cancer Institute, Fudan University, 399 Ling-Ling Rd., Shanghai, 200032, China. E-mail: zhimingshao{at}yahoo.com


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Although a number of studies have been conducted to address the relation between a gene deletion polymorphism of glutathione S-transferase M1 (GSTM1) and breast cancer, no definite conclusion has been reached and no clear risk pattern has yet to emerge for GSTM1. We first conducted case-control studies that included 1920 subjects using a genotyping method allowing the definition of GSTM1-null (–/–), homozygous wild-type (+/+), and heterozygous (+/–) genotypes. The results show that GSTM1–/– confers an increased risk for breast cancer development compared with that in GSTM1-present individuals (+/+ and +/–), which was subsequently confirmed by a meta-analysis of all of the 41 relevant studies (odds ratio: 1.10, P<0.001). Unexpectedly, we found that GSTM1+/+ is also a risk genotype compared with GSTM1+/–. Furthermore, we identified a functional polymorphism in the GSTM1 promoter region associated with breast cancer. The variant allele modifies DNA binding to the AP-2{alpha} transcription factor, resulting in reduced promoter activity and mRNA expression. However, this low-activity allele is associated with reduced breast cancer risk. It seems that ~60–70% expression from one allele of GSTM1 could suffice for protection against breast cancer; null activity and overactivity of GSTM1 are both disadvantageous. These results indicate a U-shaped association of GSTM1 with breast cancer, which challenges the linear gene-dosage effect of GSTM1 that was previously proposed. We recommend that a more complicated role for GSTM1 should be considered in breast cancer risk prediction.—Yu, K.D., Di, G.-H., Fan, L., Wu, J., Hu, Z., Shen, Z.-Z., Huang, W., Shao, Z.-M. A functional polymorphism in the promoter region of GSTM1 implies a complex role for GSTM1 in breast cancer.


Key Words: risk pattern • AP-2 • meta-analysis • functional verification


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
CUMULATIVE EVIDENCE SUPPORTS the role of steroid hormones and their catechol metabolites as carcinogens for breast cancer (1 , 2) . Breast carcinogen-metabolizing related enzymes may also affect breast cancer development by altering carcinogen metabolism (3 4 5) . Glutathione S-transferase (GST) detoxification pathways are active in normal breast tissue to protect against damage caused by reactive metabolites of chemicals, such as estrogen semiquinone and estrogen 3,4-quinone (1) . Among the GST isoforms, GSTM1 is of particular interest because it possesses a null polymorphism, referred to as GSTM1–/– (GSTM1-null), which results in a complete absence of GSTM1 enzyme activity. GSTM1–/– is present in 27–65% of the population depending on ethnicity (6 , 7) . It has attracted a significant amount of attention and has become the focus of hundreds of publications on the molecular epidemiology of cancer as well as on the epidemiology of other diseases (8) . More than 40 original studies and a few pooled analyses that alternately confirm or refute the association between GSTM1–/– and breast cancer have appeared in the literature. The underlying hypothesis of those studies was that normal or increased GSTM1 enzyme activity protects susceptible tissues from somatic DNA mutations by facilitating the detoxification of carcinogens.

It is important to note that the most commonly used analytical approach for determining the GSTM1 genotype, a short PCR-electrophoresis-based assay, has a basic flaw in that it only positively identifies the GSTM1–/– genotype but cannot distinguish GSTM1 homozygous wild-type (GSTM1+/+) from GSTM1-present individuals (GSTM1-present indicates both GSTM1+/+ and GSTM1+/– heterozygous). This flaw may be relevant because a previous case-control study pointed out that GSTM1 has a linear gene-dosage effect on the development of asthma (9) , but it is too early to assert that the effect of GSTM1+/+ is twice as much as that of GSTM1+/– in the context of breast cancer predisposition. Interestingly, in another recent case-control study using a genotyping method allowing the definition of GSTM1–/–, GSTM1+/–, and GSTM1+/+ genotypes it was proposed that Caucasian women with a GSTM1+/+ genotype have twice the risk of developing breast cancer compared with that for GSTM1–/– Caucasian women (10) . Such unexpected results indicate a more complicated role for GSTM1 in breast cancer susceptibility.

Accordingly, there are two main unresolved problems regarding the relation between GSTM1 and breast cancer. The first is whether GSTM1–/– contributes to breast cancer susceptibility based on the accumulating evidence and the second is what is the risk pattern of a GSTM1 mutation, i.e., is it a linear gene-dosage effect or a more complicated nonlinear effect, if GSTM1 is proven to be a breast cancer-susceptible gene. In the present study, we first conducted a hospital-based case-control study of Chinese women in Shanghai and subsequently performed a meta-analysis of all of the related studies carried out before December 2007 in an effort to reveal the role of GSTM1–/– in breast carcinogenesis. Then, we completed genotyping for GSTM1–/–, GSTM1+/–, and GSTM1+/+ in all subjects to scrutinize the true role of GSTM1+/+. We also explored the contributions of single nucleotide polymorphisms (SNPs) and haplotypes of GSTM1 to breast cancer susceptibility. A susceptible genetic variant was identified in the GSTM1 promoter region. This functional polymorphism affected promoter activity via altering the affinity of the variant allele for the AP-2{alpha} transcription factor, which ultimately implied that GSTM1 has a novel U-shaped effect on breast cancer susceptibility.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Study subjects and data collection
A total of 1017 patients with breast cancer and 903 female cancer-free control subjects were identified as genetically unrelated Han Chinese in Shanghai City and its surrounding regions. These subjects were recruited between January 2004 and June 2008 from the Department of Breast Surgery in the Shanghai Cancer Hospital. Primary infiltrating ductal carcinoma or ductal carcinoma in situ (DCIS) of the breast was pathologically confirmed in the patients (117 had DCIS with/without microinvasion). Patients with a previous history of other cancers (not breast cancer) and metastatic breast cancer were excluded. The control subjects were chosen from women who had come to the outpatient department of the Shanghai Cancer Hospital mainly for the purpose of breast cancer screening. They were matched to the case patients on the basis of age and geographical area. The control subjects selected were proven to be free of breast cancer by a complete physical examination, bilateral mammography, ultrasonography, and biopsy when necessary. Women who had a previous history of cancer were also excluded from the control group. A few of the control subjects had undergone a biopsy and had a diagnosis of benign breast disease that was pathologically confirmed as fibrocystic changes; such pathological changes do not confer an increased risk for breast cancer according to the current view (11) . All study subjects provided a 3- to 5-ml venous blood sample. Each participant was personally interviewed by doctors either in the outpatient department or in the inpatient department to obtain epidemiological and clincicopathological information (e.g., age, age at menarche, menstrual status, history of oral contraceptive drug use, reproductive history, and the breast cancer history of first-degree relatives). Body mass index (BMI) was also calculated for each subject according to her individual height and weight. All of the data collected were entered into a computerized database established by the Department of Breast Surgery of the Shanghai Cancer Hospital as described previously (12) . All of the subjects in the present study had available and complete epidemiological information. The case patients and control subjects were comparable in age (mean±SD: 47.6±10.6 vs. 47.6±9.3 yr for case patients and control subjects, respectively; P>0.05), age at menarche (14.8±1.6 vs. 14.8±1.6 yr, P>0.05), menopausal status (38 vs. 40% postmenopausal, P>0.05), and BMI (23.2±3.0 vs. 23.2±2.6 kg/m2, P>0.05). More women with a family history of first-degree relatives with breast cancer were enrolled in the case cohort (13 vs. 8%, P<0.001) to increase the power to detect associations (13) . All of the familial breast cancer case patients as well as the control subjects with a family history of breast cancer had been tested for BRCA1 and BRCA2 mutations, and no deleterious germ-line mutations were found, as we reported previously (14) . Because most Chinese women are nonsmokers and nondrinkers, the study population was restricted to women without the habits of cigarette smoking and alcohol consumption. Specifically, we excluded women who had smoked for a duration of >3 mo or smoked >100 cigarettes in total (current smokers were also excluded regardless of the count of cigarettes) and women who had ever drunk alcohol. This study was approved by the ethics committee of the Cancer Hospital of Fudan University, and each participant provided informed consent.

