Erratum: Genomic alterations in histopathologically normal breast tissue from BRCA1 mutation carriers may be caused by BRCA1 haploinsufficiency

GENES, CHROMOSOMES & CANCER 49:78–90 (2010)

Genomic Alterations in Histopathologically NormalBreast Tissue from BRCA1 Mutation Carriers Maybe Caused by BRCA1 Haploinsufficiency

Karin Rennstam,1 Anita Ringberg,2 Heather E. Cunliffe,3 Hakan Olsson,1 Goran Landberg,4,5

and Ingrid Hedenfalk1,3*

1Departmentof Oncology,Clinical Sciences,Lund University,Lund,Sweden2Departmentof Plastic and Reconstructive Surgery,Clinical Sciences,UMAS,Malm ,Sweden3Translational Genomics Research Institute,Phoenix,AZ4Center for Molecular Pathology,Departmentof Laboratory Medicine,Lund University,UMAS,Malm ,Sweden5Breakthrough Breast Cancer Research Unit,Paterson Institute for Cancer Research,Manchester University,UK

Multiple biopsies of normal breast tissue from 10 BRCA1 mutation carriers have been analyzed using array-based compara-tive genomic hybridization. Normal breast tissue from five age-matched control subjects without a family history of breastcancer was included for reference purposes. We repeatedly found multiple low copy number aberrations at a significantlyhigher frequency in histopathologically normal tissue from BRCA1 mutation carriers than in normal control tissue. Some ofthese aberrations were similar across samples from different patients and linked to biological functions such as transcrip-tional regulation and DNA binding. We also observed a high degree of genomic heterogeneity between samples from thesame patient, suggestive of tissue heterogeneity and etiological clonality in the breast epithelium. We show that neitherloss of heterozygosity nor promoter methylation of the wild-type BRCA1 allele is the predominant mechanistic origin ofthe observed genomic instability. Instead, we propose that haploinsufficiency of BRCA1 might be the underlying cause re-sponsible for initiation of breast cancer in these predisposed women, making cells vulnerable to mitotic recombination.We also propose that loss of ERa expression is preceded by genetic instability in the initiation of BRCA1-dependent tumor-igenesis, indicating that the breast epithelium of BRCA1 mutation carriers may initially be estrogen-responsive. Our resultsimply that genomic instability instigated by BRCA1 haploinsufficiency may be required for breast cancer initiation in BRCA1mutation carriers. Finding molecular markers of tumor initiation and progression, for the potential use in early diseasedetection, may be of great clinical importance for the improved management of at-risk women. VVC 2009 Wiley-Liss, Inc.

INTRODUCTION

Despite intense research and substantial pro-gress in understanding the disease, breast cancerremains the most common form of cancer amongwomen and is the second most common cause ofcancer-related death among women in the West-ern world. Hereditary susceptibility caused bymutations in autosomal dominant genes is respon-sible for 5–10% of all breast cancers and !10% ofall invasive epithelial ovarian cancers (Goldbergand Borgen, 2006; Guillem et al., 2006). Inheri-tance of a deleterious mutation in one allele ofeither of the breast cancer susceptibility genesBRCA1 or BRCA2 confers a lifetime risk of devel-oping breast cancer of 44–78% and a lifetime riskof ovarian cancer of 18–54% (Antoniou et al.,2003). Women who are BRCA1/2 mutation car-riers also have an increased risk of developingcontralateral breast cancer, with the 10-year riskranging from 25 to 31% (compared to 4–8% insporadic cases) (Liebens et al., 2007). A signifi-

cant proportion of these risks occur in womenunder the age of 50 years. Although the full spec-trum of functions of the BRCA proteins remainsto be elucidated, it is clear that they play a cen-tral role in the management of DNA damage,regulation of the cell cycle, and control of tran-scription, and are essential for the preservation ofnormal growth and development of epithelial

Additional Supporting Information may be found in the onlineversion of this article.

