Association of repeat polymorphisms in the estrogen receptors alpha, beta (ESR1, ESR2) and androgen receptor (AR) genes with the occurrence of breast cancer

ARTICLE IN PRESS THE BREAST The Breast 17 (2008) 159–166 www.elsevier.com/locate/breast

Original Article

Association of repeat polymorphisms in the estrogen receptors alpha, beta (ESR1, ESR2) and androgen receptor (AR) genes with the occurrence of breast cancer A. Tsezoua, M. Tzetisb,, C. Gennatasc, E. Giannatoud, A. Pampanosb, G. Malamisb, E. Kanavakisb, S. Kitsioub a

Laboratory of Cytogenetics and Molecular Genetics, Medical School, University of Thessalia, University Hospital of Larissa, Larissa, Greece b Department of Medical Genetics, Medical School, University of Athens, Thivon and Levadias, 11257 Athens, Greece c Medical School, Areteion Hospital, University of Athens, Athens, Greece d 2nd Department of Pediatrics, University of Athens, Athens, Greece Received 29 March 2007; received in revised form 10 August 2007; accepted 20 August 2007

Abstract Genetic variation in genes involved in estrogen biosynthesis, metabolism and signal transduction have been suggested to play a role in breast cancer. To determine the possible contribution of genetic variation in the ESR1 (ER-a), ESR2 (ER-b) and AR genes in breast cancer risk the 1174(TA)7–27, c. 1092+3607(CA)1026 and c. 172(CAG)640 repeat variants were studied in a case-control study of 79 women with sporadic breast cancer and 155 controls. No significant difference was observed in the frequency distribution of 1174(TA)727 in the ESR1 gene between patients and controls, while a significant difference was observed for repeat polymorphisms c. 1092+3607(CA)1026 in the ESR2 gene and c. 172(CAG)640 in the AR gene (pp0.0001). A significantly decreased odds ratio (OR) for breast cancer risk was observed in individuals having the LL and the SL genotypes for both the ESR2 (OR ¼ 0.010, 95% CI 0.003–0.036, po0.001; OR ¼ 0.013, 95% CI 0.004–0.040, po0.0001, respectively) and the AR gene (OR ¼ 0.040, 95% CI 0.011–0.138, po0.0001; OR ¼ 0.189, 95% CI 0.10–0.359, po0.0001, respectively), compared to SS genotype. The protective effect of these genotypes remained evident even after adjustment for various risk factors (BMI, age, age at menarche and menopause, family history). In conclusion, an association for breast cancer risk between short (SS) alleles for the repeat variants of the ESR2 and AR genes was found in women of Greek descent. r 2007 Elsevier Ltd. All rights reserved. Keywords: Breast cancer; Estrogen/androgen receptors; Repeat polymorphisms; Steroid hormones

Introduction Breast cancer is the most common cause of cancer death and the most common form of cancer in women with a 9% incidence of being diagnosed during a lifetime.1 Although family history is a well-known risk factor for breast cancer with estimation that 5–10% of all breast carcinomas being inherited,2 approximately 90% of cases are sporadic.3 Similar to other complex diseases, a combination of genetic factors as well as exposure to endogenous and exogenous Corresponding author. Tel.: +30 210 7467460; fax: +30 210 7795553.

E-mail address: [email protected] (M. Tzetis). 0960-9776/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.breast.2007.08.007

factors contribute to an increased risk for breast cancer development.4 Genetic variation in genes involved in estrogen synthesis, metabolism and signal transduction have been suggested to play a role. Estrogen influences the growth, differentiation and function of breast tissue, exerting its biological effect through binding to estrogen receptors (ERs).5 ERs belong to a family of transcription factors, the nuclear receptor super family, responsible for mediating the effects of steroids on development, reproduction, proliferation, cellular homeostasis and gene expression.6,7 Two isoforms of ER showing 47% identity have been cloned, ESR1 and ESR2. The ESR1 gene is located on chromosome 6q25-27, consists of eight exons

