Abstract

Background

Ionizing radiation is a known mutagen and an established breast carcinogen. The ATM gene is a key regulator of cellular responses to the DNA damage induced by ionizing radiation. We investigated whether genetic variants in ATM play a clinically significant role in radiation-induced contralateral breast cancer in women.

Methods

The Women's Environmental, Cancer, and Radiation Epidemiology Study is an international population-based case–control study nested within a cohort of 52 536 survivors of unilateral breast cancer diagnosed between 1985 and 2000. The 708 case subjects were women with contralateral breast cancer, and the 1397 control subjects were women with unilateral breast cancer matched to the case subjects on age, follow-up time, registry reporting region, and race and/or ethnicity. All women were interviewed and underwent full mutation screening of the entire ATM gene. Complete medical treatment history information was collected, and for all women who received radiotherapy, the radiation dose to the contralateral breast was reconstructed using radiotherapy records and radiation measurements. Rate ratios (RRs) and corresponding 95% confidence intervals (CIs) were estimated by using multivariable conditional logistic regression. All P values are two-sided.

Results

Among women who carried a rare ATM missense variant (ie, one carried by <1% of the study participants) that was predicted to be deleterious, those who were exposed to radiation (mean radiation exposure = 1.2 Gy, SD = 0.7) had a statistically significantly higher risk of contralateral breast cancer compared with unexposed women who carried the wild-type genotype (0.01–0.99 Gy: RR = 2.8, 95% CI = 1.2 to 6.5; ≥1.0 Gy: RR = 3.3, 95% CI = 1.4 to 8.0) or compared with unexposed women who carried the same predicted deleterious missense variant (0.01–0.99 Gy: RR = 5.3, 95% CI = 1.6 to 17.3; ≥1.0 Gy: RR = 5.8, 95% CI = 1.8 to 19.0; Ptrend = .044).

Conclusions

Women who carry rare deleterious ATM missense variants and who are treated with radiation may have an elevated risk of developing contralateral breast cancer. However, the rarity of these deleterious missense variants in human populations implies that ATM mutations could account for only a small portion of second primary breast cancers.

CONTEXT AND CAVEATS
Prior knowledge

The ataxia telangiectasia mutated ( ATM ) gene regulates cellular responses to the DNA damage induced by ionizing radiation, an established breast carcinogen. It is unclear whether individuals who carry ATM mutations and are exposed to radiation are especially susceptible to radiation-induced breast cancer.

Study design

An international population-based case–control study nested within a cohort of 52 536 unilateral breast cancer survivors that investigated whether genetic variants in ATM play a clinically significant role in radiation-induced contralateral breast cancer.

Contribution

Women who carried rare ATM missense variants that are predicted to be deleterious and who were treated with radiation had an increased risk of developing contralateral breast cancer compared with nonirradiated women who carried the same predicted deleterious missense variant.

Implications

The increased risk of radiation-related contralateral breast cancer associated with specific ATM mutations may be an important factor in the selection of treatment for breast cancer for women who have a family history of ataxia-telangiectasia, a rare autosomal recessive disorder that arises from inactivating mutations in the ATM gene.

Limitations

There were few carriers of missense variants that were predicted to be deleterious in the study population, which limited the precision of the estimates.

From the Editors

Of the nearly 185 000 US women who were expected to develop breast cancer in 2009 and the more than 2.4 million breast cancer survivors who currently reside in the United States, 5%–10% will develop a subsequent primary cancer in the contralateral breast ( www.cancer.org ). Epidemiological studies have identified a number of factors that are associated with an increased risk of developing contralateral breast cancer, including family history of breast cancer, early age at diagnosis, hormonal factors, reproductive history, body mass index, and characteristics of the first primary tumor (eg, lobular histology, stage, and estrogen receptor status) ( 1–10 ). The risk of developing contralateral breast cancer has also been associated with mutations in specific genes, including BRCA1 , BRCA2 , and CHEK2 ( 11–17 ). Importantly, the treatment a woman receives for her first primary breast cancer can also influence her risk of developing a second breast cancer: Chemotherapy is associated with a 40% reduction in the risk of developing contralateral breast cancer ( 1 , 18 , 19 ), whereas the radiation received to the contralateral breast during radiotherapy is associated with an increased risk of contralateral breast cancer ( 9 , 20 , 21 ). For example, we previously reported a threefold increase in the risk of contralateral breast cancer associated with a radiation dose of 1 Gy or more to the contralateral breast among women younger than 40 years who survived at least 5 years after treatment compared with women who received no radiotherapy; however, no excess risk was observed in women older than 40 years ( 9 ). Nevertheless, the combined effect of these factors accounts for only a portion of the contralateral breast cancers that develop each year.

Patients with ataxia-telangiectasia (A-T), a rare autosomal recessive disorder that arises from inactivating mutations in the ataxia telangiectasia mutated ( ATM ) gene, display both cellular hypersensitivity to ionizing radiation and increased incidence of cancers ( 22–24 ). ATM, the protein product of the ATM gene, plays a central role in sensing and signaling the presence of DNA double-strand breaks that are induced by ionizing radiation ( 22 , 25 , 26 ). Upon activation by ionizing radiation, ATM phosphorylates a large number of proteins that control pathways that lead to cell cycle checkpoint arrest, DNA repair, or apoptosis. Among the hundreds of known substrates for ATM phosphorylation are the products of three genes that have been implicated in the etiology of contralateral breast cancer— BRCA1 , BRCA2 , and CHEK2 ( 27 , 28 )—which suggests that genetic variation in ATM might modify the activities of those proteins and thereby affect risk of cancer, particularly in the context of radiation exposure. Studies of A-T families have consistently reported an excess of female breast cancer among obligate heterozygous carriers of ATM mutations compared with noncarriers ( 29–36 ), including a study by Swift et al. ( 35 ) that demonstrated an association between self-reported radiation exposure and breast cancer among ATM obligate heterozygotes. However, the rarity and diversity of ATM mutations, coupled with the lack of well-documented radiation exposures among carriers, have limited the statistical power of case–control studies that evaluate either main ( 37–44 ) or radiation-dependent ( 35 , 45 , 46 ) associations between ATM mutations and risk of breast cancer.

It is unclear whether, and to what extent, genetic factors and radiation exposure, individually or via interaction, contribute to the development of radiation-induced contralateral breast cancer. A key unresolved question is whether individuals who carry ATM mutations and are exposed to radiation are especially susceptible to radiation-induced breast cancer. To address this question, we initiated the Women's Environmental Cancer and Radiation Epidemiology (WECARE) Study, a population-based nested case–control study designed specifically to evaluate interactions between genetic variation in genes such as ATM and radiation exposure in the etiology of breast cancer ( 47 ).

Subjects and Methods

Study Population

The WECARE Study is a multicenter population-based case–control study nested within a cohort of 52 536 women with histologically confirmed invasive breast cancer whose cancers were reported to one of four population-based cancer registries in the United States (the Los Angeles County Cancer Surveillance Program, the Cancer Surveillance System of the Fred Hutchinson Cancer Research Center [Seattle region], the State Health Registry of Iowa, and the Cancer Surveillance Program of Orange County/San Diego-Imperial Organization for Cancer Control [Orange County/San Diego]) or the nationwide Danish Breast Cancer Cooperative Group Registry. All participants were identified, recruited, and interviewed through these registries and were known to be cancer free during the first year after breast cancer diagnosis. The outcome of interest was subsequent primary cancer in the contralateral breast at least 1 year after the initial unilateral breast cancer diagnosis. The average interval between the first and second breast cancer diagnoses was 5 years (range = 1–16 years) ( Supplementary Table 1 , available online). The study design and the characteristics of the cohort have been described in detail ( 7 , 9 , 47 ).

