No association between XRCC1 and XRCC3 gene polymorphisms and breast cancer risk: Iowa Women's Health Study

Cancer Detection and Prevention 30 (2006) 313–321 www.elsevier.com/locate/cdp

No association between XRCC1 and XRCC3 gene polymorphisms and breast cancer risk: Iowa Women’s Health Study Bharat Thyagarajan MBBS, PhD, MPHa, Kristin E. Anderson PhD, MPHa, Aaron R. Folsom MD, MPHa, David R. Jacobs Jr. PhDa, Charles F. Lynch MD, PhDb, Archana Bargaje MDc, Waseem Khaliq MD, MPHa, Myron D. Gross PhDc,* a

University of Minnesota, Division of Epidemiology, Suite 300, West Bank Office Building, Minneapolis, MN 55454, United States b University of Iowa, Department of Epidemiology, Iowa City, IA 52242, United States c Department of Laboratory Medicine and Pathology, University of Minnesota, MMC 609, 420 Delaware Street S. E., Minneapolis, MN 55455, United States Accepted 25 July 2006

Abstract Background: Genetic variation in DNA repair may contribute to differences in the susceptibility of several cancers. We evaluated two polymorphisms in the base excision repair pathway (BER) (XRCC1; Arg194Trp and Arg399Gln) and one polymorphism in the double strand DNA repair pathway (XRCC3; Thr241Met) for their association with breast cancer risk. Methods: The association was analyzed in a nested case control study of 460 breast cancer cases and 324 cancer-free controls within the Iowa Women’s Health Cohort. DNA was obtained from blood samples or paraffin embedded tissues (PET) and all samples were genotyped by one of three genotyping platforms—PCR-RFLP, PCRINVADER, or Sequenom. Results: None of the three polymorphisms studied were significantly associated with breast cancer risk (XRCC1: Arg194Trp (OR = 1.21, 95% CI: 0.78–1.88); Arg399Gln (OR = 1.20, 95% CI: 0.80–1.79); XRCC3: Thr241Met (OR = 1.04, 95% CI: 0.76– 1.41). Conclusions: These results suggest that independently these polymorphisms of XRCC1 and XRCC3 genes do not contribute significantly to the genetic susceptibility of breast cancer. # 2006 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved. Keywords: Postmenopausal breast cancer; Polymorphisms; DNA repair; Base excision repair; Double strand DNA repair; Nested case control study; IWHS; INVADER assay; Paraffin embedded tissue; Genotype; Arg399Gln; BMI; Risk factors; Statistical analysis

1. Introduction Several established risk factors for breast cancer, such as estrogen exposure and alcohol, generate reactive oxygen radicals that cause a variety of DNA damage lesions, including DNA strand breaks and oxidized DNA bases [1]. Cells repair DNA damage and maintain genomic stability with a variety of DNA repair mechanisms, which may be Abbreviations: BER, base excision repair; HR, homologous recombination; PCR, polymerase chain reaction; PET, paraffin embedded tissue; RFLP, restriction fragment length polymorphism; SNP, single nucleotide polymorphism * Corresponding author. Tel.: +1 612 624 5417; fax: +1 612 273 6994. E-mail address: [email protected] (M.D. Gross).

essential in preventing tumor initiation and delaying tumor progression. A review of epidemiological studies evaluating the association between DNA repair and a variety of cancers revealed that DNA repair is associated with an increased risk of a variety of cancers, including lung, skin, and breast cancers, in a majority of the studies [2–4]. Oxidative DNA damage and single strand breaks are repaired predominantly by the base excision repair (BER) pathway [5]. Double strand breaks are repaired predominantly by the double strand DNA repair pathway, which consists of two distinct pathways—the non-homologous repair pathway and the homologous recombination (HR) repair pathway [6]. Single nucleotide polymorphisms (SNPs) with high population frequencies have been detected

0361-090X/$30.00 # 2006 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.cdp.2006.07.002

314

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

in genes belonging to both, the BER and double strand repair pathways and are thought to contribute to the variation observed in DNA repair in the population thereby contributing to breast cancer susceptibility. The XRCC1 gene plays an important role in BER, as a scaffold protein that brings together proteins of the DNA repair complex [7–9]. Three polymorphisms, Arg194Trp, Arg280His, and Arg399Gln have been identified in the coding region of the XRCC1 gene [10]. Arg194Trp polymorphism has been associated with increased bladder and esophageal cancer risk [11,12], but has no association with breast cancer [13–17]. On the other hand, the Arg399Gln polymorphism has been associated with increased breast cancer, pancreatic cancer risk, and increased DNA adduct levels [13,18,19]. However, not all studies have shown consistently increased risk [14–16] and further studies are needed to clarify this association. In addition, there is also evidence of gene–gene interactions involving Arg194Trp polymorphism in modulating breast cancer risk [14]. In humans, XRCC3 plays an important role in homologous recombination pathway by interacting with RAD-51 and maintaining chromosomal stability, and repairing DNA damage [20]. Brenneman et al. showed that XRCC3’s function was not limited to HR initiation alone, but also extends to later stages in formation and resolution of HR intermediates, possibly by stabilizing heteroduplex DNA [21]. One polymorphism in XRCC3, Thr241Met, has been associated with increased breast cancer risk in one large case control study [22] and the Met/Met genotype has been found to be associated with a significantly higher DNA adduct level as compared to the Thr/Thr genotype in another study [23]. However, other studies have shown no association between Thr241Met polymorphism and breast cancer risk [14,24,25]. Thus, further studies are needed to confirm these associations. We hypothesized that polymorphisms in the XRCC1 and XRCC3 genes will be associated with increased breast cancer risk. The Arg280His polymorphism had a low population frequency among 200 cancer-free women in our study population (2.5%). Since this study did not have statistical power to identify modest associations of low prevalence SNPs the Arg280His polymorphism was not evaluated further in this study. The other two SNPs in XRCC1 (Agr194Trp, Agr399Gln) and the Thr241Met polymorphism in XRCC3 (Thr241Met) were evaluated for their association with breast cancer risk in a nested case control study of 460 breast cancer cases and 324 controls within the Iowa Women’s Health Cohort.

