Effect Modification by Catalase Genotype Suggests a Role for Oxidative Stress in the Association of Hormone Replacement Therapy with Postmenopausal Breast Cancer Risk

Evidence suggests that estrogens play a role in the development of breast cancer (1-3). Established risk factors for breast cancer, such as early menarche and late menopause, may affect risk by affecting lifetime exposure to endogenous estrogens (1). Numerous epidemiologic studies have shown that hormone replacement therapy (HRT) use is associated with an increased risk for breast cancer (4-7). Recently, in the Women's Health Initiative, a randomized controlled trial, increased risk for breast cancer was observed among postmenopausal women taking estrogen plus progestin hormone replacement (8), whereas unopposed estrogen given to hysterectomized women did not increase risk (9). Several mechanisms have been proposed for this increased risk, one of which is that HRT increases oxidative stress (10, 11), because estrogen exposure results in the formation of oxidative species (19).

Catalase, an antioxidant heme enzyme, helps to maintain oxidative balance by catalyzing the conversion of hydrogen peroxide, a powerful oxidative species, into water and molecular oxygen. A C/T polymorphism, in the promoter region of the CAT gene (rs1001179), has been shown to affect the transcriptional activity of the promoter (12). Several studies have shown decreased catalase activity in the CT and TT genotypes relative to the common CC genotype (13-15). The magnitude of this decrease in activity may depend on factors such as race and diet (13). The results of the only previous study to examine the association between CAT genotype and breast cancer risk suggested that women with the CC genotype are at slightly lower risk of breast cancer than those with at least one copy of the T allele (16). Several reports have suggested a connection between estrogen exposure and catalase activity. Treatment of normal human breast epithelial cells in culture with estradiol was shown to decrease cellular catalase activity (17). Further, treatment of breast cancer cell lines with estradiol resulted in decreased catalase activity in the estrogen receptor (ER)–positive but not the ER-negative cell lines (18). We hypothesized that individuals with the lower expressing forms of the CAT polymorphism may be more susceptible to the increased risk of breast cancer associated with HRT use because of increased oxidative stress, decreased CAT expression, and decreased catalase activity. We report here on the results of a population-based case-control study of HRT use and CAT genotype.

Women included in this analysis were participants in the Western New York Exposures and Breast Cancer (WEB) study, a large, population-based case-control study conducted between 1996 and 2001. This study has been described in detail elsewhere (20). Briefly, participants were females between ages 35 and 79 years, residing in either of two counties in Western New York. Cases included women with primary, incident pathologically confirmed breast cancer. Controls were randomly selected from the Department of Motor Vehicle (for those ages <65 years) and the Health Care Finance Administration (for those ages at least 65 years) records and frequency matched to cases on age and race. Of those eligible, 72% of cases and 63% of controls agreed to participate, including providing blood or oral rinse samples for genotyping. The protocol was approved by the institutional review boards of the University at Buffalo and Georgetown University Medical Center as well as by the review boards of the participating hospitals. Informed consent was obtained from all study participants.

Participants completed written questionnaires and were interviewed at the clinic for the collection of information regarding demographics, medical history, reproductive history, and family history of disease. A detailed computer-assisted questionnaire administered by trained interviewers was used to assess use of postmenopausal HRT. For women with breast cancer, ER status was obtained from medical records. Status was defined according to hospital-specific protocols and obtained from chart review by trained research nurses. Hormone receptor status was available for 483 (78.4%) of the postmenopausal cases.

DNA was extracted from blood (n = 1,645) or oral rinse (n = 53) samples collected at the time of the interview. PCR for CAT genotype was conducted using primers 5′-ACGTTGGATGAGCAATTGGAGAGCCTCGC-3′ and 5′-ACGTTGGATGCTGAAGGATGCTGATAACCG-3′. The Homogenous MassExtend reaction was used to prepare the PCR products for matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Controls of each genotype were included on each plate, and there were four no-template controls per plate. CAT genotype was successfully determined for 616 (87.5%) of eligible cases and 1,082 (89.0%) of eligible controls meeting inclusion criteria of being postmenopausal and having HRT use information available.