DNA/RNA preparation, PCR, and PCR-based allele genotyping
Genomic DNA was extracted from 3–5 ml of the study participants’ peripheral blood leukocytes using a PureGene DNA Purification kit (Gentra Systems, Inc., Minneapolis, MN, USA) according to the manufacturer’s protocol, and samples were then stored at –20°C. General PCR was performed as described previously (14) . Deletion polymorphism of the GSTM1 gene was analyzed using PCR in accordance with methods described previously (15) . Long PCR- and real-time PCR-based methods have been developed to identify GSTM1+/+ from GSTM1-present (9 , 16) . For the DNA samples from the GSTM1-present individuals recruited between January 2004 and December 2007, we initially used a long PCR assay to identify those with a GSTM1+/+ genotype (16) . The failed samples from the long PCR procedure as well as the samples from women recruited between January 2008 and June 2008 were subjected to a real-time PCR-based assay (9) . To detect the expression of GSTM1 and the AP-2 isoforms, cDNA-specific primers were designed using Primer Premier 5.00 (Premier Biosoft International, Palo Alto, CA, USA) for the human GSTM1, AP-2{alpha}, AP-2β, and AP-2{gamma} genes. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was chosen as an internal control. Typical agarose gel electrophoresis results are shown in Supplemental Fig. S1; PCR primers used in this study are shown in Supplemental Table S1.

Meta-analysis
We identified previously published related studies by a searching up to December 2007 in PubMed using the very sensitive terms, "glutathione S-transferase M1" or "GSTM1" and "breast cancer" or "breast carcinoma." We also screened relevant articles in the reference lists of review articles and key original papers. The eligibility criteria are described in Supplemental Table S2. For overlapping studies, only the first published study was selected; for republished studies, only the one with the largest sample numbers was included. The full texts of all of the eligible studies were read carefully. The following variables were abstracted from each study if available: GSTM1 genotype information, number of case patients/control subject, odds ratio (OR), 95% confidence interval (CI), year of publication, menopausal status, ethnicity, and national origin of the studied population. Two investigators (K.-D.Y. and L.F) independently extracted the relevant data. A third investigator (Z.-M.S.) adjudicated the disagreements. Meta-analysis was performed as described previously (17) . Briefly, ORs and their 95% CIs were obtained directly or calculated from the data given in the articles. Adjusted ORs were preferable. A random- or a fixed-effects model was used to calculate the pooled effect estimates in the presence (P≤0.05) or absence (P>0.05) of heterogeneity, respectively. I2 was used to describe the percentage of the total variation among studies due to heterogeneity. We also performed separate meta-analysis by omitting each study to find potential outliers. Publication bias was examined visually in a funnel plot of log [OR] against its SE, and the degree of asymmetry was tested using Egger’s test. All of the analyses were performed using Stata/SE version 10.0 (StataCorp, College Station, TX, USA).

SNP selection
The SNPs across a {approx}7-kb region spanning GSTM1 (5.9 kb) from 1 kb upstream of the 5'-flanking region to 0.5 kb downstream of the 3'-flanking region were surveyed using National Center for Biotechnology Information Single Nucleotide Polymorphism database (NCBI-dbSNP) and the International HapMap Project. The HapMap Project has genotyped a large number of SNPs in several populations. Eight SNPs available in NCBI-dbSNP were also validated in HapMap genotype database (HapMap Data Rel 21a/phaseII January 2007, on NCBIB35 assembly, dbSNPb125). We aimed to identify a set of tagging SNPs (tagSNPs) in GSTM1 that efficiently tag all the known common variants [minor allele frequencies (MAFs)>0.05] and were likely to tag most of the unknown common variants. A tagSNP is a representative SNP in a region of the genome with high linkage disequilibrium (LD). Among the above eight SNPs, only three SNPs (rs412543, rs2071487, and rs4147567) had MAFs > 0.05. Among these three SNPs, only one SNP (rs2071487, g.2640T>C) was identified as tagSNP under a restriction of MAFs > 0.05 and r2 ≥ 0.8. Therefore, we chose rs2071487 for further genotyping. In addition, SNPs with potentially functional effects (such as those causing amino acid changes, causing alternative splicing, or being in putative transcription factor binding sites) were chosen for genotyping. These potentially functional SNPs were allowed with a minimum MAF of 1%. By surveying the NCBI-dbSNP and HapMap databases and using an in silico analysis tool, we finally selected three other potentially functional SNPs with a reported MAF > 1% in the European population: an SNP (rs412543 g.–498C>G) in the promoter region, a nonsynonymous SNP (rs1065411, g.2697G>C, p.Lys173Asn), and a synonymous SNP (rs1056806, g.2706C>T). The average density was found to be 1 SNP/1.8 kb.

SNP genotyping
First, we genotyped the four SNPs in the DNA samples of the GSTM1+/– carriers. Two SNPs (rs412543 and rs1065411) were genotyped using a restriction fragment length polymorphism (RFLP) assay and direct sequencing. In the detection of rs412543, the ScrFI digestion produced 65-, 70-, and 110-bp fragments for the –498G variant and produced 70-, 110-, and 130-bp bands for the –498C genotype. For rs1065411, an allele-specific PCR and RFLP by HaeII were performed as described previously (18) . To raise the sensitivity of the RFLP assay, PCR-RFLP was performed twice, and an adequate quantity of restriction enzymes was supplied to cut the PCR amplicon effectively. The samples with inconsistent outcomes in the two independent PCR-RFLP tests were directly sequenced. The other two SNPs (rs1056806 and rs2071487) were genotyped using the GenomeLab SNPStream 12-plex Genotyping platform (Beckman Coulter, Fullerton, CA, USA), which was carried out by the Chinese National Human Genome Center (Shanghai, China). Based on the results of the four-SNP association studies, we subsequently genotyped rs412543 in the DNA samples of GSTM1+/+ carriers by direct sequencing.

Analysis of LD and haplotyping
Hardy-Weinberg equilibrium (HWE) was assessed by {chi}2 tests. The Haploview 3.32 program (Broad Institute, Cambridge, MA, USA) was used to test for LD of SNPs in the GSTM1 gene. The structure of the LD block, if any, was examined using the method of Gabriel et al. (19) . Haplotypes were constructed using all of the tested common SNPs (MAF>5%). We also applied a false discovery rate (FDR) correction to adjust the P value of the single-locus association result.

In silico analysis
The sequence flanking SNP rs412543 was screened for transcription factor binding sites by using the prediction web sites TESS (http://www.cbil.upenn.edu/cgi-bin/tess/tess) and TFSEARCH (http://www.cbrc.jp/research/db/TFSEARCH.html).

Plasmid constructs
The pGL3-Basic reporter vector was obtained to construct luciferase reporter plasmids using standard recombination techniques (20) . The fragment from –742 to +29 bp of the GSTM1 putative promoter region with a –498C allele was cloned between the KpnI and HindIII sites of the pGL3-Basic vector. The recombinant plasmid, pGL-GSTM1-C, was purified using the EndoFree plasmid purification kit (Qiagen, Valencia, CA, USA). Then, a site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) was used to generate the pGL-GSTM1-G vector. The pRL-β-actin was generated by cloning the gene promoter fraction from –535 to +71 bp of β-actin (21) into a BglII-HindIII-digested pRL-SV40 plasmid. In some experiments, pRL-β-actin replaced the pRL-SV40 plasmid to avoid the interaction between AP-2 and the SV40 enhancer (22) . Human AP-2{alpha}/β/{gamma} expression vectors were constructed using the pcDNATM3.1 Directional TOPO Expression kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. All constructs were verified before use by direct sequencing.