Table 1

Rate ratio of asynchronous contralateral breast cancer associated with ATM gene mutation carrier status *

ATM variant classification  All case subjects (n = 708) All control subjects (n = 1397) RR (95% CI)  Case subjects with dose estimates † (n = 606)   Control subjects with dose estimates † (n = 1200)  RR (95% CI) 
All variants, broadly classified 
    Wild type 271 480 1.0 (referent) 223 418 1.0 (referent) 
    Silent 88 157 1.1 (0.7 to 1.5) 78 134 1.1 (0.8 to 1.6) 
    Missense 75 129 1.2 (0.8 to 1.7) 68 113 1.2 (0.8 to 1.8) 
    Splicing 16 0.6 (0.2 to 1.8) 14 0.7 (0.2 to 2.4) 
    Truncation 11 2.0 (0.7 to 5.9) 11 2.8 (0.9 to 8.9) 
    Common ‡ 355 778 0.8 (0.6 to 1.0) 308 655 0.8 (0.6 to 1.0) 
Missense variants classified using SIFT § 
    Wild type 271 480 1.0 (referent) 223 418 1.0 (referent) 
    Tolerated 36 72 0.9 (0.6 to 1.5) 31 66 0.9 (0.5 to 1.5) 
    Deleterious 39 56 1.4 (0.8 to 2.3) 37 46 1.7 (0.9 to 2.9) 
ATM variant classification  All case subjects (n = 708) All control subjects (n = 1397) RR (95% CI)  Case subjects with dose estimates † (n = 606)   Control subjects with dose estimates † (n = 1200)  RR (95% CI) 
All variants, broadly classified 
    Wild type 271 480 1.0 (referent) 223 418 1.0 (referent) 
    Silent 88 157 1.1 (0.7 to 1.5) 78 134 1.1 (0.8 to 1.6) 
    Missense 75 129 1.2 (0.8 to 1.7) 68 113 1.2 (0.8 to 1.8) 
    Splicing 16 0.6 (0.2 to 1.8) 14 0.7 (0.2 to 2.4) 
    Truncation 11 2.0 (0.7 to 5.9) 11 2.8 (0.9 to 8.9) 
    Common ‡ 355 778 0.8 (0.6 to 1.0) 308 655 0.8 (0.6 to 1.0) 
Missense variants classified using SIFT § 
    Wild type 271 480 1.0 (referent) 223 418 1.0 (referent) 
    Tolerated 36 72 0.9 (0.6 to 1.5) 31 66 0.9 (0.5 to 1.5) 
    Deleterious 39 56 1.4 (0.8 to 2.3) 37 46 1.7 (0.9 to 2.9) 
*

The multivariable models were adjusted for factors found to be statistically significantly associated with contralateral breast cancer in the univariate models and those known to be associated with breast cancer, including the following: exact age at diagnosis of first primary breast cancer, age at menarche (<13 or ≥13 years), menopausal status (premenopausal and age at menopause <45 or ≥45 years), number of full-term pregnancies (0, 1, 2, 3, or ≥4), family history of breast cancer among first-degree relative (yes or no), lobular histology (yes or no) and Surveillance, Epidemiology, and End Results summary stage ( 47 ) (local or regional) of the first primary, and treatment history (chemotherapy or hormonal therapy and radiation where indicated). CI = confidence interval; RR = rate ratio.

Case subjects and matched control subjects for whom there were location-specific dose estimates.

ATM variants carried by 1% or more of the WECARE Study participants.

§

Variants with normalized probabilities less than .05 are predicted to be deleterious, whereas those with probabilities equal to or greater than .05 are predicted to be tolerated. Results for missense variants are adjusted for other variants. One control subject carried only one variant that lacked a SIFT classification.

The 708 case subjects were women with asynchronous contralateral breast cancer who met the following eligibility criteria: 1) was diagnosed between January 1, 1985, and December 31, 2000, with a first primary invasive breast cancer that had not spread beyond the regional lymph nodes at diagnosis and, at least 1 year after the first breast cancer diagnosis, with a second primary in situ (approximately 20% of contralateral breast cancers in this study) or invasive breast cancer in the contralateral breast; 2) resided in the same study reporting area for both diagnoses; 3) had no previous or intervening cancer diagnosis; 4) was younger than 55 years at the time of diagnosis of the first primary breast cancer; 5) was alive at the time of contact for this study; and 6) was able to provide written informed consent, complete the interview, and provide a blood sample.

The 1397 control subjects were women with unilateral breast cancer who met the following criteria: 1) was diagnosed since January 1, 1985, with a first primary in situ or invasive breast cancer while residing in one of the study reporting areas; 2) resided during the “at-risk” interval (ie, the interval between the matched case subject's first and second breast cancer diagnoses) in the same cancer reporting area as when they were diagnosed with their breast cancer; 3) was never diagnosed (as of the reference date) with a second primary breast cancer or any other cancer; 4) was diagnosed with a first primary breast cancer before the age of 55 years; 5) was able to provide written informed consent, complete the interview, and provide a blood sample; and 6) had not had a prophylactic mastectomy of the contralateral breast (as of the reference date). Two control subjects were individually matched to each case subject on the year of birth (5-year strata), year of diagnosis (4-year strata), registry region, and race and/or ethnicity. The two control subjects were also countermatched on registry-reported radiotherapy (ever or never) to improve statistical efficiency. That is, each case subject and her two matched control subjects formed a triplet, wherein two members had received radiotherapy (according to the registry records) and one member had not. Countermatching was done to assure variation in radiation exposure within case–control sets while allowing for unbiased estimation of the main effects and interactions of interest. In this study, countermatching increased the precision of the radiation main effects and the ATM gene–radiation exposure interactions compared with random sampling of the same number of control subjects ( 47 ).

Data Sources

All WECARE Study participants were interviewed by telephone with the use of a questionnaire that focused on known and suspected risk factors for breast cancer ( 47 ). Medical records, pathology reports, and hospital charts were used to collect detailed information on all radiotherapy, chemotherapy, and hormonal therapy received for the treatment of primary breast cancer, metastases, recurrences, and benign conditions as well as tumor characteristics ( 47 ). Estimated absorbed radiation doses to various specific locations in the contralateral breast were reconstructed for each treatment regimen by using tissue-equivalent phantoms. Individual radiation doses were estimated in a blinded fashion with respect to case–control status and derived for the specific contralateral breast cancer locations and treatment regimen of the patient as previously described ( 9 ) and were available for 606 triplets.

Laboratory Methods

Detailed methods for the laboratory analyses and quality control used to generate the data analyzed in this study have been previously published ( 48 , 49 ). Briefly, DNA was prepared by phenol–chloroform extraction from blood samples that were collected at the time of interview. All coding exons (exons 4–65) of the ATM gene along with flanking intronic sequences ranging from 50 to 100 nucleotides in length were screened for sequence variation with the use of denaturing high-performance liquid chromatography as previously described ( 48 , 49 ). Amplicons that yielded sequence variants by denaturing high-performance liquid chromatography analysis were evaluated by direct nucleotide sequencing. The laboratories were blinded as to the case–control status of the samples and all matching information.

Statistical Analysis

To assess the association between ATM mutation carrier status and the risk of developing contralateral breast cancer, we estimated rate ratios (RRs) with corresponding 95% confidence intervals (CIs) by using univariate and multivariable conditional logistic regression models that included an “offset” term consisting of weights to adjust appropriately for the countermatched sampling ( 47 , 50 ). The multivariable models were adjusted for factors that were found to be statistically significantly associated with the risk of contralateral breast cancer in the univariate models and included the following: exact age at diagnosis of the first primary; age at menarche (<13 and ≥13 years); menopausal status (premenopausal, menopause before 45 years, and menopause at 45 years or older); number of full-term pregnancies (0, 1, 2, 3, or ≥4); family history of breast cancer among first-degree relative (yes or no); lobular histology and Surveillance, Epidemiology, and End Results summary stage [local and regional ( 47 )] of the first primary; and treatment history (chemotherapy and hormonal and radiation therapy where indicated). Rate ratios that assess the effect of radiation dose on the risk of contralateral breast cancer were calculated by comparing, within each triplet, the dose received at the specific contralateral breast cancer location in the case subject with the dose received at the same breast location in her matched control subjects ( 51 ). Tests of homogeneity of the slopes of excess relative risk (ERR) per radiation dose in Gy across ATM variant subtypes (silent, splicing, missense, or truncating), age at diagnosis (<45 vs ≥45 years), and time since diagnosis (<5 vs ≥5 years) were performed by using likelihood ratio tests that compared the model that included a separate slope parameter for each age subgroup with one that included a single slope parameter. Cut points used for age at diagnosis and time since diagnosis were based on our prior work ( 9 ). Missing data indicators were used to account for missing covariate data ( 52 ). We calculated rate ratios that reflect two different referent groups; one rate ratio is relative to subjects with wild-type ATM who did not undergo radiotherapy, and the second is relative to subjects who carried the same class of mutation (eg, wild type, silent, missense, splicing, truncation, or common) and who did not receive radiotherapy. All analyses were conducted using the TPHREG procedure in SAS release 9.1 (SAS Institute, Inc, Cary, NC). All P values are two-sided, and those less than .05 were considered to be statistically significant.

