Epidemiology

Modifying effects of sulfotransferase 1A1 gene polymorphism on the association of

breast cancer risk with body mass index or endogenous steroid hormones

Gong Yang1, Yu-Tang Gao2, Qiu-Yin Cai1, Xiao-Ou Shu1, Jia-Rong Cheng2, and Wei Zheng11Department of Medicine, Center for Health Services Research and Vanderbilt-Ingram Cancer Center, VanderbiltUniversity Medical Center, Nashville, TN, USA; 2Department of Epidemiology, Shanghai Cancer Institute, Shanghai,China

Key words: breast cancer, estrogen, single nucleotide polymorphism/SNP, sulfotransferase 1A1/SULT1A1

Summary

Sulfotransferase (SULT) 1A1 is involved in the inactivation and elimination of estrogens and catechol estrogens.A common functional polymorphism (Arg213His) has been linked in our previous study of postmenopausalCaucasian women to an elevated risk of breast cancer and the association appeared to be modified by factors relatedto high endogenous estrogen exposures. We further evaluated this polymorphism and levels of BMI and steroidhormones in association with breast cancer risk in a population-based case–control study of Chinese women,involving 1102 incident cases aged 25–64 years and 1147 age-matched population controls. The SULT1A1 genotypewas not associated with overall breast cancer risk in this population. A possible association was suggested forpostmenopausal breast cancer (adjusted odds ratio [OR] = 1.4, 95% CI = 0.9–2.1 for subject carrying the variantHis allele). The SULT1A1 genotype was found to significantly modify postmenopausal breast cancer risk associatedwith a high BMI (‡25 kg/m2) (p for interaction = 0.02), with an adjusted OR of 3.6 (95% CI = 1.5–8.7) forwomen with the Arg/His genotype compared with 1.1 (0.8–1.5) for women with the Arg/Arg genotype (no His/Hisgenotype was identified in this study population). Similarly, the risk associated with a long duration (‡30 years) ofmenstruation also substantially differed by the SULT1A1 genotype (p for interaction = 0.05), with an OR of 4.0(95% CI = 1.3–12.8) for women with the Arg/His genotype and 1.4 (0.8–2.5) for women with the Arg/Arggenotype. Positive associations with blood levels of steroid hormones were also found generally to be more pro-nounced among women carrying the His allele. No similar effect modification was found for premenopausal breastcancer, however. These data suggest that the SULT1A1 Arg213His polymorphism may modify the effect ofendogenous sex hormone exposures on postmenopausal breast cancer risk.

Abbreviations: BMI: body mass index; CI: confidence interval; DHEA-S: dehydroepiandrosterone sulfate; estrone-S: estrone sulfate; OR: odds ratio; SNP: single nucleotide polymorphism; SHBG: sex hormone-binding globulin;SULT: sulfotransferase

Introduction

Cumulative evidence from experimental, epidemiologi-cal and clinical studies has strongly suggested a centralrole of estrogens in the development and growth ofbreast cancer [1,2]. Prolonged exposure to endogenousand exogenous estrogens is associated with a substan-tially increased risk of breast cancer [2,3]. The carcino-genicity of estrogens is, to a large extent, dependent onseveral key-metabolizing processes, such as sulfonation,O-methylation, and glucuronidation [3].

Estrogen sulfonation catalyzed by soluble sulfo-transferases (SULTs) transfers the sulfo group tonucleophilic sites of estrogens to form water-solubleand biologically inactive estrogen sulfates [4–6]. Theconjugates are excreted into the bile or urine, resulting

in reduced levels of estrogen exposure in target tissues.Ten SULT genes encoding 11 proteins have beenidentified in humans [4]. SULT1A1 is considered to bethe predominant type of SULTs due to its extensivetissue distribution, abundance, and broad substratespecificity in terms of various hormones, includingestrogens and catechol estrogens, as well as mostphenolic xenobiotics [7]. A non-synonymous SNP inthe SULT1A1 gene has been identified at nucleotide638 (a G to A transition); this transition leads to anamino-acid substitution at codon 213 (Arg to His)[8–10]. Individuals homozygous for the variant Hisallele have only about 10% of the enzyme activity(measured in platelets) of the Arg/Arg homozygotes [9].Given the role of SULTs in the metabolism of estro-gens and environmental carcinogens, it has been

Breast Cancer Research and Treatment (2005) 94: 63–70 � Springer 2005DOI 10.1007/s10549-005-7280-2

hypothesized that genetic polymorphisms in SULT1A1may relate to breast cancer susceptibility [11]. In ourprevious study of postmenopausal Iowa women, wereported the first evidence that homozygosity for theSULT1A1 His213 allele was associated with anincreased risk of breast cancer, and the associationappeared to be stronger among women with a largerBMI or other conditions related to high endogenousestrogen exposures [11]. Several other studies, however,have given inconsistent results [12–16].

Ethnic variation in allele frequencies has beenobserved [17]. The variant His213 allele was found to bemore common in Caucasian populations (30%) than inAsian populations (8% in Chinese) [17]. It is unknownwhether this variant allele is also a risk factor for breastcancer in Chinese population and whether the lowprevalence of this variant allele in Chinese can provideone potential explanation for its low incidence of breastcancer. In the present study of Chinese women inShanghai, we further investigated the association ofSULT1A1 polymorphism with breast cancer risk andevaluated, for the first time, the potential interactiveeffect of the SULT1A1 polymorphism and endogenousestrogens using measured blood levels of sex steroidhormones.