Analysis here was restricted to postmenopausal women. Postmenopausal status was defined as women who reported natural menopause or history of hysterectomy with bilateral oophorectomy or women, ages ≥50 years, who reported having had a hysterectomy with neither or one ovary removed. The Student's t test was used to compare means of continuous variables and χ² tests were used to compare values for categorical variables. Unconditional multivariable logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals (95% CI), adjusted for age, education, body mass index, age at first pregnancy, history of benign breast disease, history of breast cancer in first-degree relative, age at menarche, and natural versus surgical menopause. Environmental exposures associated with oxidative stress, including smoking and dietary fruit and vegetable consumption, were evaluated as potential confounders and did not significantly affect the model. Polytomous logistic regression was used for analyses by ER status. Ever HRT users were defined as those reporting use for at least 3 months. Duration of use was calculated by summing total duration of use of any form of HRT. Time since last use was calculated using the latest reported age stopped using HRT. Those reporting fewer than 3 months of use were considered never users for the duration of use and time since last use variables. P values for interaction were determined by including a multiplicative term in the regression model. All analyses were completed with the SAS statistical package version 9.1.3 (SAS Institute).

The characteristics of the WEB Study postmenopausal participants for whom complete information on CAT genotype and HRT use were available are shown in Table 1. Both cases (92.4%) and controls (90.0%) were predominantly Caucasians. Patterns in characteristics related to endogenous estrogen exposure were as expected, with cases more likely to be nulliparous and have earlier age at menarche and later age at menopause. The distribution of hormone receptor status is similar to that found elsewhere (21), 79.5% were ER positive and 20.5% were ER negative.

Included in Table 2 are CAT genotype distributions by race and for all races combined. CAT genotypes in controls were in Hardy-Weinberg equilibrium for both Whites (n = 598 CC, 333 CT, and 43 TT; P = 0.70) and non-Whites (n = 97 CC, 10 CT, and 1 TT; P = 0.22). Blinded quality-control duplicates were 99.3% concordant for CAT genotype (data not shown). Genotype distributions in White and non-White controls were different (P < 0.0001), with 61.4% of Whites having the common CC genotype and 89.8% of non-White controls having the CC genotype. There was no significant difference in genotype distribution between cases and controls among Whites; among non-Whites, there is a highly significant difference in distribution for cases and controls (P = 0.006), with 72.3% of non-White cases and 89.8% of non-White controls having the common CC genotype.

Additional analyses were limited to postmenopausal Caucasian women due to small numbers of non-White women in this study. Main effects for breast cancer risk by genotype and by HRT use are shown in Table 3. CAT genotype was not associated with breast cancer risk. Ever use of HRT was associated with an increased risk of breast cancer (OR, 1.39; 95% CI, 1.11-1.75). When classified by duration of use, the increased risk among HRT users appeared to be restricted to those reporting use for ≥5 years (OR, 1.68; 95% CI, 1.29-2.18); shorter duration of use was not associated with risk (OR, 1.09; 95% CI, 0.82-1.46). The increased risk was more pronounced among those who reported use of HRT within the last 5 years (OR, 1.43; 95% CI, 1.13-1.81) than among those reporting at least 5 years since their last use of HRT (OR, 1.21; 95% CI, 0.78-1.87), although the 95% CIs overlapped.

With stratification by CAT genotype (Table 4), there was an increased risk of breast cancer associated with ever use of HRT among those with at least one variant T allele (OR, 1.88; 95% CI, 1.29-2.75) but little increased risk among those with the common CC genotype (OR, 1.15; 95% CI, 0.86-1.54). We next examined duration and time since last HRT use in relation to CAT genotype and breast cancer risk. There was no significant association with HRT use for these measures among those with the common CC genotype. However, among those with at least one copy of the variant T allele, use of HRT for ≥5 years was associated with increased risk relative to never users (OR, 2.32; 95% CI, 1.50-3.59). Similarly, use of HRT within the past 5 years was associated with an increased risk of 2.01 (95% CI, 1.35-2.99) relative to never users with at least one copy of the variant T allele. The interaction with genotype was significant for both ≥5 years of use (P_interaction = 0.02) and use within the last 5 years (P_interaction = 0.01).

These associations were examined further in strata defined by ER status. Due to sample size limitations, only ever/never use was examined by ER subtype. For ER-negative breast cancer, there was no difference in breast cancer risk by CAT genotype (P_interaction = 0.28), although the sample sizes are small and the 95% CIs are wide. For ER-positive breast cancers, there is an OR of 2.26 (95% CI, 1.46-3.49) for use of HRT among those with at least one T allele; use was less strongly associated with risk among those women with the CC genotype (OR, 1.44; 95% CI, 1.01-2.05). There was a significant interaction between genotype and ever/never use (P = 0.03).

Previous studies have consistently shown an increased risk of breast cancer associated with HRT use (4-8). In this population-based case control study of women from Western New York State, we observed that the association between HRT use and breast cancer risk is modified by a CAT genetic polymorphism, the increased risk being more pronounced among women with at least one T allele. This observation is consistent with the hypothesis that the increased risk associated with HRT involves, at least in part, the formation of oxidative species. Very few studies have examined the effect of CAT polymorphisms on breast cancer risk, and no others, to date, have examined the interaction between CAT polymorphisms and HRT use.

There are several mechanisms by which estrogen exposure may increase breast cancer risk, such as increasing cell proliferation and opportunities for random errors during DNA replication (22). However, estrogen metabolism also generates reactive oxidative species (10, 23). Whereas modest levels of reactive oxidative species are necessary for cell signaling processes (24, 25), excess reactive oxidative species can damage DNA, lipids, and proteins (26-29). The metabolism of estrogens results in the generation of reactive quinones capable of forming adducts with DNA and of participating in redox cycling, thereby generating additional reactive oxygen species (19, 30). Consequently, estrogen metabolites have the ability to directly and indirectly result in oxidative damage to cellular components as well as to disrupt signaling processes such as those required for cell growth or apoptosis (24).

HRT preparations often contain equine estrogen; metabolites of some equine estrogens possess greater potential for causing oxidative damage than that caused by human estrogens (31). A metabolite of equine estrogen, 4-hydroxyequilenin, has been shown to cause oxidative damage and single-strand breaks in λ phage DNA (32) and in breast cancer cell lines, especially ER-positive cell lines (33, 34).

CAT genotype alone was not associated with increased breast cancer risk among Whites in our study. Although the number of non-Whites was small, the results suggest that, among non-Whites, those with either CT or TT genotypes may be at increased risk of breast cancer compared with those with the common CC genotype. This result warrants further exploration in a larger sample of non-Whites.

Consistent with previous epidemiologic studies (4-7), the risk associated with HRT appears to increase with prolonged use and there is evidence that current/recent use has a greater effect on risk than past use. As with ever/never use, the increase associated with longer duration of use was more pronounced among those with CT or TT CAT genotypes. Because these variant genotypes have been shown to result in decreased catalase activity (13-15), it appears that impaired ability to process the highly reactive species hydrogen peroxide may result in increased susceptibility to the effects of HRT on breast cancer risk.

On stratification by ER status, we found HRT associated with increased risk of ER-positive and not ER-negative tumors; similar findings have been observed in several other epidemiologic studies (6, 7, 35). Although we had only small numbers of subjects in each stratum, we found that the effect of CAT genotype on the association between HRT and breast cancer risk was restricted to ER-positive breast cancer, suggesting that the estrogen response and oxidative stress pathways are somehow intertwined. This result is not unexpected given the fact that binding of ER to DNA is modulated by redox status (36-38).

Although not permitting detailed subgroup analysis by ER status or race, this study was sufficiently large to allow examination of the interaction between CAT genotype and HRT use on the risk of breast cancer. Quantitative HRT use information was collected using a detailed, interviewer administered questionnaire, which, although subject to recall bias, concerned behavior in the recent past. In addition, comprehensive information regarding potential confounders was available.

The relationship between HRT use and breast cancer risk is complex. Our results suggest that the mechanism by which HRT use increases risk of breast cancer involves oxidative species. Insight into this mechanism may help predict which subsets of women are at greatest risk from HRT, thereby informing future decisions regarding short-term and long-term use. The observed association between HRT and oxidative stress contributes to our understanding of breast cancer etiology and may also have implications for other effects associated with HRT.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Shiva Krishnan for technical skills and the staff and study subjects of the WEB Study for participation in the study.