Cell culture, transient transfection, and luciferase assays
Human breast cancer cell lines (ZR-75-1, MCF7, T-47D, SK-BR-3, and MDA-MB-231), immortal normal breast epithelial HBL-100 cells, HEK-293T cells, and HeLa cells were grown in complete medium consisting of DMEM supplemented with 10% heat-inactivated FCS in a humidified, 5% CO2 incubator at 37°C. For the experiments shown in Fig. 3A , ~2 x 105 cells were transfected with 500 ng of plasmid DNA (pGL-GSTM1-C/G as test vectors and pGL3-Basic as a negative control) and cotransfected with 10 ng of pRL-SV40, which served as a transfection efficiency reference. For the experiments shown in Fig. 3D , 0.1, 0.5, or 1.0 µg of AP-2 of the expression vectors was cotransfected with 10 ng of pRL-β-actin. Transfections were performed using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s protocol. The transfected cells were maintained for 48 h, and the medium was replaced once 16 h after the transfection. After 48 h of culture, the cells were lysed in 100 µl of lysis buffer, and 20 µl of the supernatant was then measured for luciferase activity on a Veritas microplate luminometer (Turner BioSystems, Sunnyvale, CA, USA) using a dual-luciferase reporter assay system kit (Promega Corp., Madison, WI, USA) as described previously (23) . Each experiment was done in triplicate at least 3 times.


Figure 1
View larger version (32K):
[in this window]
[in a new window]

  		Figure 1. Association of the GSTM1-null genotype with breast cancer. A) Forest plot of meta-analysis regarding the association of GSTM1-null with breast cancer. Center circles represent size of the OR estimated from each study. The 95% CI for the OR is also shown. Overall OR is shown by the middle of a diamond whose left and right extremes represent the corresponding 95% CI. Refer to Supplemental Table S2 for characteristics of these eligible studies. B) Begg’s funnel plot of studies included in the meta-analysis. Horizontal line and sloping lines in funnel plot represent fixed-effects summary OR and expected 95% CI for a given SE, respectively. Area of each circle represents contribution of the study to the pooled OR. C) Subpopulation risk plot showing ORs of GSTM1–/– vs. GSTM1+/–. ORs of subpopulations are shown on the y axis; median age is shown on the x axis. For this sliding-window analysis, each subpopulation contains 400 patients, and each subsequent subpopulation is formed moving from left to right by dropping 15 patients with the youngest age and adding 15 patients with the next older age. Horizontal line indicates reference of OR = 1. {diamond}, OR estimated from each subgroup. Error bar = 95% CI.


Figure 2
View larger version (22K):
[in this window]
[in a new window]

  		Figure 2. LD of the GSTM1 region. Top horizontal bar indicates genetic region spanning the SNPs tested. Top triangle colored white to black shows LD calculated using r2 measure; bottom triangle colored white to black shows LD calculated using D' measure.


Figure 3
View larger version (24K):
[in this window]
[in a new window]

  		Figure 3. SNP –498C>G in the GSTM1 promoter region is associated with reduced promoter activity and DNA-AP-2{alpha} interaction. A) Transient reporter gene assays with constructs containing –498C>G. Luciferase units were calculated using the following formula: firefly luciferase units/Renilla luciferase units. Fold increase is measured by defining the activity of the empty pGL-3-Basic vector as 1. All luciferase assay measurements were performed in at least 3 independent experiments, and each was performed in triplicate. Data shown represent means with SE. MDA-MB-231, P = 0.136; MCF7, P = 0.009; SK-BR-3, P = 0.293; ZR-75-1, P = 0.456; 293T, P = 0.046; HeLa cells, P < 0.001. B) Allele-specific binding of the AP-2 transcription factor. Left panel: EMSA analysis of the allele-specific effect of –498C>G on the interaction of nuclear protein complexes extracted from HeLa cells. Biotin-labeled double-stranded probes (100 fmol) of –498C or –498G allele were incubated with 15 µg of nuclear extract from HeLa cells. In competition experiments, 100-fold excess of unlabeled probe was added before the addition of the labeled probes. Right panel: data confirm the site of an oligonucleotide/AP-2 complex, indicated by asterisk. EMSA was conducted with biotin-labeled/nonlabeled consensus AP-2 binding elements and nuclear extracts from HeLa, MCF7, and ZR-75-1 cells, which have been confirmed to express the AP-2 nuclear protein. C) Expression of AP-2 isoforms in breast cell lines and tool cells. Neg., no-template

(from previous page) control; N.M.T., cDNA from normal mammary tissue. D) Effects of AP-2 isoforms on genetic variant –498C>G. Luciferase expression of the constructs pGL-GSTM1-C/G in AP-2-insufficient HBL-100 cells. AP-2 isoform (top, AP-2{alpha}; middle, AP-2β; bottom, AP-2{gamma}) expression constructs with increasing concentrations and 10 ng of pRL-β-actin vector for standardizing transfection efficiency are cotransfected. Fold increase is measured by defining the activity of the empty pGL-3-Basic vector as 1. Data represent mean ± SE fold increase from 3 independent experiments. E) Binding of AP-2{alpha} transcription factor to the promoter region of GSTM1 in vivo. ChIPs were performed using either an AP-2{alpha} antibody or IgG. MCF7 cell line, which has a –498C-GSTM1+/– genotype, was used for this experiment. Presence of the GSTM1 promoter was assayed using PCR with primers 5'-ACTGTTCCTGTGTAGGCGGGG-3' and 5'-GGTCAGACTAAAAAGTGGTGG-3', which produce a 136-bp fragment of the GSTM1 promoter containing the –498C>G polymorphic site. PCR products are shown here after electrophoreses on a 1.2% agarose gel.

Electrophoretic mobility shift assay (EMSA)
Nuclear proteins were extracted from HeLa, MCF7, and ZR-75-1 cells using NE-PER nuclear and cytoplasmic extraction reagents (Pierce Biotechnology, Rockford, IL, USA). The probes and competitors for the –498C allele, –498G allele, and the AP-2 consensus binding sequence were 5'-GAGACTAAGCCCTCGGAGTAGCTTTCG-3', 5'-GAGACTAAGCCCTGGGAGTAGCTTTCG-3', and 5'-GATCGAACTGACCGCCCGCGGCCCGT-3' (24) , respectively. Probes were synthesized as single strands and end-labeled by biotin (Invitrogen). Identical unlabeled oligonucleotides with the same sequences were used as competitors. The EMSA was performed using the LightShift Chemiluminescent EMSA kit (Pierce Biotechnology).

Chromatin immunoprecipitation (ChIP) assay
A ChIP assay was performed using a ChIP assay kit (Upstate Biotechnology, Lake Placid, NY, USA) according to the manufacturer’s protocol. Briefly, 1 x 106 MCF7 cells (a cell line of –498C-GSTM1+/– genotype) were treated with 1% formaldehyde for 10 min. After washing with PBS, cells were lysed in detergent lysis buffer. Lysates were washed and sonicated to shear the DNA to fragments of 200–400 bp and subjected to immunoprecipitation with mouse monoclonal anti-AP-2{alpha} antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and incubated overnight. IgG (Santa Cruz Biotechnology, Inc.) was used as a negative control. The antibody-protein complexes were collected by protein A beads and washed 3 times with ChIP wash buffer; cross-links were removed at 65°C for 4 h in an elution buffer. DNA was isolated using phenol/chloroform extraction and ethanol precipitation. Purified DNA was analyzed by PCR with the primers 5'-ACTGTTCCTGTGTAGGCGGGG-3' and 5'-GGTCAGACTAAAAAGTGGTGG-3', which produce a 136-bp fragment of the GSTM1 promoter containing the –498C>G polymorphic site.

Small interfering RNA (siRNA)
To demonstrate the role of AP-2{alpha} in the regulation of GSTM1 expression in human breast cancer cells, we used siRNA to down-regulate the expression of AP-2{alpha} in MCF7 and HeLa cells. Double-stranded RNAs were synthesized by Shanghai GenePharma Co., Ltd. (Shanghai, China). The following siRNA oligo sequences were designed: AP-2{alpha}-siRNA-1, sense 5'-CAGAUCAAACUGUAAUUAATT-3' and antisense 5'-UUAAUUACAGUUUGAUCUGGG-3' and AP-2{alpha}-siRNA-2, sense GCUCCACCUCGAAGUACAATT-3' and antisense 5'-UUGUACUUCGAGGUGGAGCTG-3'. The targeting sequences for the human AP-2{alpha} were submitted to a basic local alignment search tool (BLAST) search against the human genome sequence to ensure that only the AP-2{alpha} gene was targeted. The siRNA duplex and a negative control scramble siRNA duplex (Shanghai GenePharma Co., Ltd.) were transfected into study cells using Lipofectamine 2000 reagent following the manufacturer’s recommended protocol. The knockdown level of AP-2{alpha} as well as that of GSTM1 was analyzed by RT-PCR and real-time PCR. The mRNA expression was determined 48 h after transfection, and the protein expression was determined 72 h after transfection.