Risk associated with carrying ATM variants was examined separately for common variants (ie, those carried by ≥1% of the WECARE Study participants) and for rare variants (ie, those with allele frequencies <1% in the WECARE Study population). Rare variants were classified on the basis of their predicted effect on the ATM protein as silent, missense, splicing, or truncating ( Supplementary Table 2 , available online). It should be noted that an individual subject may have more than one of these classes of variants and that the variant classes are not mutually exclusive (eg, some splicing variants can result in truncation). Rare missense variants were further classified as to their predicted effect on ATM protein structure or function using the SIFT program ( 53 ), which predicts whether an amino acid substation will affect protein function based on a clustal alignment of available vertebrate ATM sequences. SIFT scores range from 0 to 1, and lower scores predict variants that are most likely to be deleterious. Based on the SIFT score, rare missense variants were therefore classified as tolerated or deleterious; variants with scores less than 0.05 are predicted to be deleterious. To summarize the variants over all 62 exons, we classified women whose ATM sequence differed from wild type at more than one position based on the sum of the SIFT scores of variants across all exons. If two rare missense variants occurred at a single exon, the lowest scoring variant was selected as the score for that exon. SIFT scores range from 0 to 1, and lower scores predict variants that are most likely to be deleterious. We compared the results of analyses using SIFT with those from another similar analytic program, PolyPhen ( 54 ), and found them to be equivalent. Therefore, we present here only the analyses that used SIFT scores. We also conducted analyses using the lowest-scoring single variant at any position (instead of the sum of the scores), and similar results were obtained. In each variant-specific model (and those using SIFT), the rate ratio was adjusted for the other variant types. To assess the effect of including in situ contralateral breast cancer, analyses were conducted including and excluding triplets in which the case subject was diagnosed with in situ contralateral breast cancer. Because those risk estimates were equivalent, all triplets were retained in the analyses presented here. Lastly, we conducted analyses by excluding BRCA1 and BRCA2 mutation carriers, and the risk estimates were again equivalent; therefore, the analyses presented here include the entire study population.

Table 2

Rate ratio of developing contralateral breast cancer by radiation exposure and ATM gene mutation carrier status *

ATM variants   Radiation exposure, Gy † Case subjects (n = 606) Control subjects (n = 1200)  RR ‡ (95% CI)   RR § (95% CI)  ERR/Gy (95% CI) 
All variants, broadly classified 
    Wild type 112 72 1.0 (referent) 1.0 (referent)  
0.01–0.99 57 177 1.1 (0.7 to 1.6) 1.1 (0.7 to 1.6)  
≥1.0 54 169 1.1 (0.7 to 1.7) 1.1 (0.7 to 1.7) 0.0 (<<0 to 0.3) 
    Silent 38 29 1.1 (0.6 to 2.0) 1.0 (referent)  
0.01–0.99 25 59 1.3 (0.7 to 2.2) 1.1 (0.5 to 2.4)  
≥1.0 15 46 1.0 (0.5 to 2.0) 0.9 (0.4 to 2.2) 0.2 (−0.3 to 1.3) 
    Missense 26 30 0.6 (0.3 to 1.1) 1.0 (referent)  
0.01–0.99 21 45 1.7 (0.9 to 3.1) 2.7 (1.2 to 6.4)  
≥1.0 21 38 2.0 (1.1 to 3.9) 3.3 (1.4 to 8.0) 1.3 (0.1 to 3.9) 
    Splicing — 1.0 (referent)  
0.01–0.99 1.5 (0.4 to 6.5) —  
≥1.0 0.4 (0.0 to 3.6) — — 
    Truncation 1.6 (0.3 to 8.6) 1.0 (referent)  
0.01–0.99 2.9 (0.5 to 16.3) 1.7 (0.2 to 19.3) 2.5 (−0.4 to 36.3) 
≥1.0 — —  
    Common‖ 154 126 0.8 (0.6 to 1.2) 1.0 (referent)  
0.01–0.99 84 308 0.8 (0.5 to 1.1) 0.9 (0.6 to 1.3)  
≥1.0 70 221 0.9 (0.6 to 1.4) 1.1 (0.7 to 1.6) 0.0 (−0.2 to 0.3) 
Missense variants classified by SIFT ¶ 
    Wild type 112 72 1.0 (referent) 1.0 (referent)  
0.01–0.99 57 177 1.1 (0.7 to 1.6) 1.1 (0.7 to 1.6)  
≥1.0 54 169 1.1 (0.7 to 1.7) 1.1 (0.7 to 1.7) 0.0 (<<0 to 0.3) 
    Tolerated 12 16 0.7 (0.3 to 1.7) 1.0 (referent)  
0.01–0.99 27 1.1 (0.4 to 2.7) 1.6 (0.5 to 5.2)  
≥1.0 10 23 1.3 (0.6 to 3.2) 1.8 (0.6 to 5.8) 0.8 (−0.1 to 3.6) 
    Deleterious 14 14 0.6 (0.2 to 1.3) 1.0 (referent)  
0.01–0.99 12 17 2.8 (1.2 to 6.5) 5.3 (1.6 to 17.3)  
≥1.0 11 15 3.3 (1.4 to 8.0) 5.8 (1.8 to 19.0) 2.6 (0.0 to 10.6) 
ATM variants   Radiation exposure, Gy † Case subjects (n = 606) Control subjects (n = 1200)  RR ‡ (95% CI)   RR § (95% CI)  ERR/Gy (95% CI) 
All variants, broadly classified 
    Wild type 112 72 1.0 (referent) 1.0 (referent)  
0.01–0.99 57 177 1.1 (0.7 to 1.6) 1.1 (0.7 to 1.6)  
≥1.0 54 169 1.1 (0.7 to 1.7) 1.1 (0.7 to 1.7) 0.0 (<<0 to 0.3) 
    Silent 38 29 1.1 (0.6 to 2.0) 1.0 (referent)  
0.01–0.99 25 59 1.3 (0.7 to 2.2) 1.1 (0.5 to 2.4)  
≥1.0 15 46 1.0 (0.5 to 2.0) 0.9 (0.4 to 2.2) 0.2 (−0.3 to 1.3) 
    Missense 26 30 0.6 (0.3 to 1.1) 1.0 (referent)  
0.01–0.99 21 45 1.7 (0.9 to 3.1) 2.7 (1.2 to 6.4)  
≥1.0 21 38 2.0 (1.1 to 3.9) 3.3 (1.4 to 8.0) 1.3 (0.1 to 3.9) 
    Splicing — 1.0 (referent)  
0.01–0.99 1.5 (0.4 to 6.5) —  
≥1.0 0.4 (0.0 to 3.6) — — 
    Truncation 1.6 (0.3 to 8.6) 1.0 (referent)  
0.01–0.99 2.9 (0.5 to 16.3) 1.7 (0.2 to 19.3) 2.5 (−0.4 to 36.3) 
≥1.0 — —  
    Common‖ 154 126 0.8 (0.6 to 1.2) 1.0 (referent)  
0.01–0.99 84 308 0.8 (0.5 to 1.1) 0.9 (0.6 to 1.3)  
≥1.0 70 221 0.9 (0.6 to 1.4) 1.1 (0.7 to 1.6) 0.0 (−0.2 to 0.3) 
Missense variants classified by SIFT ¶ 
    Wild type 112 72 1.0 (referent) 1.0 (referent)  
0.01–0.99 57 177 1.1 (0.7 to 1.6) 1.1 (0.7 to 1.6)  
≥1.0 54 169 1.1 (0.7 to 1.7) 1.1 (0.7 to 1.7) 0.0 (<<0 to 0.3) 
    Tolerated 12 16 0.7 (0.3 to 1.7) 1.0 (referent)  
0.01–0.99 27 1.1 (0.4 to 2.7) 1.6 (0.5 to 5.2)  
≥1.0 10 23 1.3 (0.6 to 3.2) 1.8 (0.6 to 5.8) 0.8 (−0.1 to 3.6) 
    Deleterious 14 14 0.6 (0.2 to 1.3) 1.0 (referent)  
0.01–0.99 12 17 2.8 (1.2 to 6.5) 5.3 (1.6 to 17.3)  
≥1.0 11 15 3.3 (1.4 to 8.0) 5.8 (1.8 to 19.0) 2.6 (0.0 to 10.6) 
*

Case and matched control subjects were women with estimates of the reconstructed location-specific dose received to contralateral breast during radiotherapy. — = no estimate; CI = confidence interval; ERR = excess relative risk; RR = rate ratio.