Supported by: Mr. R. Thurston; The Swedish Cancer Society;The Swedish Research Council; The Mrs. B. Kamprad Foundation;The G. Nilsson Cancer Foundation; The A. Wiberg Foundation;The Lund University Hospital Research Foundation; TheGovernmental Funding of Clinical Research Within NationalHealth Service; KR and IH were supported by the Swedish CancerSociety.

*Correspondence to: Ingrid Hedenfalk, Department of Oncol-ogy, Clinical Sciences, Lund, Lund University, Barngatan 2B,SE-221 85 Lund, Sweden. E-mail: [email protected]

Received 18 June 2009; Accepted 8 September 2009

DOI 10.1002/gcc.20723

Published online 16 October 2009 inWiley InterScience (www.interscience.wiley.com).

VVC 2009 Wiley-Liss, Inc.

cells in the breast and other tissues (Venkitara-man, 2002; Mullan et al., 2006).Current recommendations for female BRCA1/2

mutation carriers include participation in surveil-lance programs, chemoprevention, and risk-reduc-ing surgery (prophylactic mastectomy, PM, andsalpingo-oophorectomy, PO) (Wooster andWeber, 2003; Domchek and Weber, 2006). Fur-ther, several studies have suggested that mag-netic resonance imaging (MRI) is superior toother imaging techniques for the detection ofearly stage malignancies in BRCA1/2 mutationcarriers (Warner et al., 2004; Hagen et al., 2007).Nevertheless, prophylactic removal of the breastsand ovaries is by far the most effective preven-tion strategy for these women; PO reduces therisk of developing breast and ovarian cancer by!50 and 90%, respectively, and PM reduces theoverall risk of breast cancer by as much as 90%,according to a recent ASCO/SSO review (Guillemet al., 2006).To date, very little is known about genetic

aberrations causing the initial transformation ofnormal breast epithelial cells into invasive carci-noma cells in carriers of BRCA1/2 mutations.Cytogenetic characterization of normal breast tis-sue derived from PMs from women geneticallypredisposed to breast cancer has revealed thatchromosome arms 1q and 3p are frequentlyinvolved in genetic rearrangements (Peterssonet al., 1996; Teixeira et al., 1996). However, nospecific genetic changes were linked to patientswith a known mutation in the BRCA1 gene in ei-ther of these studies. In studies of loss of hetero-zygosity (LOH), allelic imbalance (AI) has beenfound in normal tissue both adjacent to, and dis-tant from BRCA1, BRCA2, and sporadic tumors,as well as in reduction mammaplasty (RM) speci-mens (Deng et al., 1996; Lakhani et al., 1999;Cavalli et al., 2004; Clarke et al., 2006). Usinglow resolution techniques, AI was found up tothree times more frequently in normal epithelialstructures from BRCA1 mutation carriers com-pared to controls (Larson et al., 2005; Clarkeet al., 2006). Taken together, these findings indi-cate that genetic aberrations may be present inhistopathologically normal-appearing tissue in ge-netically predisposed women; however, the fre-quency and magnitude of these changes needs tobe further investigated using e.g. high-resolutionglobal molecular profiling.In the current study we have used array-based

comparative genomic hybridization (aCGH) toanalyze multiple normal breast tissue specimens

obtained from 10 BRCA1 mutation-positivepatients who underwent risk-reducing PMs. Ouraim was to identify genomic changes that may beresponsible for malignant transformation andearly tumorigenic events, and to explicatewhether these changes may represent precursorsto BRCA1-dependent invasive breast cancer.Moreover, these findings may provide informationregarding tumorigenesis in general, with possibleimplications for improved diagnosis and develop-ment of tailored treatment strategies for breastcancer as well as other malignancies. The recenttrend toward improved breast cancer mortalityrates is partly due to increased diagnosis of earlystage disease. Finding molecular markers for earlytumor development and progression could betranslated into a surveillance measure for veryearly disease detection, leading to improvedpatient management and outcome among high-risk women. These markers could also be used asa descriptor to identify the individuals (!20%),who will not develop breast cancer despite beingBRCA1 mutation carriers, thereby avoidingunnecessary risk-reducing surgery.