Materials and methods

Included in this study were subjects participating in theShanghai Breast Cancer Study, a population-basedcase–control study conducted among Chinese women inShanghai during the time period of 1996–1998. Detailsof study methods have been reported previously [18,19].Briefly, this study was designed to recruit all women,25–64 years of age, who were newly diagnosed withbreast cancer between August 1996 and March 1998,and a representative random sample of controls selectedfrom the general population. Eligible participants werepermanent residents of urban Shanghai who had noprior history of cancer and were alive at the time ofinterview. Through a rapid case-ascertainment system,supplemented by the population-based Shanghai TumorRegistry, 1602 eligible breast cancer cases were identi-fied during the study period, and in-person interviewswere completed for 1459 (91.1%) of them. The majorreasons for nonparticipation were refusal (109 cases,6.8%), death prior to interview (17 cases, 1.1%), andinability to locate (17 cases, 1.1%). Cancer diagnoses forall patients were confirmed by two senior studypathologists through the review of tumor slides. Con-trols were randomly selected from the general femalepopulation using the Shanghai Resident Registry andfrequency-matched to cases on age (5-year intervals).The number of controls in each age-specific stratum wasdetermined in advance according to the age distributionof the incident breast cancer cases reported to theShanghai Tumor Registry from 1990 to 1993. In-personinterviews were completed for 1556 (90.3%) of the 1724

eligible. Reasons for nonparticipation included refusal(166 controls, 9.6%) and death before interview(2 controls, 0.1%).

A structured questionnaire was used to elicit detailedinformation on demographic factors, dietary habits,physical activity, tobacco and alcohol use, menstrualand reproductive history, hormone use, medical history,and family history of cancer. All participants were alsomeasured for weight, height, and circumferences of thewaist and hips. Fasting blood samples, 10 ml from eachwoman, were collected in the morning from 1193 (82%)cases and 1310 (84%) controls using EDTA or heparinvacutainer tubes. Immediately after collection, thesamples were placed in portable insulated cases with icepacks (0–4�C) and transported to a core laboratory forprocessing. All samples were aliquoted and stored at)70 �C within 6 h of collection. There were no signifi-cant differences in demographic factors (such as age,education, and home income), lifestyle factors (such ascigarette smoking, alcohol drinking, and physicalactivity) and major risk factors (such as BMI, repro-ductive factors, history of fibroadenoma, and familyhistory of breast cancer) between these bio-specimendonators and the remaining participants of the ShanghaiBreast Cancer Study [20].

Genomic DNA was extracted from buffy coat frac-tions. SULT1A1 genotyping was conducted usingmatrix assisted laser desorption/ionization time-of-flightmass spectrometry (MALDI-TOFMS) (Sequenom, SanDiego, CA). This method includes an initial step of PCRamplification of the SULT1A1 gene region containingthe polymorphism followed by allele-specific primerextension using an extension primer that creates twoproducts differing by one to three nucleotides 3¢ fromthe single nucleotide polymorphism. The allele-specificextension products are then distinguished by massspectrometry. The initial PCR step of the SUL1A1 genecontaining the polymorphic regions was conductedusing 5¢-ACGTTGGATGTGCTGAACCATGAAGTCCAC-3¢ and 5¢-ACGTTGGATGGAACCCCAAAAGGGAGATTC-3¢. PCR was performed in a total volumeof 5 ll including 5 ng of DNA, 1· PCR buffer, 200 nMeach primer, 2.5 mM MgCl2, 300 lM of each dNTPs,and 0.1 unit of Taq polymerase enzyme (Solis Biodyne,Tartu, Estonia). Thermal cycling conditions were asfollows: 95�C for 15 min, 45 cycles of 95�C for 20 s,56�C for 30 s and 72�C for 1 min with a final extensionstep of 72�C for 3 min. Shrimp Alkaline Phosphatase(SAP) was added to samples to dephosphorylate anyresidual nucleotides, and prevent their future incorpo-ration and interference with the primer extension step.Purified PCR products were used for the primer exten-sion reaction after the addition of the extension mix(total reaction volume was 9 ll), containing 600 nMextension primer (5¢-ATCCTGGAGTTTGTGGGGC3¢), 50 lM each ddNTP, 1· buffer, 0.01 unit of TER-MIpol enzyme (Solis Biodyne). Primer extension reac-tions were performed at 94�C for 2 min followed by 40cycles of 94�C for 5 s, 52�C for 5 s, and 72�C for 5 s.

64 G Yang et al.

Primer extension products (Arg allele: 5¢-ATCCTGGAGTTTGTGGGGCGC)3¢; and His allele: 5¢-ATCCTGGAGTTTGTGGGGCA-3¢) were then purifiedusing SpectroCLEAN resin to remove salt ions.Approximately 15 ll of each extension product wastransferred, or ‘spotted’ onto corresponding elements ofa silicon microchip forming a crystalline matrix. Thematrix was pulsed with ultraviolet light causing ioniza-tion of the analyte molecules. The raw spectral data wasprocessed using SprectroTYPER software. The softwareassigned genotype probabilities using the programmedassay definitions and calculated mass differences.Genotyping was performed in batches containing anequal number of cases and controls.

The laboratory staff was blind to the identity of thesubjects. Quality control (QC) samples were included ingenotyping assays. Each 96-well plate of genomic DNAcontained multiple controls, including one water blank,two samples of CEPH 1347-02, two unblinded QC, andtwo blinded QC samples. QC samples were distributedacross separate 96-well plates. The SULT1A1 genotypesdetermined for the 56 blinded QC samples and those forthe 56 unblinded QC samples were in complete agree-ment with the genotypes determined for the studysamples.