Defined as reconstructed quadrant dose received to contralateral breast during radiotherapy; 0.01–0.99 Gy category: range = 0.02–0.99, mean = 0.6 (SD = 0.2); ≥1.0 Gy category: range = 1.1–6.2, mean = 1.7 (SD = 0.6).

The baseline comparison group is wild type and unexposed. Conditional logistic regression models are fully adjusted for factors found to be statistically significantly associated with contralateral breast cancer in the univariate models and those known to be associated with breast cancer, including the following: exact age at diagnosis of first primary breast cancer, age at menarche (<13 or ≥13 years), menopausal status (premenopausal and age at menopause <45 or ≥45 years), number of full-term pregnancies (0, 1, 2, 3, or ≥4), family history of breast cancer among first-degree relative (yes or no), lobular histology (yes or no) and stage (local or regional) of the first primary, and treatment history (chemotherapy or hormonal therapy and radiation where indicated).

§

The comparison group in this model is unexposed carriers of the same type of mutation. Multivariable conditional logistic regression was used to adjust for the same factors as above.

ATM variants carried by 1% or more of the WECARE Study participants.

Variants with normalized probabilities less than .05 are predicted to be deleterious, whereas those with probabilities equal to or greater than .05 are predicted to be tolerated. Results for missense variants are adjusted for other variants. One control subject carried only one variant that lacked a SIFT classification.

Results

We have previously reported individually on the association between ATM variants ( 49 ) and radiotherapy ( 9 ) and the risk of contralateral breast cancer in the WECARE Study population. In this study, we examined the interaction between these risk factors. Screening of the ATM gene in all 2105 women in the WECARE Study identified 240 unique variants, of which 147 were observed only a single time, 50 were observed two to five times, and 43 were observed six or more times. Our previous study found that when the 15 variants carried by more than 1% of WECARE Study participants were considered individually, four of these common variants were associated with a statistically significantly reduced risk of developing contralateral breast cancer ( 49 ). The remaining variants could only be considered for statistical analysis after grouping.

We first examined associations between ATM variants and the risk of asynchronous contralateral breast cancer. Rate ratios were calculated for ATM variants that were broadly classified by variant type and for missense variants that were classified using SIFT ( 53 ), for all case and control subjects, and for the case and control subjects who had location-specific radiation dose estimates ( 9 ) ( Table 1 ). Overall, we observed no statistically significant elevated risk of contralateral breast cancer among women who carried any of the different types of ATM mutations compared with those who were wild type for ATM . Furthermore, there were no substantial differences between the rate ratios for women with and without location-specific dose estimates. For comparison with Renwick et al. ( 55 ), we also examined truncating mutations that included both premature termination codons and frameshift mutations that are known to be A-T causing and found a non-statistically significant elevated risk of contralateral breast cancer (RR = 2.0, 95% CI = 0.7 to 5.9).

Among women exposed to radiation, the mean dose received to the contralateral breast was 1.2 Gy (SD = 0.7). For women who carried rare ATM missense variants, those with radiation exposure levels of 1.0 Gy or higher had an increased risk of contralateral breast cancer compared with women who were wild type for ATM and unexposed to radiation (RR = 2.0, 95% CI = 1.1 to 3.9); the dose–response trend was statistically significant (ERR/Gy = 1.3, 95% CI = 0.1 to 3.9; Ptrend = .017) ( Table 2 ). Among women who carried an ATM missense variant that was predicted to be deleterious, those who were exposed to radiation had a statistically significantly higher risk of contralateral breast cancer compared with unexposed women who were wild type for ATM (0.01–0.99 Gy: RR = 2.8, 95% CI = 1.2 to 6.5; ≥1.0 Gy: RR = 3.3, 95% CI = 1.4 to 8.0) or unexposed women who carried predicted deleterious missense variants (0.01–0.99 Gy: RR = 5.3, 95% CI = 1.6 to 17.3; ≥1.0 Gy: RR = 5.8, 95% CI = 1.8 to 19.0; ERR/Gy = 2.6, 95% CI = 0.0 to 10.6; Ptrend = .044). These trends were not evident among women who carried other types of ATM variants.

Analyses assessing variation by age at and time since breast cancer diagnosis using all case and control subjects, with radiotherapy stratified as ever or never, were conducted to provide more stable estimates ( Table 3 ). Among women who carried ATM missense variants that were predicted to be deleterious, the rate ratio for ever vs never use of radiotherapy appeared to be greater for women younger than 45 years at diagnosis (RR = 10.4, 95% CI = 2.3 to 47.2) than for older women (RR = 2.4, 95% CI = 0.6 to 9.5). However, the 95% confidence intervals on the subgroup-specific linear slope estimates were wide, and the excess relative risks per radiation dose did not differ statistically significantly from each other (ERR/Gy for women younger than 45 years = 4.9 [95% CI = −0.1 to 24.9] and for women 45 years or older = 2.2 [95% CI = −0.0 to 9.6]). Furthermore, the large rate ratio in the younger age group relative to unexposed carriers of deleterious variants must be interpreted with caution because the rate ratio for women who carried such variants compared with radiation-unexposed women who were wild type for ATM was 0.2 (95% CI = 0.1 to 0.7). Similarly, among women who carried predicted deleterious missense variants, we observed a steeper radiation dose slope for those with a longer latency (≥5 years) compared with those with a shorter latency (ERR/Gy: 4.3 [95% CI = 0.1 to 20.7] vs 1.9 [95% CI = −0.1 to 9.1], respectively), but this difference was also not statistically significant. Furthermore, the rate ratios for the ever vs never radiotherapy comparisons did not differ appreciably by time since first breast cancer diagnosis. The data were too sparse to obtain statistically meaningful results for analyses of the joint effects of ATM variant carrier status, age at diagnosis, and latency.

Table 3

Joint effects of ATM gene mutation carrier status, radiation exposure, and risk of developing asynchronous contralateral breast cancer according to age at and time since breast cancer diagnosis *

