Genetic Predisposition to Breast Cancer: Past, Present, and Future

ANRV353-GG09-17 ARI 25 July 2008 17:26

Genetic Predispositionto Breast Cancer:Past, Present, and FutureClare Turnbull and Nazneen RahmanSection of Cancer Genetics, Institute of Cancer Research, Sutton, SM2 5NG,United Kingdom; email: [email protected], [email protected]

Annu. Rev. Genomics Hum. Genet. 2008. 9:321–45

First published online as a Review in Advance onJune 10, 2008

The Annual Review of Genomics and Human Geneticsis online at genom.annualreviews.org

This article’s doi:10.1146/annurev.genom.9.081307.164339

Copyright c© 2008 by Annual Reviews.All rights reserved

1527-8204/08/0922-0321$20.00

Key Words

BRCA∗, allele, familial, susceptibility, penetrance

AbstractIn recent years, our understanding of genetic predisposition to breastcancer has advanced significantly. Three classes of predisposition fac-tors, categorized by their associated risks of breast cancer, are currentlyknown. BRCA1 and BRCA2 are high-penetrance breast cancer predis-position genes identified by genome-wide linkage analysis and posi-tional cloning. Mutational screening of genes functionally related toBRCA1 and/or BRCA2 has revealed four genes, CHEK2, ATM, BRIP1,and PALB2; mutations in these genes are rare and confer an intermedi-ate risk of breast cancer. Association studies have further identified eightcommon variants associated with low-penetrance breast cancer predis-position. Despite these discoveries, most of the familial risk of breastcancer remains unexplained. In this review, we describe the known ge-netic predisposition factors, expound on the methods by which theywere identified, and consider how further technological and intellec-tual advances may assist in identifying the remaining genetic factorsunderlying breast cancer susceptibility.

321

Click here for quick links to

Annual Reviews content online,

including:

• Other articles in this volume

• Top cited articles

• Top downloaded articles

• Our comprehensive search

FurtherANNUALREVIEWS

Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

Linkage analysis: astatistical method forevaluating linkagebetween disease andmarkers of knownlocation by followingtheir inheritance infamilies

Penetrance: theprobability that aparticularphenotype/disease isexpressed in anindividual with aparticular genotype

INTRODUCTION

Through the dynamic interplay of multipleapproaches, the past two decades have wit-nessed the gradual emergence of a clearer un-derstanding of genetic susceptibility to breastcancer. Clinical observation underpinned theseadvances through recognition of unusual famil-ial clustering and phenotypes associated withbreast cancer. Observational epidemiology hasalso provided essential foundations to our un-derstanding, both in quantifying the contri-bution of genetic factors to breast cancer andin predicting how these factors will act indi-vidually and interact. Genetic modeling hasbeen performed to predict the profiles of thegenes involved, which in turn facilitates theselection of molecular methods appropriatefor identifying them. Linkage analysis, muta-tional screening of candidate genes, and as-sociation studies have been used to identifypredisposition factors of three distinct risk-prevalence profiles: rare high-penetrance alle-les, rare intermediate-penetrance alleles, andcommon low-penetrance alleles.

OBSERVATIONALEPIDEMIOLOGY ANDSEGREGATION ANALYSIS

Epidemiological observation of the clusteringof breast cancer within families was made inRoman times and reiterated by other earlysocio-scientific commentators (22, 77). Morethan 50 studies have explored this familial ag-gregation of breast cancer using predominantlycase-control and cohort designs. Meta-analysesof these data conclude that, overall, breast can-cer is twice as common in women with anaffected first-degree relative. Simulation stud-ies suggest that an environmental factor wouldhave to confer at least a tenfold increase in riskfor shared exposure to result in even a mod-est increase to the familial risk (65). Becauseno environmental risk factors of this magnitudehave been identified for breast cancer, it seemsunlikely that shared environment accounts formuch of the familial aggregation. Heritability

studies can assist in clarifying the relative con-tribution of genetic factors and shared envi-ronment. Via demonstration that the risk to amonozygotic twin is substantially higher thanto a dizygotic twin of an affected individual,heritability studies using twins confirmed thatthe predominant component of the familial ag-gregation in breast cancer is genetic (80, 93).Patterns of disease aggregation in twins provideevidence not just for a genetic basis for breastcancer but demonstrate a markedly skewed dis-tribution of genetic liability, suggesting that themajority of genetic risk may lie within a genet-ically predisposed minority (93).

Segregation analysis (genetic modeling) in-volves the simulation of various scenarioswhereby disease occurs owing to combinationsof known or hypothetical genes of specifiedrisk, prevalence, and mode of inheritance. Thepatterns of cancers predicted by the differentmodels are compared with observed diseasefrequencies and the best-fitting model is fa-vored. Segregation analyses have strongly in-fluenced the molecular approaches with whichgenetic predisposition factors have been in-vestigated. The Cancer and Steroid HormoneStudy data and other early segregation anal-yses of families with breast cancer favored ahighly penetrant autosomal dominant geneticmodel (16, 29, 68, 121). This prediction was val-idated through linkage analysis and the identifi-cation of the highly penetrant autosomal domi-nant breast cancer predisposition genes, BRCA1and BRCA2. Subsequent complex segregationanalyses have incorporated the effects ofBRCA1 and BRCA2 and explored the evidencefor additional genes (8, 9, 35, 89). Proba-bly the most comprehensive of these analy-ses, based upon an extensive series comprisingboth population-based cases of breast cancerand large familial clusters, has strongly favoredthe polygenic model. The polygenic compo-nent of risk in this model is equivalent to mul-tiple genetic factors acting independently: Thiscomponent is log-normally distributed, has avariance that declines linearly with age, and ap-plies similarly to BRCA mutation carriers andnoncarriers (8–10). The polygenic model is also

322 Turnbull · Rahman

Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

consistent with multiple observations that theexcess of familial breast cancer is distributedacross many families, each typically comprisinga modest number of cases, rather than just a fewvery extensive families (9, 35). The polygenicmodel has largely been accepted as an appo-site explanation for the residual breast cancerpredisposition and has been validated in partby identification of a number of the incumbentpolygenes.

STRATEGIES FORIDENTIFYING BREAST CANCERPREDISPOSITION FACTORS

Three principal experimental designs have beenused in the molecular identification of geneticbreast cancer predisposition factors: genome-wide linkage analysis, mutational screening ofcandidate genes, and association studies.

Linkage Analysisand Positional Cloning

Linkage studies are used to map a disease lo-cus via analysis for cosegregation of genomicmarkers with a specified disease phenotype us-ing samples from multiple members of largefamilies. The region of the genome surround-ing the linked markers is then interrogatedfor likely causative genes (positional cloning).Linkage analysis is suitable for mapping onlyhigh-penetrance breast cancer predispositiongenes, whereby mutations result in prominentcancer families in which the majority of af-fected individuals carry the mutation. If a geneis of lower penetrance, the correlation betweenbreast cancer and mutation status in mutation-positive families may be insufficient to generatea linkage signal. Nor will a significant signal bedetectable if the breast cancer is linked to a par-ticular locus in only a small proportion of thefamilies analyzed.

Linkage maps and sets of markers that ad-equately cover the genome became availablein the late 1980s. Using large breast cancerpedigrees, genome-wide linkage analyses wereundertaken to map the high-penetrance breast

cancer susceptibility genes proposed by the seg-regation analyses, BRCA1 and BRCA2 (61, 123).Linkage analysis has also been performed ingroups of phenotypically distinct families tomap the loci underlying specific syndromes thatinclude an elevated risk of breast cancer. Po-sitional cloning was then used to identify thecausative genes, such as PTEN, STK11, andCDH1 (58, 63, 64, 70, 87, 88).

Resequencing Studies: MutationalScreening of Candidate Genes

During the 1990s, emerging understanding ofthe molecular pathogenesis of breast canceroffered insights into possible candidate genesthat predispose to breast cancer. It seemed bi-ologically plausible that proteins interactingwith BRCA1 and BRCA2 or acting in similarDNA repair pathways may also be involved inbreast cancer susceptibility. A few deleteriousmutations were found at frequencies amenableto evaluation by association studies. However,in the majority of DNA repair genes studied,disease-causing mutations (primarily leadingto premature protein truncation or nonsense-mediated RNA decay) have been individuallyvery rare. A robust demonstration that suchgenes confer predisposition to breast cancer hastherefore required mutational screening of theentire coding sequence of the gene throughlarge numbers of cases and controls to meaning-fully compare the total number of pathogenicmutations. Experiments of this magnitude havebeen published for only a handful of DNA re-pair genes. Other postulated candidate genes,such as those implicated in cell cycle regulation,checkpoint control, apoptosis, and steroid hor-mone metabolism, have rarely been evaluatedto this level.

Association Studies

In an association study, the frequency of a spec-ified variant is compared between breast can-cer cases and controls. Statistically significantdifferences in allele frequency between casesand controls are more readily demonstrable for

www.annualreviews.org • Genetic Predisposition to Breast Cancer 323

Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

Single nucleotidepolymorphism(SNP): a single-basesequence variation thatcommonly occurs at aparticular positionwithin the genome

Linkagedisequilibrium: thenon-randomassociation of alleles attwo or more loci,which suggests thatthey may be physicallyclose and thus linked

variants of population frequency >5%; henceassociation studies are most suitable for iden-tification of common breast cancer predisposi-tion variants. Initial studies focused on genesproposed by function and examined the associ-ation with breast cancer of numerous commonvariants from recognized and candidate predis-position genes. However, experiments were of-ten too small, subject to bias, and utilized toolenient levels of significance, resulting in incon-sistency and lack of replication of findings. Themajority of putative associations likely repre-sented false positives (type 1 errors) (40, 95). Toincrease power and rationalize choices of can-didates, collaborations have formed that under-take association studies across tens of thousandsof samples (19).

The emergence of comprehensive high-density maps of single nucleotide polymor-phisms (SNPs) and affordable genotypingplatforms has allowed the graduation of asso-ciation studies from the limitations of precon-ceived notions of candidacy to the agnosticismof the genome-wide approach. On account oflinkage disequilibrium, a panel of a few hundredthousand reporter SNPs can be used as tags forthe majority of the millions of common vari-ants in the genome. Accordingly, owing to thedegree of multiple testing, a genome-wide scanmust be sufficiently well-powered to ensure thatthe true associations are detected. The stagedexperimental design is a means of optimizingthe statistical power afforded by a given samplesize and has been useful in studies to date be-cause of the current high costs of whole genometag SNP panels.

BREAST CANCERPREDISPOSITIONGENES AND VARIANTS

The breast cancer predisposition factorsidentified to date can be stratified byrisk profile into three tiers: high-penetrancegenes, intermediate-penetrance genes, andlow-penetrance alleles. Three further genes areassociated with syndromes in which the inci-

dence of breast cancer is elevated but the actualrisk remains unclear (Table 1).

High-Penetrance Breast CancerPredisposition Genes

Mutations in three high-penetrance breast can-cer predisposition genes confer a greater thantenfold relative risk of breast cancer. BRCA1and BRCA2 were identified through linkageanalysis and positional cloning. TP53 wasdeemed a plausible candidate and identified asa high-risk breast cancer gene through muta-tional screening.

BRCA1 and BRCA2. The first convincingreport of linkage of breast cancer to 17q21was published in 1990 (61). In 1994, positionalcloning revealed the causative gene, accord-ingly named BRCA1 (86). Linkage analysis andpositional cloning led to the mapping and iden-tification of BRCA2 in 1994 and 1995, respec-tively (122, 123). BRCA1 and BRCA2 have im-portant roles in the maintenance of genomicstability by facilitating repair of DNA double-strand breaks. The cellular roles of BRCA1 andBRCA2 were reviewed by Gudmundsdottir &Ashworth (56).