A detailed description of the measurement of steroidhormones and SHBG has been previously reported [21].Assays were completed in two batches in 2001 and 2003,respectively. Plasma concentrations of testosterone,estradiol, estrone, estrone-S, DHEA-S and progesteronewere measured directly without extraction. The mea-surement of steroids and SHBG in our study was con-ducted by Diagnostic Systems Laboratories, Inc. (DSL,Webster, TX), a reference laboratory certified by Clini-cal Laboratory Improvement Amendments and theInternational Standard ISO 9001. Commercial radio-immunoassay kits from DSL were used for the mea-surement of steroids and an immunoradiometric assaykit from DSL was used for SHBG. Technicians whoperformed the tests were also blinded to the source ofthe specimens. Data from premenopausal women werenot included in the present study due to concerns that asingle measurement was not sufficient to capture thelarge hormonal variation that occurs during the men-strual cycle.

We used v2 statistics to analyze the distributions ofthe SULT1A1 genotypes in cases and controls. Becausethe distributions of the measures of steroid hormonesand SHBG were skewed, the Wilcoxon signed rank testwas used for comparisons of the median differencesbetween cases and controls. Odds ratios (ORs),approximators of relative risk, were used to measure theassociation of breast cancer risk with SULT1A1 geno-types [22]. Unconditional logistic regression was used toobtain maximum likelihood estimates of the ORs andtheir 95% confidence intervals (CIs) and to adjust forpotential confounders. Stratified analyses were con-ducted to evaluate the modifying effect of SULT1A1Arg213His polymorphism on the associations of breast

cancer with hormone-related conditions (BMI, calcu-lated as weight in kilograms divided by the square ofheight in meters, and years of menstruation, defined asage at menopause or age at interview for premenopausalwomen minus age at menarche) and blood steroid hor-mones (for postmenopausal women only). To minimizethe potential influence of cancer diagnosis and treatmenton sex hormone production and metabolism, cases thatdonated a blood sample after any cancer treatment wereexcluded. Batch-specific medians were used to categorizeindividuals into two groups (<median and ‡median) toeliminate between-batch variations in sex-hormonemeasures [23]. Tests for interaction were performed byintroducing a multiplicative interaction term into themain effect models. All tests of statistical significancewere based on two-sided probability. Statistical analyseswere carried out using SAS, version 8.01 (SAS InstituteInc, Cary, NC).

Results

The distributions of selected demographic characteris-tics and major risk factors for breast cancer by case–control status and menopausal status are shown inTable 1. Cases and controls were comparable in agedistribution and education attainment. Consistentlywith previous studies [24,25], the risk of breast cancerwas found to be positively and significantly associatedwith an early age at menarche ( £ 13 years), a late age atfirst live birth (‡30 years), a long duration of menstru-ation (‡30 years), and no regular physical activity. Allthese variables, as well as age at diagnosis for cases orage at interview for controls, were then included in thesubsequent unconditional logistic models to control fortheir potential confounding effects.

Overall, breast cancer patients were more likely tohave a higher level of circulating steroid hormones and alower level of SHBG than controls (Table 1). A 1.5- to1.8-fold increase in breast cancer risk was found amongpostmenopausal women who had median or higherlevels of estrone, estrone-S, testosterone, and DHEA-Scompared with those who had lower levels of thesemarkers; and a marginally significantly increased riskwas observed for women having lower SHBG levels.

The frequencies of SULT1A1 genotypes by case–control status are presented in Table 2. The genotypedistribution is inconsistent with Hardy–Weinberg equi-librium for both cases (p = 0.003) and controls(p = 0.007). No His/His genotype was identified in thisstudy population. The SULT1A1 genotype was notassociated with overall breast cancer risk. However, apossible association was suggested for postmenopausalbreast cancer (adjusted OR = 1.4, 95% CI = 0.9–2.1).

The associations of breast cancer risk with BMI andmenstruation years, as well as blood hormone levels bythe SULT1A1 genotype are presented in Tables 3 and4, respectively. The SULT1A1 genotype significantlymodified the association between BMI and

Sulfotransferase 1A1 polymorphism and breast cancer risk 65

postmenopausal breast cancer risk (p for multiplicativeinteraction = 0.02). Among women with the Arg/Hisgenotype, a high BMI (‡25 kg/m2) was associated withan OR of 3.6 (95% CI = 1.5–8.7), whereas the cor-responding OR among women with the Arg/Arg

genotype was 1.1 (95% CI = 0.8 – 1.5) (Table 3).Similarly, the association between postmenopausalbreast cancer and years of menstruation also differedmarkedly by the SULT1A1 genotype (p for interaction= 0.05), with an OR of 4.0 (95% CI = 1.3–12.8) for

Table 1. Comparison of case and control subjects by selected demographic characteristics and major risk factors for breast cancer, the Shanghai

Breast Cancer Study, 1996–1998

Subject characteristic Premenopausal women Postmenopausal women

Cases

(n = 738)

Controls

(n = 726)

OR (95%

CI)aCases

(n = 364)

Controls

(n = 421)

OR

(95% CI)a

Demographic factors

Age (25th, 75th percentile), years 43 (40–47) 42 (39–46) 57 (53–61) 57 (53–61)