2. Materials and methods The Iowa Women’s Health Study was a prospective cohort study of 41,836 women aged 55–69 years at baseline in 1986. A questionnaire was completed at baseline that focused on anthropometric, dietary, and other major risk

factors for cancer including family history of cancer, prior medical conditions, cigarette smoking, reproductive factors, and hormone use [26]. Mortality and cancer incidence data, since 1986, have been collected through a computer linkage of study participants’ identifiers with Iowa death certificate files, the National Death Index, and cancer diagnosis data collected by the Iowa Cancer Registry. All anthropometric and risk factor information was collected using the baseline questionnaire. All participants reported current height and weight, from which BMI (kg/ m2) was calculated. A paper measuring tape was sent to each participant so that a friend could measure the circumferences around the waist (1 in. above the umbilicus) and the hips (maximum). The waist-to-hip ratio (WHR) was calculated [28]. The baseline questionnaire assessed smoking status and amount of smoking. Physical leisure activity was assessed using two questions regarding how often they participated in moderate and vigorous physical activity, which were combined to form a three-level physical activity index (low, moderate, and high) [27]. Women were also asked at baseline about the age at which they first menstruated, whether they currently had menstrual periods (within the past year), and if not, the age at which menstrual periods no longer occurred. Information on hormone replacement therapy was obtained by asking women whether they had ever used pills, other than birth control pills, that contained estrogen or other female hormones. Participants were asked whether their mother, maternal and paternal grandmothers, aunts, sisters, and daughters had been diagnosed with any type of cancer. This question was used to determine a history of cancer among females in the family. Women were considered to have first-degree relatives with breast cancer if their mother, daughter, or sister had a history of breast cancer. Four hundred and sixty incident breast cancer cases (309 tissue samples and 151 blood samples) and 324 cancer-free controls (blood samples) were included in this study. Sample collection is described in Fig. 1. Eligible cases included all incident breast cancer cases (n = 765) diagnosed between January 1, 1992 and December 31, 1996. Eligible controls included all cohort members who were cancer-free through 1994. Eight hundred and seventy six cancer-free women were randomly chosen as controls from these participants in the Iowa Women’s Health Study. Initially buccal cells were requested from all 465 breast cancer cases diagnosed from 1992 to 1994 and 876 controls. Buccal cells were obtained from 273 cases (60%) and 659 controls (75%). Due to the limited amount of DNA obtained from buccal cells, these samples were not genotyped; blood samples were subsequently requested from those study participants who had donated a buccal cell sample. Blood samples were returned by 156 cases (156/273 = 57%) and 332 controls (332/ 659 = 50%) who were approached for a blood sample. DNA suitable for amplification was obtained from 324 controls. To increase the proportion of cases for whom we had biological material, paraffin embedded tissue (PET) blocks

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

315

Fig. 1. Sampling scheme for genetic analysis in IWHS. Total DNA samples: 460 cases (60%) (151 blood samples + 309 tissue samples) 324 (37%) cancer free controls. Asterisk ’*’ denotes response rates (%) given in parenthesis. N is given in brackets.

of incident breast cancer cases diagnosed between 1992 and 1994 were collected by contacting all the pathology laboratories in Iowa. This method provided biological material from an additional 196 cases. To further increase the number of cases we attempted to retrieve tissue blocks from breast cancer cases diagnosed between 1995 and 1996. We received paraffin embedded tissue from 113 cases (70%) of the 162 cases requested from 1995 to 1996, yielding a total of 309 blocks. Of the 309 cases providing blocks, benign lymph node tissue was obtained from 247 cases. Nipple or areola tissue, fatty breast cells, normal breast cells, or fibrocystic breast cells were obtained for 49 cases. Tumor breast tissue was obtained for nine cases while lymph node with tumor was obtained for four cases. Major reasons for failure to obtain samples were blocks being discarded (75%), refusal of pathology laboratories to participate in research studies (22%), and refusal of participants to provide consent to contact pathology laboratories (2%). In addition, paraffin embedded tissue was not available for two cases because one case had only a cytology specimen available and a specimen for another case could not be located. Overall blood samples or tissue samples from cancer cases were obtained from 465 subjects (156 blood samples and 309 tissue blocks). We obtained biological material from

507 (82%) of the 618 breast cancer cases, from whom biological material was requested. This response included the following biological samples, blood samples (156), tissue blocks (309), and buccal cells (42 samples). DNA suitable for amplification was obtained from 460 breast cancer cases (151 blood samples and 309 tissue blocks; no buccal cell DNA was amplified). 2.1. Laboratory methods DNA was extracted from peripheral blood leukocytes using the phenol chloroform extraction method and stored at
70 8C until analysis [29]. Tissue DNA was extracted using a PureGene DNA extraction kit using minor modifications and stored at
20 8C until analysis [30]. DNA quantity was measured by absorbance at 260 nm using a UV spectrometer, and DNA purity was evaluated by measurement of the 260/280 absorbance ratios. Three genotyping platforms, PCR-RFLP, PCR-INVADER, and Sequenom platforms were used to genotype all the samples in our study. Though the initial genotyping for determination of genotype frequencies in the Iowa population was performed using PCR-RFLP method, the PCR-INVADER methodology was used for genotype evaluation in the majority of samples.