ATM variants and age and latency factors
 
Radiation exposure Case subjects (n = 708) Control subjects (n = 1397) RR (95% CI) RR (95% CI)  ERR/Gy † (95% CI)  
Missense variants classified using SIFT ‡ 
    Wild type Never 142 82 1.0 (referent) 1.0 (referent)  
Ever 129 398 0.9 (0.6 to 1.3) 0.9 (0.6 to 1.3) 0.0 (<<0 to 0.3) 
    Tolerated Never 17 17 0.8 (0.5 to 1.0) 1.0 (referent)  
Ever 19 55 0.9 (0.5 to 1.8) 1.2 (0.5 to 3.1) 0.8 (−0.1 to 3.6) 
    Deleterious Never 14 15 0.4 (0.2 to 1.0) 1.0 (referent)  
Ever 25 41 2.1 (1.1 to 3.8) 5.0 (1.8 to 13.3) 2.6 (0.0 to 10.6) 
Age at diagnosis and SIFT classification, y 
    23–44 Wild type Never 62 36 1.0 (referent) 1.0 (referent)  
Ever 60 181 0.9 (0.6 to 1.4) 0.9 (0.6 to 1.4) 0.0 (<<0 to 0.5) 
Tolerated Never 2.0 (0.4 to 9.3) 1.0 (referent)  
Ever 25 0.8 (0.3 to 2.1) 0.4 (0.1 to 2.4) 0.7 (<<0 to 4.6) 
Deleterious Never 10 0.2 (0.1 to 0.7) 1.0 (referent)  
Ever 10 16 2.0 (0.8 to 5.0) 10.4 (2.3 to 47.2) 4.9 (−0.1 to 24.9) 
    ≥45 Wild type Never 80 46 1.0 (referent) 1.0 (referent)  
Ever 69 217 0.9 (0.6 to 1.3) 0.9 (0.6 to 1.3) −0.0 (<<0 to 0.4) 
Tolerated Never 10 14 0.5 (0.2 to 1.3) 1.0 (referent)  
Ever 10 30 1.0 (0.4 to 2.3) 1.9 (0.6 to 6.2) 0.8 (−0.1 to 4.5) 
Deleterious Never 0.9 (0.2 to 3.0) 1.0 (referent)  
Ever 15 25 2.0 (1.0 to 4.4) 2.4 (0.6 to 9.5) 2.2 (−0.0 to 9.6) 
Time since diagnosis and SIFT classification, y 
    1–4 Wild type Never 81 48 1.0 (referent) 1.0 (referent)  
Ever 77 238 0.8 (0.6 to 1.2) 0.8 (0.6 to 1.2) −0.1 (<<0 to 0.3) 
Tolerated Never 11 11 0.8 (0.3 to 2.1) 1.0 (referent)  
Ever 14 33 1.1 (0.5 to 2.4) 1.4 (0.5 to 4.4) 1.2 (−0.2 to 5.4) 
Deleterious Never 0.3 (0.1 to 1.1) 1.0 (referent)  
Ever 14 23 1.8 (0.8 to 3.9) 5.2 (1.4 to 18.8) 1.9 (−0.1 to 9.1) 
    ≥5 Wild type Never 61 34 1.0 (referent) 1.0 (referent)  
Ever 52 160 1.0 (0.7 to 1.7) 1.0 (0.7 to 1.7) 0.2 (<<0 to 0.8) 
Tolerated Never 0.8 (0.2 to 2.7) 1.0 (referent)  
Ever 22 0.5 (0.2 to 1.8) 0.7 (0.1 to 4.0) 0.3 (−0.2 to 3.1) 
Deleterious Never 0.6 (0.2 to 2.2) 1.0 (referent)  
Ever 11 18 2.5 (1.0 to 6.1) 4.3 (0.9 to 19.9) 4.3 (0.1 to 20.7) 
ATM variants and age and latency factors
 
Radiation exposure Case subjects (n = 708) Control subjects (n = 1397) RR (95% CI) RR (95% CI)  ERR/Gy † (95% CI)  
Missense variants classified using SIFT ‡ 
    Wild type Never 142 82 1.0 (referent) 1.0 (referent)  
Ever 129 398 0.9 (0.6 to 1.3) 0.9 (0.6 to 1.3) 0.0 (<<0 to 0.3) 
    Tolerated Never 17 17 0.8 (0.5 to 1.0) 1.0 (referent)  
Ever 19 55 0.9 (0.5 to 1.8) 1.2 (0.5 to 3.1) 0.8 (−0.1 to 3.6) 
    Deleterious Never 14 15 0.4 (0.2 to 1.0) 1.0 (referent)  
Ever 25 41 2.1 (1.1 to 3.8) 5.0 (1.8 to 13.3) 2.6 (0.0 to 10.6) 
Age at diagnosis and SIFT classification, y 
    23–44 Wild type Never 62 36 1.0 (referent) 1.0 (referent)  
Ever 60 181 0.9 (0.6 to 1.4) 0.9 (0.6 to 1.4) 0.0 (<<0 to 0.5) 
Tolerated Never 2.0 (0.4 to 9.3) 1.0 (referent)  
Ever 25 0.8 (0.3 to 2.1) 0.4 (0.1 to 2.4) 0.7 (<<0 to 4.6) 
Deleterious Never 10 0.2 (0.1 to 0.7) 1.0 (referent)  
Ever 10 16 2.0 (0.8 to 5.0) 10.4 (2.3 to 47.2) 4.9 (−0.1 to 24.9) 
    ≥45 Wild type Never 80 46 1.0 (referent) 1.0 (referent)  
Ever 69 217 0.9 (0.6 to 1.3) 0.9 (0.6 to 1.3) −0.0 (<<0 to 0.4) 
Tolerated Never 10 14 0.5 (0.2 to 1.3) 1.0 (referent)  
Ever 10 30 1.0 (0.4 to 2.3) 1.9 (0.6 to 6.2) 0.8 (−0.1 to 4.5) 
Deleterious Never 0.9 (0.2 to 3.0) 1.0 (referent)  
Ever 15 25 2.0 (1.0 to 4.4) 2.4 (0.6 to 9.5) 2.2 (−0.0 to 9.6) 
Time since diagnosis and SIFT classification, y 
    1–4 Wild type Never 81 48 1.0 (referent) 1.0 (referent)  
Ever 77 238 0.8 (0.6 to 1.2) 0.8 (0.6 to 1.2) −0.1 (<<0 to 0.3) 
Tolerated Never 11 11 0.8 (0.3 to 2.1) 1.0 (referent)  
Ever 14 33 1.1 (0.5 to 2.4) 1.4 (0.5 to 4.4) 1.2 (−0.2 to 5.4) 
Deleterious Never 0.3 (0.1 to 1.1) 1.0 (referent)  
Ever 14 23 1.8 (0.8 to 3.9) 5.2 (1.4 to 18.8) 1.9 (−0.1 to 9.1) 
    ≥5 Wild type Never 61 34 1.0 (referent) 1.0 (referent)  
Ever 52 160 1.0 (0.7 to 1.7) 1.0 (0.7 to 1.7) 0.2 (<<0 to 0.8) 
Tolerated Never 0.8 (0.2 to 2.7) 1.0 (referent)  
Ever 22 0.5 (0.2 to 1.8) 0.7 (0.1 to 4.0) 0.3 (−0.2 to 3.1) 
Deleterious Never 0.6 (0.2 to 2.2) 1.0 (referent)  
Ever 11 18 2.5 (1.0 to 6.1) 4.3 (0.9 to 19.9) 4.3 (0.1 to 20.7) 
*

The multivariable models were adjusted for factors found to be statistically significantly associated with contralateral breast cancer in the univariate models and those known to be associated with breast cancer, including the following: exact age at diagnosis of first primary breast cancer, age at menarche (<13 or ≥13 years), menopausal status (premenopausal and age at menopause <45 or ≥45 years), number of full-term pregnancies (0, 1, 2, 3, or ≥4), family history of breast cancer among first-degree relative (yes or no), lobular histology (yes or no) and stage (local or regional) of the first primary, and treatment history (chemotherapy or hormonal therapy and radiation where indicated). CI = confidence interval; ERR = excess relative risk; RR = rate ratio.

Dose is based on 606 triplets for whom there were dose estimates made.

Variants with normalized probabilities less than .05 are predicted to be deleterious, whereas those with probabilities equal to or greater than .05 are predicted to be tolerated. Results for missense variants are adjusted for other variants. One control subject carried only one variant that lacked a SIFT classification.

Discussion

Our findings suggest that ATM genetic variation and radiation exposure have joint etiologic roles in contralateral breast cancer in a small fraction of women who carry specific types of ATM gene variants. A current major focus of many epidemiological studies is to characterize the interactions between genetic and environmental factors ( 56 ). Although many environmental factors are not easily defined or quantified, radiation exposures in radiotherapy and in certain other settings are an exception. One advantage of our nested population-based case–control study is that it was specifically designed to maximize the statistical power to address the hypothesis that women who carry ATM gene mutations and who were exposed to radiation are at increased risk of developing radiation-induced breast cancer. In general, the low prevalence of relatively deleterious genetic abnormalities and the low levels of radiation present in most environments limit the informativeness of studies of the interaction of ATM gene and radiation exposure conducted in the general population ( 12 , 43 ). By restricting our analysis to a population of genetically high-risk women with bilateral breast cancer who were exposed to substantial and quantifiable levels of ionizing radiation, coupled with a full characterization of the ATM gene in the study population, we enhanced our ability to detect possible gene–radiation interactions. This study builds on our previously reported work demonstrating that the risk of developing contralateral breast cancer was positively associated with the dose of radiation received and inversely associated with the woman's age at radiation exposure ( 9 ) and that ATM gene carrier status alone was not associated with contralateral breast cancer ( 49 ). However, in this study, we found that among carriers of ATM missense mutations who were treated with radiotherapy, the rate ratio of radiation-induced contralateral breast cancer was twofold compared with women who were not treated with radiation and who were wild type for ATM . This effect was dose dependent, and the risk of contralateral breast cancer was greater when ATM missense variants predicted to be deleterious were considered; the effect was weakly time and age dependent.

