HUMAN MUTATION 27(11), 1122^1128, 2006

RESEARCH ARTICLE

Population-Based Estimates of Breast Cancer Risks
Associated With

ATM

Gene Variants c.7271T4G

and c.1066–6T4G (IVS10–6T4G) from the Breast
Cancer Family Registry

J.L. Bernstein,

1

S. Teraoka,

2

M.C. Southey,

3,4

M.A. Jenkins,

3

I.L. Andrulis,

5

J.A. Knight,

5

E.M. John,

6

R. Lapinski,

7

A.L. Wolitzer,

1

A.S. Whittemore,

8

D. West,

6

D. Seminara,

9

E.R. Olson,

2

A.B. Spurdle,

10

G. Chenevix-Trench,

10

G.G. Giles,

3,11

J.L. Hopper,

3

and P. Concannon

2

1

Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York;

2

Benaroya Research Institute,

Seattle, Washington;

3

University of Melbourne, Melbourne, Australia;

4

International Agency for Research on Cancer (IARC), Lyon, France;

5

Samuel Lunenfeld Research Institute, Toronto, Ontario, Canada;

6

Northern California Cancer Center, Fremont, California;

7

Mt. Sinai School of

Medicine, New York, New York;

8

Stanford University School of Medicine, Stanford, California;

9

National Cancer Institute, Bethesda, Maryland;

10

Queensland Institute of Medical Research, Brisbane, Australia;

11

The Cancer Council Victoria, Melbourne, Australia

Communicated by Dvorah Abeliovich

The

ATM gene variants segregating in ataxia-telangiectasia families are associated with increased breast cancer

risk, but the contribution of specific variants has been difficult to estimate. Previous small studies suggested two
functional variants, c.7271T4G and c.1066–6T4G (IVS10–6T4G), are associated with increased risk.
Using population-based blood samples we found that 7 out of 3,743 breast cancer cases (0.2%) and 0 out
of 1,268 controls were heterozygous for the c.7271T4G allele (P 5 0.1). In cases, this allele was more
prevalent in women with an affected mother (odds ratio [OR] 5 5.5, 95% confidence interval [CI] 5 1.2–25.5;
P 5 0.04) and delayed child-bearing (OR 5 5.1; 95% CI 5 1.0–25.6; P 5 0.05). The estimated cumulative
breast cancer risk to age 70 years (penetrance) was 52% (95% CI 5 28–80%; hazard ratio [HR] 5 8.6; 95%
CI 5 3.9–18.9; P

o0.0001). In contrast, 13 of 3,757 breast cancer cases (0.3%) and 10 of 1,268 controls

(0.8%) were heterozygous for the c.1066–6T4G allele (OR 5 0.4; 95% CI 5 0.2–1.0; P 5 0.05), and the
penetrance was not increased (P 5 0.5). These findings suggest that although the more common c.1066–6T4G
variant is not associated with breast cancer, the rare

ATM c.7271T4G variant is associated with a substantially

elevated risk. Since c.7271T4G is only one of many rare ATM variants predicted to have deleterious
consequences on protein function, an effective means of identifying and grouping these variants is essential to
assess the contribution of

ATM variants to individual risk and to the incidence of breast cancer in the

population. Hum Mutat 27(11), 1122–1128, 2006.

Published 2006 Wiley-Liss, Inc.

y

KEY WORDS:

ATM gene variants; breast cancer; DNA damage; segregation analysis; penetrance

INTRODUCTION

Mutations in the ataxia-telangiectasia (A-T) mutated gene

(

ATM; MIM] 607585), are responsible for the rare autosomal

recessive chromosomal instability disorder A-T (MIM] 208900)
[Savitsky et al., 1995]. Positional cloning of this gene revealed that
it encodes a large serine-threonine kinase that plays a central
role in sensing and signaling the presence of DNA double-strand
breaks that may be caused by exposure to ionizing radiation
or other types of DNA damaging agents. Exposure to ionizing
radiation, even at very low doses, triggers the release of a
substantial fraction of cellular

ATM from inactive homodimers to

active monomers [Bakkenist and Kastan, 2003]. This activation of
ATM requires transphosphorylation between the two members
of an

ATM dimer. Upon activation, ATM phosphorylates a wide

array of downstream targets that regulate cell cycle checkpoints,
apoptosis, and DNA repair. These substrates include the products
of several genes that, when mutated, either increase susceptibility
to cancer or regulate proteins that do, including the

BRCA1,

CHEK2, and TP53 genes, where variants are known to predispose
to breast cancer (MIM] 114480) [for review see Bakkenist and
Kastan, 2004; Kurz and Lees-Miller, 2004].

The role the

ATM protein itself plays in breast cancer

susceptibility is of keen interest, especially as most studies of

DOI 10.1002/humu.20415
Published online 6 September 2006 in Wiley InterScience (www.
interscience.wiley.com).

y

This article is a US Government work, and, as such, is in the public

domain in the United States of America.

Received 16 March 2006; accepted revised manuscript 24

July 2006.

Grant sponsor: National Institutes of Health (NIH); Grant numbers:

U01CA83178, 5U01CA69467, 5U01CA69638, 5U01CA69417.

Correspondence to: Dr. Jonine Bernstein, Department of Epide-

miology and Biostatistics, Memorial Sloan-Kettering Cancer Center,
307 East 63rd Street, 3rd Floor, NewYork, NY 10021.
E-mail: [e-mail]

PUBLISHED 2006 WILEY-LISS, INC.

