The incidence of breast cancer among premenopausal women has been increasing rapidly in recent decades in East Asia. This case–control study investigated whether estrogen-DNA adducts were associated with breast cancer risk in Taiwan. The control group (n = 146) comprised healthy female volunteers and women with non-proliferative breast disease. The case group (n = 221) comprised women either with proliferative benign breast disease or breast cancer. The ratios of estrogen-DNA adducts to their respective metabolites and conjugates in plasma were analyzed using ultraperformance LC/MS-MS. The SNPs of CYP1A1, CYP1B1, and COMT were genotyped. Logistic regression model was used to compare the estrogen-DNA adduct ratios between the two groups. The estrogen-DNA adduct ratio in the case group was significantly higher than that in the control group (median ratio: 58.52 vs. 29.36, P = 0.004). A multiple logistic regression model demonstrated that a unit increase in the natural log of the estrogen-DNA adduct ratio in premenopausal women was a significant predictor of breast cancer risk, with an estimated hazard ratio of 1.718 (1.444−2.046, P < 0.001). However, the CYP1A1, CYP1B1, and COMT SNPs were not associated with the estrogen-DNA adduct ratios. In conclusion, plasma estrogen-DNA adduct ratio was associated with the presence of breast cancer or proliferating benign breast disease in premenopausal women in Taiwan.

Prevention Relevance:

This study provides evidence that endogenous estrogen-induced genotoxicity may contribute to the carcinogenesis of breast cancer in premenopausal Asian women. This work could have important preventive implication for the emerging disease in East Asia.

The incidence of breast cancer among women in Asian countries is generally lower than that in Western countries (1). However, the incidence has increased rapidly in East Asia over the past 40 years, including in Singapore, Taiwan, Korea, Japan, and China (2). Previous age–period–cohort analyses have demonstrated a strong cohort effect on breast cancer incidence in these countries, with an increase observed particularly among women ages <50 years (3–7), and our studies have identified an unusually high prevalence of hormone receptor–positive breast tumors in patients ages ≤50 years in Taiwan and other East Asian countries (2, 8, 9).

Notably, our previous study observed that the incidences of type I endometrial cancer and endometrioid carcinoma of the ovary, which have high rates of hormonal receptor expressions and are considered estrogen-related malignancies, have increased rapidly in Taiwan in recent decades (10). These findings indicate that the frequencies of estrogen-related malignancies are increasing in premenopausal women in East Asia, and suggest that either exogenous estrogenic substances or endogenous estrogen related etiologies are linked to the rapid increase in incidence of these malignancies including breast cancer. With regard to endogenous estrogen, early studies (published between 1971 and 1991) of premenopausal women found that estradiol levels were 20% to 50% lower in Asian women (including Asian American women) than in Caucasian women of comparable age (11–15). However, a more recent study demonstrated no difference in serum estradiol levels between premenopausal Asian American women and Caucasian women (16), and another recent study even identified higher estradiol levels in premenopausal Asian American women than in Caucasian women (17). These indirect pieces of evidence suggest that serum estrogen levels have increased considerably in premenopausal Asian American women in the past two decades.

In this study, we hypothesized that the rapid increase in estrogen levels in premenopausal women in East Asia has led to the production of mutagenic estrogen metabolites. Estrogens can be oxidized to form catechol estrogens, 2-hydroxyestrone (estradiol) [2-OHE1(E2)], and 4-OHE1(E2) in the presence of cytochrome P450 (CYP) 1A1 and 1B1 in breast tissue. These catechol estrogens are further converted to semiquinones and quinones as a result of peroxidase activity. Catechol estrogen-3,4-quinones react with DNA to form depurinating estrogen-DNA adducts 4-OHE1(E2)-1-N3Ade and 4-OHE1(E2)-1-N7Gua. The apurinic sites of the reacted DNA can lead to mutations caused by error-prone DNA repair, initiating cancer (18). Depurinating estrogen-DNA adducts are shed into the blood and excreted in urine. Previous studies demonstrated that in Western countries, the ratio of estrogen-DNA adducts to estrogen metabolites and conjugates was associated with increased risk of breast cancer (19–21). The sample sizes of two studies were small (20, 21), so a confirmatory needed.

This case–control study investigated the association between estrogen-DNA adduct ratio and breast cancer or high-risk breast lesions in Taiwanese women, and the analysis was stratified by menopausal status to uncover this association among premenopausal women. In addition, we evaluated the interaction of SNPs of CYP1A1, CYP1B1, and catechol-O-methyltransferase (COMT) with estrogen-DNA adduct ratio and the interaction effect on breast cancer risk. COMT is a phase II enzyme and can mitigate the harmful effects of catechol estrogens by catalyzing the O-methylation in catechol estrogens to methyl estrogens (22).

Study population

In October 2010, we initiated a case–control study to evaluate the associations of (i) estrogenic environmental pollutants and (ii) estrogen synthesis and metabolism with breast cancer risk in Taiwanese women. The study was approved by the Ethics Committee of National Taiwan University Hospital (NTUH; 201004060R). The participants included female volunteers with no history of malignancy or breast disease (defined as healthy volunteers), biopsy-proven benign breast disease, or breast cancer patients. We obtained written informed consent from the participants, and the study was conducted in accordance with the Declaration of Helsinki. Healthy volunteers were recruited through posters and flyers at NTUH and in the community, and they received a small fee for their time (approximately US $6.30) after completing the study. The diagnosis of breast cancer (including ductal carcinoma in situ) and benign breast disease was confirmed by histologic samples. Benign breast diseases were classified into proliferative and non-proliferative diseases. Proliferative disease was indicated if specimens contained any of the following: atypical ductal hyperplasia, atypical lobular hyperplasia, ductal hyperplasia (greater than mild), papilloma, radial scars, or sclerosing adenosis. Cysts, fibroadenoma, or columnar changes were considered non-proliferative disease (23, 24). A meta-analysis demonstrated that proliferative disease was associated with breast cancer risk but that non-proliferative disease was not (25); thus, we included female volunteers with no history of malignancy and women with non-proliferative breast disease into a control group, and we included women with proliferative breast disease and breast cancer into a case group.

At enrollment, the participants provided a fasting blood sample and completed an assisted questionnaire. To reduce treatment-related bias, blood samples were collected before any treatment for breast cancer was initiated. Blood samples were centrifuged immediately after collection, and the plasma samples were stored at −80°C. The questionnaire included questions on the risk factors for breast cancer, including age, parity, breastfeeding history, body mass index (BMI), smoking and drinking habits, menopause status, education level, and family history of breast or ovarian cancer.