BRCA1 and BRCA2 are large genes inwhich multiple different loss-of-functionmutations have been detected. Some foundermutations are relatively frequent in par-ticular ethnic groups, such as BRCA1185delAG, BRCA1 5382insC, and BRCA26174delT in the Ashkenazim and BRCA2999del5 in Icelanders. However, the majorityof mutations are individually rare and manyhave been reported only in single families.Most recognized disease-associated mutationsresult in premature protein truncation and in-clude nonsense mutations, deletions/insertionsthat result in translational frameshifts, andmutations that affect splice sites. More recentlya number of exonic deletions/duplicationshave also been identified, especially in BRCA1.In addition, a large number of amino acidsubstitutions and synonymous nucleotidesubstitutions in BRCA1 and BRCA2 have been


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

Tab

le1

Sum

mar

yof

know

nbr

east

canc

erpr

edis

posi

tion

fact

ors

Gen

e/L

ocus

Rel

ativ

eR

isk

ofbr

east

canc

erC

arri

erFr

eque

ncy†

Bre

ast

canc

ersu

btyp

e

Oth

erca

ncer

sin

mon

oalle

licca

rrie

rsSy

ndro

me

inbi

alle

licca

rrie

rsM

etho

dof

iden

tific

atio

nH

igh

pene

tran

ceBR

CA

1>

100.

1%B

asal

(ER

-ne

gativ

e)

Ova

rian

Lin

kage

stud

y

BRC

A2

>10

0.1%

Ova

rian

pros

tate

Fanc

onia

naem

iaD

1L

inka

gest

udy

TP5

3>

10ra

reSa

rcom

asad

rena

lbr

ain

Can

dida

tere

sequ

enci

ngst

udy

Unc

erta

inpe

netr

ance

PTE

N2–

10ra

reT

hyro

iden

dom

etri

umL

inka

gest

udy

STK

112–

10ra

reG

asto

-int

estin

alL

inka

gest

udy

CD

H1

2–10

rare

lobu

lar

Gas

tric

(diff

use)

Lin

kage

stud

y

Inte

rmed

iate

pene

tran

ceAT

M2–

30.

4%A

taxi

ate

lang

iect

asia

Epi

dem

iolo

gy;C

andi

date

rese

quen

cing

stud

y

CH

EK

22–

30.

4%C

andi

date

rese

quen

cing

stud

y

BRIP

12–

30.

1%Fa

ncon

iana

emia

JC

andi

date

rese

quen

cing

stud

y

PALB

22–

4ra

reFa

ncon

iana

emia

NC

andi

date

rese

quen

cing

stud

y

Low pe

netr

ance

10q2

6,16

q12,

2q35

,8q

24,5

p12

1.08

–1.2

624

–50%

ER

-pos

itive

Gen

ome-

wid

eas

soci

atio

nst

udie

s

11p1

5,5q

111.

07–1

.13

28–3

0%G

enom

e-w

ide

asso

ciat

ion

stud

y

2q33

1.13

0.87

Can

dida

teas

soci

atio

nst

udy

† est

imat

edca

rrie

rfr

eque

ncy

ofm

utat

ions

/ris

kal

lele

inth

eU

K;w

here

‘rar

e’,t

heca

rrie

rfr

eque

ncy

isun

likel

yto

be>

0.1%

.


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

detected, the majority of which are innocuous.A few missense mutations, for example thosethat target cysteine residues in the BRCA1RING domain, abrogate function and are pre-sumed to confer risks comparable to truncatingmutations (43, 106).

There is evidence of genotype-phenotypecorrelations for mutations in BRCA1 andBRCA2. In both genes, the ratio of ovarian can-cers to breast cancers conferred by mutations inthe central region of the gene is higher than thatfor mutations in the 5′ or 3′ end. In BRCA1 thiscentral region is bounded by nucleotides 2401and 4191; in BRCA2 the ovarian cancer clus-ter region lies between nucleotides 3035 and6629 (114, 115). The biological mechanismsunderlying these observations remain opaque(120).

BRCA1 and BRCA2 are high penetrancebreast cancer genes. Estimates of the risks ofcancer conferred by mutations in these genesvary according to the ascertainment of the casesstudied. Early studies of large cancer familiessuggested that the risk of breast cancer by age70 may be as high as 87% [95% confidence in-terval (CI) = 72%–95%] for BRCA1 and 84%(43%–95%) for BRCA2 mutation carriers, al-though the upper estimates were based on rel-atively small numbers of families (44, 51). Inpopulation-based studies of breast cancer cases,unselected for family history, the risks are lower:65% (51%–75%) for BRCA1 and 45% (33%–54%) for BRCA2 (5). The pattern of age-relatedrisk in BRCA2 mutation carriers resembles thatof the general population (only higher). Bycontrast, the relative risk of breast cancer ismarkedly elevated in BRCA1 mutation carriersunder 40 and becomes less dramatic with ad-vancing age (5). Linkage analysis suggests thatmutations in BRCA1 and BRCA2 are responsi-ble for disease in approximately two-thirds oflarge families with site-specific female breastcancer (≥ four cases) but the attribution dimin-ishes sharply for smaller family clusters (51).The estimated population frequency of muta-tions in these genes is approximately 1/1000per gene in the United Kingdom. Overall this

equates to 15%–20% of the excess familial riskof breast cancer (4, 7, 9, 42, 91).

BRCA1 and BRCA2 are also high-penetrance ovarian cancer genes: Mutationsin BRCA1 confer a higher risk of ovariancancer than those in BRCA2, particularly forcarriers below 50 years of age (5, 7, 44, 51).BRCA1 and BRCA2 mutations account formost of the epidemiologically observed familialcoaggregation of breast and ovarian cancer (45,92). 95% of families containing four or morebreast cancers and an ovarian cancer werelinked to BRCA1 or BRCA2, whereas recentmutational screening detected mutations in83% of families containing at least two cases ofeach cancer (51, 99).

The relative risk of male breast cancer is el-evated for both genes, particularly BRCA2. Anelevated risk of prostate cancer has also beendemonstrated in BRCA2 carriers, particularlyin men aged <65 years (21, 116). Small ex-cesses of a number of other cancers have beenobserved in monoallelic (heterozygous) BRCA1and BRCA2 mutation carriers but larger stud-ies are required to clarify whether these find-ings reflect truly elevated risks of these cancers(21, 116). Biallelic mutations in BRCA2 result inFanconi anemia, subtype D1. Fanconi anemiais a rare, recessive chromosomal instability syn-drome characterized by skeletal abnormalities,bone marrow failure, and cancer predisposition.Subtype D1 has a distinctive phenotype that in-cludes a high risk of childhood solid tumors(66). Biallelic BRCA1 mutations have neverbeen reported convincingly in humans and arepresumed to be embryonically lethal (49).

BRCA1 tumors are typically high grade, in-vasive ductal carcinomas in which there is a highincidence of triple negative phenotype: nega-tive staining for ER (estrogen receptor), PR(progesterone receptor), and HER2 (ERBB2)(20, 73, 75). These tumors also frequently stainpositively for a subset of basal keratins char-acteristically expressed in the normal basalmyoepithelium of the breast. This basal phe-notype is distinctive and largely encompassesthe previous histological descriptions of BRCA1


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

tumors as medullary (69, 74). No distinctivehistopathological features have been describedin BRCA2 tumors.

TP53. Li-Fraumeni syndrome is a cancer pre-disposition syndrome in which there is a highfrequency of early onset breast cancer found inassociation with sarcomas and childhood can-cers of the adrenal cortex, brain, and other sites.Although of a penetrance sufficient for mappingthrough linkage analysis, the associated earlymortality and rarity of the condition impededcollection of sufficient familial samples. p53 wasrecognized early as a prominent transcriptionfactor central to multiple cellular pathways andis frequently somatically mutated in tumors.These observations recommended TP53 as aplausible candidate gene for Li-Fraumeni syn-drome and in 1990 mutational screening of thegene revealed causative mutations in the fivefamilies studied (82). The overall lifetime can-cer risk for women with Li-Fraumeni syndromeis grossly elevated, predominantly on accountof their high risk of breast cancer (15, 28).Li-Fraumeni syndrome is rare and mutationsin TP53 are uncommon in non-Li-Fraumenibreast cancer families (18, 48, 76). Thus, theattributable risk of TP53 mutations to familialbreast cancer is very low.

Breast Cancer Predisposition Genesof Uncertain Penetrance

Three syndromes are currently clearly associ-ated with an increased risk of breast cancer, butthe magnitude of the associated risk for eachremains uncertain: Cowden syndrome (causedby PTEN mutations), Peutz-Jeghers syndrome(caused by STK11 mutations), and Heredi-tary diffuse gastric cancer syndrome (caused byCDH1 mutations).

PTEN. Cowden syndrome is a multiplehamartoma syndrome that includes increasedrisk of benign and malignant tumors of thebreast, thyroid, and endometrium; distinctivehigher-penetrance features include mucocuta-neous lesions, macrocephaly, and hamartoma-

tous intestinal polyps. A study of 12 Cowdensyndrome families revealed linkage to chromo-some 10q, leading to identification of PTENas the causative gene (87, 88). PTEN encodes alipid phosphatase that functions as a tumor sup-pressor through negative regulation of a cell-survival signaling pathway. There is cross talkbetween this PTEN-related pathway and otherpathways, including those involving Ras, p53,and TOR (36).

STK11. Peutz-Jeghers syndrome is character-ized by hamartomatous intestinal polyps, mu-cocutaneous pigmentation, and increased inci-dence of several malignancies, including breastcancer. Directed by patterns of loss of heterozy-gosity in polyps of affected individuals, studiesin 12 Peutz-Jeghers families established link-age to chromosome 19p and positional cloningled to the identification of STK11 (LKB1) asthe causative gene. STK11 is a serine/threoninekinase that inhibits cellular proliferation, con-trols cell polarity, and interacts with the TORpathway (1, 63, 64, 70).

CDH1. Linkage analysis in a single NewZealand Maori family with multiple cases ofdiffuse gastric cancer enabled the identificationof CDH1 (ECAD) as the responsible gene (58).Many additional families have since been re-ported and an elevated frequency of lobularbreast carcinoma has been observed (71, 96).Occasional CDH1 mutations in families withcases of lobular breast cancer but no gastric can-cer have also been reported (83). CDH1 encodesE-cadherin, a transmembrane protein impor-tant in the maintenance of cell polarity.

These conditions were identified becausethe non-breast-cancer-related features weresufficiently distinctive and penetrant to reli-ably ascertain families and assign affection sta-tus, thus facilitating mapping of the causativegenes by linkage analysis. The true breast can-cer risks associated with mutations are unclearand it is possible that the published risks are in-flated through the bias of studying families withprominent phenotypes. However, the relativerisks are likely to be intermediate and in the


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

range of 2 to 10. There is no current evidencethat mutations in PTEN, STK11, or CDH1 ac-count for a substantial proportion of familial orsporadic breast cancer in the absence of theirrespective syndromes; because these syndromesare very rare, the attributable risk of mutationsin these genes to familial breast cancer is low(14, 27, 57, 83, 107).

Intermediate-Penetrance BreastCancer Predisposition Genes

Four intermediate-penetrance breast cancergenes have been identified via mutationalscreening: CHEK2, ATM, BRIP1, and PALB2.Mutations in these genes are rare and confer arelative risk of breast cancer of 2 to 4. RAD50may also be an intermediate-penetrance breastcancer predisposition gene, but convincing re-sults to date pertain only to a founder mutationdetected in the Finnish population. The rar-ity of mutations and modest associated risks aresuch that the attributable risk of mutations inthese genes is low: Together they account forapproximately 2.3% of excess familial risk (98).

CHEK2. CHEK2 encodes CHK2, a centralmediator of cellular response to DNA damagethat phosphorylates both p53 and BRCA1 toregulate repair of DNA double-strand breaks.The 1100delC mutation in CHEK2 was firstreported after mutational screening of this el-igible candidate gene in a single family withfeatures of Li-Fraumeni syndrome (13). How-ever, the population frequency of this mutationwas shown to be approximately 1% (18/1620),demonstrating that CHEK2 could not be ahigh-risk Li-Fraumeni syndrome gene. Theidentification of the CHEK2 1100delC muta-tion in 4.2% (30/718) of breast cancer fami-lies demonstrated, instead, that CHEK2 was anintermediate-penetrance breast cancer predis-position gene (P = 5 × 10−6) (85). Segregationanalysis in these families provided an indirectestimate of 1.70 (95% CI = 1.32–2.20) for therelative risk of breast cancer conferred by themutation. Combining data from ten case con-

trol studies, the CHEK2 consortium demon-strated the frequency of CHEK2 1100delC inpopulation-based breast cancer cases to be1.9% (201/10,860) compared with 0.7% in con-trols (64/9065) (P = 1 × 10−7). These fre-quencies equate to a direct odds ratio of breastcancer of 2.34 (95% CI = 1.72–3.20) (26). Ithas been proposed that CHEK2 1100delC mayconfer risks of prostate and other cancers butevidence is conflicting (37, 38, 85, 105, 117).There are also reports of other rare truncatingmutations in CHEK2, which would be antici-pated to have similar risks to CHEK2 1100delC(17, 39, 55, 103, 119). A number of rare CHEK2missense variants, such as 470T/C (I157T),have been reported in breast cancer cases buttheir significance has yet to be clarified (2, 17,55, 72, 103).