A-T families have demonstrated, on average, an excess risk of
breast cancer associated with heterozygosity for the

ATM

mutations segregating in those families ranging from three- to
seven-fold overall [Swift et al., 1987, 1991; Borresen et al., 1990;
Inskip et al., 1999; Janin et al., 1999; Athma et al., 1996; Izatt
et al., 1999; Olsen et al., 2001; Thompson et al., 2005b; Geoffroy-
Perez et al., 2001]. In contrast, mutation screening of the

ATM

gene conducted within series of breast cancer cases and controls
have produced mixed results [Bebb et al., 1999; Dork et al., 2001;
Sommer et al., 2002; Bishop and Hopper, 1997; FitzGerald et al.,
1997; Teraoka et al., 2001; Buchholz et al., 2004; Izatt et al., 1999;
Tamimi et al., 2004]. It has been hypothesized that these
differences may reflect allelic heterogeneity of

ATM and that

only a specific class of variants with the potential to dominantly
interfere with the product of the wild-type allele contribute to
breast cancer risk in heterozygotes [Gatti et al., 1999]. Neither the
identities nor the frequencies of such variants have been
established, limiting the ability to evaluate the overall impact of
ATM variation on either individual breast cancer risk or on breast
cancer incidence in the population. Some of these issues may be
resolved through large scale studies of cases, controls, and their
relatives [Bishop and Hopper, 1997].

Despite these difficulties, two functional variants of the

ATM gene have been individually associated with breast cancer
risk. One

ATM missense mutation, c.7271T4G, was originally

identified in two British A-T families with atypical clinical
presentation and an excess of breast cancer. That variant was
associated

with

a

12.7-fold

(95%

confidence

interval

[CI] 5 3.5–45.9; P 5 0.003) increase in breast cancer risk
[Stankovic et al., 1998]. An Australian study reported a third
family with this mutation, concurrently identified by testing
independently-ascertained population-based cases and multiple-
case families. By considering the history of breast cancer in the
relatives of the population-based case carrier and genotyping
family members, the risk of breast cancer was estimated to be 13.7-
fold (95% CI 5 5.1–36.6; P 5 0.001) [Chenevix-Trench et al.,
2002]. More recently, a c.7271T4G carrier was identified in a
population-based series of 1,149 women with unilateral or bilateral
breast cancer [Bernstein et al., 2003a]; the carrier had unilateral
breast cancer but no affected relatives. Subsequent published
studies of multiple-case French families [Szabo et al., 2004] and of
other women at increased risk of breast cancer have not identified
additional c.7271T4G carriers [Szabo et al., 2004]. A second,
more

common,

ATM

missense

variant,

c.1066–6T4G

(IVS10–6T4G), has also been associated with an increased risk
of breast cancer in some [Broeks et al., 2000] but not all
[Thompson et al., 2005a] studies.

The purpose of the current study was to evaluate the

associations between the c.7271T4G and c.1066–6T4G ATM
gene variants and breast cancer risk in a large population-based
case–control-family of 5,025 cases and controls and to calculate
the age-specific cumulative risk (penetrance) associated with
each variant from the breast cancer histories of the relatives of
identified carriers of these variants.

PATIENTS AND METHODS

Subjects

Cases were population-based and recruited by the Breast Cancer

Family Registry (Breast CFR; www.cfr.epi.uci.edu) in the San
Francisco Bay Area, California; in Ontario, Canada; and in
Melbourne and Sydney, Australia. The Breast CFR is described in
detail elsewhere [John et al., 2004]. The current study was limited

to the 3,757 cases and 1,268 controls who completed the family
history and epidemiologic questionnaires and from whom DNA
had been obtained. Briefly, at each of the three study sites,
incident breast cancer cases diagnosed between 1995 and 1998
were identified through population-based cancer registries. The
California and Ontario registries used a two-stage sampling design
to oversample women likely to be at increased genetic risk of breast
cancer (using a ‘‘high-risk’’ criterion based on age at diagnosis,
personal and family history of breast and ovarian cancer, and race/
ethnicity [John et al., 2004]). Cases included all those meeting the
high-risk criteria and a random sample of cases not meeting the
high-risk criteria. Age at diagnosis was 18–64 years in California
and 18–69 years in Ontario. The Australian site enrolled cases
regardless of family history, including all cases aged less than 40
years at diagnosis and a random sample of women aged 40–59 years
at diagnosis. Controls were women with no personal history of
breast cancer, randomly selected from residential telephone
numbers in Ontario, from electoral rolls in Australia, and through
random digit dialing in Northern California. Controls recruited in
San Francisco were not eligible for this study as DNA had not been
obtained when this study was initiated. Of the 2,218 Australian
cases and controls included in the current analysis, genotyping of
the c.7271T4G variant in 525 cases and 381 controls, and of the
c.1066–6T4G variant in 262 cases and 775 controls, has been
included in previous reports [Chenevix-Trench et al., 2002;
Thompson et al., 2005a]. Of those previously screened, one case
carried the c.7271T4G variant and seven controls carried the
c.1066–6T4G variant. The study protocol was approved by the
Institutional Review Boards at each participating institution and
signed informed consent was obtained from all study participants.