ARTICLE IN PRESS 160

A. Tsezou et al. / The Breast 17 (2008) 159–166

and spans more than 140 kb, while the ESR2 gene is located on chromosome 14q22-24, comprises eight exons and spans approximately 40 kb.7–9 The presence of ERs in breast tumor tissue is considered an important prognostic factor that correlates with higher survival rates and lower risk of relapse.9–11 Like ERs, the AR gene also belongs to the steroid hormone family, is located on chromosome Xq11-12 and consists of a highly polymorphic CAG repeat in exon 1 that encodes a polyglutamine stretch.12 The role of endogenous androgens in breast cancer development is less clear and it has been suggested that androgens may influence breast cancer risk indirectly through their conversion to estradiol, or directly by binding to AR and their promoting or opposing breast cell growth, or even by binding to ESR1 and directly stimulating breast cell proliferation.13,14 The association of genetic polymorphisms in the ESR1, ESR2 and AR genes and the risk of breast cancer has been of increasing interest.15 The ESR1 and ESR2 genes contain the 1174(TA)n and c. 1092+3607(CA)n repeat polymorphisms, respectively, whose association with breast cancer has not been investigated in homogeneous samples. The c. 172(CAG)n repeat polymorphism of the AR gene has been proposed to be a modulator of the penetrance of BRCA1 mutations in women,16 although it has not been associated directly with the occurrence of breast cancer in younger women.17–19 Subsequent studies report conflicting results regarding the association of the c. 172(CAG)n repeat polymorphism and breast cancer risk.18,20 In the present study, we investigated the association between 1174(TA)n (ESR1), c. 1092+3607(CA)n (ESR2) and c. 172(CAG)n (AR) repeat polymorphisms and breast cancer risk in 79 cases of women with sporadic breast cancer and in 155 controls. Materials and methods Human subjects We studied 79 women with breast cancer (mean age 57.6711.49, range 28–78 years), and 155 controls (mean age 70.977.70, range 45–87 years) both groups of Greek origin. The breast cancer cases participating in the study were recruited from the Oncology Unit of the 2nd Surgical Clinic, University of Athens. The patients were followed between the years 1999 and 2003 and were ascertained for interview at the time of their diagnosis. The controls were women who attended the outpatient clinic of the hospital for various injuries during the same period and were about 14 years on average older than cases, and free of breast cancer. There were no refusals to participate either among cases or controls. Other variables Patients and controls gave their approval for blood drawing, answered a questionnaire and took part in an

interview which included information about occupation, medication use, chronic diseases, smoking and alcohol drinking habits, family history of cancer incidence, number of children and breast feeding habits. All participants were asked about their age of menarche and menopause and about estrogen replacement therapy (ERT) (if followed). Height and weight were measured and the body mass index (BMI) was computed. None of the controls had current breast cancer disease based on physical examination or other personal history of cancer while 3.7% suffered from chronic disease such as hypertension, migraine, peptic ulcer. None of the cases and only 3/155 (1.9%) of control participants were under ERT. There was no significant difference between the cases and controls regarding number of children (cases—0 children: 21.5%, 1 child: 22.8%, X2 child: 55.7%; controls—0 children: 20%, 1 child: 25%, X2 child: 54.8%), breast-feeding habits (cases, 67.7%; controls, 69%), cigarette smoking and alcohol intake. None of the 155 controls and 17/79 (21.5%) of cases had one family member with breast/ ovarian cancer, of which 12 had a first degree relative (mother or sister) and five a second or third-degree relative (aunt, 2nd cousins). Detailed clinical data of patients and control population regarding age, BMI, age of menarche and menopause are shown in Table 1. Detailed data of patients regarding the breast cancer staging such as tumor size, grade, nodal status and metastasis are shown in Table 2. The study was approved by the ethics committee of the University of Athens, and all subjects signed an informed consent form. Isolation of genomic DNA Genomic DNA was obtained from 3 ml of peripheral blood, using the BioRobots M48 System (Qiagen, Hilden, Germany) and the commercially available kit MagAttracts DNA Blood Midi M48 Kit (Qiagen). Determination of ESR1, ESR2, and AR microsatellite allele sizes Based on the sequences of the human ESR1, ESR2, and AR genes available from the GeneBank and using Primer3 software (www.justbio.com) primers were designed for the purposes of the PCR amplification and are shown in Table 3. A multiplex PCR reaction was carried out in a final volume of 25 ml containing 50 ng of genomic DNA, 5 pmol of each primer and 12.5 ml Taq polymerase Master Mix (Multiplex PCR kit—Qiagen Science, Maryland, USA). Thermal cycling conditions were as follows: 95 1C for 15 min followed by 38 cycles of 95 1C for 30 s, 58 1C for 90 s and 72 1C for 90 s with a final extension step of 72 1C for 10 min. All forward primers were fluorescently Cy5.0 labeled (Proligo LLC, Boulder, CO, USA). An aliquot of the reaction was mixed with a loading dye and 50 and 300 bp

ARTICLE IN PRESS A. Tsezou et al. / The Breast 17 (2008) 159–166

161

Table 1 Clinical data on breast cancer cases and controls Characteristics

Total, N (%)

Controls, N (%)