Education, lower than middle school (%) 3.8 3.3 0.9 (0.5–1.5) 29.4 33.7 0.8 (0.6–1.1)

Major risk factors

Family history of breast cancer (%) 3.3 1.7 2.0 (1.0–4.0) 3.6 3.6 1.0 (0.5–2.1)

Age at menarche, £ 13 years (%) 31.3 28.1 1.3 (1.0–1.6) 30.9 22.9 1.5 (1.1–2.1)

Age at menopause, ‡50 years (%)b 40.4 37.9 1.1 (0.8–1.5)

Age at first live birth, ‡30 years (%)c 24.6 20.0 1.2 (1.0–1.6) 17.6 9.6 2.0 (1.3–3.1)

Years of menstruation, ‡30 years (%)d 55.5 44.5 0.9 (0.6–1.2) 81.8 76.4 1.5 (1.0–2.1)

No regular physical activity (%) 86.3 84.4 1.3 (0.9–1.7) 70.3 56.7 1.9 (1.4–2.5)

Waist-to-hip ratio, ‡0.85 (%) 16.8 13.5 1.2 (0.9–1.7) 29.4 24.0 1.3 (1.0–1.8)

Body mass index, ‡25 kg/m2 (%) 24.3 21.4 1.1 (0.8–1.4) 43.4 38.2 1.2 (0.9–1.7)

Measured blood sex-hormone levelsb,e

Estradiol, ‡median (%) 51.2 48.1 1.1 (0.8–1.6)

Estrone, ‡median (%) 58.9 48.1 1.5 (1.1–2.2)

Estrone sulfate, ‡median (%) 60.9 49.7 1.6 (1.1–2.2)

Testosterone, ‡median (%) 63.6 49.7 1.8 (1.2–2.5)

DHEA-S, ‡median (%) 61.2 49.7 1.6 (1.2–2.3)

SHBG, <median (%) 56.1 50.6 1.3(0.9–1.7)

aAge-adjusted ORs estimated by unconditional logistic regression models.bAmong postmenopausal women only.cAmong parous women only.dYears of menstruation = menopausal age or age at interview for premenopausal women ) menarche age.eAmong postmenopausal women, sex hormone profiles were measured in two batch assays for 205 cases and 441 controls.

Subjects were categorized into 2 groups based on the batch-specific median of sex steroid hormones or SHBG in controls. Medians in batches 1

and 2 were 5.7 and 11.2 pg/ml for estradiol, 17.7 and 17.0 pg/ml for estrone, 978 and 930 pg/ml for estrone-S, 161 and 220 pg/ml for testosterone,

611 and 650 ng/ml for DHEA-S, and 75 and 97 nmol/l for SHBG.

Table 2. Associations of SULT1A1 genotypes and breast cancer risk according to age and menopausal status, the Shanghai Breast Cancer Study,

1996–1998

SULT1A1 genotype p Adjusted OR (95% CI)a

for Arg/His versus Arg/ArgArg/Arg Arg/His

No. of cases No. of controls No. of cases No. of controls

All subjects 921 977 181 170 0.29 1.1 (0.9–1.4)

Age, yearsb

25–44 369 402 76 70 0.35 1.1 (0.8–1.7)

45–54 361 333 59 53 0.90 0.9 (0.6–1.4)

55–65 191 242 46 47 0.35 1.4 (0.8–2.2)

Menopausal status

Pre- 622 614 116 112 0.88 1.0 (0.7–1.3)

Post- 299 363 65 58 0.12 1.4 (0.9–2.1)

aOdds ratio (OR) estimated by unconditional logistic regression models and adjusted for age at diagnosis for cases or at interview for controls

(continuous), age at menarche ( £ 13 or >13), age at the first live birth(<30 or ‡30 ), physical activity (yes or no), and duration of menstruation

(<30 or ‡30).bp for SULT1A1 genotype frequency across age groups, 0.18 for cases and 0.66 for controls.

66 G Yang et al.

Table 3. Association of BMI and years of menstruation with breast cancer by SULT1A1 genotype, the Shanghai Breast Cancer Study, 1996–

1998

SULT1A1 Arg/Arg genotype SULT1A1 Arg/His genotype p for interaction

Cases/controls OR (95% CI)a Cases/controls OR (95% CI)a

All subjects

Body mass index (kg/m2)

<25 650/699 1.0 (reference) 115/132 1.0 (reference)

‡25 271/278 1.0 (0.8–1.3) 66/38 2.1 (1.3–3.6) 0.006

Year of menstruation

<30 (median) 385/467 1.0 (reference) 72/79 1.0 (reference)

‡30 535/508 1.3 (1.0–1.7) 108/91 1.6 (0.9–2.8) 0.51

Premenopausal women

Body mass index (kg/m2)

<25 475/478 1.0 (reference) 84/93 1.0 (reference)

‡25 147/136 0.9 (0.7–1.2) 32/19 1.7 (0.9–3.3) 0.16

Year of menstruation

<30 330/381 1.0 (reference) 61/66 1.0 (reference)

‡30 291/232 0.8 (0.6–1.2) 55/46 1.0 (0.4–2.7) 0.35

Postmenopausal women

Body mass index (kg/m2)

<25 175/221 1.0 (reference) 31/39 1.0 (reference)

‡25 124/142 1.1 (0.8–1.5) 34/19 3.6 (1.5–8.7) 0.02

Year of menstruation

<30 55/86 1.0 (reference) 11/13 1.0 (reference)

‡30 244/276 1.4 (0.9–2.2) 53/45 4.0 (1.3–12.8) 0.05

aORs are adjusted for age at diagnosis for cases or at interview for controls (continuous), age at menarche ( £ 13 or >13), age at the first live

birth(<30 or ‡30 ), and physical activity (yes or no).