Our finding of a possible interaction of ATM variation and radiation exposure in breast cancer etiology is plausible when considered in light of both the existing literature on the role of ATM in risk of breast cancer and our previous examination of the main effects of ATM variants in the WECARE Study population. For example, there is evidence from multiple epidemiological studies of A-T families that heterozygosity for A-T–causing mutations in ATM is associated with an increased risk of breast cancer ( 29–36 ). These mutations, which predominate among patients with A-T (A-T Mutation Database, www.LOVD.nl/ATM ), are primarily those predicted to prematurely truncate the ATM protein. Although A-T–causing mutations are relatively rare in the general population, and even in high-risk breast cancer populations, they are associated with the risk of breast cancer in families ascertained for multiple breast cancer cases ( 55 ). It seems unlikely that carriers of these rare mutations would coincidentally be enriched for radiation exposure. Thus, this risk appears to be a main effect of the presence of these inactivating ATM mutations.

ATM is a master regulator of cellular pathways that mediate responses to the most deleterious form of DNA damage induced by ionizing radiation, DNA double-strand breaks. Variants in the ATM gene that might interact with ionizing radiation need not be restricted to the rare protein-truncating mutations that cause A-T, which is precisely what we observed in this study. The risk of radiation-associated second primary breast cancers was most strongly associated with rare missense variants in ATM that are predicted to have deleterious effects on protein structure that do not involve protein truncation. How might these two different mechanisms whereby ATM affects risk of breast cancer be reconciled? We propose that in carriers of truncating mutations, ATM acts as a classic tumor suppressor gene and that loss of heterozygosity or epigenetic silencing of the normal allele occurs in tumor tissue. The loss or silencing of the normal allele could be induced by exposure to a mutagen like ionizing radiation but would not be dependent on such exposure. For carriers of the rare missense variants reported here, we propose that these variant likely act by dominant interference. Thus, the presence of rare missense variants effectively reduces the level of ATM activity below the 50% level expected in carriers of truncating mutations, which results in susceptibility to radiation-induced tumorigenesis. A key prediction of this model is that nontumor tissue, such as that in the contralateral breast, would have increased radiosensitivity in carriers of missense variants but not in carriers of truncating variants. Therefore, exposure to radiation scatter doses greater than 1 Gy may result in greater than normal amounts of DNA damage in cells of the contralateral breast, particularly in carriers of missense variants. If these cells survive, the radiation-induced DNA damage may increase the chance that a tumor will subsequently develop in the contralateral breast.

Our study included 2105 women with breast cancer who were screened for variants in all 62 coding exons in the ATM gene. However, our ability to detect associations between individual variants and the risk of contralateral breast cancer was limited because many of the unique non-silent ATM variants occurred infrequently. Therefore, we used different approaches for grouping the variants in biologically meaningful ways. First, we examined variants that were broadly classified according to the effect of the DNA change on the amino acid sequence and found no indication of an interaction with radiation exposure, including among women who carried common variants that we had previously reported were associated with a decreased risk of contralateral breast cancer ( 49 ). Second, we also examined truncating mutations that included both premature termination codons and frameshift mutations that are known to be A-T causing. In our series, however, less than 1% of the variants detected were truncating mutations, which limited our ability to examine their interaction with radiation. Our overall findings among women with a truncating or A-T–causing mutation (RR = 2.0, 95% CI = 0.7 to 5.9) are not inconsistent with the findings of Renwick et al. ( 55 ), who screened 443 breast cancer case subjects from multiple-case families and 521 control subjects for ATM mutations. They found a statistically significant association between ATM mutations that cause A-T and breast cancer (RR = 2.4, 95% CI = 1.5 to 3.8). However, they did not examine the effect of radiation, and the number of subjects with bilateral cancers was small, making a direct comparison of the results difficult.

A major strength of our study is that we had accurate estimates of radiation doses to tumor sites and simultaneously conducted a comprehensive characterization of variants in the ATM gene to highlight a gene by environment interaction. Nevertheless, we detected few carriers of missense variants that were predicted to be deleterious by SIFT, which limited the precision of our estimates. We observed large rate ratios associated with ATM missense variants and radiation exposures greater than 1 Gy; however, the estimates of radiation-induced risk for all missense variants (ERR/Gy = 1.3, 95% CI = 0.1 to 3.9) and for the predicted deleterious missense variants (ERR/Gy = 2.6, 95% CI = 0.0 to 10.6) are statistically compatible with those reported in radiation studies of other populations, such as atomic bomb survivors (ERR/Gy = 1.49, 95% CI = 1.17 to 1.85) and a pooled analysis of eight cohort studies (ERR/Gy = 0.9, 95% CI = 0.7 to 1.0) ( 57 ). Furthermore, although young age at breast cancer diagnosis and long latency are hallmarks of radiation-induced breast cancer ( 57 ), the strong radiation effect we observed in women younger than 45 years at diagnosis who carried deleterious mutations had great uncertainty because of these small numbers. Therefore, although the increased risk of radiation-related contralateral breast cancer associated with specific ATM mutations is not an important factor in the selection of treatment for breast cancer for most women, it might warrant consideration in the rare instances where a woman has a family history of A-T. Further epidemiological study of the role of ATM missense mutations among women diagnosed with breast cancer before the age of 45 years who were treated with radiation and developed contralateral breast cancer after a long latency period is necessary to better estimate these risks.

Funding

National Cancer Institute (R01 CA097397 and U01 CA083178 to J.L.B.].

References

1.
Bernstein
JL
Thompson
WD
Risch
N
Holford
TR
Risk factors predicting the incidence of second primary breast cancer among women diagnosed with a first primary breast cancer
Am J Epidemiol
 , 
1992
, vol. 
136
 
8
(pg. 
925
-
936
)
2.
Bernstein
JL
Thompson
WD
Risch
N
Holford
TR
The genetic epidemiology of second primary breast cancer
Am J Epidemiol
 , 
1992
, vol. 
136
 
8
(pg. 
937
-
948
)
3.
Hoover
RN
Curtis
RE
Freedman
DM
Ron
E
, et al.  . 
New malignancies following breast cancer
New Malignancies Among Cancer Survivors: SEER CANCER REGISTRIES, 1973–2000
  
National Cancer Institute, NIH Publ No. 05-5302. Bethesda, M . 2006:181–205
4.
Hemminki
K
Ji
J
Forsti
A
Risks for familial and contralateral breast cancer interact multiplicatively and cause a high risk
Cancer Res.
 , 
2007
, vol. 
67
 
3
(pg. 
868
-
870
)
5.
Hemminki
K
Zhang
H
Sundquist
J
Lorenzo Bermejo
J
Modification of risk for subsequent cancer after female breast cancer by a family history of breast cancer
Breast Cancer Res Treat
 , 
2008
, vol. 
111
 
1
(pg. 
165
-
169
)
6.
Innos
K
Horn-Ross
PL
Risk of second primary breast cancers among women with ductal carcinoma in situ of the breast
Breast Cancer Res Treat
 , 
2008
, vol. 
111
 
3
(pg. 
531
-
540
)
7.
Largent
JA
Capanu
M
Bernstein
L
, et al.  . 
Reproductive history and risk of second primary breast cancer: the WECARE study
Cancer Epidemiol Biomarkers Prev
 , 
2007
, vol. 
16
 
5
(pg. 
906
-
911
)
8.
Shih
HA
Nathanson
KL
Seal
S
, et al.  . 
BRCA1 and BRCA2 mutations in breast cancer families with multiple primary cancers
Clin Cancer Res.
 , 
2000
, vol. 
6
 