Cases, N (%)

p-Value

Body mass index (BMI) Underweight (BMIo18.5) Normal (BMI ¼ 18.5–24.9) Overweight (BMI ¼ 25–29.9) Obesity (BMIX30)

2 59 123 36

1 41 91 19

1 18 32 17

o0.102a

Characteristics

N (%)

Age Cases Controls

79 (33.7) 155 (66.3)

Age of menarche Cases Controls

69 (31.2) 152 (68.8)

Age of menopause Cases Controls BMI Cases Controls a

(0.9) (26.8) (55.9) (16.4)

(0.5) (27) (60) (12.5)

(1.5) (26.5) (47) (25)

Mean7SD

Range

p-Value

57.6711.5 70.977.7

28–78 45–87

o0.0001b

12.7570.2 12.8970.1

9–18 10–16

0.006b

45 (23.1) 150 (76.9)

47.670.7 48.970.3

35–55 40–57

0.034b

68 (30.9) 152 (69.1)

27.470.6 26.270.2

18–40 17–33

o0.0001b

Pearson w2. t-test.

b

Table 2 Breast cancer staging data Feature Tumor size (cm) o2.5 X2.5 p-Value (t-test) p-Value (w2-test) Grade I II III p-Value (Anova) p-Value (w2-test) Nodal status Negative Positive p-Value (t-test) p-Value (w2-test) Metastasis Yes No p-Value (t-test) p-Value (w2-test)

Number of patients (%)

ESR2 (mean7SEM)

AR (mean7SEM)

ESR1 (mean7SEM)

27 (56.3) 21 (43.8)

17.31 (70.412) 17.49 (70.379) 0.756 0.251

20.47 (70.382) 19.69 (70.460) 0.191 0.990

14.13 (70.593) 13.12 (70.685) 0.265 0.691

14 (23.3) 34 (56.7) 12 (20.0)

17.11 (70.422) 16.65 (70.318) 17.38 (70.370) 0.388 0.714

20.32 (70.530) 19.85 (70.335) 19.79 (70.654) 0.728 0.368

13.68 (70.813) 13.74 (70.543) 13.25 (70.891) 0.892 0.403

24 (43.6) 31 (56.4)

17.04 (70.411) 17.53 (70.339) 0.356 0.045

20.13 (70.377) 19.74 (70.378) 0.480 0.583

14.11 (70.643) 13.23 (70.554) 0.301 0.059

25 (36.2) 44 (63.8)

17.08 (70.401) 17.43 (70.282) 0.476 0.372

20.06 (70.423) 19.83 (70.284) 0.637 0.307

13.53 (70.632) 13.63 (70.461) 0.896 0.515

size markers (Visible Genetics, Inc.), heated at 95 1C for 5 min and cooled on ice. It was subsequently separated on a 6% denaturing polyacrylamide gel. Allele fragment sizes were determined in comparison with external size markers by the automated DNA sequencer and analyzed using the

Fragment Analysis Software (Visible Genetics, Inc). Ten cases with different known PCR product lengths for each of the three genes were directly sequenced for confirmation and were appropriately included as a quality control step for the repeat allele genotyping.

ARTICLE IN PRESS A. Tsezou et al. / The Breast 17 (2008) 159–166

162

Table 3 Sequence of primers used for microsatelite repeat PCR of ESR1, ESR2 and AR genes

body weight of 70.6711.29 kg, with a mean BMI of 27.474.90 kg/m2.

Primer name

Sequence

PCR amplicon/length

ESR1

ESR1 forward

50 -Cy5.0—AGA CGC ATG ATA TAC TTC ACC-30

ESR1 reverse ESR2 forward

50 -GTT CAC TTG GGC TAG GAT AT-30 50 -Cy5.0—GGT AAA CCA TGG TCT GTA CC-30

(TA)n repeat at 1174 bp upstream of exon 1 in the 1A promoter of the gene/176–216 bp (7–27 TA repeats)

ESR2 reverse

50 -AAC AAA ATG TTG AAT GAG TGG G-30 50 -Cy5.0—AAT CTG TTC CAG AGC GTG C-30 50 -GGA ACA GCA ACC TTC ACA G-30

The TA repeats in the ER-a gene ranged from 7 to 27 with a median length of 15. There was no significant difference in the frequency distribution of TA dinucleotide repeat polymorphism of the ER-a gene in the 79 cases (158 chromosomes) and 154 controls (308 chromosomes) (p ¼ 0.544) (Table 4). Also, no significant difference in the (TA)n repeat genotype distribution (SS, SL, and LL) was observed between cases and controls (p ¼ 0.396). When OR were adjusted for confounding factors (age, BMI, age at menarche and menopause, family history and number of children), all together no statistical significance was observed.