To our knowledge, this is the first in-depthreport of genomic instability in histopathologi-cally normal, prophylactically removed breasttissue from women carrying a mutation in theBRCA1 gene, using high resolution globalgenomic profiling.

MATERIALS AND METHODS

Clinical Specimens

Since 1995, more than 150 women with breastcancer predisposition (BRCA1/2 as well as non-BRCA1/2) have undergone PMs at the Depart-ment of Plastic and Reconstructive Surgery atMalmo University Hospital, Sweden. Immedi-ately after surgery, the removed tissues were X-rayed, and then thoroughly examined by an expe-rienced breast pathologist to discover potentialmalignancies. Four to eight normal tissue speci-mens from each quadrant of each breast werethen snap frozen and stored in the tissue bank("80#C) at the Department of Pathology, MalmoUniversity Hospital, Sweden.

Included in the present study were multiplefreshly frozen breast tissue specimens from bothbreasts (when applicable) from 10 consecutiveBRCA1 mutation carriers (1–8 samples/breast; 2–11 samples/patient). The study was approved bythe ethics committee at Lund University,

BRCA1 AND GENOMIC INSTABILITY 79

Genes, Chromosomes & Cancer DOI 10.1002/gcc

Sweden. Informed consent was obtained from allpatients included in the study. An overview ofpatients and clinical data are presented in Table1. Two of the ten patients had a previous diagno-sis of invasive ductal carcinoma (IDC; patientsno. 10 [mastectomy] and no. 15 [partial mastec-tomy]). One of the patients underwent a contra-lateral PM simultaneously with a therapeuticmastectomy for a small IDC in the left breast(patient no. 9), and in one patient a ductal carci-noma in situ (DCIS) was detected at the time ofthe bilateral PM (patient no. 7). For referencepurposes normal breast tissue from five age-matched healthy women, without a family historyof breast cancer, was included. These patientshad all undergone RM in the late 1980s, andnone of them developed breast cancer within 15years after surgery. Also included in the studywas fresh frozen tumor material (IDC) from fourother BRCA1 mutation carriers.

DNA Extraction

From each tissue specimen, frozen sectionswere obtained in the following way: one 4-lmsection followed by two 100-lm sections, one 4-lm section, two 100-lm sections, and finally one4-lm section. All sections were transferred ontomicroscope slides and stored at "80#C. The three4-lm sections from each tissue specimen werestained with H&E and visually inspected to con-firm that no microscopic malignancies were pres-ent, and that the sections contained a significantamount of normal epithelial structures. DNA wasextracted from the two 100-lm sections closest tothe 4-lm section with the highest density of nor-mal epithelium. Each 100-lm section was cutinto small pieces and transferred to a 1.5-mlEppendorf tube and DNA was extracted with theDNeasy blood and tissue kit (Qiagen GmbH,Hilden, Germany), according to the manufac-turer’s instructions.

Array-Based Comparative Genomic

Hybridization (aCGH)

For aCGH hybridizations the Human GenomeCGH Microarray 44A and B platforms (AgilentTechnologies, Santa Clara, CA) were used. Thearrays contain 42,494 distinct 60-mer oligonucleo-tide features, with an average spatial resolution of!35 kb.DNA samples were processed as follows;