ATM. The proposal of ATM as a breast can-cer predisposition gene first came in 1976 froman epidemiological study that reported an ex-cess of breast cancer in female relatives of pa-tients with ataxia telangiectasia, an autosomalrecessive syndrome characterized by progres-sive cerebellar ataxia, immune deficiency, andcancer predisposition. This astute observationpreceded the mapping of the gene by almosttwo decades and has been subsequently repli-cated in a number of large epidemiological stud-ies (41, 102, 112, 113). The candidacy of ATMas a breast cancer predisposition gene was fur-ther enhanced as the function of the encodedprotein became apparent: ATM occupies a cen-tral role in the response to double-strand DNAbreaks through initiation of a signaling cascadethat involves phosphorylation of multiple pro-teins including p53, BRCA1, and CHK2. Re-sults from initial mutational screening of ATMin breast cancer were inconclusive and/or in-consistent but convincing proof was finally pub-lished in 2006. Mutations were found in 12/443familial cases negative for mutations in BRCA1and BRCA2 and 2/521 controls (P = 0.0047).These mutations comprised truncations, splice-site abnormalities, and two missense mutationsthat were known to affect protein function and


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

cause ataxia telangiectasia. The relative risk ofbreast cancer conferred by ATM mutations wasestimated from segregation analysis to be 2.37(95% CI = 1.51–3.78, P = 0.0003) (101). Thisindirect estimate is very similar to that derivedfrom large epidemiological studies (113). Thereis some epidemiological evidence that muta-tions in ATM may predispose to other cancersbut this has yet to be confirmed or corroboratedby molecular data (113).

BRIP1. BRIP1 (BACH1) encodes a DEAHhelicase that interacts with BRCA1 and hasBRCA1-dependent roles in DNA repair andcheckpoint control (24, 90). In 2006, truncat-ing BRIP1 mutations were reported in 9/1212index breast cancer cases from families nega-tive for mutations in BRCA1 and BRCA2 com-pared with 2/2081 controls, thus providingconvincing evidence that BRIP1 is a breast can-cer predisposition gene (P = 0.0030). Segre-gation analysis found the relative risk conferredby BRIP1 mutations to be 2.0 (95% CI = 1.2–3.2, P = 0.012) (104). Of the mutations re-ported in BRIP1, approximately half are a non-sense mutation, 2392C/T (R798X), whereasthe remainder include disparate small insertionsand deletions. Concurrently, it emerged thatbiallelic mutations in BRIP1 result in Fanconianemia, although the Fanconi anemia subtypeassociated with BRIP1 (subtype J) is phenotyp-ically distinct from that associated with BRCA2(subtype D1) and has not been associated withchildhood solid tumors (78, 79, 81).

PALB2. Precipitation of BRCA2-containingcomplexes revealed a novel protein that wasshown to promote the localization and stabilityof BRCA2, thus facilitating BRCA2-mediatedDNA repair. Further study of PALB2 (partnerand localizer of BRCA2) revealed that knock-down of the gene resulted in sensitization ofcells to chromosomal damage by mitomycin C,the hallmark of Fanconi anemia (125). Biallelictruncating mutations in PALB2 were detectedin families with Fanconi anemia not caused bymutations in other genes and this subtype was

designated FA-N (100, 124). The phenotypesof FA-N and Fanconi anemia caused by BRCA2mutations (FA-D1) are very similar and canbe distinguished from classical Fanconi anemiaon account of a markedly elevated frequencyof childhood solid tumors such as Wilms tu-mor and medulloblastoma. With the recentprecedents of both BRIP1 and BRCA2 caus-ing Fanconi anemia in biallelic mutation carri-ers and conferring susceptibility to breast can-cer in monoallelic carriers, the identification ofPALB2 as a Fanconi anemia gene further rec-ommended PALB2 as an attractive candidatebreast cancer predisposition gene.

Sequencing of the gene revealed truncat-ing mutations in 10/923 index breast cancercases from families negative for mutations inBRCA1 and BRCA2 compared with 0/1084 incontrols (P = 0.0004). Segregation analysisfrom these families estimated the relative risk ofPALB2 mutations to be 2.3 (95% CI = 1.4–3.9,P = 0.0025) (98). Mutational screening ofPALB2 in a Finnish series detected one trun-cating mutation: 1592delT. This mutationwas identified in 3/113 (2.7%) familial cases,18/1918 (0.9%) unselected breast cancer cases,and 6/2501 (0.2%) controls. The directly de-rived odds ratio of breast cancer of this mu-tation is therefore 3.94 (95% CI = 1.5–12.1)(47). This mutation has not been detectedin other series and may represent a Finnishfounder. A French Canadian founder mutation,2323 C/T (Q775X), has also been reported(52).

RAD50. The highly conserved MRN com-plex (MRE11, RAD50, and NBS1) interacts withBRCA1 and plays a central role in DNA re-pair. A truncating mutation in RAD50, 657delT,was demonstrated in 8/317 consecutively ascer-tained breast cancer cases and 6/1000 controlsfrom Finland [P = 0.008, odds ratio (OR) =4.3, 95% CI = 1.5–12.5]. Other rare truncat-ing mutations have been reported in RAD50,but the contribution of RAD50 to breast can-cer predisposition outside of Finland requiresfurther clarification (62, 118).


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

Distinctive features of intermediate-penetrance breast cancer predispositiongenes. As in BRCA1 and BRCA2, multi-ple different truncating mutations occur inthese intermediate-penetrance genes, thefrequencies of which differ between popula-tions on account of founder and migrationaleffects. Monoallelic mutations in these genesconfer an approximately twofold increasein the risk of breast cancer. Although thereis some imprecision in the indirect risk es-timates derived from segregation analysis,population-based series and epidemiologicaldata have provided some corroboration forthese figures. There is currently no strongevidence that monoallelic mutations in thesegenes confer a phenotype beyond breastcancer predisposition. Biallelic mutations inATM, BRIP1, and PALB2 result in severechildhood disorders: ataxia telangiectasia andFanconi anemia types J and N, respectively(97). Because these genes function in the samepathways as BRCA1 and BRCA2, it remainsunclear why abrogation of their functionsconfers more modest risks of breast cancer.However, the lower penetrance of these genesmeans that they deviate from some of thehallmarks of pathogenicity associated withmutations of BRCA1 and BRCA2. Mutationsin these genes do not show the classical,near-complete pattern of segregation of diseasewith the family mutation. By contrast, becausebreast cancer is a common disease, in a typicalpedigree an appreciable proportion of breastcancers arise by chance and thus may manifestin either mutation carriers or noncarriers.Likewise, one should not anticipate findingthe pattern of loss of the wild-type alleledetected at high frequency in tumors in BRCA1and BRCA2 mutation-positive individuals.Firstly, the mechanism of tumorigenesis ofthese genes is unknown and may not requireinactivation of the wild-type allele. Secondly,even if the classical, two-hit model of a tumorsuppressor gene does apply, one could notexpect to demonstrate loss of heterozygosityin all tumors occurring in mutation carriers

because the mutation is not causally implicatedin a significant portion of them.

Low-Penetrance Breast CancerPredisposition Alleles

There is currently strong evidence for the asso-ciation with breast cancer of eight common al-leles, which each confer a relative risk of breastcancer of <1.5. A nonsynonymous variant inCASP8 was identified through a candidate-based approach in a consortium comprising 14studies that compared 17,109 cases and 16,423controls (34). Seven further variants have beendetected in three recent genome-wide associ-ation studies (Table 2). Easton and coworkers(46) performed a three-stage genome-wide ex-periment. In the first stage a Perlegen platformwas used to study the association of 227,876SNPs in 390 cases with a family history of breastcancer and/or bilateral disease and 364 con-trols from the United Kingdom. In the secondstage 12,711 SNPs (the most significant 5% ofSNPs from stage 1) were genotyped in 3990cases and 3916 UK controls using a custom-designed array. In the third stage the 30 mostsignificant SNPs were tested for confirmationin 21,860 cases and 22,578 controls from 22international studies (46). The study identifiedfive variants associated with breast cancer at asignificance level of P < 10−7. Stacey and col-leagues (110) conducted a genome-wide scanin 1600 Icelandic cases and 11,563 controls us-ing the Illumina HumanHap300 platform; theten best associated SNPs were studied furtherin five international replication sets comprising2954 cases and 6014 controls. This study iden-tified one of the variants reported by Eastonand coworkers (rs3803662) and a novel regionof association on 2q35. As part of the NationalCancer Institute Cancer Genetic Markersof Susceptibility (CGEMS) Project, Hunterand colleagues (67) used the Illumina Hu-manHap500 array to genotype 1145 post-menopausal invasive breast cancer cases and1142 North American controls of Europeanancestry. The six best-associated signals were


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

Tab

le2

Sum

mar

yof

know

nlo

w-p

enet

ranc

ebr

east

canc

erpr

edis

posi

tion

vari

ants

∗

Loc

us

Est

imat

edsi

zeof

LD

bloc

k(k

b)

Gen

esw

ithi

nL

Dbl

ock

SNP

MA

FH

eter

ozyg

ote

OR

(95%

CI)

Hom

ozyg

ote

OR

(95%

CI)

Per

alle

leO

R(9

5%C

I)P

-tr

end

Asc

erta

in-

men

tSt

udy

10q2

625

FGFR

2rs

2981

582

0.38

1.23

(1.1

8–1.

28)

1.63

(1.5

3–1.

72)

1.26

(1.2

3–1.

30)

2×

10−7

6G

WE

asto

net

al.

(UK

)(46

)

rs12

1964

80.

391.

20(1

.07–

1.42

)1.

64(1

.42–

1.90

)1.

1×

10−1

0G

WH

unte

ret

al.

(USA

)(67

)

16q1

216

0T

NR

C9

LOC

6437

14rs

3803

662

0.25

1.23

(1.1

8–1.

29)

1.39

(1.2

6–1.

45)

1.20

(1.1

6–1.

24)

1×

10−3

6G

WE

asto

net

al.

(UK

)

rs38

0366

20.

271.

27(1

.19–

1.36

)1.

64(1

.45–

1.85

)1.

28(1

.21–

1.35

)5.

9×

10−1

9G

WSt

acey

etal

.(I

cela

nd)(

110)

2q35

117

–rs

1338

7042

0.50

1.11

(1.0

3–1.

20)

1.44

(1.3

0–1.

58)

1.20

(1.1

4–1.

26)

1.3

×10

−13

GW

Stac

eyet

al.

(Ice

land

)(11

0)

5p12

310

MR

PS30

rs10

9416

790.

241.

19(1

.13–

1.26

)2.

9×

10−1

1G

WSt

acey

etal

.(I

cela

nd)

(111

);E

asto

net

al.

(UK

)(46

)

5q11

280

MA

P3K

1M

GC

3364

8M

IER

3

rs88

9312

0.28

1.13

(1.0

9–1.

18)

1.27

(1.1

9–1.

36)

1.13

(1.1

0–1.

16)

7×

10−2

0G

WE

asto

net

al.

(UK

)(46

)

2q33

290

CA

SP8

rs10

4548

50.

130.

89(0

.85–

0.94

)0.

74(0

.62–

0.87

)0.

88(0

.84–

0.92

)1.