Real-time PCR and Western blotting analysis
The Department of Breast Surgery at the Shanghai Cancer Hospital has an established breast tissue bank. Breast cancer tissues and normal breast tissues that had been surgically obtained from women undergoing mastectomy, lumpectomy, or reduction mammaplasty were snap-frozen in liquid nitrogen and stored at –80°C. Total RNA was extracted from frozen breast tissue specimens as well as the cultured cells using TRIzol reagent (Invitrogen), followed by a reverse transcription reaction performed according to the manufacturer’s instructions (MBI Fermentas, Hanover, MD, USA). cDNA were then subjected to quantitative real-time PCR. The expression of GSTM1 and AP-2{alpha} in normal breast tissues and cells was analyzed by an SYBR Green fluorescent-based assay (Takara Bio, Otsu, Japan) in a fluorescence temperature cycler (Opticon; MJ Research, Watertown, MA, USA) using the standard curve method as described previously (25) or a modified 2–{Delta}Ct method (26) . All samples were run in triplicate, and the mean value was used. The primers used are shown in Supplemental Table S1. The expression of the GAPDH gene was measured for normalization.

Western blotting analysis was performed according to the published method (25) . In brief, equal amounts of protein (50 µg) from different cells were separated by 10% SDS-PAGE and then incubated with mouse monoclonal antibodies against the AP-2{alpha} protein (Santa Cruz Biotechnology, Inc.). Target proteins were detected using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and exposure to Biomax ML film (Eastman Kodak, Rochester, NY, USA). Relative protein expression in different cells was normalized to the GAPDH level.

Statistical analysis
Results are expressed as percentages for the categorical variables and as means ± SD/SE or median for the continuous variables, depending on their normal distribution. Tests of association were conducted and proportions were compared using Pearson’s {chi}2 test or Fisher’s exact test as appropriate. Student’s t test or the Mann-Whitney test was used to compare continuous variables between two groups; one-way ANOVA or the Kruskal-Wallis test was used to compare continuous variables among three or more groups. The Games-Howell procedure was used for multiple comparisons tests when the variances were unequal. Logistic regression was used to adjust the association between a single locus and breast cancer risk, adjusted for age, age at menarche, menopause status, BMI, and use (or not use) of an oral contraceptive drug, parity of full-term pregnancy (null or ≥1), and family history of breast cancer. To further explore the trends in OR differences according to age, we used subpopulation analysis plot methodology by defining overlapping subgroups and subsequently studying the risk of genotypes in each subgroup (27) . P ≤ 0.05 was considered to be statistically significant. Statistical analysis was performed using Stata/SE version 10.0.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
GSTM1–/– genotype is associated with an increased risk of breast cancer in Chinese women
The frequency of a GSTM1–/– genotype in our study population was 59.0%, similar to that in previous reports of Asian populations (28 , 29) . As shown in Table 1 , GSTM1–/– is associated with a 1.21 (95% CI: 1.01–1.46, P=0.037) -fold increase in the risk of developing breast cancer compared with the GSTM1-present genotype; the difference still reached statistical significance after adjustment (P=0.044). When the study population was stratified by menopausal status, the relation between GSTM1–/– and breast cancer was more evident in premenopausal women, with an unadjusted OR of 1.33 (P=0.018) and an adjusted OR of 1.32 (P=0.020). In contrast, GSTM1–/– had a lack of association with breast cancer in postmenopausal Chinese women.


View this table:
[in this window]
[in a new window]

  		Table 1. Association between GSTM1-null genotype and breast cancer risk

Meta-analysis confirmed the association between GSTM1–/– and breast cancer risk
Of the 83 studies (77 original studies and 6 reviews) identified by searching, 53 studies presenting the data of GSTM1 in both case patients and control subjects were plausibly eligible. After carefully reading the full texts of these reports, we excluded 13 subanalyses and overlapping studies. Finally, 41 studies (40 published studies and the present study) were included in this meta-analysis. The characteristics of the 41 studies are listed in Supplemental Table S2. A fixed-effect pooling analysis yielded an overall OR of 1.10 (95% CI: 1.05–1.15, P<0.001) for GSTM1–/– compared with the GSTM1-present genotype, with a small I2 of 14.6% (heterogeneity {chi}2=46.82, P=0.213) (Fig. 1A ). The results of the stratified meta-analyses by ethnicity showed that GSTM1–/– is significantly associated with breast cancer both in Asian (OR=1.11, P=0.010) and white populations (OR=1.12, P=0.001) but not in black women (OR=0.87, P=0.242) (Supplemental Fig. S2A). In addition, we evaluated the influence of any individual study on the overall OR. No individual study could affect the overall OR dominantly, because omission of any single study made little difference (Supplemental Fig. S2B). Moreover, considering a potential publication tendency to overemphasize positive findings, we evaluated the publication bias using a funnel plot, and the result suggested the absence of bias with statistical confirmation by Egger’s test (P=0.184) (Fig. 1B ).

Unexpected association of the GSTM1+/+ genotype with increased breast cancer risk
All of the 788 GSTM1-present samples were further classified into GSTM1+/– and GSTM1+/+ genotypes using long PCR- and real-time PCR-based methods. A total of 670 women were of the GSTM1+/– genotype and 118 women were GSTM1+/+. The distribution of the three genotypes fulfilled HWE in all subjects (P>0.05) and in the control subjects (P>0.05). However, the case patient group displayed a deviation from HWE with an excess of GSTM1+/+ individuals (P=0.002). Such a phenomenon had also been observed in the Caucasian patients with breast cancer as reported previously (10) . In our study population, an association of the GSTM1+/+ genotype with increased breast cancer risk was observed in univariate analysis as well as after multivariate adjustment. The adjusted OR of breast cancer risk was 1.60 (95% CI, 1.07–2.39, P=0.022) for GSTM1+/+ and 1.29 (95% CI, 1.07–1.57, P=0.009) for the GSTM1–/– compared with the GSTM1+/– genotype, respectively. When stratified by menopausal status, the relations between the risk genotypes and breast cancer were more evident in premenopausal women, with an adjusted OR of 1.70 (P=0.035) for GSTM1+/+ and of 1.45 (P=0.004) for GSTM1–/–. On the contrary, neither GSTM1+/+ nor GSTM1–/– had an association with an increased breast cancer risk in postmenopausal Chinese women (Table 2 ).


View this table:
[in this window]
[in a new window]

  		Table 2. Distribution of GSTM1 genotypes and their risk contributions to breast cancer

On the basis of genotyping allowing the definition of GSTM1–/–, GSTM1+/–, and GSTM1+/+ genotypes, we used subpopulation analysis to explore the risk pattern of GSTM1 genotypes according to the continuous variable of age (Fig. 1C ). The plot indicated that GSTM1–/– conferred a significant increased susceptibility to breast cancer beyond GSTM1+/– in the younger subpopulations (mainly in those aged <40 yr). This result is consistent with our above findings in that the risk role of GSTM1–/– is more predominant among premenopausal Chinese women. Because of the low proportion and small sample size of the GSTM1+/+ group (n=118), the subgroup size is too small to have enough statistical power and thus would yield no meaningful statistics; therefore, we did not illustrate the GSTM1+/+ group in Fig. 1C .

SNPs and haplotypes in GSTM1 modify breast cancer risk in GSTM1+/– carriers
Because GSTM1+/– had one GSTM1 allele, the GSTM1 expression in GSTM1+/– carriers is, to a large extent, determined by the functional SNPs in the unique allele. We evaluated the risk of common SNPs in GSTM1 to breast cancer for the first time. First, four SNPs were genotyped in 583 of the 670 DNA samples from the GSTM1+/– carriers (the remaining 87 ungenotyped GSTM1+/– samples were from subjects newly recruited from January 2008). As shown in Table 3 , the frequency of the –498G allele of rs412543 (–498C>G) was higher in the control subjects than in the case patients, with a statistically significant P value even after FDR correction (P=0.048). The –498G allele still conferred protective effects on breast cancer after multivariate adjustment, with an OR of 0.50 (P=0.011). We performed genotyping for the SNP –498C>G in the remaining 87 samples. Finally, of the total 670 GSTM1+/– subjects, the frequency of the –498G allele was significantly higher in the control subjects than in the case patients, with an adjusted OR of 0.52 (P=0.007). Of the three other tested SNPs, however, no allele frequency showed a significant difference between case patients and control subjects.