Table 4. Association of hormonal measurements with postmenopausal breast cancer by SULT1A1 genotype, the Shanghai Breast Cancer Study,

1996–1998

Levels of sex-steroid hormones and SHBGa SULT1A1 Arg/Arg genotype SULT1A1 Arg/His genotype

Cases/controls OR (95% CI)b Cases/controls OR (95% CI)b

Estradiol

Low 77/177 1.0 (reference) 20/32 1.0 (reference)

High 83/170 1.1 (0.8–1.7) 12/25 0.8 (0.3–2.1)

Estrone

Low 69/180 1.0 (reference) 11/32 1.0 (reference)

High 90/164 1.4 (1.0–2.1) 21/24 2.8 (1.1–7.5)

Estrone-S

Low 63/170 1.0 (reference) 14/36 1.0 (reference)

High 96/177 1.5 (1.0–2.2) 18/21 2.2 (0.9–5.6)

Testosterone

Low 56/172 1.0 (reference) 14/32 1.0 (reference)

High 104/175 1.8 (1.2–2.6) 18/25 2.0 (0.8–5.3)

DHEA-S

Low 64/170 1.0 (reference) 11/29 1.0 (reference)

High 94/173 1.4 (0.9–2.1) 21/26 2.7 (1.02–7.2)

SHBG

High 74/170 1.0 (reference) 11/29 1.0 (reference)

Low 88/177 1.2 (0.8–1.7) 21/28 1.9 (0.7–4.8)

aSubjects grouped by batch-specific medians of sex-steroid hormones and SHBG as defined in Table 1.bORs are adjusted for age at diagnosis for cases or at interview for controls (continuous), age at menarche ( £ 13 or >13), age at the first live

birth(<30 or ‡30 ), and physical activity (yes or no).

Sulfotransferase 1A1 polymorphism and breast cancer risk 67

women with the Arg/His genotype and 1.4 (0.8–2.5) forwomen with the Arg/Arg genotype (Table 3). Positiveassociations with blood levels of steroid hormones werealso found generally to be more pronounced amongpostmenopausal women carrying the His allele, al-though most of the tests for multiplicative interactiondid not reach statistical significance, probably due tothe small sample size in this subset of data (Table 4).No interactive effect of the SULT1A1 genotype withBMI or duration of menstruation was found on pre-menopausal breast cancer risk.

In addition, we compared the concentrations of sul-fated sex hormones between subjectswith andwithout thevariant His allele (encoding a lower-activity allozyme).Due to the large variation in the levels of steroid hor-mones during the menstrual cycle, we included onlypostmenopausal women in the comparisons; and due tothe between-batch variations in sex-hormone measures,we used data from the second batch, counting for 70% ofthe total of postmenopausal controls (282 out of 404)whose samples were measured for both the SULT1A1genotype and sex hormones of interest. Wemeasured twosulfated sex hormones, estrone-S andDHEA-S, in blood.The median of estrone-S among women carrying the Hisallele (790 pg/ml) was significantly lower compared withwomen homozygous for the Arg allele (960 pg/ml;p = 0.04). Similarly, a lower concentration of DHEA-Swas also found among women with the variantHis allele(median = 580 versus 660 ng/ml; p = 0.02).

Discussion

In this large population-based case–control study, wefound that the SULT1A1 Arg213His polymorphism wasweakly associated with the risk of postmenopausalbreast cancer. However, there was a strong indicationthat this functional polymorphism may modify theassociation of endogenous sex hormone exposures withpostmenopausal breast cancer risk. In the presence ofthe low-activity His allele, women with high BMI orlong duration of menstruation had an odds of devel-oping breast cancer nearly 4 times greater than thosewith low BMI or short duration of menstruation. Nosuch associations were observed in the absence of thisvariant allele. Consistently, positive associations withlevels of circulating steroid hormones (such as estrone,estrone-S, and DHEA-S) were also found to be moreevident in the presence than in the absence of the Hisallele. The findings from this study of Chinese womenare, in general, consistent with our previous study of thepostmenopausal U.S. women [11], in which a positiveassociation was also observed between the His allele andpostmenopausal breast cancer risk, particularly amongthose with risk factors related to higher endogenousestrogen exposure, including a high BMI, earlier age atmenarche, and later age at menopause.

Our results are biologically plausible. Estrogens arebelieved to play a critical role in the pathogenesis of

breast cancer [1, 2]. Previous epidemiological studies,including the Shanghai Breast Cancer Study, haverepeatedly linked high levels of estrogens and low levelsof SHBG to an elevated risk of breast cancer, particu-larly among postmenopausal women [21, 26, 27]. Aftermenopause, adipose tissue becomes the primary sourceof estrogen production and a high BMI has been asso-ciated with an increased risk of breast cancer amongpostmenopausal women [28]. We showed in this reportthat the associations of postmenopausal breast cancerwith BMI and years of menstruation, as well as bloodhormone levels, were modified by the SULT1A1Arg213His polymorphism, which is known to influenceenzyme activity and stability. Primarily in the presenceof the low-activity His allele, higher BMI and longerexposure to endogenous hormones significantly in-creased postmenopausal breast cancer risk. Theobserved modifying effect of the SULT1A1 genotype onpostmenopausal breast cancer risk related to endoge-nous hormone exposure is consistent with the role ofSULT1A1 in the inactivation of estrogens.