restriction enzymes Alu1 and Rsa1 (Invitrogen,

Carlsbad, CA) were used to digest 900 ng each ofsample and reference DNA (Human genomicDNA; Promega, Madison, Wisconsin). TheDNAs were then labeled with Cy5-dUTP andCy3-dUTP, respectively (Amersham Biosciences,Piscataway, NJ), using the BioPrime Array CGHGenomic Labeling System (Invitrogen). Afterclean-up (Vivaspin 500; Sartorius AG, Goettingen,Germany) the differentially labeled DNAs weremixed and hybridized to the arrays according tothe Agilent Oligonucleotide Array-Based CGHProtocol (Agilent Technologies). After a 42-hrincubation, the arrays were washed according towash procedure A of the manufacturer’s instruc-tions. Arrays were scanned using the Agilentmicroarray scanner G2505A and processed usingFeature Extraction 8.1 (Agilent Technologies).Default settings were used. Nexus Copy Number3.1 (BioDiscovery, El Segundo, CA) was used tosegment the data and to estimate copy numbers.The significance threshold was set to 0.001. Thisthreshold was chosen on the basis of the numberand type of aberrations found in the control RMtissues. A small number of small genomic aberra-tions, mainly known polymorphisms, were consid-ered acceptable in the control samples.Aberrations were defined depending on the qual-ity of the hybridizations. For hybridizations witha very good quality (QC $ 0.100) an aberrationwas called when the log ratios of two or more ad-jacent and unique clones exceeded %0.20, forsamples with a good quality (0.100 < QC $0.180) this threshold was set to %0.25. Hybridiza-tions of poor quality (QC > 0.180) were disre-garded as were hybridizations that gave rise towhole genome plots with a spiky pattern of veryspecific high level amplifications mainly at chro-mosomes 11 and 19 (most likely the result ofcompromised DNA quality).

Loss of Heterozygosity (LOH)

LOH analysis for 17q21.31 was performedusing peripheral blood from each patient as anindividual reference. For patients no. 12 and 13,no peripheral blood was available, and thereforethese patients were excluded form this analysis.LOH was performed on tissue specimens from allbreasts examined by aCGH (except for the leftbreast of patient no. 14, due to insufficientDNA). Five fluorescently labeled genetic markers(D17S855, D17S1185, D17S1322, D17S1323, andD17S1326; Gibco BRL, Carlsbad, CA), situated inor next to the BRCA1 gene, were used to

80 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

TABLE

1.Overview

ofPatient

Inform

ationandClinicalData

Patient

no.

BRCA1mutation

Type

of

surgery

Age

atsurgery

Oophorectomy

prePM

Patho

logy

finding

atPM

(R)

Patho

logy

finding

atPM

(L)

Previous

breast

malignancy

Pre-PM

therapy

1RM

39–

––

2RM

38–

––

3RM

46–

––

4RM

48–

––

5RM

36–

––

6Ex11:1201

del

11BiPM

38No

multipleapocrine

cysts

intraductal

papilloma

periductal

lymphocyte

infiltration

7Intron17:IVS17-2

A-C

BiPM

484months

prior

(corpus

cancer

detectedat

PO)

DCIS,10

mm,

Ng2,

secretory

lobuli

secretory

lobuli

8Ex

5:300T-G

BiPM

35No

UDH

&multiple

apocrinecysts

UDH

&multiple

apocrinecysts

9Ex

5:300T-G

Uni

(R)PM

461year

prior

Sclerosing

adenosiswith

smallapocrine

cysts

NoPM

(simultaneous

therapeuticME)

(L) ID

C(T1N

0)ER

&,PR&,

Her2-,HG

III(detectedat

pre-PM

screening)

10Ex11:1201

del

11Uni

(R)PM

385months

prior

–NoPM

(previous

ME)

(L)IDC

(T2N

1)ER

-,PR-,

Her2-,HG

III(21months

prior)

FEC

11Ex

11:3172

ins5

BiPM

436yearsprior

UDH

ADH

fibroadenoma

withapocrine

metaplasia

HRT

12Ex

11:3172

ins5

BiPM

439yearsprior

––

HRT

13Ex

11:2594

del

CBiPM

35No

–multiplesm

all

fibroadenomas

14Ex

11:2594

del

CBiPM

30No

––

OC

15Ex

11:2594

del

CBiPM

40No

––

(L)MED

CA

(T1N

0)ER

-,PR-,HG

III(5

yearsprior)