1×

10−7

Can

dida

teC

oxet

al.(

UK

)(3

4)T

RA

K2

ALS

2CR

12A

LS2C

R2

ALS

2CR

11LO

C38

9286

LOC

7291

91

8q24

110

–rs

1328

1615

0.40

1.06

(1.0

1–1.

11)

1.18

(1.1

–1.2

5)1.

08(1

.05–

1.11

)5

×10

−12

GW

Eas

ton

etal

.(U

K)(

46)

11p1

518

0LS

P1rs

3817

198

0.30

1.06

(1.0

2–1.

11)

1.17

(1.0

8–1.

25)

1.07

(1.0

4–1.

11)

3×

10−9

GW

Eas

ton

etal

.(U

K)(

46)

TN

NT

3M

RPL

23H

19LO

C72

8008

∗ Abb

revi

atio

nsus

ed:L

D,l

inka

gedi

sequ

ilibr

ium

;MA

F,m

inor

alle

lefr

eque

ncy

inco

ntro

lpop

ulat

ion;

OR

,odd

sra

tio(m

easu

red

rela

tive

toco

mm

onho

moz

ygot

es);

CI,

confi

denc

ein

terv

al;

P-tr

end,

Pva

lue

forp

eral

lele

effe

ctun

derm

ultip

licat

ive

mod

el;G

W,g

enom

e-w

ide

asso

ciat

ion

stud

y;C

andi

date

,ass

ocia

tion

stud

ieso

fsin

gle

nucl

eotid

epo

lym

orph

ism

s(SN

Ps)

inca

ndid

ate

gene

s.


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

further studied in 1776 cases and 2072 con-trols from three North American replicationsets. The results of this study confirmed the sig-nal in intron 2 of FGFR2 reported by Eastonand coworkers (67). The final predispositionSNP on 5p12 was identified through miningand cross-referencing of associations of border-line significance from these three genome-wideassociation studies; this signal was verified byStacey and coworkers (111) in a replication se-ries of 5,028 cases and 32,090 controls.

10q26 (FGFR2 intron 2). The strongest evi-dence for association was for a signal within in-tron 2 of the gene encoding fibroblast growthfactor receptor 2 (FGFR2) [per allele OR =1.26 (1.23–1.30) P = 2 × 10−76] (46, 67, 111).Fine association mapping of the region of link-age disequilibrium has been undertaken usingUK samples and refined using an Asian series (inwhich linkage disequilibrium is weaker). Thesestudies have reduced the association to a mini-mum set of six variants within intron 2, whichare too strongly correlated for their individualeffects to be discerned using further genetic epi-demiologic approaches (46).

Somatic FGFR2 mutations have been re-ported in several cancers and result in overac-tivity of the protein. It therefore seems plausiblethat the elevated breast cancer risk is somehowmediated through increased/altered activity ofFGFR2, but the molecular basis for the associ-ation is currently unknown.

The risk-prevalence profile of this allelemeans that the power to detect it in the genome-wide scans is high. The corollary of this is thatif other alleles of a similar risk-prevalence pro-file exist, they should most likely have been de-tected. Thus, it is possible that this may be theallele of most substantial effect to exist withinthis category of risk alleles. The differencesin the risks conferred are nevertheless modest:The risk of breast cancer by age 70 in individualshomozygous for the risk allele is 10.5% com-pared with 6.7% for heterozygotes and 5.5%for nonrisk allele homozygotes (based upon UKbreast cancer incidence figures). However, therisk allele is common: 14% of the UK popu-

lation and 19% of UK breast cancer cases arehomozygous for the risk allele. Hence, this al-lele may account for approximately 1.9% of theexcess familial risk of breast cancer.

16q12. Strong evidence for association ofrs3803662 with breast cancer was reported intwo studies [per allele OR = 1.20 (1.16–1.24),P = 1 × 10−36 in the UK study and per alleleOR = 1.28 (1.21–1.35), P = 5.9 × 10−19 in theIcelandic study] (46, 110). This variant lies on16q12 and tags a region of linkage disequilib-rium containing the 5′ end of TNRC9 (TOX3)and a hypothetical gene LOC643714. TNRC9is a gene of uncertain function that contains aputative high mobility group box motif, sug-gesting that it may act as a transcription factor.Results from a previous study suggested thatTNRC9 expression is predictive of metastasis ofbreast cancer to bone (108).

2q35. The Icelandic study found evidence forassociation with breast cancer of a variant on2q35 [per allele OR = 1.20 (1.14–1.26), P =1.3 × 10−13]. The region of linkage disequi-librium does not contain any known genes.The nearest genes are TNP1 (181 kb), IGFBP5(345 kb), and IGFBP2 (376 kb) upstream andTNS1 (761 kb) downstream. Although this sig-nal was not detected in the other genome-widescans, we have replicated the association in aUK series of familial breast cancer cases [perallele OR = 1.17 (95% CI 1.07–1.27), P =0.0004] (C. Turnbull and N. Rahman, unpub-lished data) and it is being evaluated by theBreast Cancer Association Consortium.

5p12. Stacey and coworkers (111) noted thatone of the ten top-ranked SNPs in the Ice-landic genome-wide study was on chromo-some 5p12 (rs7703618). The CGEMS studyhad reported an association of borderline sig-nificance on 5p12 at rs4866929, a SNP in tightlinkage disequilibrium with rs7703618, whileEaston and coworkers (46) had reported a ten-tative signal at 5p45 (rs981782), 371 kb away,which was also in loose linkage disequilibriumwith rs7703618 (r2 = 0.10). Struck by the


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

coincidence of tentative signals, they examined21 SNPs in this region for association withbreast cancer and found rs10941679 to be themost significantly associated SNP [per alleleOR 1.19 (1.13–1.26), P = 2.9 × 10−11]. Theonly gene in the region of linkage disequilib-rium is MRPS30 (PDCD9, programmed celldeath protein 9), which encodes a componentof the small subunit of the mitochondrial ribo-some and has been implicated in apoptosis.

5q11. Association has been found for rs889312[per allele OR = 1.13 (1.10–1.16), P = 7 ×10−20]. This SNP tags a 280-kb block of linkagedisequilibrium on 5q11 that contains the genesMAP3K1 (MEKK), MGC33648, and MIER3.MAP3K1 is the most plausible candidate genetherein and encodes mitogen-activated proteinkinase kinase kinase 1, which is involved in cellsignaling (46).

2q33 (CASP8 D302H). This association wasascertained through the biological candidacy ofCASP8, which encodes caspase-8, a protein in-volved in apoptosis [per allele OR = 0.88 (0.84–0.92), P = 1.1 × 10−7] (34). However, the re-gion of linkage disequilibrium tagged by thisSNP is approximately 290 kb in length and in-cludes a number of other genes: TRAK2 (en-coding trafficking protein kinesin binding 2),three genes identified as candidates for juve-nile amyotrophic lateral sclerosis 2 (ALS2CR12,ALS2CR2, and ALS2CR11), and two hypothet-ical genes (LOC389286 and LOC729191). It iscurrently unclear whether D302H is the causalvariant or whether it tags a distinct causal vari-ant, which may or may not mediate its effectthrough CASP8. The relatively low minor al-lele frequency and risk of this allele mean thatthe power to detect its signal at the genome-wide level is comparatively low. This is con-sistent with failure of the allele to reach thearbitrary significance thresholds required forfurther investigation in the genome-wide asso-ciation scans performed to date. It is also con-sistent with the presumed existence of manyfurther alleles of comparable risk-prevalenceprofiles.

8q24. Association was found for a SNP withinan 110-kb block of linkage disequilibrium on8q24, which contains no known genes [per al-lele OR = 1.08 (1.05–1.11), P = 5 × 10−12](46). It is interesting that in the first wave ofgenome-wide association studies in commoncancers, 8q24 has also yielded multiple inde-pendent prostate cancer loci, one of which isalso a colon cancer risk allele (rs6983267) (59,60). The tag SNP associated with breast can-cer does not demonstrate association with coloncancer or prostate cancer and is 60 kb proximalto rs6983267. Clustering of these predisposi-tion alleles may be a coincidence or may indi-cate a common or related mechanism of cancerpredisposition. The nearest gene to the breastcancer predisposition locus is MYC and it isplausible that the susceptibility occurs throughsome unknown mechanism of activation of thisoncogene.

11p15. Another associated tag SNP lies inintron 10 of LSP1 (WP43), which encodeslymphocyte-specific protein 1 [per allele OR =1.07 (1.04–1.11), P = 3 × 10−9], an F-actinbundling cytoskeleton protein that is expressedin hematopoetic and endothelial cells (46).Other genes within this region of linkage dis-equilibrium include TNNT3, troponin T type3; MRPL23, mitochondrial ribosomal protein;H19, an imprinted, maternally expressed, un-translated mRNA; and LOC728008, a hypo-thetical gene.

Distinctive features of low-penetrancebreast cancer predisposition alleles. A fasci-nating aspect of the recent genome-wide scanshas been the opaqueness of the relationshipof the identified variants with known protein-coding genes. The patterns of linkage disequi-librium suggest that the causal variants neednot lie in the coding region of a gene, as evi-denced by the predisposition variants at 2q and8q that are tens of kilobases away from the near-est protein-encoding gene. Any significance tothe roles of genes upstream or downstream ofthe associated blocks of linkage disequilibriumis currently speculative.


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

The interpretation of subgroup analyses ofthe immunohistopathologic phenotypic varia-tion of risk alleles requires caution on account ofthe limited power of some analyses. Initial anal-yses by the Icelandic group (110, 111), on iden-tification of the predisposition SNPs at 16q12,2q35, and 5p12, showed the effects of thesealleles to be confined to ER-positive tumors.Although subsequent analyses of the SNPs at10q26, 16q12, 8q24, 5q11, and 11p15 fromthe United Kingdom genome-wide associationdata also revealed that for all five alleles, the es-timates of effect were stronger in ER-positivethan ER-negative disease, the difference wasonly significant for those at 10q26 and 8q24.These two SNPs were also independently asso-ciated with a lower grade of disease. Adjustedanalyses of these five SNPs did not reveal sig-nificant association with lymph node status norsurvival (53).

Because these alleles occur at a high fre-quency in the general population, their pop-ulation attributable risks (etiologic fractions)are relatively high (13%–16% for the alleles ofstronger effects) (67, 110). However, this figurerepresents only the proportion of breast can-cer cases in which the variant has played somecausal role in development of disease. The as-sociated risks are low and it is estimated that thefive loci characterized by Easton and cowork-ers (46) account for a modest 3.6% of the excessfamilial risk of breast cancer in European pop-ulations.

INTERACTIONS BETWEENBREAST CANCERPREDISPOSITION FACTORS

Interaction is an important aspect of risk cal-culation, particularly in familial disease clus-ters in which multiple predisposition factors arelikely active. The combined risk of two fac-tors is strongly dependent upon the nature ofthe interaction between them. Typically the de-fault model assumes that they act independentlyand multiplicatively (as per terms in a multi-ple regression analysis). The true situation maybe more complex and comprise a mixture of

multiplicative, additive, antagonistic, synergis-tic, and/or complex intermediate interactions.

The study of co-occurrence of genetic pre-disposition factors and exposition of gene-gene interactions has become more viable sincethe identification of convincing common pre-disposition alleles. Preliminary analyses fromgenome-wide association studies suggest thatthe low-penetrance risk alleles act multiplica-tively with each other (46, 110). However, as-sociation studies of the SNPs at 10q26 and5q11 performed within BRCA1 and BRCA2mutation-positive families suggest that theseSNPs confer additional risk in the presence ofBRCA2 but not BRCA1 mutations (12). Thisinteresting disparity may in part reflect the as-sociation of the SNPs with ER-positive tumorsbecause BRCA1 is typically associated with ER-negative tumors. In a recent candidate-basedexperiment, it was reported that homozygos-ity for a variant in the 5′ untranslated region ofRAD51 confers an increased risk of breast can-cer to BRCA2 mutation carriers [hazard ratio =3.18 (95% CI = 1.39–7.27)]. The modify-ing effect was not significant in BRCA1 mu-tation carriers or when only a single copy ofthe risk allele was present (11). By contrast,CHEK2 1100delC has not been shown to con-fer an elevated risk on the background of mu-tations in either BRCA1 or BRCA2 (85). It hasbeen proposed that this reflects the commonpathway of the encoded proteins; abrogation ofCHK2 function might have little additional im-pact on a pathway already radically subvertedby a mutation in BRCA1 or BRCA2. However,there is currently no biological proof of this hy-pothesis. Further analyses are required to estab-lish whether similar interactions are observedfor ATM, BRIP1, and PALB2, which also inter-act with BRCA1 and/or BRCA2 in DNA repairpathways.