316

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

Arg399Gln (XRCC1-10) and Thr241Met (XRCC3-7) polymorphisms were completely genotyped using PCRRFLP and PCR-INVADER methodologies. The Sequenom assay was used as a third genotyping platform for the Arg194Trp polymorphism to genotype samples that could not be genotyped by the PCR-RFLP or the PCR-INVADER assay. The Arg399Gln genotype was assessable from blood samples only (151 cases and 324 controls) while the Arg194Trp and Thr241Met genotypes were obtained from both blood and tissue samples (460 cases and 324 controls).

Table 1b PCR reaction mixtures Polymorphisms

PCR reaction mixtures (50 ml reaction)

Arg194Trp

1 ml each primer [20 pmol], 5 ml MgCl2, 10 buffer, 4 ml dNTP, 0.4 ml Taq polymerase, and 400 ng DNA 1.25 ml of each primer [20 pmol], 5 ml MgCl2, 5 ml 10 buffer, 4 ml dNTP, 0.4 ml Taq polymerase (hot start), and 400 ng DNA 1.5 ml each primer [20 pmol], 4 ml MgCl2, 5 ml 10 buffer, 3 ml dNTP, 0.4 ml Taq polymerase, 4 ml Taq antibody, and 500 ng DNA

Arg399Gln

Thr241Met

2.2. Genotyping platforms 2.2.1. PCR-RFLP The primers, PCR reaction mixtures, PCR conditions, and restriction enzymes used to determine the genotypes for the Arg194Trp, Arg399Gln and Thr241Met polymorphisms are listed in Tables 1a–1c. All restriction enzyme digestions were performed according to manufacturer’s recommendation. A 3% agarose gel was prepared, stained with ethidium bromide, agarose gel electrophoresis was performed, and the digested products were read using UV light. 2.2.2. PCR-INVADER The PCR conditions, primers, and details of the INVADER assay used to determine genotypes of the Arg194Trp and Thr241Met polymorphisms have been described previously [30]. The Arg399Gln polymorphism was amplified using a PCR reaction mixture containing 1 ml of dNTP (10 mM) (GeneAMP), 1.25 ml of 25 mmol MgCl2 (Promega), 1.25 ml of 10 buffer, 0.5 ml of each primer (20 pg/ml) (GeneSys), 0.2 ml of Taq polymerase (5 U/ml) (Promega), and 125 ng of sample DNA in a 12.5 ml reaction. The following primers were used in the PCR reaction: - Arg399Gln forward primer: 50 -CCCCAAGTACAGCCAGGTC-30 and

Table 1a PCR primers and reaction conditions Primers Arg194Trp F R Arg399Gln F R Thr241Met F R

Sequences

PCR conditions

50 -TGA AGG AGG AGG ATG AGA GC 50 -CCC TAC TCA CTC AGG ACC CA

94 8C/min, 64 8C/30 s, 728C/30 s for 33 cycles

50 -AGT AGT CTG CTG GCT CTG G 50 -CAG TGG TGC TAA CCT AAT C

94 8C/30 s, 63 8C/45 s, 728C/min for 36 cycles

50 -TGT GAA TTT GAC AGC CAG G 50 -AAA ATA CGA GCT CAG GGG TG

94 8C/min, 65 8C/30 s, 728C/30 s for 31 cycles

- Arg399Gln reverse primer: 50 -TGTCCCGCTCCTCTCAGTAG-30 The PCR reaction conditions were as follows: denaturation time—94 8C for 5 min followed by 35 cycles of 94 8C for 30 s, 57 8C for 1 min and 72 8C for 30 s. This was followed by a final extension at 72 8C for 7 min. The INVADER reaction conditions were similar to those described for the other two polymorphisms. 2.2.3. Sequenom The principle and detailed description of the Sequenom assay have been provided elsewhere [31]. The PCR reaction mixture for the Arg194Trp polymorphism contained 1 ml of dNTP (10 mM) (GeneAMP), 1.25 ml of 25 mmol MgCl2 (Promega), 1.25 ml of 10 buffer, 0.5 ml of each primer (20 pg/ml) (GeneSys), 0.1 ml of Taq polymerase (5 U/ml) (Promega), and 100 ng of sample DNA were added to a 12.5 ml reaction. PCR primers and a mass extension primer were designed using SpecroDESIGNER (SEQUENOM’s assay design software). The following primers were used in the reaction: - Arg194Trp forward primer: 50 -ACGTTGGATGATGAGAGCGCCAACTCTCTG-30 - Arg194Trp reverse primer: 50 -ACGTTGGATGACTCACTCAGGACCCACGTT-30 - Arg194Trp mass extension primer: 50 -GGATGTCTTGTTGATCC-30 Table 1c Identification of genotypes by PCR-RFLP Genotype