FEC

&RT

Patientsno

.1–5,

highlighted

ingray,arecontrolpatients

(withno

family

history

ofbreast

cancer)who

have

undergone

areductionmam

maplasty.Patientsno

.6–15

areBRCA1mutationcarriers.RM,reduc-

tionmam

maplasty;

PM,prophylacticmastectomy;

PO,prophylacticoophorectomy;

ME,mastectomy;

Bi,bilateral;Uni,unilateral;R,right;L,left;UDH,usualductalhyperplasia;

ADH,atypical

ductal

hyperplasia;DCIS,ductalcarcinomain

situ;ID

C,invasive

ductalcarcinoma;MED

CA,medullary

carcinoma;Ng,nucleargrade;

HG,histologicalgrade;

HRT,ho

rmone

replacem

enttherapy;OC,oralcon-

traceptive;FEC,po

ly-chemotherapy(5-fluo

rouracil,epirubicin,cyclophosphamide);TAM,tamoxifen;

RT,radiotherapy.

BRCA1 AND GENOMIC INSTABILITY 81

Genes, Chromosomes & Cancer DOI 10.1002/gcc

evaluate LOH of this region. PCR was performedusing a Thermo Hybaid MBS 0.2S (ThermoFisher Scientific, Waltham, MA), followed by flu-orescent quantification on a 3130xL Genetic Ana-lyzer (Applied Biosystems, Foster City, CA).

Promoter Methylation Pyrosequencing

For promoter methylation analysis, DNA sam-ples were first bisulfite converted using the EZDNA methylation gold kit (Zymo Research, Or-ange, CA), followed by pyrosequencing on aPSQTMHS 96A, using the Pyro Gold kit (bothfrom Biotage AB, Uppsala, Sweden) according tothe manufacturers’ instructions. The sequencesfor two pyrosequencing primers, together cover-ing nine methylation sites in the promoter regionof the true BRCA1 gene, were: 50-GGTAGTTTTTTGGTTTT-30 and 50-GAAT-TATAGATAAATTAAAA-30 (Biomers.netGmbH, Ulm, Germany). The software programon the PSQTMHS 96A was applied for analysis ofthe results. Positive and negative control DNAswere included in all runs (CpGenomeTM Univer-sal Methylated/Unmethylated DNA, Millipore,Billerica, MA).

Tissue Microarray (TMA) Construction

We constructed a tissue microarray (TMA) ofparaffin-embedded breast tissue specimens fromall patients undergoing a PM included in thestudy. A fresh section was cut from the originalblocks and stained with hematoxylin and eosin tolocate areas with a significant number of epithe-lial structures. From these areas, four core biop-sies (diameter 1.0 mm) were punched anddeposited in a recipient paraffin block using anautomated arrayer (Beecher Instruments, Sun

Prairie, WI). Core biopsies from the three IDCsand the DCIS previously or simultaneouslydetected in these patients were also included inthe array.

Immunohistochemistry and Gene

Expression Analysis

Immunohistochemical staining of estrogen re-ceptor (ER) a and progesterone receptor (PR)was performed on 4-lm sections using the Ven-tana Benchmark system with prediluted antibod-ies (anti-ER clone 6F11 and anti-PR clone 16;Ventana Medical Systems, Tucson, AZ).

Previously published tumor gene expressiondata (Ma et al., 2004; Nielsen et al., 2004) wereretrieved from Gene Expression Omnibus(GDS806 and GDS850). All samples were hybri-dized onto two-color arrays against Stratagene’suniversal reference as described.