Sample preparation and ultraperformance LC/MS-MS analysis of estrogen-DNA adduct ratio

Serum samples were stored at −80°C and thawed only once before analysis. Serum aliquots (0.5 mL) were partially purified by solid-phase extraction (SPE) with a small modification, as described previously (19). The serum samples were diluted with an equal volume of 0.01 M ammonium formate buffer (pH 7; loading and washing buffer) before being passed through the SPE cartridges. The cartridges were washed with methanol and distilled water and, finally, preconditioned with loading buffer. The samples were loaded and passed without vacuum, and the target compounds were eluted by using elution buffer (methanol/acetonitrile/water, 80:10:10 pH 3.5). The eluted samples were lyophilized and reconstituted in 50 μL of methanol/water (0.1% formic acid), and passed through 5,000 cutoff filters.

The ultraperformance LC/MS-MS analyses were carried out using a Waters Acquity UPLC system connected to a high-performance Quattro Micro triple quadrupole mass spectrometer (Waters; ref. 26). The analysis was performed by injecting 10 μL of each partially purified serum sample into an Acquity UPLC BEH C18 1.7-mm column (1 × 100 mm2) at a flow rate of 0.15 mL/min. Analytes were identified by their retention time and fragmentation pattern and processed using QuanLynx software to quantify the estrogen-DNA adducts, estrogen metabolites, and conjugates. The adduct ratio was defined by the following equation:
formula

The estrogen metabolites included 2-OHE1(E2) and 4-OHE1(E2). The conjugates included 2-OCH3E1(E2), 4-OCH3E1(E2), 4-OHE1(E2)-2-SG, 4-OHE1(E2)-2-Cys, and 4-OHE1(E2)-2-NAcCys. The adducts included 4-OHE1(E2)-1-N3Ade, 4-OHE1(E2)-1-N7Gua, and 2-OHE1(E2)-6-N3Ade. The concentration of each of the 20 compounds were measured, and the ratio of depurinating N3Ade and N7Gua adducts to the sum of their respective estrogen metabolites and conjugates in each serum sample was calculated to reflect the degree of imbalance in that estrogen metabolism that could lead to the initiation of cancer (20). The measurement and calculation were mainly conducted by M. Zahid, C.L. Beseler, and E.G. Rogan in a blinded fashion.

Genotyping of CYP1A1, CYP1B1, and COMT

The genomic DNAs of the blood and tumor specimens were extracted using a QIAamp DNA Mini Kit (Catalog No. 51304; Qiagen). Genotypes for the CYP1A1 Ile462Val (rs1048943), CYP1B1 Leu432Val (rs1056836), and COMT Val158Met (rs4680) polymorphisms were determined by a restriction fragment-length polymorphism assay. The PCR primers had been designed previously: CYP1A1 (rs1048943, A2455G): forward 5′-CTG TCT CCC TCT GGT TAC AGG AAGC -3′ and reverse 5′-TTC CAC CCG TTG CAG CAG GAT AGCC-3′; CYP1B1 (rs1056836, C1294G): forward 5′-CAC TGC CAA CAC CTC TGT CT-3′ and reverse 5′- GCA GGC TCA TTT GGG TTG-3′; and COMT (rs4680, G472A): forward 5′- CGA GGC TCA TCA CCA TCG AGA TC-3′ and reverse 5′- CTG ACA ACG GGT CAG GAA TGC A-3′. These PCR primers were used to generate 204 bp (CYP1A1), 294 bp (CYP1B1), and 108 bp (COMT) PCR products containing SNP sites (27–29). The PCR products were digested with restriction enzymes (FastDigest BseMI/CYP1A1 rs1048943, FastDigest Eco57I/CYP1B1 rs1056836, and FastDigest Hin1II/COMT rs4680) according to the manufacturer's protocol (FD1264, FD0344, FD1834; Thermo Fisher Scientific). The three different SNP genotypes of each gene were determined by DNA fragment length on 3% agarose gel as follows: CYP1A1 (rs1048943, A2455G): A/A homozygotes (149 and 55 bp), A/G heterozygotes (204, 149, and 55 bp), and G/G homozygotes (204 bp); CYP1B1 (rs1056836, C1294G): C/C homozygotes (187 and 107 bp), C/G heterozygotes (294, 187, and 107 bp), and G/G homozygotes (294 bp); and COMT (rs4680, G472A): G/G homozygotes (108 bp), G/A heterozygotes (108, 72, and 36 bp), and A/A homozygotes (72 bp and 36 bp).

Statistical analysis

The categorical variables were described in terms of the frequency and percentage, and the continuous variables were described in terms of the mean ± SD and median. Power transformations (xq), including the natural logarithm (q = 0), square root (q = 0.5), and square (q = 2), were applied to some continuous variables for their distributions to be more symmetrical.

A univariate analysis was performed to examine the differences in the distributions of the continuous and categorical variables between the case and control groups using a two-sample Student t test, Wilcoxon rank-sum test (or Mann–Whitney U test), chi-square test, and Fisher exact test (if the expected values in any of the cells of the contingency table was <5), as appropriate for the data type. A multivariate analysis was then conducted using a fitted linear regression model and logistic regression model to estimate the adjusted effects of potential risk factor predictors on continuous or binary outcomes. All statistical analyses were performed in R software (version 3.6.3, R Foundation for Statistical Computing). Statistical significance was indicated by a two-sided P value ≤0.05. The statistical analysis is detailed in the Supplementary Materials and Methods S1.

Data availability

The data generated in this study are available in the Supplementary Table S1.

Demographics of the study population

The control group included 88 healthy volunteers and 58 women with non-proliferative benign breast disease. The case group included 33 women with proliferative benign breast disease and 188 patients with breast cancer (n = 150) or ductal carcinoma in situ (n = 38), diagnosed at NTUH. Participants were recruited between February 2010 and November 2014. Table 1 presents the demographic characteristics of the case and control group participants. The case group had lower percentages of participants at menopause (control: 46%; case: 39%, P = 0.027) and a nonsignificantly higher BMI (BMI, 23.2 kg/m2 vs. 22.5 kg/m2, P = 0.098) than the control group. The median age, smoking habits, alcohol consumption, pregnancy history, number of births, breastfeeding history, education level, and family history of breast cancer were not significantly different between the two groups.

Table 1.

Demographic characteristics of control (healthy volunteers and benign low risk) and case (breast cancer and benign high risk).