View this table:
[in this window]
[in a new window]

  		Table 3. Associations of four common SNPs in the GSTM1 gene with breast cancer in the GSTM1-present population

LD calculated as r2 and D' are shown in Fig. 2 . There was no polymorphic pair of high-level LD (r2≥0.8); rs1065411 and rs1056806 were in a moderate LD (r2=0.64). Of the four SNPs, eight haplotypes with allele frequencies more than 1% were found, and three common haplotypes (frequency >5%) are listed in Supplemental Table S3. One haplotype, GTGC, which contained the protective –498G allele, is associated with a decreased susceptibility to breast cancer, with an adjusted OR of 0.53 (FDR corrected P=0.084 and adjusted P=0.040).

Validation of protective role of –498G allele in GSTM1+/+ carriers
To demonstrate that the effect of –498C>G observed in GSTM1+/– carriers was a true rather than a false-positive result, we further genotyped –498C>G in the 118 GSTM1+/+ carriers. Similar to what was observed in the GSTM1+/– subjects, the –498G allele had an association with decreased breast cancer risk. The GG plus GC genotype presented a protective effect against breast cancer compared with the CC genotype, with an adjusted OR of 0.38 (95% CI: 0.16–0.94, P=0.037) (Table 3) .

Polymorphism –498C>G affected GSTM1 promoter activity in human cell lines
Albeit, we had observed and reconfirmed the association of –498C>G with breast cancer in the case-control stage; however, we did not know whether it was a causative polymorphism or just a marker linked to other pathogenic markers. We generated two luciferase reporter gene constructs with identical sequences except at the –498 bp and transiently transfected them into different human cell lines. In most of the cell lines, such as MCF7, HEK-293T, and HeLa cells, the promoter activity of the minor –498G allele was significantly reduced by 30–40% compared with that of the major –498C allele. Such a tendency was also observed in MDA-MB-231 and SK-BR-3 cell lines, but without statistical significance (Fig. 3A ).

Polymorphism –498C>G affected promoter activity by allelic-specific binding of the AP-2 nuclear protein
We next performed EMSA to investigate whether the different promoter activity between the –498C and –498G alleles was attributable to their different affinity to bind transcription factor. As shown in Fig. 3B , a much clearer DNA-protein complex was detected with the –498C probe than with the –498G probe. Competition experiments highlighted the fact that the –498C-band could be competed away by a 100-fold molar excess of the unlabeled same type probe but not by the same concentration of unlabeled –498G type probe. Because in silico analysis was used to identify the AP-2 protein as a potential binding partner of the –498-containing sequence, we used a specific AP-2 probe and nuclear protein extracts from AP-2 protein-rich cells (HeLa, ZR-75-1, and MCF7) and observed the DNA-protein band on the same location as with use of the –498C/G probe to determine whether the bands we detected were an oligonucleotide-AP-2 complex,. The results clearly showed an improved affinity of the –498C rather than the –498G allele to specifically bind AP-2 and also indicated that the AP-2 protein bound to the labeled probe and formed two complexes. The two bands might correspond to a full-length AP-2 and a proteolytic cleavage product, respectively (30) or the two complexes represent different AP-2 isoforms.

Effects of different AP-2 isoforms on the genetic variant –498C>G and binding of the AP-2{alpha} transcription factor to the GSTM1 promoter in vivo
Expression levels of AP-2 were assayed at the mRNA level by RT-PCR in a panel of human cells as well as normal breast tissue. AP-2 isoforms transcripts were present to a variable degree (Fig. 3C ). Most tumor-derived lines expressed AP-2{alpha} and AP-2{gamma}. AP-2β was only detected in ZR-75-1 and T-47D. Noticeably, normal mammary tissue could express a moderate level of AP-2{alpha} and a low level of AP-2β/{gamma}. Because the expressions of the AP-2 isoforms in HBL-100 were very weak, we chose this cell line for further cotransfection assays with AP-2 expression vectors. Different AP-2 isoforms conferred varied effects on the GSTM1 promoter with –498C>G (Fig. 3D ). Generally, cotransfection of the AP-2 expression plasmids remarkably increased the GSTM1 promoter transcriptional activity in a dose-dependent manner. Treatment with an AP-2{alpha} vector could induce significant transcriptional activity of the –498C promoter as much as 2-fold over that of the –498G promoter. Unlike AP-2{alpha}, a high concentration of AP-2β achieved similar promoter activity for –498C and –498G promoters, although a low concentration of AP-2β showed a high affinity to the –498C promoter. In terms of AP-2{gamma}, no difference in the luciferase activity could be observed between the –498C and –498G promoter after AP-2{gamma} expression vector treatment.

The results from the in vitro EMSA assay and the AP-2 isoform study supported the fact that the –498C allele had a higher affinity to the nuclear protein AP-2{alpha} than the –498G allele. To further verify that the nuclear protein was indeed AP-2{alpha} and to determine the presence on the GSTM1 promoter in vivo, a ChIP assay was performed with lysates prepared from MCF7 cells (a cell line with the –498C-GSTM1+/– genotype), using either the AP-2{alpha} antibody or IgG. As shown in Fig. 3E , PCR amplification of the GSTM1 promoter from purified DNA showed that the product was detected only in the ChIP assay using the AP-2{alpha} antibody but not that using IgG. These findings suggest that AP-2{alpha} indeed binds this promoter region of GSTM1 in vivo.

Silencing of AP-2{alpha} expression by AP-2{alpha}-specific siRNA down-regulated GSTM1
As indicated above, treatment with the AP-2{alpha} vector could induce significant transcriptional activity of the GSTM1 promoter. To further corroborate whether AP-2{alpha} positively regulated GSTM1 expression, we used the AP-2{alpha}-specific siRNA approach to silence AP-2{alpha} expression. MCF7 and HeLa cells (both expressing endogenous AP-2{alpha} and GSTM1) were transiently transfected with either the negative control or the AP-2{alpha}-specific siRNA. Consistent with the finding from the promoter activity study, AP-2{alpha} siRNA blocked AP-2{alpha} expression both in mRNA and protein levels and accordingly down-regulated GSTM1 mRNA compared with that in control subjects (Fig. 4A-D ). These results suggest that there is a regulative relation between AP-2{alpha} and GSTM1 and that the GSTM1 gene is likely to be a downstream target of AP-2{alpha}. Here, we did not perform immunoblots to detect the GSTM1 protein expression because of the high degree of sequence homology within the GSTM family. Most commercial GSTM antibodies are not specifically anti-GSTM1. However, with the use of GSTM1-specific primers for PCR this difficulty can be avoided.


Figure 4
View larger version (15K):
[in this window]
[in a new window]

  		Figure 4. Silencing of AP-2{alpha} down-regulates GSTM1 expression and differential expression of GSTM1 in normal breast tissues according to GSTM1 genotypes. A–C) RT-PCR analysis of AP-2{alpha} and GSTM1 expression levels (A), and real-time analysis of AP-2{alpha} (B) and GSTM1 (C) expression levels were performed in AP-2{alpha}-specific siRNA-transfected MCF7 and HeLa cells. Expression of GAPDH was used for normalization, and relative mRNA expression of AP-2{alpha} and GSTM1 was calculated relative to GAPDH. Data are means ± SD from 3 independent experiments. *P < 0.05 vs. negative control group; Games-Howell test. MCF7 cells have a –498C-GSTM1+/– genotype; HeLa cells have a –498CG-GSTM1+/+ genotype. D) Western blot results of the AP-2{alpha}-specific siRNA-transfected HeLa cells by using the anti-AP-2{alpha} antibody. The GAPDH protein was detected for normalization. E) Differential expression of GSTM1 in normal breast tissue according to genotype. Box plots of GSTM1 mRNA expression according to GSTM1 genotype (null-present polymorphism and –498C>G polymorphism) are shown. GAPDH was used for normalization. Differences in GSTM1 expression among the four groups were analyzed by Kruskal-Wallis tests. Comparison between –498G-GSTM1+/– group and –498C-GSTM1+/– group was performed using the Games-Howell procedure.