AR forward AR reverse

(CA)n repeat at the 30 flanking region of the gene/137–169 bp (10–26 CA repeats)

(CAG)n repeat within exon 1 of the gene/234–336 bp (6–40 CAG repeats)

ESR2

Statistical analysis Allele frequencies are summarized by the mean and standard error of the mean (SEM). Differences in allele frequencies between patients and controls were tested using Student t-test. Based on previous studies, all alleles were divided into two groups of approximately equal size; those with short and those with long alleles. Splitting of groups was undertaken with the median as the cut-off. As a result three subgroups were generated: those with two long alleles (LL), those with one short and one long allele (SL) and those with two short alleles (SS). The cut-off limits were 15 (shortp15 repeats) for ESR1, 20 (shortp20 repeats) for ESR2 and 22 (shortp22) for AR gene. Odds ratios (OR) were calculated by logistic regression analysis with 95% confidence intervals (CIs). We computed ORs using subjects homozygous for the short TA, CA and CAG alleles as the reference. We estimated unadjusted (univariate) OR for various genotypes and subsequently included in the logistic regression models variables known to be associated with breast cancer. These variables were age, BMI, age at menarche and menopause, family history, ERT, breast feeding and number of children. In all cases, a p-value of o0.05 was taken as significant and OR, 95% CI were calculated. All statistical analysis was performed by using SPSS 10.0 statistical package. Results Mean body weight for the controls was 69.879.85 kg, with a mean BMI of 26.273.05 kg/m2. Patients had a mean

The CA repeats in the ESR2 gene ranged from 10 to 27 with a median length of 20 (Supplementary Data, Fig. 1). The mean number of (CA)n repeats was higher in the control group compared to cases (po0.0001) (Table 4). A significant difference in the (CA)n repeat genotype distribution (SS, SL, and LL) was observed between patients and controls (po0.0001) (Table 5). The crude OR for breast cancer was significantly decreased in individuals with SL and LL genotypes compared to individuals with the SS genotype (OR ¼ 0.013, 95% CI 0.004–0.040, po0.0001 and OR ¼ 0.010, 95% CI 0.003–0.036, po0.001) (Table 5). When OR were adjusted for confounding factors, individuals with SL and LL genotypes had significantly decreased risk to develop breast cancer compared to the ones with SS genotype (OR ¼ 0.002, 95% CI 0.000–0.022, po0.0001 and OR ¼ 0.001, 95% CI 0.000–0.023, po0.0001) (Table 6). Table 4 Repeat polymorphism values for ESR1, ESR2 and AR genes for cases and controls Repeat polymorphisms

Mean7SEM

Repeats

p-Valuea

CA (ESR2) Cases (n ¼ 79) Controls (n ¼ 155)

17.470.21 20.970.14

10–27 13–27

o0.0001

TA (ESR1) Cases (n ¼ 79) Controls (n ¼ 154)

13.870.35 17.3870.26

7–24 9–27

0.544

CAG (AR) Cases (n ¼ 79) Controls (n ¼ 155)

20.170.22 23.770.26

10–27 6–40

o0.0001

a

2

Pearson w .

ARTICLE IN PRESS A. Tsezou et al. / The Breast 17 (2008) 159–166 Table 5 Unadjusted OR for various genotypes in ESR1, ESR2 and AR genes for cases and controls Controls (n)

OR (95% CI)

p-Value

ESR2 gene SS 71 LL 4 SL 3

16 71 67

1 (reference) 0.010 (0.003–0.036) 0.013 (0.004–0.040)

o0.0001 o0.0001

ESR1 gene SS 22 LL 15 SL 41

51 36 66

1 (reference) 0.97 (0.44–2.11) 1.44 (0.76–2.71)

0.931 0.259

AR gene SS LL SL

31 46 77

1 (reference) 0.040 (0.011–0.138) 0.189 (0.100–0.359)

o0.0001 o0.0001

Genotype

Cases (n)

51 3 24

Table 6 Adjusted OR (BMI, age, age at menarche and menopause, family history) for various genotypes Genotype

OR (95% CI)

p-Value

ESR1 gene SS LL SL

1 (reference) 0.001 (0.000–0.023) 0.002 (0.000–0.022)

o0.0001 o0.0001

AR gene SS LL SL

1 (reference) 0.089 (0.016–0.486) 0.220 (0.072–0.670)