RESULTS

Analysis of Normal Breast Tissue from Women

Without a Family History of Breast Cancer

A small number of genomic changes weredetected in the normal tissue samples from fivecontrol RMs from healthy women without a fam-ily history of breast cancer (range 7–14; mean '12; median ' 13; Table 2). These changes weretypically small, ranging from 39 to 183 clones(mean ' 112; median ' 123) or 3–10 Mb (meanand median ' 7 Mb). The majority of thesechanges were found in genomic regions wherepolymorphisms have previously been reportedaccording to The Centre for Applied Genomics(TCAG) Database of Genomic Variants (humangenome build 36; http://projects.tcag.ca/variation/).Several of the aberrations identified in regions

TABLE 2. Overview of Genomic Aberrations by aCGH

Tissue type Control (n ' 5) PM tissue (n ' 44) Tumor (n ' 4) P-valuea

No. of aberrations Mean 12 29 109Median 13 28 103 0.0006Range 7–14 5–69 93–136

No. of genes Mean 81 281 8,387Median 95 187 8,203 0.0009Range 23–119 29–1,978 7,539–9,602

No. of aberrant clones Mean 112 517 19,060Median 123 361 18,819 0.0006Range 39–183 23–3,261 16,085–22,519

No. of aberrant base pairs (Mb) Mean 7 31 1,237Median 7 24 1,222 0.0006Range 3–10 3–158 997–1,507

aKruskal-Wallis equality-of-populations rank test, based on the median value for each PM patient.

82 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

without previously known polymorphisms werefound in more than one control sample, includinga small gain at 9q12 (with no known genes orpolymorphisms) found in four out of five samples.In two out of five samples we found a loss at4q34.2 with five known genes, a small loss at5q33.2-q33.3 harboring one known gene, a loss at12q23.3 with four known genes, a gain at 16p13.3with five known genes, and a loss at 19q13.31with eight known genes. We also found a gain inthe telomeric region of chromosome arm 22q(22q13.33) in three out of five samples. Thisregion includes two previously known polymor-phisms. These genomic regions most likely repre-sent previously unreported sites of polymorphism;none of the genes found therein have been impli-cated in neoplasia. To verify that the aberrationsin the normal RM control samples are indeedpolymorphisms, patient-matched peripheral bloodshould preferably be used as a reference in theaCGH experiments; unfortunately, however, bloodsamples from these patients were not available atthe time when the aCGH analyses wereperformed.

Genomic Aberrations in Histopathologically

Normal PM Tissue from BRCA1 Mutation Carriers

The number of genomic aberrations found inthe 44 PM samples, which varied considerablybetween samples (range 5–69 changes/sample;

mean ' 29; median ' 28), was in most caseshigher than in the five control RM samples butsignificantly lower than in the four invasive tumorspecimens (P ' 0.0006, Kruskal-Wallis equality-of-population rank test; range 93–136; mean '109; median ' 103; Table 2). Several of the sam-ples were run in duplicate and showed identicalaberration patterns (data not shown). The aberra-tions found in PM samples were typically lowcopy number changes; they were generally largerthan aberrations found in control samples, butsmaller than those found in malignant tissues.The number and size of genomic aberrationswithin each group is presented in Table 2. Themajority of these aberrations have not previouslybeen described in non-malignant tissue, althoughseveral of them are found in known polymorphicregions.

Genomic Heterogeneity

As reported in the previous section, the levelof genomic heterogeneity within this cohort isvast; this holds true for both between-patient andwithin-patient comparisons. The mean number ofgenomic aberrations within one individual variedfrom 17 (patient no. 15) to 56 (patient no. 10).The number of aberrations also varied greatlybetween samples from the same individual, withthe largest variation observed for patients no. 6(range 5–61) and no. 7 (range 6–35). Figure 1

Figure 1. Number of genomic aberrations found in control RM (n ' 5), PM (n ' 44), and invasivetumor (n ' 4) tissues, respectively. Each bar represents individual samples. Samples from the samepatient have been colored identically. Note the variation between samples from the same patient andeven within the same breast. R, right breast; L, left breast.

BRCA1 AND GENOMIC INSTABILITY 83

Genes, Chromosomes & Cancer DOI 10.1002/gcc

illustrates the number of genomic aberrationsfound in all tissue samples analyzed.When comparing the pattern of genomic aberra-

tions between samples from the same individual,we found that samples with an increased numberof aberrations compared with the control RM tis-sues often displayed similar genomic patterns.Interestingly, this holds true also when comparinggenomically aberrant samples from both breastsfrom an individual. The level of intra-individualheterogeneity is shown in Supporting InformationFigure 1, where the seven samples analyzed frompatient no. 6 are displayed.