Control groupCase group
Healthy volunteers (n = 88)Non-proliferative disease (n = 58)Whole-controls (n = 146)Proliferative disease (n = 33)Breast cancer or carcinoma in situ (n = 188)Whole-cases (n = 221)Pa
Age (median, range) 48 (27–69) 48 (26–67) 48 (26–69) 45 (27–63) 49 (28–77) 48 (27–77) 0.909 
Menopause (n, %) 42 (48%) 25 (43%) 67 (46%) 7 (21%) 69 (37%) 76 (39%) 0.027 
Body mass index (kg/m2) (mean, SD) 22.4 ± 3.4 22.5 (3.1) 22.5 (3.3) 22.6 (3.6) 23.3 (3.8) 23.2 (3.7) 0.098 
Cigarette smoking (n, %) 0 (0%) 4 (7%) 4 (3%) 4 (12%) 10 (5%) 14 (6%) 0.143 
Alcohol drinking (n, %) 4 (5%) 4 (7%) 8 (6%) 1 (3%) 16 (9%) 17 (8%) 0.527 
Pregnant history (n, %) 62 (70%) 49 (84%) 111 (76%) 21 (64%) 149 (79%) 170 (77%) 0.469 
No. of birth       0.818 
 0 28 (32%) 10 (17%) 38 (26%) 14 (42%) 46 (24%) 60 (27%)  
 1 11 (13%) 9 (16%) 20 (14%) 5 (15%) 31 (16%) 36 (16%)  
 2 29 (33%) 27 (47%) 56 (38%) 9 (27%) 75 (40%) 84 (38%)  
 ≥3 20 (23%) 12 (21%) 32 (22%) 5 (15%) 36 (19%) 41 (19%)  
Breastfeeding (n, %) 39 (44%) 29 (50%) 68 (47%) 10 (30%) 85 (45%) 95 (43%) 0.498 
Education level ≥ 12 years (n, %) 88 (100%) 51 (88%) 139 (95%) 31 (94%) 176 (94%) 207 (94%) 0.534 
Family history of breast or ovarian cancer (n, %) 9 (10%) 6 (10%) 15 (10%) 7 (21%) 25 (13%) 32 (14%) 0.238 
CYP1A1 (rs1048943) (n, %) 
 Ile/Ile 45 (51%) 30 (52%) 75 (51%) 20 (61%) 97 (52%) 117 (53%) 0.571 
 Ile/Val 36 (41%) 23 (40%) 59 (40%) 11 (33%) 81 (43%) 92 (42%)  
 Val/Val 7 (8%) 5 (9%) 12 (8%) 2 (6%) 10 (5%) 12 (5%)  
CYP1B1 (rs1056836) 
 leu/leu 67 (76%) 46 (79%) 113 (77%) 25 (76%) 152 (81%) 177 (80%) 0.249 
 leu/Val 21 (24%) 12 (21%) 33 (23%) 8 (24%) 33 (18%) 41 (19%)  
 Val/Val 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 (2%) 3 (1%)  
COMT (rs4680) 
 Val/Val 44 (50%) 35 (60%) 79 (54%) 14 (42%) 102 (54%) 116 (52%) 0.825 
 Val/Met 32 (36%) 22 (38%) 54 (37%) 15 (45%) 73 (39%) 88 (40%)  
 Met/Met 12 (14%) 1 (2%) 13 (9%) 4 (12%) 13 (7%) 17 (8%)  
Control groupCase group
Healthy volunteers (n = 88)Non-proliferative disease (n = 58)Whole-controls (n = 146)Proliferative disease (n = 33)Breast cancer or carcinoma in situ (n = 188)Whole-cases (n = 221)Pa
Age (median, range) 48 (27–69) 48 (26–67) 48 (26–69) 45 (27–63) 49 (28–77) 48 (27–77) 0.909 
Menopause (n, %) 42 (48%) 25 (43%) 67 (46%) 7 (21%) 69 (37%) 76 (39%) 0.027 
Body mass index (kg/m2) (mean, SD) 22.4 ± 3.4 22.5 (3.1) 22.5 (3.3) 22.6 (3.6) 23.3 (3.8) 23.2 (3.7) 0.098 
Cigarette smoking (n, %) 0 (0%) 4 (7%) 4 (3%) 4 (12%) 10 (5%) 14 (6%) 0.143 
Alcohol drinking (n, %) 4 (5%) 4 (7%) 8 (6%) 1 (3%) 16 (9%) 17 (8%) 0.527 
Pregnant history (n, %) 62 (70%) 49 (84%) 111 (76%) 21 (64%) 149 (79%) 170 (77%) 0.469 
No. of birth       0.818 
 0 28 (32%) 10 (17%) 38 (26%) 14 (42%) 46 (24%) 60 (27%)  
 1 11 (13%) 9 (16%) 20 (14%) 5 (15%) 31 (16%) 36 (16%)  
 2 29 (33%) 27 (47%) 56 (38%) 9 (27%) 75 (40%) 84 (38%)  
 ≥3 20 (23%) 12 (21%) 32 (22%) 5 (15%) 36 (19%) 41 (19%)  
Breastfeeding (n, %) 39 (44%) 29 (50%) 68 (47%) 10 (30%) 85 (45%) 95 (43%) 0.498 
Education level ≥ 12 years (n, %) 88 (100%) 51 (88%) 139 (95%) 31 (94%) 176 (94%) 207 (94%) 0.534 
Family history of breast or ovarian cancer (n, %) 9 (10%) 6 (10%) 15 (10%) 7 (21%) 25 (13%) 32 (14%) 0.238 
CYP1A1 (rs1048943) (n, %) 
 Ile/Ile 45 (51%) 30 (52%) 75 (51%) 20 (61%) 97 (52%) 117 (53%) 0.571 
 Ile/Val 36 (41%) 23 (40%) 59 (40%) 11 (33%) 81 (43%) 92 (42%)  
 Val/Val 7 (8%) 5 (9%) 12 (8%) 2 (6%) 10 (5%) 12 (5%)  
CYP1B1 (rs1056836) 
 leu/leu 67 (76%) 46 (79%) 113 (77%) 25 (76%) 152 (81%) 177 (80%) 0.249 
 leu/Val 21 (24%) 12 (21%) 33 (23%) 8 (24%) 33 (18%) 41 (19%)  
 Val/Val 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 (2%) 3 (1%)  
COMT (rs4680) 
 Val/Val 44 (50%) 35 (60%) 79 (54%) 14 (42%) 102 (54%) 116 (52%) 0.825 
 Val/Met 32 (36%) 22 (38%) 54 (37%) 15 (45%) 73 (39%) 88 (40%)  
 Met/Met 12 (14%) 1 (2%) 13 (9%) 4 (12%) 13 (7%) 17 (8%)  

aThe sample statistics presented in this table were mean ± SD for continuous variables and frequency (percentage, %) for categorical variables. The listed P values of statistical tests were calculated using the Wilcoxon rank-sum test for continuous variables and the Pearson Chi-squared test for categoric variables.