In addition to its role in the metabolism of estrogens,SULT1A1 is also known to be involved in the bioacti-vation of certain pro-carcinogens, including polycyclicaromatic hydrocarbons and N-hydroxylated heterocy-clic amines, which are present in well-done meat andtobacco smoke [4, 7]. In keeping with the latter role ofSULT1A1, only among carriers of high-activity Argallele, well-done meat and cigarette smoking wereassociated with increased cancer risk in previous studies[11, 13, 29]. Because of the dual effect of SULT1A1 onboth activation and inactivation reactions, the associa-tion between the SULT1A1 genotype and breast cancerrisk could be complicated. In the present study, less than5% of women smoked cigarettes or consumed roastedmeat/chicken regularly, and adjustment for these factorsas well as occupation did not alter the positive associa-tion with the variant allele.

Several other groups have also investigated theSULT1A1 Arg213His polymorphism in relation tobreast cancer. In a hospital-based case–control study(103 cases, 97 benign breast disease controls, and 133healthy controls), Tang and colleagues found a positiveassociation between breast cancer risk and the numberof the His allele regardless of control groups being used(benign breast disease controls or healthy controls) [12].Similarly, an increased risk of breast cancer amongsubjects carrying two copies of the His allele was alsonoted in another recently published case–control study[14]. A case-only study evaluating the effect of gene-environment interaction on breast cancer risk lentadditional support to the association of the SULT1A1polymorphism with breast cancer risk [13]. In contrast,an earlier study reported no association with breastcancer risk at younger ages [15]. Recently, anotherstudy also reported a null association but suggested alink between the SULT1A1 polymorphism and pres-ence of lymph node metastases in breast cancer [30].Taken together, these data, in line with ours, suggested

68 G Yang et al.

a potential role for the SULT1A1 polymorphism inbreast caner risk.

Striking ethnic differences in SULT1A1 allele fre-quencies have been noted. Compared with Caucasiansand African-Americans, the His allele is less common inChinese [17]. In the present study, no homozygous Hisgenotype was identified and the genotype distributionsin both cases and controls deviated from Hardy–Weinberg equilibrium. We do not have a clear expla-nation for this finding. Since the majority of our studysubjects belong to a single (Han Chinese) ethnic group,ethnic admixture is unlikely to be attributor. It is alsounlikely to be explained by small population variation(genetic drift), given the large sample size we had. Torule out the possibility of genotyping error, we repeatedthe assay among 112 (�5%) randomly selected samples,and found the same genotype distribution. The sameblood samples have been used to genotype for manyother polymorphisms in genes involving in estrogensynthesis or metabolism and the data are consistent withHardy–Weinberg equilibrium [20,23,34–37]. Studiesconducted by several other independent groups alsoobserved no or extremely low prevalence of His/Hisgenotype in Chinese populations [31–33]. For example,two studies in Taiwan all found absence of the His/Hisgenotype among both a small (n = 308) [32] and a largenumber (n = 740) [31] of controls. The allele frequen-cies reported in these studies are similar to those seen inour study, suggesting that our study subjects are repre-sentative of Chinese population.

One limitation of the study is that postdiagnosticblood samples were used in the assays of sex hormoneprofiles. However, most of the blood samples were col-lected within 1–2 weeks of cancer diagnosis and beforethe initiation of any cancer therapy, and thus thepotential influence related to cancer diagnosis andtreatment on biomarker measurement was minimized.Potential errors in hormone measurements might exist,particularly for estradiol due to the very low bloodconcentrations among these postmenopausal women.However, all the assays were performed with laboratorypersonnel blinded to case–control status and levels ofthese hormones were shown, consistently with the lit-erature, to be correlated with BMI and other knownbreast cancer risk factors in our previous study [38].These suggest that the potential measurement errors areunlikely to systematically bias the overall findings on thehormones. The strengths of our study include a largesample size, the use of population-based controls, and ahigh participation rate, which minimized selectionbiases.

In summary, this large population-based case–con-trol study provides support for a potential role of theSULT1A1 polymorphism in postmenopausal breastcancer risk. Postmenopausal women carrying thevariant SULT1A1 His allele may be more susceptible toestrogens- and their metabolites-induced carcinogenesisin mammary tissue. These findings suggest that theSULT1A1 genotype may provide additional informa-

tion beyond that conveyed by known breast cancer riskfactors in identifying individuals at high risk for post-menopausal breast cancer.

Acknowledgements

We are grateful to the patients and research staff whoparticipated in the Shanghai Breast Cancer Study. Thisresearch was supported by USPHS Grant RO1CA64277and RO1CA90899 from the National Caner Institute.

References

1. Clemons M, Goss P: Estrogen and the risk of breast cancer. N

Engl J Med 344: 276–285, 2001

2. Parl FF: Estrogens as breast cancer risk factor In: Parl FF (ed)

Estrogens, Estrogen Receptor and Breast Cancer. IOS Press

Amsterdam, Netherlands, 2000, pp 205–242.