Common Genomic Aberrations

We found several genomically aberrant regionsin common between PM samples. The regionsmost commonly found to harbor a statistically sig-nificant number of aberrant gene copies (P $0.001; the frequency P value is defined in theSTAC algorithm at http://genome.cshlp.org/con-tent/16/9/1149.abstract) across all PM samples,and the genes included in these regions are listedin Table 3.

Performing a gene ontology enrichment analy-sis on the significantly aberrant regions revealedthat when correcting for multiple analyses only

TABLE 3. Significantly Aberrant Regions (P $ 0.001) Occurring in More than 5% of All PM Tissue Samples and the GenesLocated in These Regions

Cytoband Start Stop Gain/loss CNV Frequency (%) Genes

1p35.3 28237143 28836370 Loss No 14 DNAJC8, ATPIF1, SESN2,MED18, PHACTR4,SNHG3-RCC1, RCC1,TRSPAP1, RAB42, TAF12,GMEB1

1p34.1 45470087 45635082 Loss Yes 12 MUTYH, TOE1, TESK2,LOC126661, MMACHC

1q23.1 153583409 154021519 Gain No 12 SH2D2A, NTRK1, INSRR,PEAR1, C1orf92, ARH-GEF11, ETV3L, ETV3

1q23.2 156632952 157009148 Gain No 12 VSIG8, CCDC19, TAGLN2,IGSF9, SLAMF9, PIGM,KCNJ10, KCNJ9, IGSF8,ATP1A2, ATP1A4, CASQ1,PEA15

1q23.3 161866286 162273688 Gain No 14 LMX1A, RXRG, LRRC521q44 245338349 245522847 Gain Yes (telomere) 21 ZNF672, ZNF692, PGBD22q14.1 115267854 118299243 Loss No 12 DPP10, DDX182q21.3 134916297 135166174 Gain No 12 –2q24.2 160093053 160300462 Loss No 26 BAZ2B2q31.1 176787632 176871951 Gain No 9 HOXD12, HOXD11,

HOXD10, HOXD9,HOXD8, HOXD4, HOXD3

3p25.3 9936071 10304111 Loss No 16 IL17RC, CRELD1, PRRT3,TMEM111, LOC401052,FANCD2, C3orf24,C3orf10, VHL, IRAK2,TATDN2, GHRL

3q13.33-q21.1 123062292 123609902 Loss Yes 12 EAF2, SLC15A2, ILDR1,CD86, CASR, CSTA,CCDC58, FAM162A

4p16.1-p15.33 9623378 11252194 Gain Partial 16 WDR1, ZNF518B, MIST,HS3ST1

4q12 57174067 57570031 Loss No 16 SRP72, ARL9, HOPX, SPINK24q34.3 179738319 181159302 Gain Partial 12 –5q13.1-q13.2 67896847 68899970 Loss Partial (distal part) 26 SLC30A5, CCNB1, CENPH,

MRPS36, CDK7, CCDC125,TAF9, RAD17, MARVELD2,OCLN, LOC730394,LOC728340

Regions covered by known polymorphisms are highlighted in gray. The gain on 19q13.31 may represent a new polymorphism as it was also foundin two out of five control RM samples.

84 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

two biological processes—DNA-dependent regu-lation of transcription (GO:0006355) and multi-cellular organismal development (GO:0007275)—were significantly correlated to these commonaberrations and sequence-specific DNA binding(GO:0043565) and transcription factor activity(GO:0003700) were the two molecular functionslinked to the aforementioned aberrations (Fisher’sexact test; P $ 0.001). Significant for all of theseprocesses/functions were the HOXD and HOXAgene clusters (at 2q31.1 and 7p15.2, respectively),several zinc finger genes, TATA-binding proteinassociated factor 9 (TAF9), mediator of RNA poly-merase II transcription subunit 18 (MED18),cyclin dependent protein kinase 7 (CDK7),RAD17, the Fanconi anemia complementationgroup D2 (FANCD2), cyclins B1 and E1 (CCNB1and CCNE1), sestrin 2 (SES2), caspase 8 associ-ated protein 2 (CASP8AP2), and the tumor proteinp53-inducible nuclear protein 1 (TP53INP1).