Correlations of estrogen-DNA adduct ratios with breast disease status and demographic variables

The ratios of estrogen-DNA adduct to these selected categorical variables are listed in Table 2. The estrogen-DNA adduct ratios were significantly higher in the case group than in the control group (mean ratios, 58.52 vs. 29.36, P = 0.004 using the Student t test, and P < 0.001 using the Mann–Whitney U test, Fig. 1A; Supplementary Table S2). Healthy volunteers had lower estrogen-DNA ratio than the women with non-proliferative breast disease by using the Mann–Whitney U test (mean ratio 27.52 vs. 32.76, P = 0.590 using the Student t test, and P = 0.005 using the Mann–Whitney U test). No significant difference was demonstrated between the women with proliferative breast disease and those with breast cancer (mean ratio 47.97 vs. 60.38, P = 0.565 using the Student t test, and P = 0.885 using the Mann–Whitney U test; Supplementary Table S2).

Table 2.

The associations of estrogen-DNA adduct ratio with selected variables.

Estrogen-DNA adduct ratio (mean ± SD)P valueaP valueb
Study group  0.004 <0.001 
Control group (n = 146) 29.36 ± 4.19   
Case group (n = 221) 58.52 ± 7.67   
Age group (years)  0.006 0.013 
<40 (n = 112) 72.14 ± 10.98   
40–49 (n = 97) 32.17 ± 4.87   
50–59 (n = 101) 43.44 ± 11.92   
≥60 (n = 157) 28.67 ± 3.37   
Menopausal status  0.275 0.417 
Premenopausal (n = 224) 51.25 ± 6.04   
Postmenopausal (n = 143) 40.14 ± 8.52   
Alcohol consumption    
No (n = 342) 44.27 ± 4.24 0.048 0.427 
Yes (n = 25) 83.17 ± 44.17   
No. of birth  0.019 0.467 
0 (n = 98) 60.47 ± 13.89   
1 (n = 56) 58.91 ± 15.81   
2 (n = 140) 41.56 ± 5.25   
≥ 3 (n = 73) 29.83 ± 4.64   
CYP1A1 (rs1048943)  0.692 0.676 
Ile/Ile (n = 192) 44.19 ± 5.68   
Ile/Val (n = 151) 52.13 ± 9.44   
Val/Val (n = 24) 36.02 ± 12.87   
CYP1B1 (rs1056836)  0.690 0.296 
leu/leu (n = 290) 46.67 ± 5.38   
leu/Val (n = 74) 47.01 ± 12.67   
Val/Val (n = 3) 68.73 ± 46.21   
COMT (rs4680)  0.694 0.025 
Val/Val (n = 194) 41.28 ± 4.97   
Val/Met (n = 142) 54.32 ± 9.99   
Met/Met (n = 29) 48.61 ± 20.32   
Breast cancer subgroup  0.578 0.645 
ER+/HER2− (n = 99) 56.22 ± 13.13   
ER+/HER2+ (n = 25) 63.57 ± 19.91   
ER−/HER2− (n = 7) 66.72 ± 34.97   
ER−/HER2− (n = 19) 103.22 ± 43.25   
Estrogen-DNA adduct ratio (mean ± SD)P valueaP valueb
Study group  0.004 <0.001 
Control group (n = 146) 29.36 ± 4.19   
Case group (n = 221) 58.52 ± 7.67   
Age group (years)  0.006 0.013 
<40 (n = 112) 72.14 ± 10.98   
40–49 (n = 97) 32.17 ± 4.87   
50–59 (n = 101) 43.44 ± 11.92   
≥60 (n = 157) 28.67 ± 3.37   
Menopausal status  0.275 0.417 
Premenopausal (n = 224) 51.25 ± 6.04   
Postmenopausal (n = 143) 40.14 ± 8.52   
Alcohol consumption    
No (n = 342) 44.27 ± 4.24 0.048 0.427 
Yes (n = 25) 83.17 ± 44.17   
No. of birth  0.019 0.467 
0 (n = 98) 60.47 ± 13.89   
1 (n = 56) 58.91 ± 15.81   
2 (n = 140) 41.56 ± 5.25   
≥ 3 (n = 73) 29.83 ± 4.64   
CYP1A1 (rs1048943)  0.692 0.676 
Ile/Ile (n = 192) 44.19 ± 5.68   
Ile/Val (n = 151) 52.13 ± 9.44   
Val/Val (n = 24) 36.02 ± 12.87   
CYP1B1 (rs1056836)  0.690 0.296 
leu/leu (n = 290) 46.67 ± 5.38   
leu/Val (n = 74) 47.01 ± 12.67   
Val/Val (n = 3) 68.73 ± 46.21   
COMT (rs4680)  0.694 0.025 
Val/Val (n = 194) 41.28 ± 4.97   
Val/Met (n = 142) 54.32 ± 9.99   
Met/Met (n = 29) 48.61 ± 20.32   
Breast cancer subgroup  0.578 0.645 
ER+/HER2− (n = 99) 56.22 ± 13.13   
ER+/HER2+ (n = 25) 63.57 ± 19.91   
ER−/HER2− (n = 7) 66.72 ± 34.97   
ER−/HER2− (n = 19) 103.22 ± 43.25   

aStatistical significance by two-tailed Student t test or ANOVA test.

bThe sample statistics presented in this table were mean ± SD for continuous variables and frequency (percentage, %) for categoric variables. The listed P values of statistical tests were calculated using the Wilcoxon rank-sum test.

Figure 1.

The estrogen-DNA adduct ratios in women without breast disease, with non-proliferative breast disease, with proliferative breast disease, and with breast cancer in whole participants (A), premenopausal subgroup (B), and postmenopausal subgroup (C). The P values indicated the statistical differences of estrogen-DNA adduct ratios between the case and control groups by Mann–Whitney U test.

Figure 1.