0.005 0.008

AR gene The GAG repeats in the AR gene ranged in women from 6 to 40 with a median length of 20. The mean number of GAG repeats was higher in control women than in cases (p ¼ o0.0001) (Table 4), (Supplementary Data, Fig. 2). Comparison of the SS, LL and SL genotype groups showed that women with LL and SL genotypes had a 25and 5-fold lower risk to have breast cancer compared to individuals with SS genotype (OR ¼ 0.040, 95% CI 0.011– 0.138, po0.0001 and OR ¼ 0.189, 95% CI 0.10–0.359, po0.0001, respectively) (Table 5). When OR were adjusted for risk increasing variables, the ones with LL and SL genotypes had significantly decreased risk for breast cancer compared to the ones with SS genotype (OR ¼ 0.089, 95% CI 0.016–0.486, p ¼ 0.005 and OR ¼ 0.220, 95% CI 0.072–0.670, p ¼ 0.008, respectively) (Table 6). The frequency distribution of SS, LL and SL genotypes of the ESR1, ESR2 and AR genes was in Hardy–Weinberg equilibrium for both cases and controls. Discussion Breast cancer is the most prevalent non-skin cancer of the world and the second leading cause of cancer deaths

163

among women.1–3 Hormone related cancers such as breast, endometrial, and ovarian share the same mechanism of carcinogenesis with endogenous and exogenous hormones driving cell proliferation, and thus increasing the opportunity of accumulation of somatic mutations that occur during cell division.15,21 Approximately half of the hereditary breast cancer cases are the result of diseasecausing mutations in BRCA1 and BRCA2, or other identified cancer susceptibility genes while a large body of epidemiological and experimental data implicates estrogens and their receptors (ESR1, ESR2) and AR in the etiology of human breast cancer.2,7,15 Animal studies have demonstrated that estrogens can induce and promote mammary tumors in rodents.22 The polymorphic 1174(TA)n repeat in promoter region of the ESR1 gene has been associated with coronary heart disease, bone mineral density, familiar premature ovarian failure and endometriosis.7 While the CA repeat in the 30 non-coding portion of the ESR2 gene, c. 1092+3607(CA)n, was also suggested to be associated with bone mineral density in women and Osteoarthritis.2,7,22–24 The functional importance of these repeat nucleotide sequences has not been clarified although reports have suggested that these sequences, even when situated in the non-translated regions may influence the expression of a gene, including their potential to form alternative DNA structures—such as Z DNA—that may modulate transcriptional activity.15,25 The CAG polyglutamine stretch in the amino-terminal domain of the AR gene appears to inversely influence the function of the receptor as a transcription factor, and is assumed to be involved in interactions between the AR and different coactivators, with long repeats being inhibitory to these interactions which could explain the lower activity of the receptor. Especially among postmenopausal women circulating androgen levels appear to be associated with breast cancer risk and the hypothesis is that the shorter the length of the CAG polyglutamine stretch the greater the affinity of androgens to the AR.15 Data on the functional importance of the CAG repeats in women is sparse, but associations have been found with bone mass density and breast cancer.16,26 A recent study27 has suggested that serum androgen levels in women are associated with both the AR and ESR2 gene repeat sequences while no association was found with the TA repeats in the ESR1 gene. Fewer than 22CAG repeats were associated with higher levels of androgens, opposite of what has been reported for men, suggesting that the major influence of AR on androgen production in women is stimulatory rather than inhibitory. Androgens could provide a large pool of substrate for conversion to estrogen via the action of aromatase in breast tissue.15,21 Less than 20CA repeats in the ESR2 gene lead to a less active receptor displayed by higher serum levels of free testosterone and low levels of sex steroid hormone-binding globulin (SHBG).27 Lower SHBG would result in increased risk of breast cancer through higher estrogen bioavailability.7