3. Gruber CJ, Tschugguel W, Schneeberger C, Huber JC: Production

and actions of estrogens. N Engl J Med 346: 340–352, 2002

4. Glatt H, Meinl W: Pharmacogenetics of soluble sulfotransferases

(SULTs). Naunyn Schmiedebergs. Arch Pharmacol (online pub-

lication), 2003, Nov 5

5. Weinshilboum RM, Otterness DM, Aksoy IA, Wood TC, Her C,

Raftogianis RB: Sulfation and sulfotransferases 1: sulfotransferase

molecular biology: cDNAs and genes. FASEB J 11: 3–14, 1997

6. Falany CN: Enzymology of human cytosolic sulfotransferases.

FASEB J 11: 206–216, 1997

7. Coughtrie MW: Sulfation through the looking glass – recent

advances in sulfotransferase research for the curious. Pharmac-

ogenomics J 2: 297–308, 2002

8. Raftogianis RB, Wood TC, Otterness DM, Van Loon JA, Wein-

shilboum RM: Phenol sulfotransferase pharmacogenetics in

humans: association of common SULT1A1 alleles with TS PST

phenotype. Biochem Biophys Res Commun 239: 298–304, 1997

9. Raftogianis RB, Wood TC, Weinshilboum RM: Human phenol

sulfotransferases SULT1A2 and SULT1A1: genetic polymor-

phisms, allozyme properties, and human liver genotype–phenotype

correlations. Biochem Pharmacol 58: 605–616, 1999

10. Coughtrie MW, Gilissen RA, Shek B, Strange RC, Fryer AA,

Jones PW, Bamber DE: Phenol sulphotransferase SULT1A1

polymorphism: molecular diagnosis and allele frequencies in

Caucasian and African populations. Biochem J 337: 45–49, 1999

11. Zheng W, Xie D, Cerhan JR, Sellers TA, Wen W, Folsom AR:

Sulfotransferase 1A1 polymorphism, endogenous estrogen expo-

sure, well-done meat intake, and breast cancer risk. Cancer Epi-

demiol Biomarkers Prev 10: 89–94, 2001

12. Tang D, Rundle A, Mooney L, Cho S, Schnabel F, Estabrook A,

Kelly A, Levine R, Hibshoosh H, Perera F: Sulfotransferase 1A1

(SULT1A1) polymorphism, PAH-DNA adduct levels in breast

tissue and breast cancer risk in a case–control study. Breast Cancer

Res Treat 78: 217–222, 2003

13. Saintot M, Malaveille C, Hautefeuille A, Gerber M: Interactions

between genetic polymorphism of cytochrome P450-1B1, sulfo-

transferase 1A1, catechol-o-methyltransferase and tobacco expo-

sure in breast cancer risk. Int J Cancer 107: 652–657, 2003

14. Han DF, Zhou X, Hu MB, Wang CH, Xie W, Tan XD, Zheng F,

Liu F: Sulfotransferase 1A1 (SULT1A1) polymorphism and breast

cancer risk in Chinese women. Toxicol Lett 150: 167–177, 2004

15. Seth P, Lunetta KL, Bell DW, Gray H, Nasser SM, Rhei E, Kaelin

CM, Iglehart DJ, Marks JR, Garber JE, Haber DA, Polyak K:

Phenol sulfotransferases: hormonal regulation, polymorphism,

and age of onset of breast cancer. Cancer Res 60: 6859–6863, 2000

16. Lilla C, Risch A, Kropp S, Chang-Claude J: SULT1A1 genotype,

activeandpassivesmoking,andbreastcancerriskbyage50 years ina

German case–control study. Breast Cancer Res 7: R229–R237, 2005

Sulfotransferase 1A1 polymorphism and breast cancer risk 69

17. Carlini EJ, Raftogianis RB, Wood TC, Jin F, Zheng W, Rebbeck

TR, Weinshilboum RM: Sulfation pharmacogenetics: SULT1A1

and SULT1A2 allele frequencies in Caucasian, Chinese and Afri-

can-American subjects. Pharmacogenetics 11: 57–68, 2001

18. ZhengW,XieDW, JinF, Cheng JR,DaiQ,WenWQ, ShuXO,Gao

YT: Genetic polymorphism of cytochrome P450-1B1 and risk of

breast cancer. Cancer Epidemiol Biomarkers Prev 9: 147–150, 2000

19. Yang G, Lu G, Jin F, Dai Q, Best R, Shu XO, Chen JR, Pan XY,

Shrubsole M, Zheng W: Population-based, case–control study of

blood C-peptide level and breast cancer risk. Cancer Epidemiol

Biomarkers Prev 10: 1207–1211, 2001

20. Zheng W, Gao YT, Shu XO, Wen W, Cai Q, Dai Q, Smith JR:

Population-based case–control study of CYP11A gene polymor-

phism and breast cancer risk. Cancer Epidemiol Biomarkers Prev

13: 709–714, 2004

21. Yu H, Shu XO, Shi R, Dai Q, Jin F, Gao YT, Li BD, Zheng W:

Plasma sex steroid hormones and breast cancer risk in Chinese

women. Int J Cancer 105: 92–97, 2003

22. Breslow NE, Day NE: Statistical Methods in Cancer Research:

Analysis of Case–control Studies (IARC Sci Publ 32). IARC,

Lyon, France, 1980

23. Zheng SL, Zheng W, Chang BL, Shu XO, Cai Q, Yu H, Dai Q, Xu

J, Gao YT: Joint effect of estrogen receptor beta sequence variants

and endogenous estrogen exposure on breast cancer risk in Chi-

nese women. Cancer Res 63: 7624–7629, 2003

24. Gao YT, Shu XO, Dai Q, Potter JD, Brinton LA, Wen W, Sellers

TA, Kushi LH, Ruan Z, Bostick RM, Jin F, ZhengW: Association

of menstrual, reproductive factors with breast cancer risk: results

from the Shanghai Breast Cancer Study. Int J Cancer 87: 295–300,

2000

25. McPherson K, Steel CM, Dixon JM: ABC of breast diseases

Breast cancer-epidemiology, risk factors, and genetics. BMJ 321:

