CYP2D6 phenotype, tamoxifen, and risk of contralateral breast cancer in the WECARE Study

The Women’s Environmental Cancer and Radiation Epidemiology (WECARE) Study is a multicenter, population- based case–control study where cases are women with asynchronous CBC and controls are women with unilateral breast cancer (UBC). Case–control recruitment for the WECARE Study was conducted in two phases: WECARE I (2001–2004) and WECARE II (2009–2012). Participants were identified through eight population-based cancer registries: Los Angeles County Cancer Surveillance Program; Cancer Surveillance System of the Fred Hutchinson Cancer Research Center (Seattle, WA, USA); State Health Registry of Iowa; The Cancer Surveillance Program of Orange County/San Diego—Imperial Organization for Cancer Control (Orange County/San Diego, CA, USA); the Greater Bay Area Cancer Registry (San Francisco Bay Area Region and Santa Clara Region, CA, USA); and the Sacramento and Sierra Center Registries (Sacramento Region, CA, USA). These cancer registries all contribute to the National Cancer Institute (NCI) Surveillance, Epidemiology and End Results (SEER) program in the United States. Patients were also recruited from the Ontario (Canada) Cancer Registry and the Danish Breast Cancer Cooperative Group Registry, supplemented by data from the Danish Cancer Registry. All study participants provided written informed consent, and the study protocol was reviewed and approved by the Institutional Review Board at each recruitment site and the ethical committee systems in Denmark and Ontario. Across the eight cancer registries a total of 2354 CBC cases and 3599 UBC controls were identified as being eligible and were approached for the study. The final number of participants who completed the interview and provided a biospecimen for DNA analysis was 1518 (64%) cases and 2208 (61%) controls. Reasons for nonparticipation have been described in detail [28]. An additional nine participants were excluded from the current analysis due to low quality or quantity of DNA (four cases, five controls), resulting in a final sample size of 1514 CBC cases and 2203 UBC controls.

Details of recruitment procedures, eligibility, and the study questionnaire have been described previously, and were nearly identical for the two study phases [29]. Briefly, all women were diagnosed prior to age 55 years between 1985 and 2010 with a first primary invasive breast cancer (stage I–III). Cases were diagnosed with a second primary CBC (in situ or invasive for WECARE I and invasive only for WECARE II) at least 1 year later. Controls also had no history of any second cancer diagnosis up to their reference date. The reference date for cases was the CBC diagnosis date, while for controls this was defined by adding the interval between the first breast cancer and the CBC for the matched case to the date of (first) breast cancer diagnosis for the control. Cases must also have been living in the same study reporting area for both diagnoses, while controls were required to be living in the cancer-reporting area of their (first) breast cancer diagnoses on their reference dates. Additionally, controls must not have undergone prophylactic mastectomy of the unaffected contralateral breast. Study eligibility was restricted to women who were alive when contacted and were able to provide informed consent, complete a telephone interview, and donate a blood or saliva sample for DNA extraction. Controls were matched to cases (2:1 for WECARE I and 1:1 for WECARE II) on year of birth in 5-year strata, year of diagnosis in 4-year strata, cancer registry region, and race/ethnicity. In WECARE I, cases and controls were further counter-matched based on cancer registry-reported radiation treatment such that two members of each case–control trio had received radiation treatment for their first breast cancer and the other member had not [29]. Counter-matching was not used in WECARE II; this was taken into account in all statistical analyses, as detailed in the following. The selection of case–control trios in this way for WECARE I is reflected in the frequency of radiation treatment by case–control status such that the frequencies cannot be directly compared. For this reason, weighted frequencies are provided (Table 1).

WECARE Study participants were interviewed by telephone using a structured questionnaire that was designed to obtain information about events occurring before the diagnosis of the first primary breast cancer, as well as events that occurred during the at-risk period. The at-risk period was defined as beginning at least 1 year after diagnosis with a first breast cancer and ending at the second diagnosis in CBC cases, or the corresponding reference date for UBC controls. The study questionnaire included questions about personal demographics, medical history, family and reproductive history, hormone use, body size, smoking status, and alcohol intake. Additionally, medical records, pathology reports, and hospital charts were used to collect detailed treatment information (i.e., chemotherapy, hormonal therapy, and radiation therapy) for the first primary breast cancer, any recurrences experienced prior to the reference date, and tumor characteristics of the first primary tumor (e.g., ER status, histology).

Germline DNA was isolated from blood using standard phenol–chloroform extraction methods or from saliva using the manufacturer’s recommendations (Genotek, ON, Canada). Samples were genotyped for seven single nucleotide polymorphisms (SNPs), one three-nucleotide insertion-deletion (indel), and a full gene deletion (Table 2), accounting for most of the clinically significant variants in CYP2D6 [30]. Genotyping was conducted in the Molecular Epidemiology Laboratory at Memorial Sloan Kettering Cancer Center using PCR-based methods.