ARTICLE IN PRESS 164

A. Tsezou et al. / The Breast 17 (2008) 159–166

A study on three common sequence variants in the ESR2 gene in cases of sporadic and familial cancer found no overall significant association for any of the SNPs studied.28 While among Chinese women with high levels and long term exogenous estrogen exposure an association was found with one SNP (rs1256054; L392L; C/G), located in exon 7 of ESR2 gene in an exonic splicing enhancer motif.4 Several epidemiological studies have examined the association between the length of the CAG repeat polymorphism in the AR gene and breast cancer risk. An increased breast cancer risk has been observed among women with the long CAG allele in three previous studies,16,29,30 whereas only a weak, non-significant association was observed in four.17,31–33 The three positive studies were conducted among BRCA1 mutation carriers,16 postmenopausal women,29 and women with a first-degree family history of breast cancer.30 Positive association with repeat length genotypes of AR and breast cancer risk was found in the Philippines particularly among women 450 years old.34 While no association was found between exon 1 AR CAG repeat and risk of breast cancer in a large population of postmenopausal Caucasian women or amongst BRCA1 or BRCA2 mutation carriers.14,18 On the other hand, short CAG repeat length polymorphism of the AR gene has been strongly associated with breast carcinomas and atypical hyperplasias.20 Two other studies have reported that the length of CAG repeats either in one allele or in both alleles was inversely correlated with the histological grade of breast cancer with every CAG repeat increase resulting in a reduction in the risk of breast cancer death.35,36 To our knowledge, there are no previous reports studying the combined association between ESR1, ESR2 and AR genes microsatellite repeat polymorphisms with breast cancer risk. Our case-control study included women with sporadic breast cancer, both pre- and postmenopausal, with the majority (60/79) being postmenopausal and an age range between 28 and 78 years (mean 57.6711.5). The controls were randomly recruited from the same hospital, specifically from the outpatient clinic due to other problems not related to breast cancer. They completed the same questionnaire as the cases, including questions related to family history of cancer incidence. The cases and controls are of the same ethnic origin (both Greek). The controls are about 14 years older than the cases and the fact that they have not developed breast cancer could be taken as an additional indicator of the protective effect of the LL and SL microsatellite repeat genotypes for both the ESR2 and AR genes. Amongst the cases no significant difference was found between mean microsatellite repeat length of ESR1, ESR2 and AR genes when they were separated by tumor size (o2.5 and X2.5 cm), tumor grade, nodal status or metastases (Table 2). Additionally none of the cases and only 3/155 (1.9%) of control participants were under ERT. There was no significant difference between the cases and controls regarding other confound-

ing factors such as number of children, breast-feeding habits, cigarette smoking and alcohol intake. None of the 155 controls and 17/79 (21.5%) of cases had one family member with breast/ovarian cancer. Specifically amongst the cases seven postmenopausal (455 years of age) and three premenopausal (ages 30, 45 and 43) had one firstdegree relative (sister or mother with breast/ovarian cancer), while three postmenopausal (54, 62 and 72 years old) and three premenopausal (29, 30 and 50 years old) had a second-degree relative with breast cancer. Finally, one case (66 years old, postmenopausal) had a third-degree relative with breast cancer. The percentage of cases in our study with first- or second-degree relative with breast/ ovarian cancer agrees with other similar case-control studies.34–36 We have shown that women with LL and SL CAG genotypes in the AR gene had a 25- and 5-fold lower risk, respectively to have breast cancer compared to those with SS genotype (po0.0001 and o0.001) (Table 5), retaining significance even after adjustment for confounding factors (po0.005 and p ¼ 0.008) (Table 6). While women with LL and SL genotypes for the (CA)n alleles in the ESR2 gene had 100 and 77 times decreased risk respectively for breast cancer compared to those with the SS genotype (po0.0001 and o0.001) retaining significance even after adjustment for confounding factors, compared to the SS genotype (p ¼ 0.003) (Tables 5 and 6). A possible explanation for the protective role of LL and SL genotypes in the AR and ESR2 genes could be reduced transcriptional activity of AR gene and more active ESR2 gene resulting in lower levels of androgens and lower estrogen bioavailability, in combination with a more active role of the tumor suppressor activity of ESR2 gene. Our results agree with three previous studies indicating the SS (CAG) repeat genotype of AR gene being a risk allele for breast cancer in sporadic cases calculating a 6% reduction in the risk of death with every CAG repeat increase.20,35,36 More specifically our results are in agreement with the only other publication that takes into account the combined genotype microsatellite profile arguing that it has greater predictive value for breast cancer than the (CAG)n genotype alone. Combined effects of the two low risk polymorphisms (ESR2 and AR) might confer significant genetic predisposition to a polygenic/ complex disease and furthermore these effects might be ethnic/population specific as has been reported for other multigenic/complex disorders and repeat length polymorphisms.37 The strengths of our study are the inclusion of a homogeneous population, and the availability of detailed questionnaire information allowing us to consider potential confounding factors. The power of our study reaches 100% even though the numbers might be low; the statistical difference is quite substantial (alpha ¼ 0.05; two sided). One limitation of our study is not investigating the X-inactivation pattern in cases and controls. The AR gene

ARTICLE IN PRESS A. Tsezou et al. / The Breast 17 (2008) 159–166

is located on the X chromosome and approximately 10% of breast cancer cases between the ages of 27 and 45 show preferential X-inactivation, although no difference was found for middle aged and older cases compared to controls of a similar age.38 The age of both our case and control population are predominately older and middle aged women and should therefore have comparable levels of X-inactivation skewing. However, taking into account the X-inactivation hypothesis, only the S or the L allele will be expressed in subjects that have one X chromosome inactivated, the heterozygote subjects with preferential X-inactivation would therefore be more accurately grouped as SS or LL, even in that case the major statistical significance of the protective nature of the c.172(CAG)n LL genotype and risk associated effect of the SS genotype should be retained. In conclusion, we found that the 1174(TA)n VNTR in the promoter of ESR1 gene had no significant effect on breast cancer risk while a significant correlation was found between c. 1092+3607(CA)n and c. 172(CAG)n SS genotypes of the ESR2 and AR genes and breast cancer risk in individuals of Greek descent. A combination of short CAG repeats of the AR gene resulting in higher levels of androgens and short CA repeats of ESR2 resulting in less active receptor and higher estrogen bioavailability agrees with the hypothesis that the most widely accepted risk factor for breast cancer is increased cumulative ‘‘dose’’ of estrogen exposure to the breast epithelium. Additional data from a larger cohort of breast cancer patients is needed for definite evaluation of the risk impact of these allelic variants. Conflict of Interest statement None declared. Acknowledgment This work was funded by the European Union Pythagoras II research grant (70/3/7939), Ministry of Education. Appendix A. Supplementary Materials Supplementary data associated with this article can be found in the online version at doi:10.1016/j.breast. 2007.08.007.

References 1. Wiemann RA. Breast cancer hypothesis: a single cause for the majority of cases. J Epidemiol Community Health 2000;54:851–8. 2. Lacroix M, Leclercq G. The ‘‘portrait’’ of hereditary breast cancer. Breast Cancer Res Treat 2005;89:297–304. 3. Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable factors in the causation of cancer-analyses of cohorts of twins from Sweden, Denmark and Finland. N Eng J Med 2000; 343:78–85.

165

4. Zheng SL, Zheng W, Chang B, et al. Joint effect of estrogen receptor b sequence variants and endogenous estrogen exposure on breast cancer risk in Chinese women. Cancer Res 2003;63:7624–9. 5. Roodi N, Bailey LR, Cao WY, et al. Estrogen receptor gene analysis in estrogen receptor positive and receptor negative primary breast cancer. J Natl Cancer Inst 1995;87:446–51. 6. Ingles SA, Ross RK, Yu MC, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1977;89:166–70. 7. Ascenzi P, Bocedi A, Marino M. Structure–function relationship of estrogen receptor a and b: impact on human health. Mol Aspects Med 2006;27:299–402. 8. Ponglikitmongkol M, Green S, Chambon P. Genomic organisation of the human oestrogen receptor gene. EMBO J 1988;7:3385–8. 9. Enmark E, Pelto-Huikko M, Grandien K, et al. Human estrogen receptor b-gene structure, chromosomal localization and expression pattern. J Clin Endocrinol Metab 1977;82:4258–65. 10. Iwase H, Greenman JM, Barnes DM, Hodgson S, Bobrow L, Matthew CG. Sequence variants of the estrogen receptor (ER) gene found in breast cancer patients with ER negative and progesterone receptor positive tumors. Cancer Lett 1996;108:179–84. 11. Picard D, Bunone G, Liu JW, Donzw O. Steroid independent activation of steroid receptors in mammalian and yeast cells and in breast cancer. Biochem Soc Trans 1997;25:597–602. 12. Menin C, Banna GL, De Salvo G, et al. Lack of association between androgen receptor CAG polymorphism and familial breast/ovarian cancer. Cancer Lett 2001;168:31–6. 13. Maggiolini M, Donze O, Jeannin E, et al. Adrenal androgens stimulate the proliferation of breast cancer cells as direct activation of estrogen receptor a. Cancer Res 1999;59:4864–9. 14. Cox D, Blanche H, Pearce CL, et al. A comprehensive analysis of the androgen receptor gene and risk of breast cancer: results from the National Cancer Insitute Breast and Prostate Cancer Cohort Consortium (BPC3). Breast Cancer Res 2006;8:R54. 15. Zhang W, Yu YY. Polymorphisms of short tandem repeat of genes and breast cancer susceptibility. EJSO 2007;33:529–34. 16. Rebbeck TR, Kantoff PW, Krithivas K, et al. Modification of the BRCA1 associated breast cancer risk by the polymorphic androgen receptor CAG repeat. Am J Hum Genet 1999;64:1371–7. 17. Spurdle AB, Dite GS, Chen X, et al. Androgen receptor exon 1 CAG repeat length and breast cancer in women before age forty years. J Natl Cancer Inst 1999;91:961–6. 18. Sprudle AB, Antoniou A, Duffy DL, et al. The androgen receptor CAG repeat polymorphism and modification of breast cancer risk in BRCA1 and BRCA2 mutation carriers. Breast Cancer Res 2005;7:R176–83. 19. Given HF, Radbourne R, Oag H, et al. The androgen receptor exon 1 trinucleotide repeat does not act as a modifier of the age of presentation in breast cancer. Eur J Cancer 2000;36:533–4. 20. DeAbreu FB, Picolo LJ, Canevari RA, et al. Shorter CAG repeat in the AR gene is associated with atypical hyperplasia and breast carcinoma. Anticancer Res 2007;27:1199–205. 21. Henderson BE, Feigelson HS. Hormonal carcinogenesis. Carcinogenesis 2000;21:427–33. 22. Russo J, Hu YF, Yang X, Russo IH. Developmental, cellular, and molecular basis of human breast cancer. J Natl Cancer Inst Monogr 2000;27:17–37. 23. Becherini L, Gennari L, Masi L, et al. Evidence of a linkage disequilibrium between polymorphisms in the human estrogen receptor alpha gene and their relationship to bone mass variation in postmenopausal Italian women. Hum Mol Genet 2000;9:2043–50. 24. Fytili P, Giannatou E, Papanikolaou V, et al. Association of repeat polymorphisms in the estrogen receptors alpha, beta, and androgen receptor genes with knee osteoarthritis. Clin Genet 2005; 68:268–77. 25. Rothenburg S, Koch-Nolte F, Rich A, Haag F. A polymorphic dinucleotide repeat in the rat nucleolin gene forms Z-DNA and inhibits promoter activity. Proc Natl Acad Sci 2001;98:8985–90.

ARTICLE IN PRESS 166

A. Tsezou et al. / The Breast 17 (2008) 159–166

26. Yeap BB, Wilce JA, Leedman PJ. The androgen receptor mRNA. BioEssays 2004;26:672–82. 27. Westberf L, Baghaei F, Rosmond R, et al. Polymorphisms of the Androgen receptor gene and the estrogen receptor b gene are associated with androgen levels in women. J Clin Endocrinol Metabol 2001;86:2562–8. 28. Maguire P, Margolin S, Skoglund J, et al. Estrogen receptor beta (ESR2) polymorphisms in familial and sporadic breast cancer. Breast Cancer Res Treat 2005;94:145–52. 29. Giguere Y, Dewailly E, Brisson J, et al. Short polyglutamine tracts in the androgen receptor are protective against breast cancer in the general population. Cancer Res 2001;61:5869–74. 30. Haiman CA, Bernstein L, Berg D, Ingles SA, Salane M, Ursin G. Genetic determinants of mammographic density. Breast Cancer Res 2002;4:R5. 31. Dunning AM, Healey CS, Pharoah PD, Teaie MD, Ponder BA, Easton DF. A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1999;8:843–54. 32. Kaduri L, Easton DF, Edwards S, et al. CAG and GGC repeat polymorphisms in the androgen receptor gene and breast cancer susceptibility in BRCA1/2 carriers and non-carriers. Br J Cancer 2001;85:36–40.

33. Suter NM, Malone KE, Daling JR, Doody DR, Ostrander EA. Androgen receptor (CAG)n and (GGC)n polymorphisms and breast cancer risk in population-based case-control study of young women. Cancer Epidemiol Biomarkers Prev 2003;12:127–35. 34. Liede A, Zhang W, De Leon Matsuda ML, Tan A, Narod SA. Androgen receptor gene polymorphism and breast cancer susceptibility in the Philippines. Cancer Epidemiol Biomarkers Prev 2003; 12:848–52. 35. Yu H, Bharaj B, Vassilikos EJK, Giai M, Diamandis E. Shorter CAG repeat length in the androgen receptor gene is associated with more aggressive forms of breast cancer. Breast Cancer Res Treat 2000; 59:153–61. 36. Iobagiu C, Lambert C, Normand M, Genin C. Microsatellite profile in hormonal receptor genes associated with breast cancer. Breast Cancer Res Treat 2006;95:153–9. 37. Nakamura T, Shi D, Tzetis M, Rodriguez-Lopez J, et al. Metaanalysis of association between the ASPN D-repeat and osteoarthritis. Hum Mol Genet 2007;16:1676–81. 38. Kristiansen M, Langered A, Knudsen GP, Weber BL, Berresen-Dale AL, Orstavik KH. High frequency of skewed X inactivation in young breast cancer patients. J Med Genet 2002;39:30–3.