Clarity regarding gene-environment inter-actions is even more limited. Recognized en-vironmental risk factors for breast cancer inthe general population are well establishedand predominantly relate to estrogen expo-sure. Endogenous estrogen-related risk factorsinclude timing of menarche and menopause,


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

parity, age of first live birth, and breast-feeding,whereas exogenous factors include adminis-tration of contraceptives and hormone re-placement therapy (30, 31, 33). Studies ofhormonal/reproductive factors in BRCA1 andBRCA2 mutation carriers are challenging, notjust in the assembly of sufficient families, butbecause of the biases inherent in observa-tion of the behaviors of a group of individ-uals who know themselves to be at elevatedrisk. However, well-powered studies of gene-environment interactions are becoming possi-ble through collaboration and are allowing theeffects of BRCA1 and BRCA2 to be studiedindependently. Significant alteration of breastcancer risk in BRCA1 and BRCA2 mutation car-riers has been demonstrated for oral contra-ceptive usage, age at first pregnancy, and de-gree of parity, whereas significant effects werenot demonstrated for age of menarche, ageof natural menopause, nulliparity, and breastfeeding (3, 23, 25). These observations war-rant further investigation and confirmation inprospective studies. Other large studies will berequired to explore gene-environment interac-tions for genetic predisposition factors of lowerpenetrance.

IDENTIFICATION OF FURTHERGENETIC BREAST CANCERPREDISPOSITION FACTORS

There is a clear discontinuity in the risks asso-ciated with the three categories of breast can-cer predisposition factors identified to date:The high-penetrance genes confer a risk thatis elevated >10-fold, the known intermediate-penetrance genes 2–4-fold, and the low-penetrance alleles <1.5-fold. There may besome biological significance to these strata orthey may be an artifact of the limited meth-ods of ascertainment. More than 70% of ge-netic predisposition to breast cancer remainsunaccounted for and as further genetic predis-position factors are identified, new categoriesmay emerge and/or the apparent distinctionsbetween extant classes may blur or disappear.Nevertheless, risk and prevalence provide a use-

ful framework for considering how technolog-ical and intellectual advances may allow exten-sion of the repertoire of breast cancer predis-position factors.

High-Penetrance Genes

Any further high-penetrance dominant predis-position genes are likely to be very rare causesof familial breast cancer. Segregation analysesgenerated no evidence for further dominantgenes of a risk-penetrance profile comparableto BRCA1 or BRCA2 and this has been corrobo-rated by linkage studies. In a recently publishedexample, genome-wide linkage analysis was un-dertaken in 149 nonsyndromic breast cancerfamilies negative for mutations in BRCA1 andBRCA2 and the number of linkage peaks de-tected under parametric (dominant and reces-sive) and nonparametric (allele-sharing) modelsdid not differ significantly from that expected bychance (109). However, linkage analyses suchas this have been undertaken in families largelyascertained on the basis of multiple-generationbreast cancer pedigrees. A highly penetrant re-cessive predisposition gene would not producethis pattern of disease in families. Two inde-pendent segregation analyses found evidenceof a recessive pattern of inheritance (8, 35).Thus, there may be utility in genome-wide re-cessive linkage analyses in more suitable series,in particular families from understudied popu-lation isolates, and/or in families with higherlevels of consanguinity. Further, unidentifiedbreast cancer-associated syndromes may alsoexist. However, if these syndromes have not yetcome to medical/scientific attention, they arelikely to be very rare; linkage is likely to repre-sent the optimal method for mapping any suchunderlying predisposition genes.

Intermediate-Penetrance Genes

Further intermediate-penetrance breast cancerpredisposition genes likely exist, although it isdifficult to predict how many there may beand what proportion of the total excess famil-ial risk is attributable to them. Resequencing


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

is currently the optimal approach by whichto identify genes of this nature. The knownintermediate-penetrance genes are all involvedin DNA repair and function in pathwayswith BRCA1 and/or BRCA2; this common-ality may reflect the underlying biology ofthis class or may just be an artifact of thegroups of genes that have been most in-tensely investigated. Further genes involvedin DNA repair and other relevant path-ways represent plausible candidates and arebeing investigated by us and other groups.However, technological advances are facilitat-ing increasingly high-throughput mutationalscreening such that genome-wide resequencingis becoming viable. This will allow interroga-tion of regions of the genome not previously in-vestigated and may reveal further intermediate-penetrance genes that function in pathways notpredictable from current paradigms.

Intermediate-penetrance predisposition tobreast cancer may also be an unrecognized com-ponent of known pleomorphic cancer predis-position syndromes. Large, collaborative epi-demiological studies of rare syndromes canoptimize power and minimize bias and mayrepresent the optimal strategy by which to es-tablish accurate estimates of these breast can-cer risks. Epidemiological studies of reces-sive syndromes, particularly those associatedwith childhood cancer, may reveal elevated fre-quency of breast cancers in relatives of affectedindividuals and may lead to the identification offurther intermediate-penetrance breast cancergenes, as in the case of ATM.

Low-Penetrance Variants

Common low-penetrance variants. Ex-tant genome-wide association data offerclear evidence that many further commonlow-penetrance breast cancer predispositionvariants exist. Further risk alleles may be iden-tified from these data by mining the variantsof borderline significance for further signalsand through the use of imputation techniques.However, comprehensive study of this (likely)extensive repertoire of low-penetrance alleles

will require further, larger, genome-wideassociation experiments. The effect sizes ofsubsequent rounds of risk alleles are likely tobe of progressively diminishing magnitudeand will require commensurate increases inpower for detection. It is currently unclearwhat proportion of common low-penetrancealleles will be detectable by the feasible studiesin the immediate future.

Strongly associated tag SNPs from genome-wide scans may be utilized for genetic epidemi-ologic analyses and clinical risk estimation intheir own right. However, the way in whichthese common variants contribute to cancer islargely unknown and exploration of the biolog-ical mechanisms that underlie these signals of-fers exciting new avenues of study. Resequenc-ing and fine-association mapping of the regiontagged by a reporter SNP can refine the asso-ciation to a minimum set of SNPs in a fixedblock of linkage disequilibrium. The cancer-causing components within these blocks arecryptic. Attempts to identify them may chal-lenge current paradigms of the relationship be-tween genes and disease and require innovativemethods.

To date, genome-wide studies have been un-dertaken in simple series of breast cancer casesof European descent. Novel risk alleles maybe identifiable through genome-wide studiesof different ethnic groups in which the minorallele frequencies and phenotypic spectrum ofdisease differ from Europeans. For example, arisk locus at 6q22 was identified in a recentgenome-wide study performed exclusively inAshkenazi Jewish breast cancer cases and con-trols; verification of this association in otherpopulations is awaited (54). Broader insight intogenetic etiology may be gained through the ex-pansion of genome-wide studies into subgroupsof breast cancer cases. For example, analysis ofBRCA-positive individuals may allow identifi-cation of modifier alleles. Studies focusing onwell-characterized histopathological subgroupsof breast cancers may advance our understand-ing of the genetic basis of disease heterogene-ity. The use of quantitative intermediate phe-notypes, such as mammographic density, as the


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

outcome for genome-wide studies may also rep-resent an alternative approach.

Rare low-penetrance variants. The risk-prevalence profiles already identified suggestthat rare variants of low penetrance representanother plausible category of breast cancer pre-disposition factor. However, there are currentlyno reliable methods by which to quantify therisk of breast cancer associated with individual,rare variants. Therefore, although it is possi-ble that some of the rare variants detected dur-ing mutational screening experiments are low-penetrance predisposition factors, our ability todemonstrate this is limited. Genome-wide re-sequencing will inevitably detect many furtherrare variants, some of which will likely be asso-ciated with small increases in risk of breast can-cer. Developing robust methods for the iden-tification of the cancer-associated rare variantsamong the numerous innocuous variants will bea major challenge for the future.

Optimizing the Power of GeneIdentification Studies

A priority for all future studies will likely bethe harnessing of sufficient statistical power atacceptable cost. An important tool for gain-ing power without increasing the experimentsize/cost is to assemble a case series enriched forgenetic predisposition. The best-recognized ofthese enrichment parameters, successfully usedin many experiments to date, include familyhistory, bilaterality, and early age of onset ofdisease. Their relative efficacies have been com-pared through modeling the sample size re-quired to demonstrate association with diseaseof a variant of specified frequency and effect.The use of cases with a single affected first-degree relative affords a greater than twofoldreduction in the required sample size; for caseswith two affected first-degree relatives the re-duction is more than fourfold. The sample sizereduction when using bilateral cases is also four-fold (and thus equivalent to using cases with twofirst-degree relatives). However, the sample sizerequired for an association study that uses cases

diagnosed at age 35 is only 40% less than onethat uses cases diagnosed at 65 years. Relativeefficiencies are all magnified for experiments inwhich rarer alleles are studied (6).

An alternative strategy is geographical en-richment, namely to first screen genes in pop-ulation isolates in which a higher prevalence offounder mutations might be anticipated, such asthe Ashkenazim, the Finnish, or the Icelanders.This has proved very successful for the iden-tification and characterization of known genes.However, as is currently the case for RAD50, thegeneralizability of findings from specific popu-lations may prove challenging.

There has been longstanding expectationthat evolution in molecular profiling may re-sult in pathological classifications that moredirectly reflect the genetic etiology of tu-mors. Although the basal phenotype clearly en-riches for cases that arise because of BRCA1mutations, it remains unclear whether novelimmunohistopathological profiles or otherphenotypic surrogates may emerge that can dis-tinguish further subsets of breast cancers thatoccur because of genetic predisposition.

CLINICAL TRANSLATION

Risk Estimation and Management

Risk estimation is currently the primary clini-cal application for genetic factors that predis-pose to breast cancer. Ongoing improvementof clinical risk assessment tools is necessary toensure that individuals at truly elevated riskare identified so that they might benefit fromadvances in surveillance techniques and pro-phylactic interventions. Advances in the under-standing of the polygenic basis of breast cancermay have different effects on three groups: fam-ilies positive for mutations in BRCA1 or BRCA2,other breast cancer families, and the generalpopulation.

Risk estimation in BRCA-positive breastcancer families. Detection in a family of a mu-tation in BRCA1 or BRCA2 has afforded relativeclarity in risk estimation. Unaffected females,


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

at high prior risk of cancer on account of theirfamily history, can undertake predictive testingto determine whether they carry the high-riskmutation. Risk-reducing prophylactic surgeryand/or intensive surveillance may be offered tomutation carriers whereas noncarriers can bereassured that their risk is not as high. How-ever, studies have found wide variation in thepenetrance of mutations in BRCA1 and BRCA2(see above) and this has raised questions regard-ing the appropriate estimates of risk for clinicalusage. Some of this apparent variation in pen-etrance between families may be the result ofdiffering doses of additional lower-risk modi-fying variants; recently discovered examples in-clude SNPs in RAD51, FGFR2, and at 5q (11,12). Thus, one of the early clinical applicationsof low-penetrance alleles may be to offer in-dividualized refinement of risk to BRCA1 andBRCA2 mutation carriers.

Risk estimation in BRCA-negative breastcancer families. The majority of breast cancerfamilies do not harbor mutations in BRCA1 orBRCA2. Clinical breast cancer risk estimationin these families is currently based empiricallyupon family history of cancer. For unaffectedindividuals in breast cancer families negativefor BRCA1 and BRCA2 mutations, it would beclinically useful to have better discriminationof risk to distinguish high-risk individuals inwhom radical surgical prophylaxis may be jus-tified and individuals at lower/population riskwho could be spared some anxiety and inter-vention. It remains unclear whether the un-derlying genetic architecture is such that thisdiscrimination could be provided by genotyp-ing multiple genetic predisposition factors (pre-suming sufficient numbers were identified). Ifthe clustering of breast cancer in a family hasoccurred because of multiple low-penetrancegenetic factors, typing of these factors wouldresult in many possible risk categories. Geno-typing may place the majority of unaffectedmembers of the family into intermediate riskcategories that differ little from their risk basedon family history. If family clustering is theresult of fewer factors of greater penetrance,

genotyping these may adjust risk estimation suf-ficiently to alter clinical management for unaf-fected individuals.