–498C>G genotype altered GSTM1 gene expression ex vivo
Thereafter, we tried to scrutinize the association between GSTM1 genotypes and GSTM1 expressive phenotypes in 75 normal breast tissue specimens. In our samples, 43 were GSTM1–/–, 4 had the –498G allele GSTM1+/–, 23 had the –498C allele GSTM1+/–, and 5 were GSTM1+/+. No individual was identified as –498GG GSTM1+/+ (3 had –498CC and 2 had –498GC), and we combined the GSTM1+/+ samples as a whole. We observed that there is a gene-dosage relation between GSTM1 expression and genotypes: GSTM1 expression increases from GSTM1–/–, –498G-allele GSTM1+/–, and –498C-allele GSTM1+/– to GSTM1+/+ (global P<1x10–10) (Fig. 4E ). Among the samples from women with a GSTM1+/– genotype, the individuals with a –498G-GSTM1+/– genotype displayed reduced expression of GSTM1 mRNA compared with those with a –498C-GSTM1+/– genotype (Games-Howell test, P=0.048) (Fig. 4E ).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
A large number of studies have been conducted to scrutinize the association between breast cancer and a GSTM1-null polymorphism but with conflicting results (7 , 28 , 31) . The inconsistency may be caused by the underpowered sample size and selection bias. In the present hospital-based study, we observed that a GSTM1-null genotype confers an increased risk of breast cancer in the Chinese population. When stratified by menopausal status, a more pronounced increased risk is observed in premenopausal women rather than in postmenopausal women. These findings are in line with the previous reports from Asia in Korean (29) and Indian (15) women. It seems that a significantly increased risk of premenopausal breast cancer is associated with the GSTM1 genotype in Asian women. Epidemiological studies have shown the relation between earlier age of onset and hereditary breast cancers, implying that genetically susceptible individuals may develop cancer at an earlier age than those without the susceptibility. For breast cancer in particular, which may be related to exogenous or endogenous exposures resulting in electrophilic intermediates or free radicals, lack of GSTM1 or an aberrant function of GSTM1 may increase susceptibility, with greater DNA damage leading to carcinogenesis at an earlier age (32) . This conjecture is supported by our current study and others reporting an increased frequency of the GSTM1–/– genotype in premenopausal breast cancer patients who, on average, are younger than most women who develop breast cancer.

Interestingly, in white women, the significant association between GSTM1–/– and breast cancer is frequently observed in postmenopausal women rather than in premenopausal women (33 , 34) . The reverse findings between the two ethnicities suggest an ethnicity-specific heterogeneity regarding the role of GSTM1 in breast cancer. There are also other explanations. For instance, the distinct inclusion/exclusion criteria between studies could have affected the outcomes. We considered the estrogen-related environmental confounders in our analysis but excluded the influence of cigarette smoking and alcohol because most Asian women are nonsmokers and nondrinkers, whereas women’s exposure to cigarette smoking and alcohol is more common in the West, and cigarette smoke and alcohol contain a wide variety of potentially carcinogenic compounds, some of which are substrates of GSTM1. In addition, recombination events in different populations and epistasis, e.g., due to degree of expression of other metabolizing enzymes, may modify risk conferred by the GSTM1-null genotype. The potential ethnicity-specific role of GSTM1 in breast cancer also raises a new question of whether the dynamic expression, expressing abundance, and biological function of GSTM1 between Asian women and white women are the same. Data from future in-depth research may further elucidate this issue.

Other investigators have conducted meta-analyses to obtain a global view, and the results have alternately confirmed or refused the association between GSTM1–/– and breast cancer (7 , 28 , 31) . By integrating the accumulating evidence, we updated the meta-analysis, and our newest results offer reliable evidence that GSTM1–/– does confer susceptibility to breast cancer. The results of this meta-analysis are convincing because no significant heterogeneity among the 41 studies and no potential publication bias are observed. Similar GSTM1–/– genotyping approaches (39 of 41 used a PCR-electrophoresis-based method and 2 of 41 used a quantitative PCR-based method) warrant the comparability as well. Moreover, our stratified subanalysis indicated that GSTM1–/– is significantly associated with breast cancer in white and Asian women but not in the black population. This finding again suggests a population/ethnicity-specific role for GSTM1. If this is true, the overall OR for all the ethnicities could be underestimated and inappropriate.

Notably, the prevalent genotyping assay of GSTM1 only distinguishes GSTM1–/– from GSTM1-present individuals. Roodi et al. (10) reported an unexpected risk for GSTM1+/+ in breast carcinogenesis, although their study was underpowered because of the small sample size, insufficient statistical power, and a low-sensitivity genotyping approach. In our study of a large number of samples, we observed the elevated likelihood of breast cancer risk both in GSTM1+/+ and GSTM1–/– carriers in a comparison with GSTM1+/– carriers. In addition, we found that the HWE criteria were fulfilled in the control group but not in the case group. Because our genotyping strategy warrants the genotyping data to be accurate to the largest extent, the deviation in the case group may provide further evidence for the existence of an association between disease and polymorphism (35 , 36) . Given that GSTM1+/+ is a risk genotype, combining it with the protective GSTM1+/– genotype would surely compromise the true results, which might explain the obscure and weak association between GSTM1 and breast cancer in most one-institute small sample size studies.

The paradoxical findings that both low-level GSTM1 (GSTM1–/–) and high-level GSTM1 (GSTM1+/+) are risk factors require an explanation. It is speculated that glutathione (GSH), the substrate of GSTM1, may be a causal force (10) . Because GSH depletion to ~30% of the total GSH level can impair the conjugation defense against toxic actions (37) , the overactivity of GST conjugation may lead to GSH depletion and thereby become counterproductive. We conjecture that appropriate expression of GSTM1 is, at least partially, involved in the balance of breast cancer risk and protection. Not only the null activity of GSTM1 leading to insufficient metabolism of carcinogen but also the overactivity of GSTM1 leading to GSH depletion would result in carcinogenesis. Regrettably, because most investigators take for granted that a higher expression of GSTM1 will offer a more protective effect, most studies do not specifically distinguish GSTM1+/+ from GSTM1-present individuals. In view of this fact, the complicated role of GSTM1 in breast tumorigenesis as well as in other malignancies has been addressed rarely in case-control studies.

Is there any evidence to support the U-shaped effect of GSTM1 on breast carcinogenesis as we observed in the association study? The functional genetic variants within GSTM1, if any, might provide a potential clue. Thus far, no researchers have systemically evaluated the GSTM1 genetic variants in the context of breast cancer susceptibility. Only a missense SNP in exon 7, referred to as GSTM1*A/*B (rs1065411, Lys173Asn), has been investigated, although it does not affect the enzyme function (38) . In the present study, a new SNP, rs412543 (–498C>G), was identified as being remarkably associated with breast cancer in GSTM1+/– carriers as well as GSTM1+/+ carriers; –498C confers a risk relative to –498G. As SNPs located in the promoter region usually modulate gene transcription by interacting with trans-acting elements, we subsequently evaluated the biological function of –498C>G on GSTM1 expression. Bioinformatics indicates that –498C>G is located at the sixth base in the sequence GCCCT(C/G)GG, which is in the core recognition site of AP-2 (22 , 24) . This SNP thereby may modify the binding capability of AP-2 to the GSTM1 promoter region. We demonstrated that the –498G allele decreases gene transcription by 30–40% via reducing the DNA-binding affinity of AP-2. Our luciferase assays in vitro and GSTM1 mRNA analysis in human normal breast tissues consistently showed a significantly decreased transcriptional activity in the –498G allele compared with the –498C allele. It would be expected that individuals carrying the different –498C>G alleles have differential expression of the GSTM1 enzyme over their lifetimes and therefore may have a distinct susceptibility to develop cancer. The –498C allele with enhanced GSTM1 expression is, however, associated with increased breast cancer susceptibility, suggesting a deleterious effect of overactivity of GSTM1. Moreover, AP-2{alpha} predominantly provides differential regulation of –498C>G, whereas AP-2β and AP-2{gamma} show similar effects on the two alleles of this polymorphism. For this reason, we speculate that –498C>G would exhibit its maximum biological effect in normal mammary tissue, which expresses a high level of AP-2{alpha} and weak levels of AP-2β and AP-2{gamma}. Although the in vitro EMSA assay and AP-2 isoforms research identify the AP-2{alpha} protein as the key moderator of –498C>G, little is known of the regulative role of AP-2{alpha} in GSTM1 expression. We further demonstrated that AP-2{alpha} either binds directly or as part of a heterogeneous complex to the endogenous GSTM1 promoter using a ChIP assay. Furthermore, down-regulation of AP-2{alpha} by siRNA inhibits GSTM1 expression. These data as summarized above provide strong evidence that AP-2{alpha} has an important function in the regulation of GSTM1 expression as well as in the modulation of –498C>G biological function.