624–628, 2000

26. Key TJ, Verkasalo PK: Endogenous hormones and the aetiology

of breast cancer. Breast Cancer Res 1: 18–21, 1999

27. Key TJ, Appleby PN, Reeves GK, Roddam A, Dorgan JF,

Longcope C, Stanczyk FZ, Stephenson HE Jr, Falk RT, Miller R,

Schatzkin A, Allen DS, Fentiman IS, Key TJ, Wang DY, Dowsett

M, Thomas HV, Hankinson SE, Toniolo P, Akhmedkhanov A,

Koenig K, Shore RE, Zeleniuch-Jacquotte A, Berrino F, Muti P,

Micheli A, Krogh V, Sieri S, Pala V, Venturelli E, Secreto G,

Barrett-Connor E, Laughlin GA, Kabuto M, Akiba S, Stevens

RG, Neriishi K, Land CE, Cauley JA, Kuller LH, Cummings SR,

Helzlsouer KJ, Alberg AJ, Bush TL, Comstock GW, Gordon GB,

Miller SR, Longcope C: Endogenous Hormones Breast Cancer

Collaborative Group: Body mass index, serum sex hormones, and

breast cancer risk in postmenopausal women. J Natl Cancer Inst

95: 1218–1226, 2003

28. Shu XO, Jin F, Dai Q, Shi JR, Potter JD, Brinton LA, Hebert JR,

Ruan Z, Gao YT, Zheng W: Association of body size and fat

distribution with risk of breast cancer among Chinese women. Int

J Cancer 94: 449–455, 2001

29. Hung RJ, Boffetta P, Brennan P, Malaveille C, Hautefeuille A,

Donato F, Gelatti U, Spaliviero M, Placidi D, Carta A, Scotto di

Carlo A, Porru S: GST, NAT, SULT1A1, CYP1B1 genetic poly-

morphisms, interactions with environmental exposures and

bladder cancer risk in a high-risk population. Int J Cancer 110:

598–604, 2004

30. Langsenlehner U, Krippl P, Renner W, Yazdani-Biuki B, Eder T,

Wolf G, Wascher TC, Paulweber B, Weitzer W, Samonigg H:

Genetic variants of the sulfotransferase 1A1 and breast cancer risk.

Breast Cancer Res Treat 87: 19–22, 2004

31. Cheng TC, Chen ST, Huang CS, Fu YP, Yu JC, Cheng CW, Wu

PE, Shen CY: Breast cancer risk associated with genotype poly-

morphism of the catechol estrogen-metabolizing genes: a multi-

genic study on cancer susceptibility. Int J Cancer 113: 345–353,

2005

32. Wu MT, Wang YT, Ho CK, Wu DC, Lee YC, Hsu HK, Kao EL,

Lee JM: SULT1A1 polymorphism and esophageal cancer in males.

Int J Cancer 103: 101–104, 2003

33. Liang G, Miao X, Zhou Y, Tan W, Lin D: A functional poly-

morphism in the SULT1A1 gene (G638A) is associated with risk of

lung cancer in relation to tobacco smoking. Carcinogenesis 25:

773–778, 2004

34. Adegoke OJ, Shu XO, Gao YT, Cai Q, Breyer J, Smith J, Zheng

W: Genetic polymorphisms in uridine diphospho-glucuronosyl-

transferase 1A1 (UGT1A1) and risk of breast cancer. Breast

Cancer Res Treat 85: 239–245, 2004

35. Egan KM, Cai Q, Shu XO, Jin F, Zhu TL, Dai Q, Gao YT, Zheng

W: 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, 2004

36. Fowke JH, Shu XO, Dai Q, Jin F, Cai Q, Gao YT, Zheng W: Oral

contraceptive use and breast cancer risk: modification by

NAD(P)H:quinine oxoreductase (NQO1) genetic polymorphisms.

Cancer Epidemiol Biomarkers Prev 13: 1308–1315, 2004

37. Wen W, Cai Q, Shu XO, Cheng JR, Parl F, Pierce L, Gao YT,

Zheng W: Cytochrome P450 1B1 and catechol-O-methyltransfer-

ase genetic polymorphisms and breast cancer risk in Chinese wo-

men: results from the shanghai breast cancer study and a meta-

analysis. Cancer Epidemiol Biomarkers Prev 14: 329–335, 2005

38. Boyapati SM, Shu XO, Gao YT, Dai Q, Yu H, Cheng JR, Jin F,

Zheng W: Correlation of blood sex steroid hormones with body

size, body fat distribution, and other known risk factors for breast

cancer in post-menopausal Chinese women. Cancer Causes Con-

trol 15: 305–311, 2004

Address for offprints and correspondence: Dr Gong Yang, Center forHealth Services Research, Vanderbilt University Medical Center,Medical Center East, Suite 6000, 1215 21st Avenue South, Nashville,TN 37232-8300, USA; Tel.: +1-615-936-0748; Fax: +1-615-936-1269;E-mail: Gong.Yang@vanderbilt.edu

70 G Yang et al.