Genotyping

Samples

were

screened

for

the

c.7271T4G

and

the

c.1066–6T4G variants using several different methods with
blinded rescreening between techniques to evaluate consistency.
These methods were: 1) primer extension with the AcycloPrime
SNP Detection Kit (PerkinElmer Life Sciences, Boston, MA)
and extension primers 5

0

CTGAAAAGAGCCAAAGAGGAAG for

c.7271T4G and 5

0

TGGTATCTTCATTAAAAACCTGTA for

c.1066–6T4G; 2) DHPLC on the WAVE

TM

platform (Transge-

nomic, Inc., Omaha, NE) as previously described [Bernstein et al.,
2003b]; 3) DHPLC on the Varian Helix System (Varian, Walnut
Creek, CA) as previously described [Thompson et al., 2005a]; 4)
the MGB Eclipse

TM

Probe System that discriminates allele-based

temperature-induced dissociation of allele specific probes (Epoch
Biosciences, Inc., Bothell, WA/Nanogen, Inc., San Diego, CA); 5)
an RFLP approach using primers and PCR conditions as described
in Chenevix-Trench et al. [2002]; and 6) using a Taqman probe
analyzed on a Rotogene 2000 (Corbett Research, Australia;
www.corbettlifescience.com) with allelic discrimination software
(as described in Chenevix-Trench et al. [2002]). All positive
findings were confirmed by sequencing. All laboratory work was
performed blind to case–control status. Inter- and intralaboratory
quality control exchanges were conducted in the U.S. and Australian
laboratories to confirm all positive findings. The reference sequence
for ATM is NM

_000051.3 with the ‘‘A’’ residue of the ATG initiation

codon occurring at position 386 of this sequence. Mutation
numbering is with 11 as this ‘‘A’’ residue.

Statistical Methods

Fisher’s exact test and Yate’s corrected tests were used to

compare the crude allele frequencies between cases and controls.

HUMAN MUTATION 27(11), 1122^1128, 2006

1123

Human Mutation DOI 10.1002/humu

Unconditional logistic regression was used to estimate age-
adjusted ORs and 95% CIs for: 1) disease risk by comparing cases
and controls; and 2) predictors of disease in cases by comparing
exposures for carriers and noncarriers. Offsets for cases were set to
the natural log of the sampling fraction, and were included in all
regression models to accommodate the two-stage sampling designs
used by two of the sites [Piegorsch et al., 1994]. These calculations
were performed using the Statistical Analysis System software
(SAS Institute, Cary, NC).

Using the breast cancer history of the first- and second-degree

relatives of the cases who carried the variant, the age-specific
cumulative risk (penetrance) associated with each variant was
estimated by a modified segregation analysis [Hopper et al., 1999;
Antoniou et al., 2003], fitted under maximum likelihood theory
using the statistical package MENDEL [Lange et al., 1988]. The
hazard ratio (HR), defined as the ratio of the incidence in carriers
to that of the general population, was estimated on the log scale.
The associated estimated cumulative risks to age 50 years, and to
age 70 years, and the corresponding 95% CIs were based on U.S.
Surveillance, Epidemiology, and End Results (SEER) age-specific
breast cancer incidence rates for 1983–2001 combined (SEER

-

Stat Database: Incidence – SEER 11; National Cancer Institute,
Bethesda, MD). The likelihood of the observed pedigree was
conditioned on the (population-based) proband being a carrier and
affected at the age at diagnosis. Standard errors for calculation of
confidence intervals were estimated using the robust Huber/
White/sandwich estimator [Halbert, 1980]. Table 1 gives a
summary of the data for the seven cases who carried the

ATM

c.7271T4G variant and their affected first- and second- degree
relatives that were used to estimate the hazard ratios and
cumulative breast cancer risks. Three case carriers (the probands
for Families 1–3) were selected for family history, with a sampling
fraction of four times that of the unselected carriers. Therefore,
each of these pedigrees was given a weight one-fourth of that given
to the pedigrees of the other four unselected case carriers. The
population-based case carrier from Family 7 was the daughter
(ID 5 0013) of the clinic-based index patient of Family A
diagnosed at age 31 years, shown in Figure 1 of Chenevix-Trench
et al. [2002]. We also conducted analyses in which we included
pedigree information from two published multiple-case families
[Stankovic et al., 1998], by conditioning on the phenotype of
breast cancer in family members, and one published population-

based family from a population-based study [Bernstein et al.,
2003a]. For relatives reporting multiple primary breast cancers,
only the first was used in the modeling. True carrier status
was unknown for all relatives of all except the Australian
c.7271T4G–carrying family [Chenevix-Trench et al., 2002].

RESULTS

We found that 7 out of 3,743 cases and 0 out of 1,268 controls

carried the c.7271T4G variant (Fisher’s exact test P 5 0.2; age-
adjusted P 5 0.1); four of the carriers were from Ontario, two were
from the San Francisco Bay Area, and one (previously published)
was from Australia (Table 2).

Within cases, we estimated the ORs for carrier status associated

with age at diagnosis, type of disease, and a variety of breast cancer
risk factors, including family history, reproductive and menstrual
history, demographic and lifestyle factors, exogenous hormone use,
diagnostic radiation, and tumor characteristics. Carrier status was
statistically associated with having a family history of breast cancer,
with three carriers who had an affected mother (OR 5 5.5; 95%
CI 5 1.2–25.5; P 5 0.05); later age at first full-term pregnancy,
with three carriers who had their first pregnancy after age 30 years
(OR 5 5.1; 95% CI 5 1.0–25.6; P

r0.05); and type of disease,

with two carriers who had in situ disease (OR 5 6.6; 95%
CI 5 1.2–36.6; P 5 0.01).