The estrogen-DNA adduct ratios in women without breast disease, with non-proliferative breast disease, with proliferative breast disease, and with breast cancer in whole participants (A), premenopausal subgroup (B), and postmenopausal subgroup (C). The P values indicated the statistical differences of estrogen-DNA adduct ratios between the case and control groups by Mann–Whitney U test.

Close modal

The estrogen-DNA adduct ratio varied between age groups (heterogeneity among age groups: P = 0.006 using an ANOVA and P = 0.013 using the Kruskal–Wallis test), and the group aged <40 years had the highest ratio (mean ratios: age < 40 years: 72.14; age 40–49 years: 32.17; age 50–59 years: 43.44; age ≥60 years: 28.67). Alcohol consumption was significantly associated with higher estrogen-DNA adduct ratios (mean ratios: 83.17 vs. 44.27) in the Student t test (P = 0.048) but not in the Mann–Whitney U test (P = 0.427). A lower number of births was significantly associated with increased estrogen-DNA adduct ratios (mean ratios: no parity: 60.47; one birth: 58.91; two births: 41.56; three births or more: 29.83) in the ANOVA (heterogeneity between age groups: P = 0.019), but the association was not significant in the Kruskal–Wallis test (P = 0.467).

Pearson and Spearman rank correlation coefficients were used to analyze the correlation of estrogen-DNA adduct ratios with continuous variables. The Pearson correlation coefficient revealed that the decreases of age or number of birth was associated with increase of estrogen-DNA adduct ratio, but the Spearman rank correlation coefficient did not reveal any significant association of these variables with estrogen-DNA adduct ratio (Supplementary Table S3).

Multivariate analysis of predictors for breast cancer or proliferative breast disease

A multivariate analysis demonstrated that premenopausal women and per a natural log unit increase in the estrogen-DNA adduct ratio (estimated OR = 1.718, P <0.001), premenopausal women and women aged 39–52 years (estimated OR = 0.293, P <0.001), postmenopausal women and women ages ≥61 years (estimated OR = 3.047, P = 0.002), and per unit increase in BMI (estimated OR = 1.073, P = 0.040) were associated with the risk of breast cancer or proliferative breast disease (Table 3).

Table 3.

Logistic regression analysis of the predictors for cases (breast cancer and proliferative breast disease).

CovariateEstimated regression coefficientEstimated SEz valueP valueEstimated Odds Ratio (95% CI)
Intercept −1.974 0.819 −2.410 0.016 0.139 (0.028−0.691) 
Premenopausal and per unit increase of natural log of the estrogen-DNA adduct ratio 0.541 0.089 6.088 <0.001 1.718 (1.444–2.046) 
Premenopausal and age 39–52 years −1.229 0.321 −3.826 <0.001 0.293 (0.156–0.549) 
Postmenopausal and age ≥ 61 years 1.114 0.365 3.055 0.002 3.047 (1.491–6.227) 
Per unit increase of BMI 0.071 0.034 2.052 0.040 1.073 (1.003–1.248) 
CovariateEstimated regression coefficientEstimated SEz valueP valueEstimated Odds Ratio (95% CI)
Intercept −1.974 0.819 −2.410 0.016 0.139 (0.028−0.691) 
Premenopausal and per unit increase of natural log of the estrogen-DNA adduct ratio 0.541 0.089 6.088 <0.001 1.718 (1.444–2.046) 
Premenopausal and age 39–52 years −1.229 0.321 −3.826 <0.001 0.293 (0.156–0.549) 
Postmenopausal and age ≥ 61 years 1.114 0.365 3.055 0.002 3.047 (1.491–6.227) 
Per unit increase of BMI 0.071 0.034 2.052 0.040 1.073 (1.003–1.248) 

Goodness-of-fit assessment: n = 367, adjusted generalized R2 = 0.180 < 0.3, the estimated area under the ROC curve = 0.715 > 0.7 (95% CI, 0.661−0.768), and the modified Hosmer and Lemeshow goodness-of-fit F test P = 0.502 > 0.05 (df = 9, 346), which indicated a fair fit.

Prediction: To calculate the estimated probability of Case group (i.e., the predicted value, |${\rm{\ }}{\hat{P}}_i$|⁠) given the observed covariate values of subject i, one can use the following formula. According to the above fitted multiple logistic regression model.

formula

where the indicator function I(˙) = 1 when the condition˙inside the parentheses is true. A conditional effect plot can also be drawn to visualize the estimated probability of Case group according to the specified covariate values for making predictions in clinical practice.

Considering the interaction of menopausal status with estrogen-DNA adduct ratios in predicting disease, we performed a stratified analysis to compare the estrogen-DNA adduct ratios between the case and control groups by menopausal status. Among the premenopausal participants, the estrogen-DNA adduct ratio was significantly higher in the case group than in the control group (mean ratios: 65.65 vs. 24.82, P = 0.001 with the Student t test and P < 0.001 using the Mann–Whitney U test; Fig. 1B). Premenopausal healthy volunteers had lower estrogen-DNA ratio than the premenopausal women with non-proliferative breast disease (mean ratio 16.57 vs. 36.33, P = 0.018 using the Student t test, and P < 0.001 using the Mann–Whitney U test; Supplementary Table S2). In the postmenopausal participants, the estrogen-DNA adduct ratios were not significantly different between the case group and the control group (mean ratios: 44.92 vs. 34.72, P = 0.552 with the Student t test and P = 0.285 using the Mann–Whitney U test; Fig. 1C). No significant difference was demonstrated between the postmenopausal women with proliferative breast disease and those with breast cancer (mean ratio 34.03 vs. 46.02, P = 0.814 using the Student t test, and P = 0.921 using the Mann–Whitney U test; Supplementary Table S2).

Correlations of estrogen-DNA adduct ratios with estrogen metabolism-related SNPs or breast cancer pathologic features

The distributions of the CYP1A1, CYP1B1, and COMT SNPs were not significantly different between the case and control groups (Table 1). The CYP1A1 and CYP1B1 SNPs were not significantly associated with estrogen-DNA adduct ratios. The estrogen-DNA adduct ratios varied between the three COMT SNPs (rs4680; heterogeneity among SNP groups: P = 0.025 using the Kruskal–Wallis test), and the Val/Met group had the highest ratio (mean ratios: Val/Val: 41.28, Val/Met: 54.32, Met/Met: 48.61). Among the patients with breast cancer, the estrogen-DNA adduct ratios were not associated with the distributions of positivity of estrogen receptor (ER) or HER2 in tumor cells. (Table 2).