11
(pg. 
4259
-
4264
)
9.
Stovall
M
Smith
SA
Langholz
BM
, et al.  . 
Dose to the contralateral breast from radiotherapy and risk of second primary breast cancer in the WECARE study
Int J Radiat Oncol Biol Phys.
 , 
2008
, vol. 
72
 
4
(pg. 
1021
-
1030
)
10.
Yadav
BS
Sharma
SC
Patel
FD
Ghoshal
S
Kapoor
RK
Second primary in the contralateral breast after treatment of breast cancer
Radiother Oncol
 , 
2008
, vol. 
86
 
2
(pg. 
171
-
176
)
11.
Antoniou
AC
Cunningham
AP
Peto
J
, et al.  . 
The BOADICEA model of genetic susceptibility to breast and ovarian cancers: updates and extensions
Br J Cancer
 , 
2008
, vol. 
98
 
8
(pg. 
1457
-
1466
)
12.
Begg
CB
Haile
RW
Borg
A
, et al.  . 
Variation of breast cancer risk among BRCA1/2 carriers
JAMA
 , 
2008
, vol. 
299
 
2
(pg. 
194
-
201
)
13.
Fletcher
O
Johnson
N
Dos Santos Silva
I
, et al.  . 
Family history, genetic testing, and clinical risk prediction: pooled analysis of CHEK2 1100delC in 1,828 bilateral breast cancers and 7,030 controls
Cancer Epidemiol Biomarkers Prev
 , 
2009
, vol. 
18
 
1
(pg. 
230
-
234
)
14.
Malone
KE
Begg
CB
Haile
RW
, et al.  . 
A population-based study of the relative and absolute risks of second primary contralateral breast cancer associated with carrying a mutation in BRCA1 or BRCA2
J. Clin. Oncol
 , 
2009
15.
Mellemkjaer
L
Dahl
C
Olsen
JH
, et al.  . 
Risk for contralateral breast cancer among carriers of the CHEK2*1100delC mutation in the WECARE Study
Br J Cancer
 , 
2008
, vol. 
98
 
4
(pg. 
728
-
733
)
16.
Metcalfe
K
Lynch
HT
Ghadirian
P
, et al.  . 
Contralateral breast cancer in BRCA1 and BRCA2 mutation carriers
J Clin Oncol
 , 
2004
, vol. 
22
 
12
(pg. 
2328
-
2335
)
17.
Rogozinska-Szczepka
J
Utracka-Hutka
B
Grzybowska
E
, et al.  . 
BRCA1 and BRCA2 mutations as prognostic factors in bilateral breast cancer patients
Ann Oncol
 , 
2004
, vol. 
15
 
9
(pg. 
1373
-
1376
)
18.
Bertelsen
L
Bernstein
L
Olsen
JH
, et al.  . 
Effect of systemic adjuvant treatment on risk for contralateral breast cancer in the Women's Environment, Cancer and Radiation Epidemiology Study
J Natl Cancer Inst
 , 
2008
, vol. 
100
 
1
(pg. 
32
-
40
)
19.
Hooning
MJ
Aleman
BM
Hauptmann
M
, et al.  . 
Roles of radiotherapy and chemotherapy in the development of contralateral breast cancer
J Clin Oncol
 , 
2008
, vol. 
26
 
34
(pg. 
5561
-
5568
)
20.
Boice
JD
Jr
Harvey
EB
Blettner
M
Stovall
M
Flannery
JT
Cancer in the contralateral breast after radiotherapy for breast cancer
N Engl J Med
 , 
1992
, vol. 
326
 
12
(pg. 
781
-
785
)
21.
Gao
X
Fisher
SG
Emami
B
Risk of second primary cancer in the contralateral breast in women treated for early-stage breast cancer: a population-based study
Int J Radiat Oncol Biol Phys.
 , 
2003
, vol. 
56
 
4
(pg. 
1038
-
1045
)
22.
Savitsky
K
Bar-Shira
A
Gilad
S
, et al.  . 
A single ataxia telangiectasia gene with a product similar to PI-3 kinase
Science
 , 
1995
, vol. 
268
 
5218
(pg. 
1749
-
1753
)
23.
Savitsky
K
Sfez
S
Tagle
DA
, et al.  . 
The complete sequence of the coding region of the ATM gene reveals similarity to cell cycle regulators in different species
Hum Mol Genet.
 , 
1995
, vol. 
4
 
11
(pg. 
2025
-
2032
)
24.
Morrell
D
Cromartie
E
Swift
M
Mortality and cancer incidence in 263 patients with ataxia-telangiectasia
J Natl Cancer Inst
 , 
1986
, vol. 
77
 
1
(pg. 
89
-
92
)
25.
Bakkenist
CJ
Kastan
MB
DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation
Nature
 , 
2003
, vol. 
421
 
6922
(pg. 
499
-
506
)
26.
Taylor
AM
Harnden
DG
Arlett
CF
, et al.  . 
Ataxia telangiectasia: a human mutation with abnormal radiation sensitivity
Nature
 , 
1975
, vol. 
258
 
5534
(pg. 
427
-
429
)
27.
Cortez
D
Wang
Y
Qin
J
Elledge
SJ
Requirement of ATM-dependent phosphorylation of brca1 in the DNA damage response to double-strand breaks
Science
 , 
1999
, vol. 
286
 
5442
(pg. 
1162
-
1166
)
28.
Matsuoka
S
Ballif
BA
Smogorzewska
A
, et al.  . 
ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage
Science
 , 
2007
, vol. 
316
 
5828
(pg. 
1160
-
1166
)
29.
Cavaciuti
E
Lauge
A
Janin
N
, et al.  . 
Cancer risk according to type and location of ATM mutation in ataxia-telangiectasia families
Genes Chromosomes Cancer
 , 
2005
, vol. 
42
 
1
(pg. 
1
-
9
)
30.
Janin
N
Andrieu
N
Ossian
K
, et al.  . 
Breast cancer risk in ataxia telangiectasia (AT) heterozygotes: haplotype study in French AT families
Br J Cancer
 , 
1999
, vol. 
80
 
7
(pg. 
1042
-
1045
)
31.
Olsen
JH
Hahnemann
JM
Borresen-Dale
AL
, et al.  . 
Breast and other cancers in 1445 blood relatives of 75 Nordic patients with ataxia telangiectasia
Br J Cancer
 , 
2005
, vol. 
93
 
2
(pg. 
260
-
265
)
32.
Pippard
EC
Hall
AJ
Barker
DJ
Bridges
BA
Cancer in homozygotes and heterozygotes of ataxia-telangiectasia and xeroderma pigmentosum in Britain
Cancer Res.
 , 
1988
, vol. 
48
 
10
(pg. 
2929
-
2932
)
33.
Stankovic
T
Kidd
AM
Sutcliffe
A
, et al.  . 
ATM mutations and phenotypes in ataxia-telangiectasia families in the British Isles: expression of mutant ATM and the risk of leukemia, lymphoma, and breast cancer
Am J Hum Genet.
 , 
1998
, vol. 
62
 
2
(pg. 
334
-
345
)
34.
Swift
M
Morrell
D
Massey
RB
Chase
CL
Incidence of cancer in 161 families affected by ataxia-telangiectasia
N Engl J Med
 , 
1991
, vol. 
325
 
26
(pg. 
1831
-
1836
)
35.
Swift
M
Reitnauer
PJ
Morrell
D
Chase
CL
Breast and other cancers in families with ataxia-telangiectasia
N Engl J Med
 , 
1987
, vol. 
316
 
21
(pg. 
1289
-
1294
)
36.
Swift
M
Sholman
L
Perry
M
Chase
C
Malignant neoplasms in the families of patients with ataxia-telangiectasia
Cancer Res.
 , 
1976
, vol. 
36
 
1
(pg. 
209
-
215
)
37.
Baynes
C
Healey
CS
Pooley
KA
, et al.  . 
Common variants in the ATM, BRCA1, BRCA2, CHEK2 and TP53 cancer susceptibility genes are unlikely to increase breast cancer risk
Breast Cancer Res.
 , 
2007
, vol. 
9
 