Risk estimation in the general population.By contrast, in the absence of a family historyof breast cancer, genotyping is the only meansof identifying individuals at increased geneticrisk. The discrimination of risk that could be af-forded by population-level genotyping of com-mon low-penetrance variants is influenced bytwo factors: firstly, the true extent to whichbreast cancer risk varies across the population(which is currently unclear) and secondly, theproportion of the risk variants that is availablefor genotyping (8, 9, 94). Although the logisti-cal, ethical, social, and economic considerationsare numerous, in principle it seems plausiblethat population-based genotyping could even-tually be used to aid stratification and resourceallocation. Surveillance, primary antiestrogenchemoprophylaxis, prophylactic surgery, andother emerging interventions may be appropri-ate for the high-risk upper tail of the popula-tion. The low-risk tail may require little or noadditional intervention.

Targeted Therapies

Understanding of the biological mechanismsunderlying breast cancer predisposition genesis beginning to offer exciting opportunitiesfor new therapies. The roles of BRCA1 andBRCA2 in DNA repair and the resultant sen-sitivity of BRCA1- and BRCA2-deficient cellsto agents that cross-link DNA are being ex-ploited by studies in which mutation carriersare treated with platinum-based drugs. The in-herent DNA repair defect in BRCA-deficientcells has also provided a rationale for a furthertherapeutic approach. Poly (ADP-ribose) poly-merase (PARP) is an enzyme involved in baseexcision repair. Inhibition of PARP results inan increase in DNA lesions that are normallyrepaired through homologous recombination,which requires BRCA1 and BRCA2. In a back-ground deficient for either BRCA1 or BRCA2protein, cells are profoundly sensitive to


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

inhibition by PARP, which results in cell cyclearrest, chromosome instability, and cell death.Thus in BRCA mutation carriers, PARP in-hibitors are synthetically lethal to tumor cellsbut confer no demonstrable toxicity to normalheterozygous cells. Studies are also underwayto investigate whether PARP inhibitors havesimilar effects in individuals with basal tumorssimilar to those occurring in BRCA1 mutationcarriers and/or cancers associated with defectsin other DNA repair proteins (50, 84). Under-standing the molecular basis of the increasedcancer risk associated with low-penetrance al-leles is in its infancy. However, this understand-ing may offer opportunities for novel therapies.For example, if some risk alleles drive tumori-genesis through the upregulation of oncogenes,such as FGFR2 and/or MYC, these genes mayrepresent potential new targets for therapeuticinterventions.

CONCLUSION

Although recent breakthroughs have resulted inthe identification of distinct new groups of ge-netic breast cancer predisposition factors, morethan 70% of the genetic predisposition to breastcancer remains unexplained. Technologies areemerging rapidly that may allow us to iden-tify many further predisposition factors withinthese classes and of other risk-prevalence pro-files. The identification of novel predispositionfactors offers exciting challenges, but ongoingclarification and characterization of known ge-netic risk factors is also important. A formidableand ongoing challenge is to marshal disparateexperimental results to capture, validate, qual-ify, and organize into an integrated schema allthe emerging components of risk. This is par-ticularly important if we are to advance clinicalrisk estimation to optimally apportion surveil-lance and prophylactic interventions.

SUMMARY POINTS

1. Multiple strategies have been used to identify breast cancer–predisposing genetic factors.

2. Linkage analysis has been successful in mapping high-penetrance genes such as BRCA1and BRCA2.

3. Mutational screening of candidate genes has been used to identify genes such as CHEK2,ATM, BRIP1, and PALB2, mutations in which are rare and confer intermediate penetranceof breast cancer.

4. Genome-wide association studies have revealed several low-penetrance alleles.

5. More than 70% of breast cancer predisposition remains unexplained.

6. The outstanding predisposition is likely to be polygenic and may include multiple fur-ther common low-penetrance alleles, rare intermediate-penetrance genes, and rare low-penetrance alleles.

7. BRCA1 and BRCA2 mutation testing allows identification of individuals at elevated riskof breast cancer who can be offered risk-reducing interventions.

8. Targeted therapies are being developed that exploit the biological functions of BRCA1and BRCA2.

FUTURE ISSUES

1. Further genome-wide association studies are required to identify further commonvariants.


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

2. Genome-wide resequencing is likely to detect numerous novel rare variants, some ofwhich may predispose to breast cancer.

3. Understanding how novel variants result in breast cancer predisposition may requireinnovative strategies.

4. Quantification of gene-gene interactions and gene-environment interactions, which maybe heterogeneous in nature, may be used to improve risk estimation.

5. Judicious clinical translation of genetic factors in addition to BRCA1 and BRCA2 mayassist in risk estimation, optimization of management, and development of therapies.

DISCLOSURE STATEMENT

The authors are not aware of any biases that might be perceived as affecting the objectivity of thisreview.

ACKNOWLEDGMENTS

We are grateful to Mike Stratton, Anthony Renwick, and Richard Scott for their critical readingof the manuscript and to Peter Donnelly, Julian Maller, and Paul Pharoah for their assistance inanalyses of the low penetrance alleles.

LITERATURE CITED

1. Alessi DR, Sakamoto K, Bayascas JR. 2006. LKB1-dependent signaling pathways. Annu. Rev. Biochem.75:137–63

2. Allinen M, Huusko P, Mantyniemi S, Launonen V, Winqvist R. 2001. Mutation analysis of the CHK2gene in families with hereditary breast cancer. Br. J. Cancer 85:209–12

3. Andrieu N, Goldgar DE, Easton DF, Rookus M, Brohet R, et al. 2006. Pregnancies, breast-feeding, andbreast cancer risk in the International BRCA1/2 Carrier Cohort Study (IBCCS). J. Natl. Cancer Inst.98:535–44

4. Anglian Breast Cancer Study Group. 2000. Prevalence and penetrance of BRCA1 and BRCA2 mutationsin a population-based series of breast cancer cases. Br. J. Cancer 83:1301–8

5. Antoniou A, Pharoah PD, Narod S, Risch HA, Eyfjord JE, et al. 2003. Average risks of breast and ovariancancer associated with BRCA1 or BRCA2 mutations detected in case series unselected for family history:a combined analysis of 22 studies. Am. J. Hum. Genet. 72:1117–30

6. Antoniou AC, Easton DF. 2003. Polygenic inheritance of breast cancer: Implications for design ofassociation studies. Genet. Epidemiol. 25:190–202

7. Antoniou AC, Gayther SA, Stratton JF, Ponder BA, Easton DF. 2000. Risk models for familial ovarianand breast cancer. Genet. Epidemiol. 18:173–90

8. Antoniou AC, Pharoah PD, McMullan G, Day NE, Ponder BA, Easton D. 2001. Evidence for furtherbreast cancer susceptibility genes in addition to BRCA1 and BRCA2 in a population-based study. Genet.Epidemiol. 21:1–18

9. Antoniou AC, Pharoah PD, McMullan G, Day NE, Stratton MR, et al. 2002. A comprehensive modelfor familial breast cancer incorporating BRCA1, BRCA2 and other genes. Br. J. Cancer 86:76–83

10. Antoniou AC, Pharoah PP, Smith P, Easton DF. 2004. The BOADICEA model of genetic susceptibilityto breast and ovarian cancer. Br. J. Cancer 91:1580–90


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

11. Antoniou AC, Sinilnikova OM, Simard J, Leone M, Dumont M, et al. 2007. RAD51 135G–>C modifiesbreast cancer risk among BRCA2 mutation carriers: results from a combined analysis of 19 studies. Am.J. Hum. Genet. 81:1186–200

12. Antoniou AC, Spurdle AB, Sinilnikova OM, Healey S, Pooley KA, et al. 2008. Common breast cancer-predisposition alleles are associated with breast cancer risk in BRCA1 and BRCA2 mutation carriers.Am. J. Hum. Genet. 82:937–48

13. Bell DW, Varley JM, Szydlo TE, Kang DH, Wahrer DC, et al. 1999. Heterozygous germ line hCHK2mutations in Li-Fraumeni syndrome. Science 286:2528–31

14. Bignell GR, Barfoot R, Seal S, Collins N, Warren W, Stratton MR. 1998. Low frequency of somaticmutations in the LKB1/Peutz-Jeghers syndrome gene in sporadic breast cancer. Cancer Res. 58:1384–86

15. Birch JM, Alston RD, McNally RJ, Evans DG, Kelsey AM, et al. 2001. Relative frequency and morphol-ogy of cancers in carriers of germline TP53 mutations. Oncogene 20:4621–28

16. Bishop DT, Cannon-Albright L, McLellan T, Gardner EJ, Skolnick MH. 1988. Segregation and linkageanalysis of nine Utah breast cancer pedigrees. Genet. Epidemiol. 5:151–69

17. Bogdanova N, Enssen-Dubrowinskaja N, Feshchenko S, Lazjuk GI, Rogov YI, et al. 2005. Associationof two mutations in the CHEK2 gene with breast cancer. Int. J. Cancer 116:263–66

18. Borresen AL, Andersen TI, Garber J, Barbier-Piraux N, Thorlacius S, et al. 1992. Screening for germline TP53 mutations in breast cancer patients. Cancer Res. 52:3234–36

19. Breast Cancer Assoc. Consort. 2006. Commonly studied single-nucleotide polymorphisms and breastcancer: results from the Breast Cancer Association Consortium. J. Natl. Cancer Inst. 98:1382–96

20. Breast Cancer Link. Consort. 1997. Pathology of familial breast cancer: differences between breastcancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases. Lancet 349:1505–10

21. Breast Cancer Link. Consort. 1999. Cancer risks in BRCA2 mutation carriers. J. Natl. Cancer Inst.91:1310–16

22. Broca P. 1866. Traite des Tumeurs. Paris: Asselin23. Brohet RM, Goldgar DE, Easton DF, Antoniou AC, Andrieu N, et al. 2007. Oral contraceptives

and breast cancer risk in the international BRCA1/2 carrier cohort study: a report from EMBRACE,GENEPSO, GEO-HEBON, and the IBCCS Collaborating Group. J. Clin. Oncol. 25:3831–36

24. Cantor SB, Bell DW, Ganesan S, Kass EM, Drapkin R, et al. 2001. BACH1, a novel helicase-like protein,interacts directly with BRCA1 and contributes to its DNA repair function. Cell 105:149–60

25. Chang-Claude J, Andrieu N, Rookus M, Brohet R, Antoniou AC, et al. 2007. Age at menarche andmenopause and breast cancer risk in the International BRCA1/2 Carrier Cohort Study. Cancer Epidemiol.Biomark. Prev. 16:740–46

26. CHEK2 Breast Cancer Case-Control Consort. 2004. CHEK2∗1100delC and susceptibility to breastcancer: a collaborative analysis involving 10,860 breast cancer cases and 9,065 controls from 10 studies.2004. Am. J. Hum. Genet. 74:1175–82 (WAS 8)

27. Chen J, Lindblom P, Lindblom A. 1998. A study of the PTEN/MMAC1 gene in 136 breast cancerfamilies. Hum. Genet. 102:124–25

28. Chompret A, Brugieres L, Ronsin M, Gardes M, Dessarps-Freichey F, et al. 2000. p53 germline muta-tions in childhood cancers and cancer risk for carrier individuals. Br. J. Cancer 82:1932–37

29. Claus EB, Risch N, Thompson WD. 1991. Genetic analysis of breast cancer in the cancer and steroidhormone study. Am. J. Hum. Genet. 48:232–42

30. Collab. Group Horm. Factors Breast Cancer. 1996. Breast cancer and hormonal contraceptives: collab-orative reanalysis of individual data on 53,297 women with breast cancer and 100,239 women withoutbreast cancer from 54 epidemiological studies. Lancet 347:1713–27

31. Collab. Group Horm. Factors Breast Cancer. 1997. Breast cancer and hormone replacement therapy:collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and108,411 women without breast cancer. Lancet 350:1047–59

32. Collab. Group Horm. Factors Breast Cancer. 2001. Familial breast cancer: collaborative reanalysis ofindividual data from 52 epidemiological studies including 58,209 women with breast cancer and 101,986women without the disease. Lancet 358:1389–99


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

33. Collab. Group Horm. Factors Breast Cancer. 2002. Breast cancer and breastfeeding: collaborative re-analysis of individual data from 47 epidemiological studies in 30 countries, including 50,302 women withbreast cancer and 96,973 women without the disease. Lancet 360:187–95

34. Cox A, Dunning AM, Garcia-Closas M, Balasubramanian S, Reed MW, et al. 2007. A common codingvariant in CASP8 is associated with breast cancer risk. Nat. Genet. 39:352–58

35. Cui J, Antoniou AC, Dite GS, Southey MC, Venter DJ, et al. 2001. After BRCA1 and BRCA2-what next?Multifactorial segregation analyses of three-generation, population-based Australian families affected byfemale breast cancer. Am. J. Hum. Genet. 68:420–31

36. Cully M, You H, Levine AJ, Mak TW. 2006. Beyond PTEN mutations: the PI3K pathway as an integratorof multiple inputs during tumorigenesis. Nat. Rev. Cancer 6:184–92

37. Cybulski C, Gorski B, Huzarski T, Masojc B, Mierzejewski M, et al. 2004. CHEK2 is a multiorgan cancersusceptibility gene. Am. J. Hum. Genet. 75:1131–35

38. Dong X, Wang L, Taniguchi K, Wang X, Cunningham JM, et al. 2003. Mutations in CHEK2 associatedwith prostate cancer risk. Am. J. Hum. Genet. 72:270–80

39. Dufault MR, Betz B, Wappenschmidt B, Hofmann W, Bandick K, et al. 2004. Limited relevance of theCHEK2 gene in hereditary breast cancer. Int. J. Cancer 110:320–25

40. Dunning AM, Healey CS, Pharoah PD, Teare MD, Ponder BA, Easton DF. 1999. A systematic reviewof genetic polymorphisms and breast cancer risk. Cancer Epidemiol. Biomark. Prev. 8:843–54

41. Easton DF. 1994. Cancer risks in A-T heterozygotes. Int. J. Radiat. Biol. 66:S177–8242. Easton DF. 1999. How many more breast cancer predisposition genes are there? Breast Cancer Res.

1:14–4743. Easton DF, Deffenbaugh AM, Pruss D, Frye C, Wenstrup RJ, et al. 2007. A systematic genetic assessment

of 1433 sequence variants of unknown clinical significance in the BRCA1 and BRCA2 breast cancer-predisposition genes. Am. J. Hum. Genet. 81:873–83

44. Easton DF, Ford D, Bishop DT. 1995. Breast and ovarian cancer incidence in BRCA1-mutation carriers.Breast Cancer Linkage Consortium. Am. J. Hum. Genet. 56:265–71

45. Easton DF, Matthews FE, Ford D, Swerdlow AJ, Peto J. 1996. Cancer mortality in relatives of womenwith ovarian cancer: the OPCS Study. Office of Population Censuses and Surveys. Int. J. Cancer 65:284–94

46. Easton DF, Pooley KA, Dunning AM, Pharoah PD, Thompson D, et al. 2007. Genome-wide associationstudy identifies novel breast cancer susceptibility loci. Nature 447:1087–93

47. Erkko H, Xia B, Nikkila J, Schleutker J, Syrjakoski K, et al. 2007. A recurrent mutation in PALB2 inFinnish cancer families. Nature 446:316–19

48. Evans DG, Birch JM, Thorneycroft M, McGown G, Lalloo F, Varley JM. 2002. Low rate of TP53germline mutations in breast cancer/sarcoma families not fulfilling classical criteria for Li-Fraumenisyndrome. J. Med. Genet. 39:941–44

49. Evers B, Jonkers J. 2006. Mouse models of BRCA1 and BRCA2 deficiency: past lessons, current under-standing and future prospects. Oncogene 25:5885–97

50. Farmer H, McCabe N, Lord CJ, Tutt AN, Johnson DA, et al. 2005. Targeting the DNA repair defect inBRCA mutant cells as a therapeutic strategy. Nature 434:917–21

51. Ford D, Easton DF, Stratton M, Narod S, Goldgar D, et al. 1998. Genetic heterogeneity and pene-trance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer LinkageConsortium. Am. J. Hum. Genet. 62:676–89

52. Foulkes WD, Ghadirian P, Akbari MR, Hamel N, Giroux S, et al. 2007. Identification of a novel trun-cating PALB2 mutation and analysis of its contribution to early-onset breast cancer in French Canadianwomen. Breast Cancer Res. 9:R83

53. Garcia-Closas M, Hall P, Nevanlinna H, Pooley K, Morrison J, et al. 2008. Heterogeneity of breastcancer associations with five susceptibility loci by clinical and pathological characteristics. PLoS Genet.4:e1000054

54. Gold B, Kirchhoff T, Stefanov S, Lautenberger J, Viale A, et al. 2008. Genome-wide association studyprovides evidence for a breast cancer risk locus at 6q22.33. Proc. Natl. Acad Sci. USA 105:4340–5

55. Gorski B, Cybulski C, Huzarski T, Byrski T, Gronwald J, et al. 2005. Breast cancer predisposing allelesin Poland. Breast Cancer Res. Treat. 92:19–24


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

56. Gudmundsdottir K, Ashworth A. 2006. The roles of BRCA1 and BRCA2 and associated proteins in themaintenance of genomic stability. Oncogene 25:5864–74

57. Guenard F, Labrie Y, Ouellette G, Beauparlant CJ, Bessette P, et al. 2007. Germline mutations in thebreast cancer susceptibility gene PTEN are rare in high-risk non-BRCA1/2 French Canadian breastcancer families. Fam. Cancer 6:483–90

58. Guilford P, Hopkins J, Harraway J, McLeod M, McLeod N, et al. 1998. E-cadherin germline mutationsin familial gastric cancer. Nature 392:402–5

59. Haiman CA, Le ML, Yamamato J, Stram DO, Sheng X, et al. 2007. A common genetic risk factor forcolorectal and prostate cancer. Nat. Genet. 39:954–56

60. Haiman CA, Patterson N, Freedman ML, Myers SR, Pike MC, et al. 2007. Multiple regions within 8q24independently affect risk for prostate cancer. Nat. Genet. 39:638–44

61. Hall JM, Lee MK, Newman B, Morrow JE, Anderson LA, et al. 1990. Linkage of early-onset familialbreast cancer to chromosome 17q21. Science 250:1684–89

62. Heikkinen K, Rapakko K, Karppinen SM, Erkko H, Knuutila S, et al. 2006. RAD50 and NBS1 are breastcancer susceptibility genes associated with genomic instability. Carcinogenesis 27:1593–99

63. Hemminki A, Markie D, Tomlinson I, Avizienyte E, Roth S, et al. 1998. A serine/threonine kinase genedefective in Peutz-Jeghers syndrome. Nature 391:184–87

64. Hemminki A, Tomlinson I, Markie D, Jarvinen H, Sistonen P, et al. 1997. Localization of a susceptibilitylocus for Peutz-Jeghers syndrome to 19p using comparative genomic hybridization and targeted linkageanalysis. Nat. Genet. 15:87–90

65. Hopper JL, Carlin JB. 1992. Familial aggregation of a disease consequent upon correlation betweenrelatives in a risk factor measured on a continuous scale. Am. J. Epidemiol. 136:1138–47

66. Howlett NG, Taniguchi T, Olson S, Cox B, Waisfisz Q, et al. 2002. Biallelic inactivation of BRCA2 inFanconi anemia. Science 297:606–9

67. Hunter DJ, Kraft P, Jacobs KB, Cox DG, Yeager M, et al. 2007. A genome-wide association studyidentifies alleles in FGFR2 associated with risk of sporadic postmenopausal breast cancer. Nat. Genet.39:870–74

68. Iselius L, Slack J, Littler M, Morton NE. 1991. Genetic epidemiology of breast cancer in Britain. Ann.Hum. Genet. 55:151–59

69. Jacquemier J, Padovani L, Rabayrol L, Lakhani SR, Penault-Llorca F, et al. 2005. Typical medullarybreast carcinomas have a basal/myoepithelial phenotype. J. Pathol. 207:260–68

70. Jenne DE, Reimann H, Nezu J, Friedel W, Loff S, et al. 1998. Peutz-Jeghers syndrome is caused bymutations in a novel serine threonine kinase. Nat. Genet. 18:38–43

71. Keller G, Vogelsang H, Becker I, Hutter J, Ott K, et al. 1999. Diffuse type gastric and lobular breastcarcinoma in a familial gastric cancer patient with an E-cadherin germline mutation. Am. J. Pathol.155:337–42

72. Kilpivaara O, Vahteristo P, Falck J, Syrjakoski K, Eerola H, et al. 2004. CHEK2 variant I157T may beassociated with increased breast cancer risk. Int. J. Cancer 111:543–47

73. Lakhani SR. 1999. The pathology of familial breast cancer: Morphological aspects. Breast Cancer Res.1:31–35

74. Lakhani SR, Reis-Filho JS, Fulford L, Penault-Llorca F, van der Vijver M, et al. 2005. Prediction ofBRCA1 status in patients with breast cancer using estrogen receptor and basal phenotype. Clin. CancerRes. 11:5175–80

75. Lakhani SR, van der Vijver M, Jacquemier J, Anderson TJ, Osin PP, et al. 2002. The pathology offamilial breast cancer: predictive value of immunohistochemical markers estrogen receptor, progesteronereceptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2. J. Clin. Oncol. 20:2310–18

76. Lalloo F, Varley J, Moran A, Ellis D, O’dair L, et al. 2006. BRCA1, BRCA2 and TP53 mutations in veryearly-onset breast cancer with associated risks to relatives. Eur. J. Cancer 42:1143–50

77. Le Dran H. 1757. Memoire avec un precis de plusieurs observations sur le cancer. Mem. Acad. R. Chir.3:1–54

78. Levitus M, Waisfisz Q, Godthelp BC, de Vries Y, Hussain S, et al. 2005. The DNA helicase BRIP1 isdefective in Fanconi anemia complementation group. J. Nat. Genet. 37:934–35


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

79. Levran O, Attwooll C, Henry RT, Milton KL, Neveling K, et al. 2005. The BRCA1-interacting helicaseBRIP1 is deficient in Fanconi anemia. Nat. Genet. 37:931–33

80. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, et al. 2000. Environmental and heritablefactors in the causation of cancer–analyses of cohorts of twins from Sweden, Denmark, and Finland. N.Engl. J. Med. 343:78–85

81. Litman R, Peng M, Jin Z, Zhang F, Zhang J, et al. 2005. BACH1 is critical for homologous recombinationand appears to be the Fanconi anemia gene product FANCJ. Cancer Cell 8:255–65

82. Malkin D, Li FP, Strong LC, Fraumeni JF Jr, Nelson CE, et al. 1990. Germ line p53 mutations in afamilial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250:1233–38

83. Masciari S, Larsson N, Senz J, Boyd N, Kaurah P, et al. 2007. Germline E-cadherin mutations in familiallobular breast cancer. J. Med. Genet. 44:726–31

84. McCabe N, Turner NC, Lord CJ, Kluzek K, Bialkowska A, et al. 2006. Deficiency in the repair of DNAdamage by homologous recombination and sensitivity to poly(ADP-ribose) polymerase inhibition. CancerRes. 66:8109–15

85. Meijers-Heijboer H, van den Ouweland A, Klijn J, Wasielewski M, de Snoo A, et al. 2002. Low-penetrance susceptibility to breast cancer due to CHEK2∗1100delC in noncarriers of BRCA1 or BRCA2mutations. Nat. Genet. 31:55–59

86. Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harshman K, et al. 1994. A strong candidate for thebreast and ovarian cancer susceptibility gene BRCA1. Science 266:66–71

87. Nelen MR, Padberg GW, Peeters EA, Lin AY, van den Helm B, et al. 1996. Localization of the genefor Cowden disease to chromosome 10q22–23. Nat. Genet. 13:114–16

88. Nelen MR, van Staveren WC, Peeters EA, Hassel MB, Gorlin RJ, et al. 1997. Germline mutations inthe PTEN/MMAC1 gene in patients with Cowden disease. Hum. Mol. Genet. 6:1383–87

89. Parmigiani G, Berry D, Aguilar O. 1998. Determining carrier probabilities for breast cancer-susceptibility genes BRCA1 and BRCA2. Am. J. Hum. Genet. 62:145–58

90. Peng M, Litman R, Jin Z, Fong G, Cantor SB. 2006. BACH1 is a DNA repair protein supporting BRCA1damage response. Oncogene 25:2245–53

91. Peto J, Collins N, Barfoot R, Seal S, Warren W, et al. 1999. Prevalence of BRCA1 and BRCA2 genemutations in patients with early-onset breast cancer. J. Natl. Cancer Inst. 91:943–49

92. Peto J, Easton DF, Matthews FE, Ford D, Swerdlow AJ. 1996. Cancer mortality in relatives of womenwith breast cancer: the OPCS Study. Office of Population Censuses and Surveys. Int. J. Cancer 65:275–83

93. Peto J, Mack TM. 2000. High constant incidence in twins and other relatives of women with breastcancer. Nat. Genet. 26:411–14

94. Pharoah PD, Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BA. 2002. Polygenic suscepti-bility to breast cancer and implications for prevention. Nat. Genet. 31:33–36

95. Pharoah PD, Dunning AM, Ponder BA, Easton DF. 2004. Association studies for finding cancer-susceptibility genetic variants. Nat. Rev. Cancer 4:850–60

96. Pharoah PD, Guilford P, Caldas C. 2001. Incidence of gastric cancer and breast cancer in CDH1 (E-cadherin) mutation carriers from hereditary diffuse gastric cancer families. Gastroenterology 121:1348–53

97. Rahman N, Scott RH. 2007. Cancer genes associated with phenotypes in monoallelic and biallelicmutation carriers: new lessons from old players. Hum. Mol. Genet. 16(Spec. No. 1):R60–66

98. Rahman N, Seal S, Thompson D, Kelly P, Renwick A, et al. 2007. PALB2, which encodes a BRCA2-interacting protein, is a breast cancer susceptibility gene. Nat. Genet. 39:165–67

99. Ramus SJ, Harrington PA, Pye C, Dicioccio RA, Cox MJ, et al. 2007. Contribution of BRCA1 and BRCA2mutations to inherited ovarian cancer. Hum. Mutat. 28:1207–15

100. Reid S, Schindler D, Hanenberg H, Barker K, Hanks S, et al. 2007. Biallelic mutations in PALB2 causeFanconi anemia subtype FA-N and predispose to childhood cancer. Nat. Genet. 39:162–64

101. Renwick A, Thompson D, Seal S, Kelly P, Chagtai T, et al. 2006. ATM mutations that cause ataxia-telangiectasia are breast cancer susceptibility alleles. Nat. Genet. 38:873–75

102. Savitsky K, Bar-Shira A, Gilad S, Rotman G, Ziv Y, et al. 1995. A single ataxia telangiectasia gene witha product similar to PI-3 kinase. Science 268:1749–53


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

ANRV353-GG09-17 ARI 25 July 2008 17:26

103. Schutte M, Seal S, Barfoot R, Meijers-Heijboer H, Wasielewski M, et al. 2003. Variants in CHEK2 otherthan 1100delC do not make a major contribution to breast cancer susceptibility. Am. J. Hum. Genet.72:1023–28

104. Seal S, Thompson D, Renwick A, Elliott A, Kelly P, et al. 2006. Truncating mutations in the Fanconianemia J gene BRIP1 are low-penetrance breast cancer susceptibility alleles. Nat. Genet. 38:1239–41

105. Seppala EH, Ikonen T, Mononen N, Autio V, Rokman A, et al. 2003. CHEK2 variants associate withhereditary prostate cancer. Br. J. Cancer 89:1966–70

106. Sharan SK, Wims M, Bradley A. 1995. Murine Brca1: sequence and significance for human missensemutations. Hum. Mol. Genet. 4:2275–78

107. Shugart YY, Cour C, Renard H, Lenoir G, Goldgar D, et al. 1999. Linkage analysis of 56 multiplexfamilies excludes the Cowden disease gene PTEN as a major contributor to familial breast cancer. J. Med.Genet. 36:720–21

108. Smid M, Wang Y, Klijn JG, Sieuwerts AM, Zhang Y, et al. 2006. Genes associated with breast cancermetastatic to bone. J. Clin. Oncol. 24:2261–67

109. Smith P, McGuffog L, Easton DF, Mann GJ, Pupo GM, et al. 2006. A genome wide linkage search forbreast cancer susceptibility genes. Genes Chromosomes Cancer 45:646–55

110. Stacey SN, Manolescu A, Sulem P, Rafnar T, Gudmundsson J, et al. 2007. Common variants on chro-mosomes 2q35 and 16q12 confer susceptibility to estrogen receptor-positive breast cancer. Nat. Genet.39:865–69

111. Stacey SN, Manolescu A, Sulem P, Thorlacius S, Gudjonsson SA, et al. 2008. Common variantson chromosome 5p12 confer susceptibility to estrogen receptor-positive breast cancer. Nat. Genet.doi:10.1038/ng.131

112. Swift M, Sholman L, Perry M, Chase C. 1976. Malignant neoplasms in the families of patients withataxia-telangiectasia. Cancer Res. 36:209–15

113. Thompson D, Duedal S, Kirner J, McGuffog L, Last J, et al. 2005. Cancer risks and mortality inheterozygous ATM mutation carriers. J. Natl. Cancer Inst. 97:813–22

114. Thompson D, Easton D. 2001. Variation in cancer risks, by mutation position, in BRCA2 mutationcarriers. Am. J. Hum. Genet. 68:410–19

115. Thompson D, Easton D. 2002. Variation in BRCA1 cancer risks by mutation position. Cancer Epidemiol.Biomark. Prev. 11:329–36

116. Thompson D, Easton DF. 2002. Cancer incidence in BRCA1 mutation carriers. J. Natl. Cancer Inst.94:1358–65

117. Thompson D, Seal S, Schutte M, McGuffog L, Barfoot R, et al. 2006. A multicenter study of cancerincidence in CHEK2 1100delC mutation carriers. Cancer Epidemiol. Biomark. Prev. 15:2542–45

118. Tommiska J, Seal S, Renwick A, Barfoot R, Baskcomb L, et al. 2006. Evaluation of RAD50 in familialbreast cancer predisposition. Int. J. Cancer 118:2911–16

119. Walsh T, Casadei S, Coats KH, Swisher E, Stray SM, et al. 2006. Spectrum of mutations in BRCA1,BRCA2, CHEK2, and TP53 in families at high risk of breast cancer. JAMA 295:1379–88

120. Ware MD, DeSilva D, Sinilnikova OM, Stoppa-Lyonnet D, Tavtigian SV, Mazoyer S. 2006. Doesnonsense-mediated mRNA decay explain the ovarian cancer cluster region of the BRCA2 gene? Oncogene25:323–28

121. Williams WR, Anderson DE. 1984. Genetic epidemiology of breast cancer: segregation analysis of 200Danish pedigrees. Genet. Epidemiol. 1:7–20

122. Wooster R, Bignell G, Lancaster J, Swift S, Seal S, et al. 1995. Identification of the breast cancersusceptibility gene BRCA2. Nature 378:789–92

123. Wooster R, Neuhausen SL, Mangion J, Quirk Y, Ford D, et al. 1994. Localization of a breast cancersusceptibility gene, BRCA2, to chromosome 13q12–13. Science 265:2088–90

124. Xia B, Dorsman JC, Ameziane N, de Vries Y, Rooimans MA, et al. 2007. Fanconi anemia is associatedwith a defect in the BRCA2 partner PALB2. Nat. Genet. 39:159–61

125. Xia B, Sheng Q, Nakanishi K, Ohashi A, Wu J, et al. 2006. Control of BRCA2 cellular and clinicalfunctions by a nuclear partner, PALB2. Mol. Cell 22:719–29


Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

AR353-GG09-FM ARI 25 July 2008 5:19

Annual Review ofGenomics andHuman Genetics

Volume 9, 2008Contents

Human Telomere Structure and BiologyHarold Riethman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Infectious Disease in the Genomic EraXiaonan Yang, Hongliang Yang, Gangqiao Zhou, and Guo-Ping Zhao � � � � � � � � � � � � � � � � � � �21

ENU Mutagenesis, a Way Forward to Understand Gene FunctionAbraham Acevedo-Arozena, Sara Wells, Paul Potter, Michelle Kelly,

Roger D. Cox, and Steve D.M. Brown � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �49

Clinical Utility of Contemporary Molecular CytogeneticsBassem A. Bejjani and Lisa G. Shaffer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �71

The Role of Aminoacyl-tRNA Synthetases in Genetic DiseasesAnthony Antonellis and Eric D. Green � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �87

A Bird’s-Eye View of Sex Chromosome Dosage CompensationArthur P. Arnold, Yuichiro Itoh, and Esther Melamed � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 109

Linkage Disequilibrium and Association MappingB. S. Weir � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 129

Positive Selection in the Human Genome: From Genome Scansto Biological SignificanceJoanna L. Kelley and Willie J. Swanson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 143

The Current Landscape for Direct-to-Consumer Genetic Testing:Legal, Ethical, and Policy IssuesStuart Hogarth, Gail Javitt, and David Melzer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 161

Transcriptional Control of SkeletogenesisGerard Karsenty � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 183

A Mechanistic View of Genomic ImprintingKy Sha � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 197

Phylogenetic Inference Using Whole GenomesBruce Rannala and Ziheng Yang � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 217

v

Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

AR353-GG09-FM ARI 25 July 2008 5:19

Transgenerational Epigenetic EffectsNeil A. Youngson and Emma Whitelaw � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 233

Evolution of Dim-Light and Color Vision PigmentsShozo Yokoyama � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 259

Genetic Basis of Thoracic Aortic Aneurysms and Dissections: Focuson Smooth Muscle Cell Contractile DysfunctionDianna M. Milewicz, Dong-Chuan Guo, Van Tran-Fadulu, Andrea L. Lafont,

Christina L. Papke, Sakiko Inamoto, and Hariyadarshi Pannu � � � � � � � � � � � � � � � � � � � � � � � � � � � 283

Cohesin and Human DiseaseJinglan Liu and Ian D. Krantz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 303

Genetic Predisposition to Breast Cancer: Past, Present, and FutureClare Turnbull and Nazneen Rahman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 321

From Linkage Maps to Quantitative Trait Loci: The Historyand Science of the Utah Genetic Reference ProjectStephen M. Prescott, Jean Marc Lalouel, and Mark Leppert � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 347

Disorders of Lysosome-Related Organelle Biogenesis: Clinicaland Molecular GeneticsMarjan Huizing, Amanda Helip-Wooley, Wendy Westbroek,

Meral Gunay-Aygun, and William A. Gahl � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 359

Next-Generation DNA Sequencing MethodsElaine R. Mardis � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 387

African Genetic Diversity: Implications for Human DemographicHistory, Modern Human Origins, and Complex Disease MappingMichael C. Campbell and Sarah A. Tishkoff � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 403

Indexes

Cumulative Index of Contributing Authors, Volumes 1–9 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 435

Cumulative Index of Chapter Titles, Volumes 1–9 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 438

Errata

An online log of corrections to Annual Review of Genomics and Human Genetics articlesmay be found at http://genom.annualreviews.org/

vi Contents

Ann

u. R

ev. G

enom

. Hum

an G

enet

. 200

8.9:

321-

345.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

isso

uri -

Col

umbi

a on

05/

13/1

3. F

or p

erso

nal u

se o

nly.

Documents

Genetic Predisposition to Breast Cancer: Past, Present, and Future