Therefore, it is reasonable to describe a U-shaped effect of GSTM1 on breast cancer susceptibility. GSTM1–/–, –498G-allele GSTM1+/–, –498C-allele GSTM1+/–, –498GC GSTM1+/+, and –498CC GSTM1+/+ represent ~0, 60–70, 100, 160, and 200% expression of GSTM1, respectively (expression of –498C-allele GSTM1+/– as reference). Among these, GSTM1–/– and –498CC GSTM1+/+ are both high-risk genotypes; the –498G-allele GSTM1+/– exerts the most protective effect. From null to 60–70% expression, we observe a gene-dosage effect with decreasing breast cancer risk, whereas from 60–70 to 200% expression, we observe a tendency for the reverse gene-dosage effect with increasing breast cancer risk. The novel U-shaped effect of GSTM1 challenges the old research methodology and conclusion, leading to a more objective conception of GSTM1 in breast carcinogenesis.

As a hospital-based study, this investigation has inevitable limitations, such as population selection bias. Our initial findings, especially the role of –498C>G in breast cancer, should be further verified in other populations. It is also possible that other functional cis-acting elements upstream or downstream of the –498C>G-containing sequences cooperate with this SNP (39) . Moreover, because GSTs have overlapping substrate specificities, simultaneous determination of the combined GST genotypes seems to be required for more reliable interpretation. Despite these limitations, the results of the present study are particularly interesting. We propose a novel two-sided role for GSTM1 in breast cancer susceptibility and emphasize that the moderate activity of GSTM1 (~60–70% expression of one-allele GSTM1) could suffice for protection against breast cancer; null activity and overactivity of GSTM1 are both disadvantageous.


   ACKNOWLEDGMENTS
 
This research was supported in part by grants from the National Basic Research Program of China (2006CB910501), National Natural Science Foundation of China (30371580 and 30572109), and the Shanghai Science and Technology Committee (06DJ14004 and 06DZ19504) We thank the women in the study for their willingness to cooperate. We also greatly appreciate Lei Wang (Pharmacology Research Center, Shanghai Medical College) for his excellent technical assistance in the chromatin immunoprecipitation assay.