The family-based analyses showed evidence of a substantial and

highly statistically significant risk associated with the c.7271T4G
variant. Table 3 shows that, using data from the seven population-
based families, the carriers of the variant were estimated to have a
risk of breast cancer 8.6-fold greater than the U.S. population
(95% CI 5 3.9–18.9; P

o0.0001). This equates to a cumulative

risk (penetrance) of 52% (95% CI 5 28–80%) to age 70 years.
Evidence for increased risk was evident even when the previously
reported Australian family was excluded, with a 6.2-fold increased
risk (95% CI 5 1.9–19.5; P 5 0.002) equivalent to a penetrance
of 41% (95% CI 5 15–81%) to age 70 years. After including
the three published families, the increased risk was 13.9-fold (95%
CI 5 6.2–30.8; P

o0.0001), equivalent to a penetrance of 69%

(95% CI 5 41–93%) to age 70 years.

We found that 13 out of 3,757 cases and 10 out of 1,268

controls carried the c.1066–6T4G variant (Yate’s corrected
P 5 0.08;

age-adjusted

OR 5 0.44;

95%

CI 5 0.19–1.00;

TABLE 1.

Families Found to Carry the ATM c.72714G Variant

Family

Proband

A¡ected relatives

Ascertainment

Age at Dx (range)

Cancer site

Number (degree relation)

b

Age at Dx (range)

1

Multiple-case

50^59

Breast

1 (1)

50^59

2 (2)

30^39, 30^39

2

Multiple-case

40^49

Breast

1 (1)

80^89

Uterine

1 (2)

70^79

3

Multiple-case

40^49

Breast

2 (1)

40^49, 60^69

Uterine

1 (1)

90^99

Stomach

1 (2)

50^59

4

Sporadic

50^59

Breast

1 (1)

50^59

Colon

1 (2)

50^59

5

Sporadic

40^49

Breast

1 (2)

70^79

6

Sporadic

40^49

Throat

1 (2)

50^59

7

a

Unselected case

30^39

Breast

1 (1)

30^39

1 (2)

70^79

Age at breast cancer diagnosis. Age ranges are provided per HIPAA guidelines to protect the identity of the women included in this series.

a

Chenevix-Trench et al. [2002].

b

First- or second-degree relative.

Dx, diagnosis.

1124

HUMAN MUTATION 27(11), 1122^1128, 2006

Human Mutation DOI 10.1002/humu

P 5 0.04). For cases, carriers were significantly more likely to have
a younger age at diagnosis; the OR for

o45 vs. 451 was 0.2 (95%

CI 5 0.1–0.7; data not shown). We also analyzed the risk factors
shown in Table 3 and found no statistically significant associations
(data not shown).

In family-based analyses using the 11 North American families

(two from Northern California, nine from Ontario) and the 12
Australian families in which the proband was a carrier of the
c.1066–6T4G variant, there was no evidence that relatives of
the carriers were at increased risk of breast cancer compared

TABLE 2.

Frequency of c.7271T4G Mutations in the ATM Gene inWomen With Breast Cancer

Category

Variable

Positive cases

(N 5 7)

Negative cases

(N 5 3,736)

Age-adjusted OR

and 95% CI

n (%)

n (%)

Demographics

Registry

Australia

1 (0.1)

1,431 (99.9)

0.10 (0.01^1.03)

Northern California

2 (0.2)

1,110 (99.8)

0.44 (0.08^2.42)

Ontario

4 (0.3)

1,195 (99.7)

1

Age at diagnosis (years)

o45

4 (0.2)

1,815 (99.8)

0.65 (0.15^2.93)

451

3 (0.2)

1,921 (99.8)

1

Education

Post-high school

4 (0.2)

2,456 (99.8)

0.81 (0.18^3.66)

High school or less

3 (0.2)

1,259 (99.8)

1

Race

Non-Caucasian

2 (0.3)

652 (99.7)

1.95 (0.38^10.09)

Caucasian

5 (0.2)

3,057 (99.8)

1

Marital status

Never married

0 (0.0)

303 (100.0)

Had married

6 (0.2)

3,022 (99.8)

Family history

a

Mother

Yes

3 (0.2)

1,921 (99.8)

5.46 (1.17^25.48)

No

4 (0.2)

1,815 (99.8)

1

Any ¢rst-degree relative

Yes

4 (0.4)

897 (99.6)

3.85 (0.79^18.70)

No

3 (0.1)

2,839 (99.9)

1

Any second-degree relative

Yes

2 (0.3)

649 (99.7)

1.89 (0.32^11.36)

No

3 (0.2)

1,977 (99.8)

1

Medical and reproductive

history

Menopausal status

Postmenopausal

0 (0.0)

1,077 (100.0)

Premenopausal

4 (0.2)

1,972 (99.8)

Pregnancy history

Parous

6 (0.2)

3,095 (99.8)

1.29 (0.15^10.81)

Nulliparous

1 (0.2)

637 (99.8)

1

Age at ¢rst full-term pregnancy

301

3 (0.6)

470 (99.4)

5.11 (1.02^25.63)

o30

3 (0.1)

2,459 (99.9)

1

History of benign breast disease

Yes

2 (0.2)

1,005 (99.8)

1.43 (0.25^8.14)

No

5 (0.2)

2,731 (99.8)

1

Exogenous hormone use

Age at ¢rst OC use

201

2 (0.1)

1,657 (99.9)

0.21 (0.03^1.32)

o20

4 (0.5)

826 (99.5)

1

Alcohol use

Ever

5 (0.3)

1,580 (99.7)

5.45 (0.64^46.79)

Never

1 (0.1)

1,728 (99.9)

1

Cigarette use

Ever

5 (0.3)

1,800 (99.7)

2.56 (0.50^13.22)

Never

2 (0.1)

1,924 (99.9)

1

History of X-rays and

stage at diagnosis

Total number of diagnostic

chest X-rays
21

2 (0.4)

491 (99.6)

3.19 (0.53^19.27)

o2

3 (0.1)

2,302 (99.9)

1

Diagnostic X-ray in chest or

abdominal area
Ever

4 (0.3)

1,327 (99.7)

5.07 (0.56^46.22)

Never

1 (0.1)

1,604 (99.9)

1

Tumor behavior

In situ

2 (0.9)

223 (99.1)

6.57 (1.18^36.57)

Invasive

5 (0.1)

3,513 (99.9)

1

A total of 5,020 women were evaluated. Of those, 3,743 were cases (seven positive for c.7271T4G) and 1,277 were controls (none positive for

c.7271T4G).

a

Relatives with reported breast cancer.

OC, oral contraceptives; OR, odds ratio, adjusted for age at reference date greater than or equal to 45 years; 95% CI, 95% con¢dence interval;
***, odds ratios are not estimable.

HUMAN MUTATION 27(11), 1122^1128, 2006

1125

Human Mutation DOI 10.1002/humu

with the population (HR 5 1.6; 95% CI 5 0.4–6.2; P 5 0.5; data
not shown).

DISCUSSION

Although the majority of

ATM mutations detected in A-T

patients are predicted to truncate the

ATM protein, the initial

studies that examined the role of truncating

ATM mutations and

breast cancer risk failed to reveal statistically significant disease
associations [Bebb et al., 1999; FitzGerald et al., 1997]. This was
despite the evidence of excess breast cancer incidence within A-T
families [Swift et al., 1987, 1991; Borresen et al., 1990; Inskip
et al., 1999; Janin et al., 1999; Athma et al., 1996; Izatt et al.,
1999; Olsen et al., 2001; Thompson et al., 2005b; Geoffroy-Perez
et al., 2001]. This apparent dichotomy between the results
obtained with these two different study designs may simply reflect
inadequate power rather than true disagreement [Bishop and
Hopper, 1997]. Most recent studies have focused on missense
mutations in

ATM that have the potential to produce some

protein, albeit in a mutated form [Dork et al., 2001; Teraoka et al.,
2001; Stankovic et al., 1998; Chenevix-Trench et al., 2002], that
might have the potential to dominantly interfere with ATM
function. Activation of

ATM requires transphosphorylation by the

ATM molecules in inactive dimers [Bakkenist and Kastan, 2003].
This specific requirement for transphosphorylation provides a
mechanism for dominant interference in

ATM heterozygotes in

which mutant protein can potentially dimerize with wild-type and
interfere with activation. Nevertheless, while numerous putative
ATM missense variants have now been identified by sequence-
based screening approaches, a deleterious effect on

ATM function

has been established for only a few of these variants [Scott et al.,
2002; Chenevix-Trench et al., 2002; Gutierrez-Enriquez et al.,
2004] and several studies have failed to detect any difference in
breast cancer risk among different types of

ATM mutations

[Cavaciuti et al., 2005; Thompson et al., 2005b]. In this study
we focused on two such variants for which there are both
epidemiologic data indicating an association with breast cancer
risk and functional data suggesting dominant negative activity in
heterozygote cell lines.

Despite the promising nature of some of the previously

published data for both of these variants, c.1066–6T4G and
c.7271T4G, our study produced different results. Using a large
population-based study we found that c.1066–6T4G was not
associated with an increased risk of breast cancer. Though rare,
this variant was more common in controls (0.8%) than in cases
(0.3%), although the difference was not statistically significant,
and was not associated with family history. The within-family
analyses did not reveal evidence for increased risks. This variant
had been demonstrated to affect normal splicing of

ATM

transcripts and to reduce the in vivo

ATM kinase activity in cell

lines from heterozygotes by 60 to 75% when assayed on p53 or
BRCA1 [Chenevix-Trench et al., 2002].

There are several possible reasons why we may have failed to

detect an association with breast cancer for the c.1066–6T4G
variant despite the published evidence of a functional effect of this
allele in heterozygous cell lines. For example, there may be
differences in splicing efficiency or stability of the mutant protein
between the cell lines that were studied and breast tissue.
Increased risk for breast cancer may result from a very specific
deficit in

ATM function; although kinase activity is a fundamental

feature of

ATM function, it may not be the most informative with

regard to breast cancer risk. The effects on

ATM kinase activity

observed might reflect the presence of a second unrecognized
variant that occurs on only a fraction of haplotypes along with
c.1066–6T4G. Finally, there may be no real effect, or the effect
may be too small for us to detect by a case–control comparison
with the current sample size, which had 80% power at the 0.05
level of significance to detect an effect of 2.5-fold or more.

The second variant studied, c.7271T4G, is predicted to result

in a valine to glycine substitution at position 2424 of the

ATM

protein (p.Val2424Gly) [Stankovic et al., 1998]. While this
substitution does not affect any readily-recognizable functional
domain, programs such as SIFT [Ng and Henikoff, 2002] and
PolyPhen [Sunyaev et al., 2001], which attempt to classify
mutations based on sequence conservation and structural predic-
tion, suggest that this substitution is deleterious. This conclusion is
consistent with published functional studies of cell lines from a
limited number of heterozygotes that suggest the variant has
dominant negative activity [Chenevix-Trench et al., 2002].

Few studies have reported identifying any c.7271T4G carriers.

The large size of our study allowed us to identify six additional
carriers, more than doubling the number of known carriers of this
variant. Combining all known c.7271T4G heterozygote carriers,
we estimated the cumulative risk of breast cancer to be 69% to age
70 years. However, the probability of carrying the variant is
difficult to estimate precisely, but is very small. The original
description of the c.7271T4G variant was in two British families,
one of which included several homozygous individuals. In our
study, most carriers reported origins in the United Kingdom and/or
Scotland, suggesting strong historical ties consistent with a
possible founder effect in the United Kingdom population for this
variant. This is consistent with, but does not necessarily prove,
that the population attributable risk may be higher in the United
Kingdom, where this variant may be more prevalent.