PCR product (bp)

Restriction enzyme

Bands seen on gel electrophoresis after restriction digestion (bp)

Arg194Trp

116

HpaII

Arg/Arg: 58, 36, 20 Arg/Trp: 78, 50, 36, 20 Trp/Trp: 78, 36

Arg399Gln

852

NciI

Arg/Arg: 460, 276, 116 Arg/Gln: 576, 460, 276, 116 Gln/Gln: 576, 276

Thr241Met

155

Hsp92II

Thr/Thr: 155 Thr/Met: 155, 92, 63 Met/Met: 92, 63

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

317

genotyping this polymorphism. In addition, eight paraffin embedded blocks without tissue (blanks) did not show genotype for all three SNPs.

The PCR reaction conditions were as follows: denaturation time—94 8C for 5 min followed by 40 cycles of 94 8C for 1 min, 46 8C for 2 min and 72 8C for 1 min. This was followed by a final extension at 72 8C for 7 min.

2.4. Statistical analysis 2.3. Quality control Individual genotypes were divided into three categories—homozygous for the majority allele, heterozygous, and homozygous for the minority allele. Since there were few women who were homozygous for the minority allele, the women who were heterozygotes and homozygous for the minority allele were combined into a single category and compared to the women who were homozygous for the majority allele during data analysis. A Chi-square test was performed for each SNP to determine if the control sample demonstrated Hardy–Weinberg equilibrium. Genotype frequencies were calculated for the cases and controls to determine their association with breast cancer. Each genotype was tested for its association with traditional breast cancer risk factors, namely age, age at menarche, age at menopause, BMI, waist-to-hip ratio, use of exogenous hormones, smoking status, alcohol intake, physical activity, and age at first live birth. The association between individual genotypes and breast cancer risk was analyzed using unconditional logistic regression. Crude odds ratios were initially calculated (Table 2). The final model (Table 2) was adjusted for age (continuous variable), family history of breast cancer among firstdegree relatives (categorical variable coded as yes/no), physical activity (categorical variable coded as low, moderate, high levels of activity), and BMI (continuous

2.3.1. Blood samples Several sets of DNA samples were genotyped by PCRRFLP as well as PCR-INVADER. Both genotyping platforms yielded identical genotypes of the Thr241Met polymorphisms in a set of 28 samples. Similarly, all 34 samples genotyped for the Arg399Gln polymorphism also yielded identical genotypes using both methods. For the Arg194Trp polymorphism, 20 out of 21 samples, genotyped using both methods, yielded identical genotypes. Ten blood samples previously genotyped using PCR-RFLP and PCRINVADER were genotyped using Sequenom technology. All 10 samples yielded identical genotypes in all three genotyping platforms for the Arg194Trp genotype. 2.3.2. Tissue samples Seven pairs of blinded duplicate samples were used as positive controls and eight paraffin blocks without any tissue were used as negative controls. There was 100% concordance between the seven pairs of blinded duplicate tissue samples for the Arg194Trp polymorphism. Six pairs of blinded duplicate tissue samples were concordant for the Thr241Met polymorphism. One blinded sample could not be genotyped for the Thr241Met polymorphism and so the seventh pair was not used to determine the accuracy of

Table 2 Frequencies of three DNA repair genotypes and odds ratios of incident breast cancer cases between 1992 and 1996 in the Iowa Women’s Health Cohort**,++ Gene

Case, N (%)

Control, N (%)

Model 1 a

Model 2b

Odds ratio (OR)

95% CI

Odds ratio (OR)

95% CI

(87.6) (11.5) (0.9) (12.4)

1.00 1.2 1.02 1.18

REF 0.77–1.87 0.23–4.57 0.77–1.81

1.00 1.23 0.98 1.21

REF 0.21–4.63 0.21–4.63 0.78–1.88

135 140 47 187

(41.9) (43.5) (14.6) (58.1)

1.00 1.29 0.91 1.19

REF 0.85–1.95 0.49–1.70 0.80–1.77

1.00 1.29 0.91 1.20

REF 0.84–1.97 0.48–1.72 0.80–1.79

126 157 40 197

(39.0) (48.6) (12.4) (61.0)

1.00 0.96 1.32 1.04

REF 0.70–1.32 0.84–2.08 0.77–1.40

1.00 0.96 1.37 1.04

REF 0.69–1.32 0.85–2.19 0.76–1.41

Arg194Trp (XRCC1-6) Arg/Arg Arg/Trp Trp/Trp Arg/Trp + Trp/Trp

370 58 4 62

(85.7) (13.4) (0.9) (14.3)

282 37 3 40

Arg399Gln (XRCC1-10) Arg/Arg Arg/Gln Gln/Gln Arg/Gln + Gln/Gln

57 76 60 136

(37.8) (50.3) (11.9) (62.2)

Thr241Met (XRCC3-7) Thr/Thr Thr/Met Met/Met Thr/Met + Met/Met

160 192 67 259

(38.2) (45.8) (16.0) (61.8)

Other traditional risk factors such as age at menarche, age at menopause, number of pregnancies, age at first live birth, hormone replacement therapy, and oral contraceptive pill use were neither significant independent risk factors nor confounders in this dataset. Hence, they were excluded from final model. ++Odds ratios presented are (heterozygous + homozygous mutant) vs. homozygous normal. **460 cases and 324 controls were analyzed for XRCC3-7 and XRCC1-6. 151 cases and 324 controls were analyzed for XRCC1-10. XRCC1-10 genotypes were obtained for 151 cases and 322 controls. XRCC1-6 genotypes were obtained for 432 cases and 322 controls, while XRCC3-7 genotypes were available for 419 cases and 323 controls. a Crude model. b Model adjusted for age, physical activity, BMI, and family history of breast cancer.

318

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

variable). Other variables including age at menarche, age at menopause, hormone replacement therapy, female history of cancer, age at first live birth, alcohol intake, smoking status, and waist-to-hip ratio were neither statistically significant confounders nor had a significant independent association with breast cancer in our study. Hence, these variables were not included in the final model for any of the three SNPs.

3. Results DNA was extracted from 326 PET samples (including study samples and blinded controls). Seventy-one PET samples (22%) had low DNA quantity (A260 OD < 0.1) while 78 PET samples (24%) had low DNA quality (A260/ A280 < 1.5). Since PET samples were not genotyped for the Arg399Gln polymorphism, genotyping of PET samples will be described for Arg194Trp and Thr241Met polymorphisms. In general, DNA quantity, rather than DNA quality, was correlated with ability to obtain a genotype for a sample. However, the effect of DNA quantity on the ability to genotype a sample was locus dependent. For example, of the 45 samples that could not be genotyped for the Arg194TRp (XRCC1-6) polymorphism 58% (26/45) of these samples had low DNA quantity. In a similar manner, 90% (45/50) of the samples that could not be genotyped for the Thr241Met polymorphism had low DNA quantity. In contrast, DNA quality showed a weaker correlation with inability to genotype samples (42% of samples that could not be genotyped for Arg194Trp had low DNA quality while 58% of samples that could not be genotyped for Thr241Met had low DNA quality). A comparison between the three genotyping platforms revealed that PCR-INVADER and Sequenom assays had a higher sensitivity in genotyping samples as compared to PCR-RFLP in that samples that could not be genotyped using the PCR-RFLP system could be reproducibly genotyped using the other two genotyping platforms.

All three genotypes were in Hardy-Weinberg equilibrium. The crude model for each genotype was very similar to its respective fully adjusted model. Hence, only the results for the fully adjusted model are presented in the text for each genotype. However, both crude and fully adjusted models are presented in Table 2. Comparison of risk factor and demographic data between participants in the study and cases and controls who refused to participate did not show a significant difference. Various risk factors such as age, waist-to-hip ratio, BMI, age at menopause, age at menarche, use of hormone replacement therapy, alcohol consumption, family history of breast cancer, and physical activity were not associated with breast cancer risk (Table 3). The Trp allele for the Arg194Trp polymorphism (XRCC1-6) had an allele frequency of 6.6%, which was similar to reported frequency in Caucasian populations [15]. The Arg194Trp polymorphism was not associated with any traditional breast cancer risk factors. Since very few women were homozygous for Trp allele (0.9%), the Trp/Trp genotype and Arg/Trp were combined together. This combined genotype was not associated with breast cancer risk (OR = 1.21, 95% CI: 0.78–1.88) when compared with the Arg/Arg genotype. The Gln allele frequency for the Arg399Gln polymorphism (XRCC1-10) was 36.3%, which was comparable to the frequency published earlier [14,17]. The Gln/Gln genotype or the Arg/Gln genotype were not associated with an increased risk of breast cancer as compared to the Arg/Arg genotype (OR 1.29, 95% CI: 0.84–1.97 for the Arg/Gln genotype and OR 0.91, 95% CI: 0.48–1.72 for the Gln/Gln genotype). A statistically significant interaction was observed between smoking status and the Arg399Gln polymorphism ( p = 0.01). Non-smokers with the combined Arg/Gln and Gln/Gln genotypes were at significantly higher breast cancer risk as compared to the Arg/Arg genotype (OR = 3.13, 95% CI: 1.23–7.97). However, smokers with Arg399Gln polymorphism did not have a significantly increased breast cancer risk as compared to non-smokers with the Arg/Arg genotype ( p  0.26).

Table 3 Comparison of traditional breast cancer risk factors between cases and controls of the Iowa Women’s Health Study Study participants

Age at menopause (52 years) Age at menarche (12 years) Women on HRTa First-degree relative with breast cancer Family history of cancer Alcohol consumption (2.6 g) Age (64 years) Waist-to-hip ratio (0.77) Age at first pregnancy (25 years) BMI (>26.6 kg/m2) Physical activity (high activity) Ever smoker a

HRT: hormone replacement therapy.