Variants were tested using Agena iPlex chemistry and the MassArray system (Agena Bioscience, San Diego CA, USA), with assays designed using the Agena Cx online tools (currently https://​agenacx.​com), and a method similar to that reported by Falzoi et al. [31], with an independent preamplification of CYP2D6 in a long-range PCR. The PCR products were verified on a 1% agarose gel. One microliter of the amplified product was used for genotyping using specific primers and cycling conditions (Additional file 1). Products were purified enzymatically, extended and desalted, and then spotted onto 384 SpectroCHIP bioarrays (Agena Bioscience). Cluster plots were evaluated with the TyperAnalyzer application (MassARRAY v3.4). Assays were considered optimal according to degree of clustering, absence of signal in the blanks, and reproducibility. Quality control (QC) procedures were similar to those published elsewhere [32, 33], and included 5% intraplate duplicates for each plate, 0.5% interplate duplicates, and at least two independent readers for review and interpretation of cluster plots and results. Samples with known genotypes from the Coriell Institute were also included as controls during the assay development. Samples with weak signals were repeated.

Full CYP2D6 deletion (CYP2D6*5) analysis was conducted with a modified three-primer-based, long-range PCR using primers described by Okubo et al. [34] to ensure gene specificity. Genomic DNA (30 ng) was amplified in a reaction mix containing 1× Takara LA-Taq PCR buffer with 2.5 mM Mg²⁺ (Clontech, Mountain View, CA, USA), 0.5 mM dNTPs (Life Technologies, Carlsbad, CA, USA), 0.5 μM of each primer, and 0.4 units of Takara LA Taq DNA Polymerase (Clontech), in a total volume of 8 μl. Cycling included denaturation at 94 °C for 1 min, 25 cycles of 98 °C for 20 s and 70 °C for 6 min, annealing at 70 °C for 6 min, and a final hold at 4 °C. PCR products were run on 1% e-gels (Invitrogen, Carlsbad, CA, USA), and band sizes were determined against molecular weight markers to evaluate CYP2D6 presence (4.7 kb band) or deletion (CYP2D6*5, 3.5 kb). QC procedures included use of control samples from the Coriell Institute, 5% intraplate and 0.5% interplate duplicates, as well as negative (sterile distilled water) controls.

CYP2D6 nomenclature came from the American College of Medical Genetics and Genomes (ACMG) Standards and Guidelines [35] and PharmVar (Pharmacogene Variation Database, www.​Pharmvar.​org/​gene/​CYP2D6 (Table 1). Alleles were classified into three categories based on their anticipated impact on CYP2D6 activity (normal, CYP2D6*1 and CYP2D6*2; reduced, CYP2D6*9, CYP2D6*10, and CYP2D6*41; and inactive, CYP2D6*3, CYP2D6*4, CYP2D6*5, and CYP2D6*6) [9] and assigned activity scores (AS) of 1, 0.5, and 0, respectively. The scores for the two alleles carried by each woman were summed in order to assign an overall AS which could range from 0 to 2. One exception to this is the instance of individuals carrying two IM (reduced activity) alleles. In order to distinguish between two combinations that would otherwise have had the same overall AS of 1.0, we left individuals classified as EM/PM with an AS of 1 and assigned IM/IM women an AS of 0.75 (instead of the AS of 1.0 they would otherwise have received) [26]. In this way, women were classified as: EM/EM (AS = 2), EM/IM (AS = 1.5), EM/PM (AS = 1), IM/IM (AS = 0.75), IM/PM (AS = 0.5), and PM/PM (AS = 0) [26, 36, 37]. Analyses were also conducted collapsing across some categories of the AS to generate an overall inferred phenotype. Women were classified as EM if they had at least one EM allele (i.e., AS ≥ 1), IM if they had no EM alleles and at least one IM allele (e.g., AS = 0.5–0.75), or PM if they had two PM alleles (i.e., AS = 0) [27].

Rate ratios (RRs) and 95% confidence intervals (CIs) for the association between tamoxifen treatment and risk of CBC in our study population were estimated fitting conditional logistic regression models, stratified by CYP2D6 AS. Rate ratios can be interpreted as a relative risk (e.g., ratio of risks) of CBC in tamoxifen-treated versus untreated women in our study population. Models were adjusted for: age at first diagnosis, histology, stage, and ER status of first diagnosis, first-degree family history of breast cancer, chemotherapy for first breast cancer, radiation treatment for first breast cancer, hormonal therapy other than tamoxifen for a first breast cancer, number of full-term pregnancies before first diagnosis, and age at menarche. Because treatment for a first breast cancer could induce menopause, menopausal status and age at menopause 2 years prior to first diagnosis were used. To account for the counter-matching in WECARE I, we included a log-weight covariate offset term. For WECARE II participants who were matched in pairs (without counter-matching on radiation treatment), we assigned the value of 1 to be the offset term. We used the likelihood ratio test to test for homogeneity of treatment effect across AS/phenotypes. A sensitivity analysis was conducted restricting model fitting to Caucasian women (N = 1335 CBC cases and N = 1980 UBC controls) to address the issue of ancestral differences in genotype frequencies. All statistical analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).