For cases, we also noted a statistically significant association

between the c.7271T4G variant and later ages at first pregnancy.
ATM is believed to play a role in meiosis. ATM knockout mice are
infertile and human A-T patients have a number of features
suggesting reduced fertility. Interestingly, the only A-T patient
documented to have had a child was homozygous for the
c.7271T4G variant [Stankovic et al., 1998]. This patient had

TABLE 3.

Estimates of Hazard Ratios and Cumulative Risks of Breast Cancer for Heterozygote Carriers of the ATM c.7271T4G Variant

Families

Hazard ratio

(95% CI)

P value

Cumulative risk to age

50 years (95% CI)

a

Cumulative risk to age

70 years (95% CI)

a

Australia, Ontario and Northern
California (1^7)

8.6 (3.9^18.9)

o0.0001

18% (8^35)

52% (28^80)

Excluding Australia (1^6)

6.2 (1.9^19.5)

o0.002

13% (4^36)

41% (15^81)

Including all published families (1^10)

b

13.9 (6.2^30.8)

o 0.0001

27% (13^50)

69% (41^93)

95% CI denotes the 95% con¢dence interval.

a

Based on estimated hazard ratio and comparing to U.S. incidence rate.

b

Includes the 7 families from this series as well as the published pedigrees from Bernstein et al. [2003a] and Stankovic et al. [1998].

1126

HUMAN MUTATION 27(11), 1122^1128, 2006

Human Mutation DOI 10.1002/humu

an atypically mild clinical course, perhaps reflecting the benefit of
producing even a modest amount of a mutant protein as compared
to the undetectable levels of ATM that are typically observed in
A-T. It is therefore possible that our observed association with later
age at first pregnancy reflects a direct biologic effect of this allele
previously demonstrated to reduce

ATM kinase activity in

heterozygotes well below what would be expected for a simple
loss of function allele. In the Breast CFR, the only data collected
on fertility was the use of infertility medicine. Of the seven
carriers, one had never been pregnant and one who had been
pregnant, reported taking medication for infertility. Further
exploration of a possible association of c.7271T4G carrier status
with infertility is warranted.

In conclusion, we evaluated in a large population-based study

the effects of two

ATM functional variants, c.1066–6T4G and

c.7271T4G, on risk of breast cancer. We found no evidence for
c.1066–6T4G being associated with increased breast cancer risk.
However, our results suggest that c.7271T4G is a high-penetrant
breast cancer susceptibility variant. Although the infrequent
occurrence of c.7271T4G limits its overall contribution to
population risk, it should be recognized that this is only one of
many missense substitutions identified within the

ATM gene that

are predicted to have functional consequences. Thus, while it may
be difficult to associate individual

ATM variants with breast

cancer risk, in aggregate, the risk associated with

ATM variants

could be clinically, as well as individually, significant. As illustrated
by the contrasting results obtained from the two variants we
studied, identification of breast cancer predisposing variants
in

ATM by mutation screening alone is challenging and might

require testing very large numbers of cases to find a few carriers,
and then studying their family histories. If

ATM-mediated risk

for breast cancer results from the effects of many infrequent alleles
such as c.7271T4G, then instead of mutation screening for
individual variants, a more efficient approach may be to develop
functional assays that reliably measure the activity of

ATM that

is relevant to breast cancer risk.

ACKNOWLEDGMENTS

We acknowledge the technical support provided by Xiaoqing

Chen, and Margaret McCredie, and we especially thank the study
subjects who made this valuable work possible. This study was
supported by National Institutes of Health grants U01CA83178
(to J.L.B.), 5U01CA69467 (to I.L.A.), 5U01CA69638 (to J.L.H.),
and 5U01CA69417 (to D.W. and E.M.J.).

REFERENCES

Antoniou A, Pharoah PD, Narod S, Risch HA, Eyfjord JE, Hopper JL,

Loman N, Olsson H, Johannsson O, Borg A, Pasini B, Radice P,
Manoukian S, Eccles DM, Tang N, Olah E, Anton-Culver H, Warner E,
Lubinski J, Gronwald J, Gorski B, Tulinius H, Thorlacius S, Eerola H,
Nevanlinna H, Syrjakoski K, Kallioniemi OP, Thompson D, Evans C,
Peto J, Lalloo F, Evans DG, Easton DF. 2003. Average risks of breast and
ovarian cancer associated with BRCA1 or BRCA2 mutations detected in
case series unselected for family history: a combined analysis of 22
studies. Am J Hum Genet 72:1117–1130.

Athma P, Rappaport R, Swift M. 1996. Molecular genotyping shows that

ataxia-telangiectasia heterozygotes are predisposed to breast cancer.
Cancer Genet Cytogenet 92:130–134.

Bakkenist CJ, Kastan MB. 2003. DNA damage activates ATM through

intermolecular autophosphorylation and dimer dissociation. Nature 421:
499–506.

Bakkenist CJ, Kastan MB. 2004. Initiating cellular stress responses. Cell

118:9–17.

Bebb DG, Yu Z, Chen J, Telatar M, Gelmon K, Phillips N, Gatti RA,

Glickman BW. 1999. Absence of mutations in the ATM gene in forty-
seven cases of sporadic breast cancer. Br J Cancer 80:1979–1981.

Bernstein JL, Bernstein L, Thompson WD, Lynch CF, Malone KE,

Teitelbaum SL, Olsen JH, Anton-Culver H, Boice JD, Rosenstein BS,
Borresen-Dale AL, Gatti RA, Concannon P, Haile RW. 2003a. ATM
variants 7271T4G and IVS10–6T4G among women with unilateral
and bilateral breast cancer. Br J Cancer 89:1513–1516.

Bernstein JL, Teraoka S, Haile RW, Borresen-Dale AL, Rosenstein BS,

Gatti RA, Diep AT, Jansen L, Atencio DP, Olsen JH, Bernstein L,
Teitelbaum SL, Thompson WD, Concannon P. 2003b. Designing and
implementing quality control for multi-center screening of mutations
in the ATM gene among women with breast cancer. Hum Mutat 21:
542–550.

Bishop D, Hopper J. 1997. AT-tributable risks? Nat Genet 15:226.
Borresen A, Andersen T, Tretli S, Heilberg A, Moller P. 1990. Breast cancer

and other cancers in Norwegian families with ataxia-telangiectasia.
Genes Chromosomes Cancer 2:339–340.

Broeks A, Urbanus JH, Floore AN, Dahler EC, Klijn JG, Rutgers EJ,

Devilee P, Russell NS, van Leeuwen FE, ’t Veer LJ. 2000. ATM-
heterozygous germline mutations contribute to breast cancer-suscept-
ibility. Am J Hum Genet 66:494–500.

Buchholz TA, Weil MM, Ashorn CL, Strom EA, Sigurdson A, Bondy M,

Chakraborty R, Cox JD, McNeese MD, Story MD. 2004. A Ser49Cys
variant in the ataxia telangiectasia, mutated, gene that is more common
in patients with breast carcinoma compared with population controls.
Cancer 100:1345–1351.

Cavaciuti E, Lauge A, Janin N, Ossian K, Hall J, Stoppa-Lyonnet D,

Andrieu N. 2005. Cancer risk according to type and location of ATM
mutation in ataxia-telangiectasia families. Genes Chromosomes Cancer
42:1–9.

Chenevix-Trench G, Spurdle AB, Gatei M, Kelly H, Marsh A, Chen X,

Donn K, Cummings M, Nyholt D, Jenkins MA, Scott C, Pupo GM, Dork
T, Bendix R, Kirk J, Tucker K, McCredie MR, Hopper JL, Sambrook J,
Mann GJ, Khanna KK. 2002. Dominant negative ATM mutations in
breast cancer families. J Natl Cancer Inst 94:205–215.

Dork T, Bendix R, Bremer M, Rades D, Klopper K, Nicke M, Skawran B,

Hector A, Yamini P, Steinmann D, Weise S, Stuhrmann M, Karstens JH.
2001. Spectrum of ATM gene mutations in a hospital-based series of
unselected breast cancer patients. Cancer Res 61:7608–7615.

FitzGerald MG, Bean JM, Hegde SR, Unsal H, MacDonald DJ, Harkin DP,

Finkelstein DM, Isselbacher KJ, Haber DA. 1997. Heterozygous ATM
mutations do not contribute to early onset of breast cancer. Nat Genet
15:307–310.

Gatti RA, Tward A, Concannon P. 1999. Cancer risk in ATM

heterozygotes: a model of phenotypic and mechanistic differences
between missense and truncating mutations. Mol Genet Metab 68:
419–423.

Geoffroy-Perez B, Janin N, Ossian K, Lauge A, Croquette MF, Griscelli C,

Debre M, Bressac-de-Paillerets B, Aurias A, Stoppa-Lyonnet D, Andrieu
N. 2001. Cancer risk in heterozygotes for ataxia-telangiectasia. Int J
Cancer 93:288–293.

Gutierrez-Enriquez S, Fernet M, Dork T, Bremer M, Lauge A, Stoppa-

Lyonnet D, Moullan N, Angele S, Hall J. 2004. Functional consequences
of ATM sequence variants for chromosomal radiosensitivity. Genes
Chromosomes Cancer 40:109–119.

Halbert H. 1980. A heteroskedasticity-consistent covariance matrix

estimator and a direct test for heteroskedasticity. Econometrica 48:
817–838.

Hopper JL, Southey MC, Dite GS, Jolley DJ, Giles GG, McCredie MR,

Easton DF, Venter DJ. 1999. Population-based estimate of the average
age-specific cumulative risk of breast cancer for a defined set of protein-
truncating mutations in BRCA1 and BRCA2. Australian Breast Cancer
Family Study. Cancer Epidemiol Biomarkers Prev 8:741–747.

Inskip H, Kinlen L, Taylor A, Woods C, Arlett C. 1999. Risk of breast

cancer and other cancers in heterozygotes for ataxia telangiectasia. Br J
Cancer 79:1304–1307.

Izatt L, Greenman J, Hodgson S, Ellis D, Watts S, Scott G, Jacobs C,

Liebmann R, Zvelebil MJ, Mathew C, Solomon E. 1999. Identification of

HUMAN MUTATION 27(11), 1122^1128, 2006

1127

Human Mutation DOI 10.1002/humu

germline missense mutations and rare allelic variants in the ATM gene
in early-onset breast cancer. Genes Chromosomes Cancer 26:286–294.

Janin N, Andrieu N, Ossaian K, Lauge A, Croquette M, Griscelli C, Debre

M, Bressac-de-Paillertes B, Aurias A, Stoppa-Lyonnet D. 1999. Breast
cancer risk in ataxia telangiectasia (AT) heterozygotes: haplotype study
in French AT families. Br J Cancer 80:1042–1045.

John EM, Hopper JL, Beck JC, Knight JA, Neuhausen SL, Senie RT, Ziogas

A, Andrulis IL, Anton-Culver H, Boyd N, Buys SS, Daly MB, O’Malley
FP, Santella RM, Southey MC, Venne VL, Venter DJ, West DW,
Whittemore AS, Seminara D. 2004. The Breast Cancer Family Registry:
an infrastructure for cooperative multinational, interdisciplinary and
translational studies of the genetic epidemiology of breast cancer. Breast
Cancer Res 6:R375–R389.

Kurz EU, Lees-Miller SP. 2004. DNA damage-induced activation of ATM

and ATM-dependent signaling pathways. DNA Repair (Amst) 3:
889–900.

Lange K, Weeks D, Boehnke M. 1988. Programs for pedigree analysis:

MENDEL, FISHER, and dGENE. Genet Epidemiol 5:471–472.

Ng PC, Henikoff S. 2002. Accounting for human polymorphisms predicted

to affect protein function. Genome Res 12:436–446.

Olsen JH, Hahnemann JM, Borresen-Dale AL, Brondum-Nielsen K,

Hammarstrom L, Kleinerman R, Kaariainen H, Lonnqvist T, Sankila R,
Seersholm N, Tretli S, Yuen J, Boice JD Jr, Tucker M. 2001. Cancer in
patients with ataxia-telangiectasia and in their relatives in the Nordic
countries. J Natl Cancer Inst 93:121–127.

Piegorsch W, Weinberg C, Taylor J. 1994. Non-hierarchical logistic models

and case-only designs for assessing susceptibility in population-based
case-control studies. Stat Med 13:153–162.

Savitsky K, Bar-Shira A, Gilad S, Rotman G, Ziv Y, Vanagaite L, Tagle DA,

Smith S, Uziel T, Sfez S, Ashkenazi M, Pecker I, Frydman M, Harnik R,
Patanjali SR, Simmons A, Clines GA, Sartiel A, Gatti RA, Chessa L,
Sanal O, Lavin MF, Jaspers NGJ, Taylor MR, Arlett CF, Miki T,
Weissman SM, Lovett M, Collins FS, Shiloh Y. 1995. A single ataxia
telangiectasia gene with a product similar to PI-3 kinase. Science
268:1749–1753.

Scott SP, Bendix R, Chen P, Clark R, Dork T, Lavin MF. 2002. Missense

mutations but not allelic variants alter the function of ATM by
dominant interference in patients with breast cancer. Proc Natl Acad Sci
USA 99:925–930.

Sommer SS, Buzin CH, Jung M, Zheng J, Liu Q, Jeong SJ, Moulds J,

Nguyen VQ, Feng JN, Bennett WP, Dritschilo A. 2002. Elevated
frequency of ATM gene missense mutations in breast cancer relative to
ethnically matched controls. Cancer Genet Cytogenet 134:25–32.

Stankovic T, Kidd A, Sutcliffe A, McGuire G, Robinson P, Weber P,

Bedenham T, Bradwell A, Easton D, Lennox G, Haites N, Byrd P, Taylor
A. 1998. 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 62:334–345.

Sunyaev S, Ramensky V, Koch I, Lathe W, III, Kondrashov AS, Bork P.

2001. Prediction of deleterious human alleles. Hum Mol Genet 10:
591–597.

Swift M, Reitnauer P, Morrell D, Chase C. 1987. Breast and other cancers

in families with ataxia-telangiectasia. N Engl J Med 316:1289–1294.

Swift M, Morrell D, Massey R, Chase C. 1991. Incidence of cancer

in 161 families affected by ataxia-telangiectasia. N Engl J Med 325:
1831–1836.

Szabo CI, Schutte M, Broeks A, Houwing-Duistermaat JJ, Thorstenson YR,

Durocher F, Oldenburg RA, Wasielewski M, Odefrey F, Thompson D,
Floore AN, Kraan J, Klijn JG, van den Ouweland AM, Wagner TM,
Devilee P, Simard J, ’t Veer LJ, Goldgar DE, Meijers-Heijboer H. 2004.
Are ATM mutations 7271T–4G and IVS10-6T–4G really high-risk
breast cancer-susceptibility alleles? Cancer Res 64:840–843.

Tamimi RM, Hankinson SE, Spiegelman D, Kraft P, Colditz GA, Hunter

DJ. 2004. Common ataxia telangiectasia mutated haplotypes and risk
of breast cancer: a nested case-control study. Breast Cancer Res 6:
R416–R422.

Teraoka SN, Malone KE, Doody DR, Suter NM, Ostrander EA, Daling JR,

Concannon P. 2001. Increased frequency of ATM mutations in breast
carcinoma patients with early onset disease and positive family history.
Cancer 92:479–487.

Thompson D, Antoniou AC, Jenkins M, Marsh A, Chen X, Wayne T,

Tesoriero A, Milne R, Spurdle A, Thorstenson Y, Southey M, Giles GG,
Khanna KK, Sambrook J, Oefner P, Goldgar D, Hopper JL, Easton D,
Chenevix-Trench G. 2005a. Two ATM variants and breast cancer risk.
Hum Mutat 25:594–595.

Thompson D, Duedal S, Kirner J, McGuffog L, Last J, Reiman A, Byrd P,

Taylor M, Easton DF. 2005b. Cancer risks and mortality in heterozygous
ATM mutation carriers. J Natl Cancer Inst 97:813–822.

1128

HUMAN MUTATION 27(11), 1122^1128, 2006

Human Mutation DOI 10.1002/humu