This case–control study demonstrated that plasma estrogen-DNA adduct ratios were associated with breast cancer in Taiwan. The multiple logistic regression model indicated that a unit increase in the natural log of the estrogen-DNA adduct ratio was significantly associated with the presence of breast cancer or proliferative benign breast disease in premenopausal women, with an estimated hazard ratio of 1.718. Age, alcohol consumption, and the number of births were inversely associated with estrogen-DNA adduct ratios with both a Student t test and ANOVA, but only age was statistically significant with the Mann–Whitney U test or Kruskal–Wallis test. Of the three estrogen metabolism-related genes, only the COMT SNP was associated with estrogen-DNA adduct ratios using the Kruskal–Wallis test. With regard to breast cancer, the ER/HER2 subtypes were not associated with estrogen-DNA adduct ratio.

In postmenopausal women, studies have demonstrated a positive association between breast cancer risk and circulating concentrations of estrogens (30). For premenopausal women, evidence is limited because hormone measurements are complicated by the dynamic change of serum estrogen levels across menstrual cycles. A collaborative reanalysis of seven prospective studies demonstrated that breast cancer risk was associated with a doubling in concentrations of estradiol (OR, 1.19; 95% CI, 1.06–1.35) in premenopausal women, but the association was modest (31). This study and three prior studies (19–21) did not have a fixed timing of sampling during the menstrual cycle. This study demonstrated an association between estrogen-DNA adduct ratios and presence of breast cancer in premenopausal women.

This study is the largest study evaluating the association of estrogen-DNA adduct ratio with presence of breast cancer, and differed from the previous three studies (19–21) with respect to participants, sample type, and the subgroup analysis by menopausal status. With regard to participants, the previous three studies enrolled women from Italy and the United States whereas the present study enrolled women from Taiwan. The differences in estrogen profiles and metabolism between Asian, Asian American, and Caucasian women have been demonstrated in previous studies (32–35). In addition, a 5-year Gail Model score of ≥1.66% was used to define a high breast cancer risk in the previous three studies, but in this study, histology was used to identify proliferative disease, which is a more stringent criteria for defining high risk groups. Regarding sample type, plasma samples were used in the present study whereas urine samples or serum samples were used in the previous three studies. Furthermore, we conducted a subgroup analysis according to menopausal status and demonstrated that estrogen-DNA adduct ratios were significantly associated with presence of breast cancer in premenopausal women. The previous three studies did not identify significant interactions between estrogen-DNA adduct ratios and menopausal status in the risk prediction for breast cancer. In addition, we showed that parity, a known protective factor, was inversely associated with estrogen-DNA adduct ratios. This may be explained by a prior study which showed refractoriness to chemical carcinogenesis in the involuted mammary gland of parous mice (36).

This study has some limitations. First, this study included imbalanced sample sizes between case and control groups because of different recruitment efficiency of participants. Second, this study had heterogenous histology types in women with benign breast diseases. We grouped women with non-proliferating benign breast disease and healthy volunteers into control group and grouped proliferating benign breast disease and women with breast cancer into case group according to the previous meta-analysis (25), but this study showed that women with non-proliferative breast disease had higher estrogen-DNA ratio than healthy volunteers by using the Mann–Whitney U test. However, when women with non-proliferating and proliferating benign breast disease were excluded, women with breast cancer or carcinoma in situ still had significantly higher estrogen-DNA ratio than healthy volunteers (mean ratios, 60.38 vs. 27.52, P = 0.016 using the Student t test, and P < 0.001 using the Mann–Whitney U test). The difference was also significant in premenopausal subgroup (mean ratios, 68.70 vs. 16.57, P = 0.002 using the Student t test, and P < 0.001 using the Mann–Whitney U test; Supplementary Table S2). Third, we cannot exclude the possibility of instability of the adducts, metabolites and conjugates during storage. However, the aliquots of plasma samples were stored at −80°C until analyzed, and the samples were thawed only once prior to analysis.

In contrast to the three earlier studies (19–21), we examined three gene polymorphisms related to estrogen metabolism. Of these, only COMT polymorphism exhibited a marginal association with estrogen-DNA adduct ratios. COMT catalyzes an inactivation pathway for catechol estrogen. The previous in vitro study indicated that a pretreatment of 2,3,7,8-tetrachlorodibenzo-p-dioxin followed by the inhibition of COMT activity increased the formation of depurinating 4-OHE1(E2)-1-N3Ade and 4-OHE1(E2)-1-N7Gua adducts in human breast epithelial cells (37). With regard to the SNP, the COMT gene Met/Met homozygotes yielded a 3-to-4-fold reduction in COMT activity relative to the Val/Val homozygotes, and Val/Met heterozygotes demonstrated intermediate activity (38). The frequency of the variant (low-activity) COMT allele was lower in Asian women than in Caucasian women (summarized in a previous article; ref. 39). In this study, the estrogen-DNA adduct ratios varied in the three COMT SNPs (heterogeneity of the SNP groups: P = 0.025 using the Kruskal–Wallis test), and the Val/Met allele with intermediate activity had the highest ratio (mean ratios: Val/Val, 41.28; Val/Met, 54.32; Met/Met, 48.61). The finding can be purely by chance, or explained by confounding of certain environmental pollutant exposure which decreased COMT activity.

To evaluate whether the formation of estrogen-DNA adducts preferentially contributes to ER-positive breast cancer, we examined the association between estrogen-DNA adduct ratios and ER/HER2 status. However, we were unable to identify a significant association. Estrogen-DNA adducts might mediate the initial steps of breast carcinogenesis in various pluripotent cell types, and therefore, estrogen-DNA adduct ratios could differentiate between the control group participants and the high breast cancer risk participants regardless of ER/HER2 status.

In summary, our study suggests that the rapidly increasing incidence of breast cancer in young women in Taiwan or other East Asian countries is related to the formation of estrogen-DNA adducts. The estrogen-DNA adduct ratio as a predictive biomarker for breast cancer risk in this population warrants further investigation.

Y. Lu reports grants and personal fees from MSD, AstraZenica, Novartis; personal fees from Eisai, Roche, Daiichi Sankyo, Eli Lilly, and Pfizer outside the submitted work. A. Cheng reports personal fees from Bristol-Myers Squibb, BAYER Healthcare, Eisai, Ono Pharmaceutical, AstraZeneca, Genentech/Roche, Merck Sharp Dohme, BeiGene, Ltd., EXELIXIS Ltd., IPSEN Innovation, F. Hoffmanna-La ROCHE Ltd., Amge Taiwan, Novartis, Bayer Yakuhin, IQVIA, and Chugai Pharmaceutical outside the submitted work. No disclosures were reported by the other authors.

C. Lin: Conceptualization, resources, data curation, software, formal analysis, supervision, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. M. Zahid: Conceptualization, resources, data curation, software, formal analysis, supervision, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. W. Kuo: Conceptualization, resources, data curation, software, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. F. Hu: Data curation, software, formal analysis, methodology, writing–original draft. M. Wang: Resources, data curation, software, formal analysis, validation, investigation, methodology, writing–original draft. I. Chen: Conceptualization, resources, validation, investigation, methodology, writing–original draft. C.L. Beseler: Conceptualization, resources, formal analysis, investigation, methodology, writing–original draft, project administration. B. Mondal: Resources, data curation, formal analysis, investigation, methodology, project administration. Y. Lu: Conceptualization, supervision, investigation, writing–original draft, writing–review and editing. E.G. Rogan: Conceptualization, resources, data curation, formal analysis, supervision, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. A. Cheng: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing.

We would like to thank the participants for providing their samples and clinical information for use in our research. This work was supported by grants from the Yonglin Foundation (to A.L. Cheng), the Ministry of Health and Welfare (MOHW111-TDU-B-221–114016 to C.H. Lin), the Ministry of Education (NTU-109L901403 to C.H. Lin), and the Ministry of Science and Technology (MOST 110–2314-B-002–214 to C.H. Lin) in Taiwan. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Prevention Research Online (http://cancerprevres.aacrjournals.org/).