   FOOTNOTES
 
1 These authors contributed equally to this work.

Received for publication October 19, 2008. Accepted for publication January 29, 2009.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Yager, J. D., Davidson, N. E. (2006) Estrogen carcinogenesis in breast Cancer N. Engl. J. Med. 354,270-282[Free Full Text]
  2. Cavalieri, E. L., Stack, D. E., Devanesan, P. D., Todorovic, R., Dwivedy, I., Higginbotham, S., Johansson, S. L., Patil, K. D., Gross, M. L., Gooden, J. K., Ramanathan, R., Cerny, R. L., Rogan, E. G. (1997) Molecular origin of cancer: catechol estrogen-3,4-quinones as endogenous tumor initiators. Proc. Natl. Acad. Sci. U. S. A. 94,10937-10942[Abstract/Free Full Text]
  3. Dunning, A. M., Healey, C. S., Pharoah, P. D., Teare, M. D., Ponder, B. A., Easton, D. F. (1999) A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol. Biomarkers Prev. 8,843-854[Abstract/Free Full Text]
  4. Xiong, P., Bondy, M. L., Li, D., Shen, H., Wang, L. E., Singletary, S. E., Spitz, M. R., Wei, Q. (2001) Sensitivity to benzo(a)pyrene diol-epoxide associated with risk of breast cancer in young women and modulation by glutathione S-transferase polymorphisms: a case-control study. Cancer Res. 61,8465-8469[Abstract/Free Full Text]
  5. Xu, X., Gammon, M. D., Zeisel, S. H., Lee, Y. L., Wetmur, J. G., Teitelbaum, S. L., Bradshaw, P. T., Neugut, A. I., Santella, R. M., Chen, J. (2008) Choline metabolism and risk of breast cancer in a population-based study. FASEB J. 22,2045-2052[Abstract/Free Full Text]
  6. Parl, F. F. (2005) Glutathione S-transferase genotypes and cancer risk. Cancer Lett. 221,123-129[CrossRef][Medline]
  7. Vogl, F. D., Taioli, E., Maugard, C., Zheng, W., Pinto, L. F., Ambrosone, C., Parl, F. F., Nedelcheva-Kristensen, V., Rebbeck, T. R., Brennan, P., Boffetta, P. (2004) Glutathione S-transferases M1, T1, and P1 and breast cancer: a pooled analysis. Cancer Epidemiol. Biomarkers Prev. 13,1473-1479[Abstract/Free Full Text]
  8. Wilson, M. H., Grant, P. J., Hardie, L. J., Wild, C. P. (2000) Glutathione S-transferase M1 null genotype is associated with a decreased risk of myocardial infarction. FASEB J. 14,791-796[Abstract/Free Full Text]
  9. Brasch-Andersen, C., Christiansen, L., Tan, Q., Haagerup, A., Vestbo, J., Kruse, T. A. (2004) Possible gene dosage effect of glutathione-S-transferases on atopic asthma: using real-time PCR for quantification of GSTM1 and GSTT1 gene copy numbers. Hum. Mutat. 24,208-214[CrossRef][Medline]
  10. Roodi, N., Dupont, W. D., Moore, J. H., Parl, F. F. (2004) Association of homozygous wild-type glutathione S-transferase M1 genotype with increased breast cancer risk. Cancer Res. 64,1233-1236[Abstract/Free Full Text]
  11. Santen, R. J., Mansel, R. (2005) Benign breast disorders. N. Engl. J. Med. 353,275-285[Free Full Text]
  12. Yu, K. D., Di, G. H., Wu, J., Lu, J. S., Shen, K. W., Shen, Z. Z., Shao, Z. M. (2007) Development and trends of surgical modalities for breast cancer in china: a review of 16-year data. Ann. Surg. Oncol. 14,2502-2509[CrossRef][Medline]
  13. Antoniou, A. C., Easton, D. F. (2003) Polygenic inheritance of breast cancer: Implications for design of association studies. Genet. Epidemiol. 25,190-202[CrossRef][Medline]
  14. Li, W. F., Hu, Z., Rao, N. Y., Song, C. G., Zhang, B., Cao, M. Z., Su, F. X., Wang, Y. S., He, P. Q., Di, G. H., Shen, K. W., Wu, J., Lu, J. S., Luo, J. M., Liu, X. Y., Zhou, J., Wang, L., Zhao, L., Liu, Y. B., Yuan, W. T., Yang, L., Shen, Z. Z., Huang, W., Shao, Z. M. (2008) The prevalence of BRCA1 and BRCA2 germline mutations in high-risk breast cancer patients of Chinese Han nationality: two recurrent mutations were identified. Breast Cancer Res. Treat. 110,99-109[CrossRef][Medline]
  15. Chacko, P., Joseph, T., Mathew, B. S., Rajan, B., Pillai, M. R. (2005) Role of xenobiotic metabolizing gene polymorphisms in breast cancer susceptibility and treatment outcome. Mutat. Res. 581,153-163[Medline]
  16. Harada, S., Tachikawa, H., Kawanishi, Y. (2001) Glutathione S-transferase M1 gene deletion may be associated with susceptibility to certain forms of schizophrenia. Biochem. Biophys. Res. Commun. 281,267-271[CrossRef][Medline]
  17. Liu, J. B., Li, M., Chen, H., Zhong, S. Q., Yang, S., Du, W. D., Hao, J. H., Zhang, T. S., Zhang, X. J., Zeegers, M. P. (2007) Association of vitiligo with HLA-A2: a meta-analysis. J. Eur. Acad. Dermatol. Venereol. 21,205-213[Medline]
  18. Fryer, A. A., Zhao, L., Alldersea, J., Pearson, W. R., Strange, R. C. (1993) Use of site-directed mutagenesis of allele-specific PCR primers to identify the GSTM1 A, GSTM1 B, GSTM1 A,B and GSTM1 null polymorphisms at the glutathione S-transferase, GSTM1 locus. Biochem. J. 295(Pt. 1),313-315[Medline]
  19. Gabriel, S. B., Schaffner, S. F., Nguyen, H., Moore, J. M., Roy, J., Blumenstiel, B., Higgins, J., DeFelice, M., Lochner, A., Faggart, M., Liu-Cordero, S. N., Rotimi, C., Adeyemo, A., Cooper, R., Ward, R., Lander, E. S., Daly, M. J., Altshuler, D. (2002) The structure of haplotype blocks in the human genome. Science 296,2225-2229[Abstract/Free Full Text]
  20. Moshynska, O., Moshynskyy, I., Misra, V., Saxena, A. (2005) G125A single-nucleotide polymorphism in the human BAX promoter affects gene expression. Oncogene 24,2042-2049[CrossRef][Medline]
  21. Danilition, S. L., Frederickson, R. M., Taylor, C. Y., Miyamoto, N. G. (1991) Transcription factor binding and spacing constraints in the human β-actin proximal promoter. Nucleic Acids Res. 19,6913-6922[Abstract/Free Full Text]
  22. Mitchell, P. J., Wang, C., Tjian, R. (1987) Positive and negative regulation of transcription in vitro: enhancer-binding protein AP-2 is inhibited by SV40 T antigen. Cell 50,847-861[CrossRef][Medline]
  23. Zhang, X., Miao, X., Tan, W., Ning, B., Liu, Z., Hong, Y., Song, W., Guo, Y., Zhang, X., Shen, Y., Qiang, B., Kadlubar, F. F., Lin, D. (2005) Identification of functional genetic variants in cyclooxygenase-2 and their association with risk of esophageal cancer. Gastroenterology 129,565-576[Medline]
  24. Imagawa, M., Chiu, R., Karin, M. (1987) Transcription factor AP-2 mediates induction by two different signal-transduction pathways: protein kinase C and cAMP. Cell 51,251-260[CrossRef][Medline]
  25. Li, D. Q., Wang, L., Fei, F., Hou, Y. F., Luo, J. M., Zeng, R., Wu, J., Lu, J. S., Di, G. H., Ou, Z. L., Xia, Q. C., Shen, Z. Z., Shao, Z. M. (2006) Identification of breast cancer metastasis-associated proteins in an isogenic tumor metastasis model using two-dimensional gel electrophoresis and liquid chromatography-ion trap-mass spectrometry. Proteomics 6,3352-3368[CrossRef][Medline]
  26. Pfaffl, M. W. (2001) A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29,e45[Abstract/Free Full Text]
  27. Bonetti, M., Gelber, R. D. (2000) A graphical method to assess treatment-covariate interactions using the Cox model on subsets of the data. Stat. Med. 19,2595-2609[CrossRef][Medline]
  28. Egan, K. M., Cai, Q., Shu, X. O., Jin, F., Zhu, T. L., Dai, Q., Gao, Y. T., Zheng, W. (2004) Genetic polymorphisms in GSTM1, GSTP1, and GSTT1 and the risk for breast cancer: results from the Shanghai Breast Cancer Study and meta-analysis. Cancer Epidemiol. Biomarkers Prev. 13,197-204[Abstract/Free Full Text]
  29. Park, S. K., Yim, D. S., Yoon, K. S., Choi, I. M., Choi, J. Y., Yoo, K. Y., Noh, D. Y., Choe, K. J., Ahn, S. H., Hirvonen, A., Kang, D. (2004) Combined effect of GSTM1, GSTT1, and COMT genotypes in individual breast cancer risk. Breast Cancer Res. Treat. 88,55-62[CrossRef][Medline]
  30. Jiang, J. G., DeFrances, M. C., Machen, J., Johnson, C., Zarnegar, R. (2000) The repressive function of AP2 transcription factor on the hepatocyte growth factor gene promoter. Biochem. Biophys. Res. Commun. 272,882-886[CrossRef][Medline]
  31. Sull, J. W., Ohrr, H., Kang, D. R., Nam, C. M. (2004) Glutathione S-transferase M1 status and breast cancer risk: a meta-analysis. Yonsei Med. J. 45,683-689[Medline]
  32. Ambrosone, C. B., Freudenheim, J. L., Graham, S., Marshall, J. R., Vena, J. E., Brasure, J. R., Laughlin, R., Nemoto, T., Michalek, A. M., Harrington, A., Ford, T. D., Shields, P. G. (1995) Cytochrome P4501A1 and glutathione S-transferase (M1) genetic polymorphisms and postmenopausal breast cancer risk. Cancer Res. 55,3483-3485[Abstract/Free Full Text]
  33. Helzlsouer, K. J., Selmin, O., Huang, H. Y., Strickland, P. T., Hoffman, S., Alberg, A. J., Watson, M., Comstock, G. W., Bell, D. (1998) Association between glutathione S-transferase M1, P1, and T1 genetic polymorphisms and development of breast cancer. J. Natl. Cancer Inst. 90,512-518[Abstract/Free Full Text]
  34. Mitrunen, K., Jourenkova, N., Kataja, V., Eskelinen, M., Kosma, V. M., Benhamou, S., Vainio, H., Uusitupa, M., Hirvonen, A. (2001) Glutathione S-transferase M1, M3, P1, and T1 genetic polymorphisms and susceptibility to breast cancer. Cancer Epidemiol. Biomarkers Prev. 10,229-236[Abstract/Free Full Text]
  35. Gyorffy, B., Kocsis, I., Vasarhelyi, B. (2004) Missed calculations and new conclusions: re-calculation of genotype distribution data published in Journal of Investigative Dermatology, 1998–2003. J. Invest. Dermatol. 122,644-646[CrossRef][Medline]
  36. Deng, H. W., Chen, W. M., Recker, R. R. (2001) Population admixture: detection by Hardy-Weinberg test and its quantitative effects on linkage-disequilibrium methods for localizing genes underlying complex traits. Genetics 157,885-897[Abstract/Free Full Text]
  37. Reed, D. J. (1990) Glutathione: toxicological implications. Annu. Rev. Pharmacol. Toxicol. 30,603-631[CrossRef][Medline]
  38. Widersten, M., Pearson, W. R., Engstrom, A., Mannervik, B. (1991) Heterologous expression of the allelic variant mu-class glutathione transferases mu and psi. Biochem. J. 276(Pt 2),519-524[Medline]
  39. Moyer, A. M., Salavaggione, O. E., Hebbring, S. J., Moon, I., Hildebrandt, M. A., Eckloff, B. W., Schaid, D. J., Wieben, E. D., Weinshilboum, R. M. (2007) Glutathione S-transferase T1 and M1: gene sequence variation and functional genomics. Clin. Cancer Res. 13,7207-7216[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
K. W. Reding, N. S. Weiss, C. S. Carlson, C. Chen, C. I. Li, K. E. Malone, K. E. Thummel, and F. M. Farin

Cancer Epidemiol. Biomarkers Prev., October 1, 2009; 18(10): 2793 - 2793.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.08-124073v1
23/7/2274    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yu, K.-D.
Right arrow Articles by Shao, Z.-M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yu, K.-D.
Right arrow Articles by Shao, Z.-M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS