Polymorphisms in ER-α Gene Interact with Estrogen Receptor Status in Breast Cancer Survival

Estrogens play a crucial role in the pathogenesis and progression of breast cancer. The effects of estrogens are mediated primarily through intracellular estrogen receptors (ER). To date, there are two known ERs, ER-α and ER-β. Both are nuclear receptor proteins that act as ligand-inducible transcription factors (1). ER-α expression has been used in clinical practice as an indicator for selecting hormone therapy and loss of ER-α expression is often associated with poor survival (2).

There are several known polymorphisms in the ER-α gene, some of which alter the function of the receptor (1). Of the known polymorphisms, PvuII, XbaI, and the recently identified (GT)_n polymorphisms have been reported to be associated with breast cancer risk (3-8). PvuII and XbaI are located on intron 1, whereas the (GT)_n repeats polymorphism is located at 2.8 kb 5′ to exon 1D (9). To our knowledge, no study has been published on the relation between ER-α polymorphisms and breast cancer survival. The association of ER polymorphisms with breast cancer risk suggests that these polymorphisms may also affect breast cancer survival. However, it is difficult to extrapolate the findings from etiologic research to cancer prognosis for several reasons: (a) there is not enough evidence to conclusively determine the direction of the polymorphism association with breast cancer risk and (b) even if the associations with breast cancer risk were conclusive, it is not certain whether the polymorphisms would play the same role in breast cancer risk as breast cancer survival. In this report, we evaluated the association of three common polymorphisms in the ER-α gene, PvuII, XbaI, and GT dinucleotide repeats with breast cancer survival in a cohort of breast cancer patients who were recruited as part of the Shanghai Breast Cancer Study, a population-based case-control study.

Participants and Study Design. Through a rapid case ascertainment system and supplemented by the population-based Shanghai Cancer Registry, we identified 1,602 women who were between the ages of 25 to 64 years and were diagnosed with a primary breast cancer between August 1996 and March 1998. Of them, 1,459 (91.1%) completed an in-person interview and were included as cases in the Shanghai Breast Cancer Study. Reasons for nonparticipation included refusal (109 cases, 6.8%), death before interview (17 cases, 1.1%), and inability to locate (17 cases, 1.1%). The median time interval between diagnosis and interview was 66 days. The initial institutional cancer diagnoses were confirmed by independent review of pathologic slides by two senior pathologists. Blood samples (10 mL from each woman) were obtained from 1,193 (82%) cases who participated in the study. These samples were typically processed within 6 hours of collection and were stored at −70°C until the relevant bioassays were conducted. Information on cancer diagnosis, disease stage (tumor-node-metastasis stage), cancer treatments, and ER/progesterone receptor (PR) status was abstracted from medical charts using a standard protocol.

Patients were followed up through January 2003 with a combination of active follow-up and record linkage to the death certificates kept by the Vital Statistics Unit of the Shanghai Center for Disease Control and Prevention. Of the 1,459 patients included in the original study, 1,290 (88.4%) were successfully contacted either by home visit (n = 1,241, 85.0%) or by telephone (n = 49, 3.4%) from March 2000 to December 2002. Among them, 200 patients were deceased. Through interview of patients, or next of kin for deceased patients, we obtained information on disease progress, recurrence, quality of life, and cause of death (if deceased). Survival status for the remaining 169 participants who could not be contacted through a home visit or by telephone call was established in June 2003 by linkage to the mortality registry. Of them, 40 deaths were identified; information on the date of death and cause of death was obtained. Subjects (n = 126) had no match in the mortality registry and were assumed to be still living. Their censoring date was assigned to be December 31, 2002, 6 months before our search of the vital statistics registry, to allow for a possible delay of entry of the death certificates into the registry. Three subjects had insufficient information for the record linkage and were excluded from the current analysis. This study was approved by the Institutional Review Board of all participating institutes.

Laboratory Protocols. Determination of PvuII and XbaI Polymorphisms. Genomic DNA was extracted from buffy coat fractions. PvuII and XbaI genotypes were determined with PCR-RFLP method described previously (7). Briefly, the primers for analysis were 5′-CTGCCACCCTATCTGTATCTTTTCCTATTCTCC-3′ (forward) and 5′-TCTTTCTCTGCCACCCTGGCGTCGATTATCTGA-3′ (reverse). These primers generated a 1.3-kb fragment, and contain a part of intron 1 and exon 2 of the ER-gene. The PCR products were digested by the PvuII and XbaI restriction endonucleases, respectively. The DNA fragments were then separated using 1.5% agarose gel and detected by ethidium bromide staining. PP and XX, signifying the absence of restriction sites, gave one 1.3-kb fragment. pp, signifying the presence of PvuII restriction sites on both alleles, was digested into two fragments (0.85 and 0.45 kb). The xx genotype was revealed by XbaI digestion into two fragments (0.9 and 0.4 kb).

Determination of (GT)_n Polymorphism. Genotyping for the (GT)_n polymorphism was done by detection of fluorescent amplimers on an ABI PRISM 3700 automated DNA analyzer. Details of genotyping method have been described elsewhere (8). Briefly, the primers were forward 5′-gtgtCTGCTCAAATCTCCTCTG-3′ and reverse 5′-GTTAAGAAGGGCCTTTAC-3′. The forward primer was labeled with 6-carboxyfluorescein. Allele fragment size estimation was accomplished using the internal size standard Genescan 400HD ROX and the Local Southern algorithm of GENESCAN software. Allele binding and adjustment of run mobility according to control alleles of Centre d'Etude du Polymorphisme Humain 1347-02 were accomplished by custom software. The number of (GT)_n repeats were confirmed by direct sequencing using BigDye Terminator Chemistry on an ABI PRISM 3700 automated DNA Analyzer.

The laboratory staff was blind to the identity of the subject. QC samples were included in genotyping assays. Each 96-well plate contains one water, two Centre d'Etude du Polymorphisme Humain 1347-02 DNA, two blinded QC DNA, and two unblinded QC DNA samples. The blinded QC samples were taken from the second tube of study samples included in the study.

Genotyping data for the (GT)_n polymorphism were obtained from 947 cases who gave blood samples. For the PvuII and XbaI polymorphisms, genotyping data were obtained from 1,069 cases who gave blood samples. The major reasons for incomplete genotyping were insufficient DNA used in the particular assays and unsuccessful PCR amplification.

Statistical Analysis. The primary outcome for this study was overall survival. Survival time was calculated as the time from cancer diagnosis to death, censoring at the date of last contact. The Kaplan-Meier method was used to compute 5-year survival rates, and the log-rank test was applied to test the differences in survival across different genotypes. The proportional hazard assumption was checked by log (−log) plots. The Cox regression model was applied to evaluate the effect of the ER-α genotype on overall survival with adjustments for age at diagnosis and known prognostic factors for breast cancer, including tumor-node-metastasis status and cancer treatment. Stratified analyses by ER status and traditional breast cancer prognostic factors were done to examine potential interactive effects of these variables on the association between ER-α genotype and breast cancer survival. Test for multiplicative interaction was conducted by including the main term and product of two study variables in the Cox regression model. All statistical tests are based on a two-sided probability.

Among subjects with genotype data (n = 1,063), ER status was obtained from 739 breast cancer cases. Of these, ∼62.4% were ER+ and 37.6% were ER−. Approximately 30.5% of cases were missing information on ER status. ER status was not significantly associated with 5-year survival in this study population. As expected, stage at diagnosis was an important predictor for overall survival (Table 1). Having radiotherapy was related to a higher mortality. Patients who received radiotherapy had a more advanced cancer; however, the inverse association persisted even after adjustment for stage of disease. Nearly all subjects received surgery, six subjects with missing surgery information were excluded from the subsequent analyses. The vast majority (95%) of participants received chemotherapy. Older age at diagnosis was related to lower but not statistically significant survival. Clinical prognostic factors were found to be similar for all subjects as well as those who were included in the genotyping analyses (data not shown).

Unadjusted and multiple-adjusted associations of breast cancer survival with PvuII, XbaI, and (GT)_n polymorphisms are shown in Table 2. No overall association was observed between PvuII or XbaI polymorphisms and breast cancer survival. For the (GT)_n polymorphism, analyses were focused on the five most common alleles: (GT)₁₅, (GT)₁₆, (GT)₁₇, (GT)₁₈, and (GT)₂₃. A significant increased risk for death was observed among participants carrying one (GT)₁₈ allele [multiple-adjusted hazard ratio (HR), 1.61; 95% confidence interval (95% CI), 1.15-2.27]. The risk, however, was not elevated among subjects homozygous for this allele (HR, 1.03; 95% CI, 0.33-3.27).

Genotype associations were modified by ER status in breast cancer tissues (Table 3). Among participants with an ER+ cancer, the adjusted HRs associated with Pp and PP genotypes compared with a pp genotype were 0.41 (95% CI, 0.25-0.67) and 0.54 (95% CI, 0.24-1.23), respectively. These genotypes were associated with HRs of 1.27 (95% CI, 0.61-2.61) and 3.30 (95% CI, 1.42-7.69) for ER-negative breast cancer. Similarly, among those carrying one or two (GT)₂₃ alleles the HR was 0.25 (95% CI, 0.09-0.69) compared with those with no (GT)₂₃ alleles for participants with ER-positive breast cancer, whereas the corresponding HR for ER-negative cancer was 1.48 (95% CI, 0.77-2.87). Tests for multiplicative interaction were both highly significant. For the XbaI polymorphisms, the genotype-breast cancer survival associations did not vary according to ER status.

The effects of ER status on breast cancer survival by ER-α genotypes are shown in Table 4. Among participants with a PP genotype, ER-negative cancer was associated with a worse prognosis than ER-positive breast cancer (HR, 3.93; 95% CI, 1.51-10.25), whereas ER negativity was not related to survival among those participants with pp or Pp genotypes. ER negativity was related to poor survival among patients who had one or two (GT)₂₃ alleles of the GT repeat polymorphism (HR, 5.27; 95% CI, 2.08-13.36) but was unrelated to prognosis among subjects carrying no (GT)₂₃ allele. We did not observe any interaction between XbaI polymorphism and ER status on breast cancer survival.

Additional analyses were conducted stratified by ER/PR status, disease stage, and age at diagnosis. Results from the analyses stratifying by ER/PR status were similar to those found when considering ER and PR status individually (data not shown). There were no differences in the genotype-survival association according to disease stage or age at diagnosis (data not shown).

In this study, we found a multiplicative interaction between PvuII and (GT)₂₃ polymorphisms in the ER-α gene and ER status of the breast cancer on breast cancer survival. Although ER status is known to modify the outcome of hormonal treatment for breast cancer (10), the biological mechanism behind its interaction with ER-α genotypes remains to be clarified. To our knowledge, no previous study has investigated the association between polymorphisms in the ER-α gene and breast cancer survival alone or with ER status.

The three polymorphisms investigated in this study are located in different areas of the ER-α gene. The GT repeat polymorphism is located at 2.8 kb 5′ to exon 1D, and PvuII and XbaI are located on intron 1. These polymorphisms are not in close linkage disequilibrium in the Chinese population (7, 8).

Although polymorphisms in the ER gene have been linked to altered tissue responsiveness to estrogens, the functional impact of these ER-α polymorphisms is not well understood (11). Noncoding short tandem repeats may act as protein binding sites. It has been reported that protein binding may depend on the ability of short tandem repeats to form specific DNA structures and an increasing number of short tandem repeats in the promoter region or introns may interfere with transcription processes by their effect on secondary DNA structure (9). Previous studies investigating ER polymorphisms and breast cancer risk have produced mixed results, which may be explained by ethnicity of the populations under study (3-8). We have linked presence of the (GT)₁₇ or (GT)₁₈ allele of GT polymorphism, or PP allele of Pvull polymorphism to an reduced risk of breast cancer (7, 8) in Chinese population. We did not find that the associations between ER-α genotypes and breast cancer risk vary according to ER/PR status (7, 8).

Previous studies have shown that common ER-α polymorphisms, including the one found at the PvuII restriction site, are associated with cardiovascular disease risk (13) and possibly modify the effects of estrogens on HDL cholesterol level (14) and changes in bone mineral density and vertebral fractures (15). The latter two studies showed that women with the pp genotype who received hormone replacement therapy had a significantly more pronounced response in outcome than women with other genotypes (i.e., a greater increase in HDL cholesterol levels and bone mineral density; refs. 14, 15), suggesting that this polymorphism may have an interactive effect with hormones.

ER status is the main determinant for tamoxifen use and modulates its effect. According to an ongoing study of Chinese breast cancer patients conducted by our group, we found that 83% of ER-positive breast cancer versus 28% of ER-negative breast cancer patients used tamoxifen up to 18 months post-diagnosis. The interaction between genotype and ER status on breast cancer prognosis found in the current study, therefore, could reflect a differential effect of tamoxifen by genotype. Unfortunately, information regarding tamoxifen use in the current study was collected ∼3 to 4 years after cancer diagnosis. This information was missing for most of the deceased cases and subjects whose survival status was established via linkage to the death registry. Therefore, direct evaluation of the interaction between ER-α gene polymorphism and tamoxifen use was not possible. Nevertheless, we did include tamoxifen use in the model that evaluated the interaction between genotype and ER status, treating those with missing tamoxifen information as a subgroup. We found that patterns of interactive effect remained but the point estimates attenuated (data not shown), suggesting tamoxifen use may play a role in the ER genotype and breast cancer survival association. Future studies are needed to evaluate the possible interactive effect between ER genotype and tamoxifen use on breast cancer survival.

Our study has several strengths that should be considered in interpreting its results, including (a) a population-based patient cohort, (b) the relatively long follow-up period of 5.2 years, (c) a large sample size, (d) relatively homogeneous genetic background (∼98% of the study population are Han Chinese), and (e) detailed information on ER/PR status, disease stage, and treatment information.

There are also a few limitations that must be considered in evaluating these results. As mentioned above, information regarding tamoxifen use was only collected for a subset of the women participating in this study, making the evaluation of the interaction of ER genotype and tamoxifen impossible. In addition, ER status was abstracted from hospital record and missing for ∼30% of the study participants, thus compromising the statistical power of this study to evaluate interactions between genotype and ER status.

In summary, our findings of an interaction between ER status and ER-α polymorphisms on breast cancer survival are new and need to be confirmed in future studies. If confirmed, these findings could have important clinical implications in breast cancer treatment.

We thank the patients and research staff who participated in the Shanghai Breast Cancer Study.