Odds ratio

Case (N = 460)

Control (N = 324)

140 120 201 83 307 117 171 355 122 244 105 157

106 89 135 37 198 83 82 256 74 154 91 97

(31.5) (26.3) (43.8) (18.0) (66.7) (25.4) (37.3) (77.2) (28.7) (53.0) (23.1) (34.6)

(33.9) (27.6) (41.7) (11.4) (61.1) (25.6) (25.3) (79.3) (24.5) (47.5) (28.4) (30.4)

0.90 0.94 1.09 1.71 1.28 0.99 1.76 0.89 1.24 1.25 0.76 1.21

(0.66–1.22) (0.68–1.29) (0.82–1.46) (1.13–2.58) (0.95–1.72) (0.72–1.37) (1.29–2.40) (0.63–1.25) (0.89–1.74) (0.94–1.66) (0.55–1.05) (0.89–1.65)

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

The Met allele frequency for the Thr241Met polymorphism (XRCC3-7) was 36.7%. This allele frequency was similar to those reported earlier [24,25]. The Thr241Met polymorphism was not associated with any of the traditional breast cancer risk factors. The Thr/Met genotype or the Met/ Met genotype was not associated with breast cancer risk (OR = 1.37, 95% CI: 0.85–2.19 for the Met/Met genotype and OR = 0.96, 95% CI: 0.69–1.32 for the Thr/Met genotype). Other than the interaction between smoking status and Arg399Gln polymorphism no other significant interaction was observed between any of the three polymorphisms and any of risk factors for breast cancer such as oral contraceptive use, physical activity, BMI, and other environmental exposures such as cigarette smoking and alcohol intake. In addition, no significant gene–gene interactions were noted between the three polymorphisms.

4. Discussion Our study shows that DNA obtained from PET can be reliably used in epidemiological studies to determine SNPs at various loci. Both PCR-INVADER and Sequenom genotyping platforms were successfully used to genotype PET samples and since both PCR-INVADER and Sequenom are more suited to automation as compared to PCR-RFLP they can be more readily adapted to large scale genotyping studies. No association was found between the Arg194Trp and Arg399Gln polymorphisms in the XRCC1 gene and the Thr241Met polymorphism in the XRCC3 gene and breast cancer risk. Six prior studies found no association between Arg194Trp polymorphism and breast cancer risk [13– 17,32]. Overall, findings from our study were consistent with previous reports that showed no association between the Arg194Trp polymorphism and breast cancer risk. However, the interaction between the Arg194Trp polymorphism with environmental factors and other genes remains incompletely investigated as a majority of studies were either hospital based or population based case control studies that either did not collect information regarding various environmental factors or did so retrospectively from breast cancer cases and controls. Two studies have shown significant interactions between Arg194Trp polymorphisms and other SNPs or environmental factors. Smith et al. showed a significant interaction between 194Trp allele and 241Met allele (Thr241Met polymorphism) and increased breast cancer risk (OR = 8.74, 95% CI: 1.13–67.53) [14]. Han et al. showed a significant interaction between plasma carotene and Arg194Trp polymorphism ( p = 0.02) [15]. Since plasma carotene levels were not available in the Iowa Women’s Health Cohort this finding could not be verified in our study. Our study found no evidence of interaction between Arg194Trp polymorphism and other polymorph-

319

isms such as Arg399Gln or Thr241Met. In addition, no significant interactions were seen with environmental factors such as cigarette smoking, physical activity, alcohol intake, BMI, and estrogen exposure. Four studies did not show an association between breast cancer risk and the Arg399Gln polymorphism [14–16,32]. One study of a Korean population has shown a significant positive association between the Gln/Gln genotype (XRCC1-10) and breast cancer risk (OR = 2.4, CI: 1.20– 4.72) as compared to the Arg/Arg genotype [13] Duell et al. found the combined Arg/Gln and Gln/Gln genotype to be associated with increased breast cancer risk among African– Americans (OR = 1.7, 95% CI: 1.1–2.4), but not in Caucasians (OR = 1.0, 95% CI: 0.8–1.4) [17]. Thus, some studies suggest a possible association between the Arg399Gln polymorphism and breast cancer risk. As noted for the Arg194Trp polymorphism interactions between Arg399Gln and other environmental factors have been inadequately studied except in the study by Han et al. [15]. Our study found a statistically significant interaction between smoking and the Arg399Gln polymorphism with the combined Arg/Gln and Gln/Gln genotypes being at statistically increased breast cancer risk among non-smokers while there was no evidence of increased breast cancer risk among the smokers. Band et al. showed a decreased risk of postmenopausal breast cancer among women who started smoking after their first full-term pregnancy as compared to non-smokers and attributed the decreased risk among smokers to the anti-estrogenic effects of cigarette smoke [33]. Thus, the decreased postmenopausal breast cancer risk observed in smokers may account for the lack of statistically significant increased breast cancer risk in smokers with the Arg/Gln and Gln/Gln genotypes. Duell et al. showed that the African–American women with Arg/Arg genotype were at increased breast cancer risk with increasing duration of smoking (trend test; p < 0.001) [17]. However, no evidence of interaction was seen among Caucasian women [17]. A large prospective study by Han et al. also showed no evidence of interaction between smoking and Arg399Gln polymorphism [15]. Another study showed that smokers who were carriers of 399Gln allele showed increased DNA damage as compared to smokers with the Arg/Arg genotype while no increased DNA damage was observed in nonsmokers [34]. Given these, conflicting results, the positive interaction observed in our study needs to be interpreted with caution because of the relatively small size of the study for the Arg399Gln polymorphism (151 cases and 324 controls) and the large number of interactions that were tested in our study. The Met/Met genotype of the Thr241Met polymorphism has been associated with increased risk of breast cancer in a Caucasian population (1.3 [1.1–1.6]) as compared to the Thr/Thr genotype [22]. However, three other studies have shown no association between Thr241Met polymorphism in XRCC3 and breast cancer risk [14,24,25]. Similar to studies on XRCC1 gene polymorphisms, only the Nurses’ Health