2
pg. 
R27
 
38.
Bernstein
JL
Teraoka
S
Southey
MC
, et al.  . 
Population-based estimates of breast cancer risks associated with ATM gene variants c.7271T>G and c.1066-6T>G (IVS10-6T>G) from the Breast Cancer Family Registry
Hum Mutat
 , 
2006
, vol. 
27
 
11
(pg. 
1122
-
1128
)
39.
Broeks
A
Urbanus
JH
Floore
AN
, et al.  . 
ATM-heterozygous germline mutations contribute to breast cancer-susceptibility
Am J Hum Genet.
 , 
2000
, vol. 
66
 
2
(pg. 
494
-
500
)
40.
Chenevix-Trench
G
Spurdle
AB
Gatei
M
, et al.  . 
Dominant negative ATM mutations in breast cancer families
J Natl Cancer Inst
 , 
2002
, vol. 
94
 
3
(pg. 
205
-
215
)
41.
FitzGerald
MG
Bean
JM
Hegde
SR
, et al.  . 
Heterozygous ATM mutations do not contribute to early onset of breast cancer
Nat Genet.
 , 
1997
, vol. 
15
 
3
(pg. 
307
-
310
)
42.
Szabo
CI
Schutte
M
Broeks
A
, et al.  . 
Are ATM mutations 7271T–>G and IVS10-6T–>G really high-risk breast cancer-susceptibility alleles?
Cancer Res.
 , 
2004
, vol. 
64
 
3
(pg. 
840
-
843
)
43.
Thompson
D
Duedal
S
Kirner
J
, et al.  . 
Cancer risks and mortality in heterozygous ATM mutation carriers
J Natl Cancer Inst
 , 
2005
, vol. 
97
 
11
(pg. 
813
-
822
)
44.
Vorechovsky
I
Rasio
D
Luo
L
, et al.  . 
The ATM gene and susceptibility to breast cancer: analysis of 38 breast tumors reveals no evidence for mutation
Cancer Res.
 , 
1996
, vol. 
56
 
12
(pg. 
2726
-
2732
)
45.
Broeks
A
Braaf
LM
Huseinovic
A
, et al.  . 
Identification of women with an increased risk of developing radiation-induced breast cancer: a case only study
Breast Cancer Res.
 , 
2007
, vol. 
9
 
2
pg. 
R26
 
46.
Su
Y
Swift
M
Outcomes of adjuvant radiation therapy for breast cancer in women with ataxia-telangiectasia mutations
JAMA
 , 
2001
, vol. 
286
 
18
(pg. 
2233
-
2234
)
47.
Bernstein
JL
Langholz
B
Haile
RW
, et al.  . 
Study design: evaluating gene-environment interactions in the etiology of breast cancer—the WECARE study
Breast Cancer Res.
 , 
2004
, vol. 
6
 
3
(pg. 
R199
-
R214
)
48.
Bernstein
JL
Teraoka
S
Haile
RW
, et al.  . 
Designing and implementing quality control for multi-center screening of mutations in the ATM gene among women with breast cancer
Hum Mutat
 , 
2003
, vol. 
21
 
5
(pg. 
542
-
550
)
49.
Concannon
P
Haile
RW
Borresen-Dale
AL
, et al.  . 
Variants in the ATM gene associated with a reduced risk of contralateral breast cancer
Cancer Res.
 , 
2008
, vol. 
68
 
16
(pg. 
6486
-
6491
)
50.
Langholz
B
Borgan
Ø
Counter-matching: a stratified nested case-control sampling method
Biometrika
 , 
1995
, vol. 
82
 
1
(pg. 
69
-
79
)
51.
Langholz
B
Thomas
DC
Stovall
M
, et al.  . 
Statistical methods for analysis of radiation effects with tumor and dose location-specific information with application to the WECARE study of asynchronous contralateral breast cancer
Biometrics
 , 
2009
, vol. 
65
 
2
(pg. 
599
-
608
)
52.
Huberman
M
Langholz
B
Application of the missing-indicator method in matched case–control studies with incomplete data
Am J Epidemiol
 , 
1999
, vol. 
150
 
12
(pg. 
1340
-
1345
)
53.
Ng
PC
Henikoff
S
Predicting deleterious amino acid substitutions
Genome Res.
 , 
2001
, vol. 
11
 
5
(pg. 
863
-
874
)
54.
Sunyaev
S
Ramensky
V
Koch
I
Lathe
W
III
Kondrashov
AS
Bork
P
Prediction of deleterious human alleles
Hum Mol Genet.
 , 
2001
, vol. 
10
 
6
(pg. 
591
-
597
)
55.
Renwick
A
Thompson
D
Seal
S
, et al.  . 
ATM mutations that cause ataxia-telangiectasia are breast cancer susceptibility alleles
Nat Genet.
 , 
2006
, vol. 
38
 
8
(pg. 
873
-
875
)
56.
Le Marchand
L
Wilkens
LR
Design considerations for genomic association studies: importance of gene-environment interactions
Cancer Epidemiol Biomarkers Prev
 , 
2008
, vol. 
17
 
2
(pg. 
263
-
267
)
57.
Preston
DL
Mattsson
A
Holmberg
E
Shore
R
Hildreth
NG
Boice
JD
Jr.
Radiation effects on breast cancer risk: a pooled analysis of eight cohorts
Radiat Res.
 , 
2002
, vol. 
158
 
2
(pg. 
220
-
235
)
J. L. Bernstein and R. W. Haile contributed equally to this work.
The ideas and opinions expressed in this article are those of the authors, and no endorsement by the study sponsor is intended or should be inferred.
WECARE Study Collaborative Group: Memorial Sloan Kettering Cancer Center (New York, NY): J. L. Bernstein, PhD (WECARE Study Principle Investigator), Colin B. Begg, PhD, M. Capanu, PhD, Irene Orlow, PhD, X. Liang, MD, A. S. Reiner, MPH, and Tracy M. Layne, MPH. City of Hope (Duarte, CA) (some work was performed at University of Southern California in Los Angeles, CA): L. Bernstein and L. Donnelly-Allen. Danish Cancer Society (Copenhagen, Denmark): J. H. Olsen, MD, DMSc, Michael Andersson, MD, DMSc, Lisbeth Bertelsen, MD, PhD, Per Guldberg, PhD, and Lene Mellemkjær, PhD. Fred Hutchinson Cancer Research Center (Seattle, WA): K. E. Malone, PhD, and Noemi Epstein. International Epidemiology Institute (Rockville, MD) and Vanderbilt University (Nashville, TN): J. D. Boice Jr, ScD. Lund University (Lund, Sweden): Åke Borg, PhD, Therese Törngren, MSc, and Lina Tellhed, BSc. Mount Sinai School of Medicine (New York, NY): B. S. Rosenstein, PhD, and David P. Atencio, PhD. National Cancer Institute (Bethesda, MD): Daniela Seminara, PhD, MPH. New York University (New York, NY): Roy E. Shore, PhD, DrPH. Norwegian Radium Hospital (Oslo, Norway): A.-L. Børresen-Dale, PhD, and Laila Jansen. University of California at Irvine (Irvine, CA): Hoda Anton-Culver, PhD, and Joan Largent, PhD, MPH. University of California at Los Angeles (Los Angeles, CA): R. A. Gatti, MD. University of Iowa (Iowa City, IA): C. F. Lynch, MD, PhD, and Jeanne DeWall, MA. University of Southern California (Los Angeles, CA): R. W. Haile, DrPH, B. Langholz, PhD, D. C. Thomas, PhD, S. Xue, MD, N. Zhou, MD, A. T. Diep, BS and E. Ter-Karapetova, BS. University of Southern Maine (Portland, ME): W. Douglas Thompson, PhD. University of Texas, M.D. Anderson Cancer Center (Houston, TX): M. Stovall, PhD, Thomas Buchholz, MD, and S. A. Smith, MPH. University of Virginia (Charlottesville, VA) (some was work performed at Benaroya Research Institute in Seattle, WA): P. Concannon, PhD, S. N. Teraoka, PhD, Eric R. Olson, PhD, and Nirasha Ramchurren, PhD.