1.
Sung
H
,
Ferlay
J
,
Siegel
RL
,
Laversanne
M
,
Soerjomataram
I
,
Jemal
A
, et al
.
Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J Clin
2021
;
71
:
209
49
.
2.
Lin
CH
,
Yap
YS
,
Lee
KH
,
Im
SA
,
Naito
Y
,
Yeo
W
, et al
.
Contrasting epidemiology and clinicopathology of female breast cancer in Asians vs the US population
.
J Natl Cancer Inst
2019
;
111
:
1298
306
.
3.
Sim
X
,
Ali
RA
,
Wedren
S
,
Goh
DL
,
Tan
CS
,
Reilly
M
, et al
.
Ethnic differences in the time trend of female breast cancer incidence: Singapore, 1968–2002
.
BMC Cancer
2006
;
6
:
261
.
4.
Yoo
KY
,
Kim
Y
,
Park
SK
,
Kang
D
.
Lifestyle, genetic susceptibility and future trends of breast cancer in Korea
.
Asian Pac J Cancer Prev
2006
;
7
:
679
82
.
5.
Matsuno
RK
,
Anderson
WF
,
Yamamoto
S
,
Tsukuma
H
,
Pfeiffer
RM
,
Kobayashi
K
, et al
.
Early- and late-onset breast cancer types among women in the United States and Japan
.
Cancer Epidemiol Biomarkers Prev
2007
;
16
:
1437
42
.
6.
Shen
YC
,
Chang
CJ
,
Hsu
C
,
Cheng
CC
,
Chiu
CF
,
Cheng
AL
.
Significant difference in the trends of female breast cancer incidence between Taiwanese and Caucasian Americans: implications from age-period-cohort analysis
.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
1986
90
.
7.
Fan
L
,
Strasser-Weippl
K
,
Li
JJ
,
St Louis
J
,
Finkelstein
DM
,
Yu
KD
, et al
.
Breast cancer in China
.
Lancet Oncol
2014
;
15
:
e279
289
.
8.
Lin
CH
,
Liau
JY
,
Lu
YS
,
Huang
CS
,
Lee
WC
,
Kuo
KT
, et al
.
Molecular subtypes of breast cancer emerging in young women in Taiwan: evidence for more than just westernization as a reason for the disease in Asia
.
Cancer Epidemiol Biomarkers Prev
2009
;
18
:
1807
14
.
9.
Lin
CH
,
Chuang
PY
,
Chiang
CJ
,
Lu
YS
,
Cheng
AL
,
Kuo
WH
, et al
.
Distinct clinicopathological features and prognosis of emerging young-female breast cancer in an East Asian country: a nationwide cancer registry-based study
.
Oncologist
2014
;
19
:
583
91
.
10.
Lin
CH
,
Chen
YC
,
Chiang
CJ
,
Lu
YS
,
Kuo
KT
,
Huang
CS
, et al
.
The emerging epidemic of estrogen-related cancers in young women in a developing Asian country
.
Int J Cancer
2012
;
130
:
2629
37
.
11.
MacMahon
B
,
Cole
P
,
Brown
JB
,
Aoki
K
,
Lin
TM
,
Morgan
RW
, et al
.
Oestrogen profiles of Asian and North American women
.
Lancet
1971
;
2
:
900
2
.
12.
Goldin
BR
,
Adlercreutz
H
,
Gorbach
SL
,
Woods
MN
,
Dwyer
JT
,
Conlon
T
, et al
.
The relationship between estrogen levels and diets of Caucasian American and Oriental immigrant women
.
Am J Clin Nutr
1986
;
44
:
945
53
.
13.
Shimizu
H
,
Ross
RK
,
Bernstein
L
,
Pike
MC
,
Henderson
BE
.
Serum oestrogen levels in postmenopausal women: comparison of American whites and Japanese in Japan
.
Br J Cancer
1990
;
62
:
451
3
.
14.
Bernstein
L
,
Yuan
JM
,
Ross
RK
,
Pike
MC
,
Hanisch
R
,
Lobo
R
, et al
.
Serum hormone levels in pre-menopausal Chinese women in Shanghai and white women in Los Angeles: results from two breast cancer case-control studies
.
Cancer Causes Control
1990
;
1
:
51
58
.
15.
Key
TJ
,
Chen
J
,
Wang
DY
,
Pike
MC
,
Boreham
J
.
Sex hormones in women in rural China and in Britain
.
Br J Cancer
1990
;
62
:
631
6
.
16.
Falk
RT
,
Fears
TR
,
Hoover
RN
,
Pike
MC
,
Wu
AH
,
Nomura
AM
, et al
.
Does place of birth influence endogenous hormone levels in Asian-American women?
Br J Cancer
2002
;
87
:
54
60
.
17.
Pinheiro
SP
,
Holmes
MD
,
Pollak
MN
,
Barbieri
RL
,
Hankinson
SE
.
Racial differences in premenopausal endogenous hormones
.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
2147
53
.
18.
Cavalieri
EL
,
Stack
DE
,
Devanesan
PD
,
Todorovic
R
,
Dwivedy
I
,
Higginbotham
S
, et al
.
Molecular origin of cancer: catechol estrogen-3,4-quinones as endogenous tumor initiators
.
Proc Natl Acad Sci U S A
1997
;
94
:
10937
42
.
19.
Pruthi
S
,
Yang
L
,
Sandhu
NP
,
Ingle
JN
,
Beseler
CL
,
Suman
VJ
, et al
.
Evaluation of serum estrogen-DNA adducts as potential biomarkers for breast cancer risk
.
J Steroid Biochem Mol Biol
2012
;
132
:
73
79
.
20.
Gaikwad
NW
,
Yang
L
,
Muti
P
,
Meza
JL
,
Pruthi
S
,
Ingle
JN
, et al
.
The molecular etiology of breast cancer: evidence from biomarkers of risk
.
Int J Cancer
2008
;
122
:
1949
57
.
21.
Gaikwad
NW
,
Yang
L
,
Pruthi
S
,
Ingle
JN
,
Sandhu
N
,
Rogan
EG
, et al
.
Urine biomarkers of risk in the molecular etiology of breast cancer
.
Breast Cancer (Auckl)
2009
;
3
:
1
8
.
22.
Dawling
S
,
Roodi
N
,
Mernaugh
RL
,
Wang
X
,
Parl
FF
.
Catechol-O-methyltransferase (COMT)-mediated metabolism of catechol estrogens: comparison of wild-type and variant COMT isoforms
.
Cancer Res
2001
;
61
:
6716
22
.
23.
Dupont
WD
,
Page
DL
.
Risk factors for breast cancer in women with proliferative breast disease
.
N Engl J Med
1985
;
312
:
146
51
.
24.
Hartmann
LC
,
Sellers
TA
,
Frost
MH
,
Lingle
WL
,
Degnim
AC
,
Ghosh
K
, et al
.
Benign breast disease and the risk of breast cancer
.
N Engl J Med
2005
;
353
:
229
37
.
25.
Dyrstad
SW
,
Yan
Y
,
Fowler
AM
,
Colditz
GA
.
Breast cancer risk associated with benign breast disease: systematic review and meta-analysis
.
Breast Cancer Res Treat
2015
;
149
:
569
75
.
26.
Zahid
M
,
Mondal
B
,
LeVan
TD
,
Rogan
EG
.
Estrogen metabolism in African-American women with and without breast cancer: a pilot study
.
Chem Res Toxicol
2019
;
32
:
190
4
.
27.
Martinez-Ramirez
OC
,
Perez-Morales
R
,
Castro
C
,
Flores-Diaz
A
,
Soto-Cruz
KE
,
Astorga-Ramos
A
, et al
.
Polymorphisms of catechol estrogens metabolism pathway genes and breast cancer risk in Mexican women
.
Breast
2013
;
22
:
335
43
.
28.
Yadav
S
,
Singhal
NK
,
Singh
V
,
Rastogi
N
,
Srivastava
PK
,
Singh
MP
.
Association of single nucleotide polymorphisms in CYP1B1 and COMT genes with breast cancer susceptibility in Indian women
.
Dis Markers
2009
;
27
:
203
10
.
29.
Lajin
B
,
Hamzeh
AR
,
Ghabreau
L
,
Mohamed
A
,
Al Moustafa
AE
,
Alachkar
A
.
Catechol-O-methyltransferase Val 108/158 Met polymorphism and breast cancer risk: a case control study in Syria
.
Breast Cancer
2013
;
20
:
62
66
.
30.
Key
T
,
Appleby
P
,
Barnes
I
,
Reeves
G
,
Endogenous
H
, et al
.
Breast Cancer Collaborative G
.
Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies
.
J Natl Cancer Inst
2002
;
94
:
606
16
.
31.
Endogenous
H
,
Breast Cancer Collaborative
G
,
Key
TJ
,
Appleby
PN
,
Reeves
GK
,
Travis
RC
, et al
.
Sex hormones and risk of breast cancer in premenopausal women: a collaborative reanalysis of individual participant data from seven prospective studies
.
Lancet Oncol
2013
;
14
:
1009
19
.
32.
Adlercreutz
H
,
Gorbach
SL
,
Goldin
BR
,
Woods
MN
,
Dwyer
JT
,
Hamalainen
E
.
Estrogen metabolism and excretion in Oriental and Caucasian women
.
J Natl Cancer Inst
1994
;
86
:
1076
82
.
33.
Dickinson
LE
,
MacMahon
B
,
Cole
P
,
Brown
JB
.
Estrogen profiles of oriental and caucasian women in Hawaii
.
N Engl J Med
1974
;
291
:
1211
3
.
34.
Moore
SC
,
Matthews
CE
,
Ou Shu
X
,
Yu
K
,
Gail
MH
,
Xu
X
, et al
.
Endogenous estrogens, estrogen metabolites, and breast cancer risk in postmenopausal chinese women
.
J Natl Cancer Inst
2016
;
108
:
djw103
.
35.
Visvanathan
K
,
Yager
JD
.
Ethnic variations in estrogen and its metabolites: sufficient to explain differences in breast cancer incidence rates?
J Natl Cancer Inst
2016
;
108
:
djw223
.
36.
Medina
D
,
Smith
GH
.
Chemical carcinogen-induced tumorigenesis in parous, involuted mouse mammary glands
.
J Natl Cancer Inst
1999
;
91
:
967
9
.
37.
Lu
F
,
Zahid
M
,
Saeed
M
,
Cavalieri
EL
,
Rogan
EG
.
Estrogen metabolism and formation of estrogen-DNA adducts in estradiol-treated MCF-10F cells. The effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin induction and catechol-O-methyltransferase inhibition
.
J Steroid Biochem Mol Biol
2007
;
105
:
150
8
.
38.
Weinshilboum
RM
,
Otterness
DM
,
Szumlanski
CL
.
Methylation pharmacogenetics: catechol O-methyltransferase, thiopurine methyltransferase, and histamine N-methyltransferase
.
Annu Rev Pharmacol Toxicol
1999
;
39
:
19
52
.
39.
Huang
CS
,
Lin
CH
,
Lu
YS
,
Shen
CY
.
Unique features of breast cancer in Asian women–breast cancer in Taiwan as an example
.
J Steroid Biochem Mol Biol
2010
;
118
:
300
3
.