320

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

Study collected environmental exposures prior to occurrence of disease [25] and gene–gene or gene–environment interactions remain inadequately studied. Our study found no significant interactions between the Thr241Met polymorphism and other genetic or breast cancer risk factors. Thus, no association is apparent in four out of five studies (including the present study). Our study had 80% power at a significance level of alpha = 0.05 to detect an odds ratio of 1.5 for the Thr241Met polymorphism, an OR of 1.7 for the Arg194Trp polymorphism, and an odds ratio of 1.75 for the Arg399Gln polymorphism when individuals who were heterozygous and homozygous for the minority allele were combined into a single ‘‘at risk’’ category and compared to the individuals who were homozygous for the majority allele. However, our study had very limited power to evaluate breast cancer risk in individual genotype categories. Thus, power was good for moderate to large associations, but a weak association could have been missed. Even though biological samples were obtained from 507 (82%) of cases who were contacted and 659 (75%) of all eligible controls, only 460 breast cancer cases and 324 controls were available for genetic analysis due to a variety a technical issues. Thus the sample size is lower for technical reasons, unrelated to patient characteristics and so unlikely to be a source of selection bias. Furthermore, comparison of risk factor and demographic data between participants in the study and cases and controls who refused to participate showed no significant differences indicating no significant selection bias in our study. A potential limitation of our study is that cases and controls were selected from different time periods in that breast cancer cases were selected from 1992 to 1996 while controls were selected from 1992 to 1994. We evaluated whether any of the selected controls changed their disease status between 1995 and 1996. Four controls developed breast cancer between 1995 and 1996. Re-analysis of the data after excluding these participants or treating these participants as breast cancer cases did not significantly change the results of our study. An important strength of our study is that all environmental exposure information was collected prior to occurrence of breast cancer minimizing recall bias in ascertaining exposure status. However, the relatively small size of our study does not provide sufficient power for evaluation of gene– environment interactions. However, consistent with the results published by Han et al. in the Nurses’ Health Study no significant gene–environment interactions were identified in our study with respect to the major breast cancer risk factors [15,25]. Replication of the interaction between Arg399Gln polymorphism and smoking status in other studies is necessary to ensure that this interaction is not observed due to chance in our study. In conclusion, we found no association between the Arg194Trp and Arg399Gln polymorphisms in the XRCC1 gene and the Thr241Met polymorphism in the XRCC3 gene with breast cancer risk.

Acknowledgements This work was supported by NCI (Grant Number: RO1CA39742). Contributions. BT designed the PCR-RFLP, PCRINVADER, and Sequenom assay, assisted in PCR amplification, and performed statistical analysis and manuscript preparation. CFL was involved in study design and collection of paraffin embedded blocks from pathology laboratories while AB and WK were involved in DNA extraction and performing the PCR-INVADER and Sequenom reactions. DJ was involved in statistical analysis. KEA, MDG, and ARF were involved in study design, coordination, and data analysis. All authors read and approved the final manuscript.

References [1] Caldecott KW. Mammalian DNA single-strand break repair: an Xra(y)ted affair. Bioessays 2001;23:447–55. [2] Rao NM, Pai SA, Shinde SR, Ghosh SN. Reduced DNA repair capacity in breast cancer patients and unaffected individuals from breast cancer families. Cancer Genet Cytogenet 1998;102: 65–73. [3] Smith TR, Miller MS, Lohman KK, Case LD, Hu JJ. DNA damage and breast cancer risk. Carcinogenesis 2003;24:883–9. [4] Berwick M, Vineis P. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J Natl Cancer Inst 2000;92:874–97. [5] Cline SD, Hanawalt PC. Who’s on first in the cellular response to DNA damage? Nat Rev Mol Cell Biol 2003;4:361–72. [6] West SC. Molecular views of recombination proteins and their control. Nat Rev Mol Cell Biol 2003;4:435–45. [7] Nash RA, Caldecott KW, Barnes DE, Lindahl T. XRCC1 protein interacts with one of two distinct forms of DNA ligase III. Biochemistry 1997;36:5207–11. [8] Kubota Y, Nash RA, Klungland A, Schar P, Barnes DE, Lindahl T. Reconstitution of DNA base excision-repair with purified human proteins: interaction between DNA polymerase beta and the XRCC1 protein. EMBO J 1996;15:6662–70. [9] Masson M, Niedergang C, Schreiber V, Muller S, Menissier-de Murcia J, de Murcia G. XRCC1 is specifically associated with poly(ADPribose) polymerase and negatively regulates its activity following DNA damage. Mol Cell Biol 1998;18:3563–71. [10] Shen MR, Jones IM, Mohrenweiser H. Nonconservative amino acid substitution variants exist at polymorphic frequency in DNA repair genes in healthy humans. Cancer Res 1998;58:604–8. [11] Stern MC, Umbach DM, van Gils CH, Lunn RM, Taylor JA. DNA repair gene XRCC1 polymorphisms, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 2001;10:125–31. [12] Xing D, Qi J, Miao X, Lu W, Tan W, Lin D. Polymorphisms of DNA repair genes XRCC1 and XPD and their associations with risk of esophageal squamous cell carcinoma in a Chinese population. Int J Cancer 2002;100:600–5. [13] Kim SU, Park SK, Yoo KY, Yoon KS, Choi JY, Seo JS, et al. XRCC1 genetic polymorphism and breast cancer risk. Pharmacogenetics 2002;12:335–8. [14] Smith TR, Miller MS, Lohman K, Lange EM, Case LD, Mohrenweiser HW, et al. Polymorphisms of XRCC1 and XRCC3 genes and susceptibility to breast cancer. Cancer Lett 2003;190:183–90. [15] Han J, Hankinson SE, De Vivo I, Spiegelman D, Tamimi RM, Mohrenweiser HW, et al. A prospective study of XRCC1 haplotypes

B. Thyagarajan et al. / Cancer Detection and Prevention 30 (2006) 313–321

[16]

[17]

[18]

[19]

[20] [21]

[22]

[23]

[24]

and their interaction with plasma carotenoids on breast cancer risk. Cancer Res 2003;63:8536–41. Moullan N, Cox DG, Angele S, Romestaing P, Gerard JP, Hall J. Polymorphisms in the DNA repair gene XRCC1, breast cancer risk, and response to radiotherapy. Cancer Epidemiol Biomarkers Prev 2003;12:1168–74. Duell EJ, Millikan RC, Pittman GS, Winkel S, Lunn RM, Tse CK, et al. Polymorphisms in the DNA repair gene XRCC1 and breast cancer. Cancer Epidemiol Biomarkers Prev 2001;10:217–22. Lunn RM, Langlois RG, Hsieh LL, Thompson CL, Bell DA. XRCC1 polymorphisms: effects on aflatoxin B1-DNA adducts and glycophorin A variant frequency. Cancer Res 1999;59:2557–61. Duell EJ, Holly EA, Bracci PM, Wiencke JK, Kelsey KT. A population-based study of the Arg399Gln polymorphism in X-ray repair cross-complementing group 1 (XRCC1) and risk of pancreatic adenocarcinoma. Cancer Res 2002;62:4630–6. Karran P. DNA double strand break repair in mammalian cells. Curr Opin Genet Dev 2000;10:144–50. Brenneman MA, Wagener BM, Miller CA, Allen C, Nickoloff JA. XRCC3 controls the fidelity of homologous recombination: roles for XRCC3 in late stages of recombination. Mol Cell 2002;10:387– 95. Kuschel B, Auranen A, McBride S, Novik KL, Antoniou A, Lipscombe JM, et al. Variants in DNA double-strand break repair genes and breast cancer susceptibility. Hum Mol Genet 2002;11: 1399–407. Matullo G, Peluso M, Polidoro S, Guarrera S, Munnia A, Krogh V, et al. Combination of DNA repair gene single nucleotide polymorphisms and increased levels of DNA adducts in a population-based study. Cancer Epidemiol Biomarkers Prev 2003;12:674–7. Jacobsen NR, Nexo BA, Olsen A, Overvad K, Wallin H, Tjonneland A, et al. No association between the DNA repair gene XRCC3 T241M polymorphism and risk of skin cancer and breast cancer. Cancer Epidemiol Biomarkers Prev 2003;12:584–5.

321

[25] Han J, Hankinson SE, Ranu H, De Vivo I, Hunter DJ. Polymorphisms in DNA double-strand break repair genes and breast cancer risk in the Nurses’ Health Study. Carcinogenesis 2004;25:189–95. [26] Folsom AR, Zhang S, Sellers TA, Zheng W, Kushi LH, Cerhan JR. Cancer incidence among women living on farms: findings from the Iowa Women’s Health Study. J Occup Environ Med 1996;38:1171–6. [27] Anderson JP, Ross JA, Folsom AR. Anthropometric variables, physical activity, and incidence of ovarian cancer: the Iowa Women’s Health Study. Cancer 2004;100:1515–21. [28] Prineas RJ, Folsom AR, Kaye SA. Central adiposity and increased risk of coronary artery disease mortality in older women. Ann Epidemiol 1993;3:35–41. [29] Zheng W, Deitz AC, Campbell DR, Wen WQ, Cerhan JR, Sellers TA, et al. N-acetyltransferase 1 genetic polymorphism, cigarette smoking, well-done meat intake, and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1999;8:233–9. [30] Thyagarajan B, Anderson KE, Kong F, Selk FR, Lynch CF, Gross MD. New approaches for genotyping paraffin wax embedded breast tissue from patients with cancer: the Iowa Women’s Health Study. J Clin Pathol 2005;58:955–61. [31] Jurinke C, van den Boom D, Cantor CR, Koster H. The use of MassARRAY technology for high throughput genotyping. Adv Biochem Eng Biotechnol 2002;77:57–74. [32] Shu XO, Cai Q, Gao YT, Wen W, Jin F, Zheng W. A population-based case-control study of the Arg399Gln polymorphism in DNA repair gene XRCC1 and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 2003;12:1462–7. [33] Band PR, Le ND, Fang R, Deschamps M. Carcinogenic and endocrine disruptiing effects of cigarette smoke and risk of breast cancer. Lancet 2002;360:1044–9. [34] Duell EJ, Wiencke JK, Cheng TJ, Varkonyi A, Zuo ZF, Ashok TD, et al. Polymorphisms in the DNA repair genes XRCC1 and ERCC2 and biomarkers of DNA damage in human blood mononuclear cells. Carcinogenesis 2000;21:965–71.