Genomic Aberrations and Clinical Parameters

No statistically significant correlations wereestablished between the number of changes pres-ent in PM tissues and the age of the patient atPM surgery, previous PO (i.e., menopausal sta-tus), previous or simultaneous breast cancer diag-nosis or previous local or systemic therapy (datanot shown).

LOH Analysis and Promoter Methylation

Pyrosequencing

LOH at the BRCA1 locus was indicated in theaCGH analysis (Fig. 2A) and confirmed in multi-ple samples from both breasts from one of the 10PM patients in the study (Fig. 2B). This woman(patient no. 7) was diagnosed with DCIS in theright breast at the time of the prophylactic sur-gery. Results from the pyrosequencing analysisrevealed that none of the samples included in thestudy were methylated in the promoter region ofthe BRCA1 gene (Fig. 2C).

Analysis of Hormone Receptor Levels

Immunohistochemical quantification of ERaand PR was possible in the normal breast tissuefrom all patients (all but one breast). All of thesetissues showed moderate to strong positive ERastaining in !15% of the epithelial cells (Fig. 3A).Breast tissue from one patient was negative for PRin all samples from both breasts, and for onepatient only one biopsy, which was PR negative,

could be evaluated. All other biopsies were weaklyto moderately positive in !15% of the cells.

mRNA levels of ERa, ERb, and PR from geneexpression microarray analyses of the PM tissuesand RM control specimens (data not shown) werecompared with previously published transcrip-tional profiling data of ER positive and negativebreast cancers (Gene Expression OmnibusGDS806 and GDS850 (Ma et al., 2004; Nielsenet al., 2004)). The expression of ERa differed sig-nificantly between the three groups, such thatPM tissues displayed levels intermediatebetween ERa positive and negative cancers (Fig.3B), while the expression of ERb and PR did notdiffer (data not shown).

DISCUSSION

For individuals predisposed to breast cancer,especially by mutations in the BRCA1 gene, accu-mulation of somatic genetic changes seems to fol-low a unique pathway during tumorigenesis, andthe overall number of genomic changes (asassessed by metaphase CGH) is much higher inBRCA1-dependent tumors compared to sporadictumors (Tirkkonen et al., 1997; van Beers et al.,2005). The specific genomic profiles of BRCA1,BRCA2, and sporadic tumors have been refinedusing aCGH, implicating loss of 4p, 4q, and 5q,and gain of 3q27.1-q27.3 as specific for BRCA1-de-pendent tumors (as compared to BRCA2-depend-ent and sporadic cases) (Jonsson et al., 2005).Distinct genomic profiles have also been estab-lished for two subgroups of familial non-BRCA1/2-dependent breast tumors (Hedenfalk et al., 2003).

Our findings clearly show that genomic aberra-tions can be found in histopathologically normalbreast tissue from BRCA1 mutation carriers.These aberrations were typically low copy num-ber gains and losses, which were more frequentin PM tissues from BRCA1 mutation carriers thanin normal control tissues, but much rarer (andinvolving a smaller part of the genome) thanaberrations found in BRCA1-dependent invasivetumors. Losses were more frequent in PM tissuesthan gains, which was also the case in BRCA1-de-pendent invasive cancers. The number ofgenomic aberrations varied greatly between sam-ples, both between and within individuals. Thismay be explained by the way a cancer is believedto arise, whether by clonal evolution or from can-cer stem cells (Campbell and Polyak, 2007). Ineither case the genomic instability leading to tu-mor development is thought to arise in a single

BRCA1 AND GENOMIC INSTABILITY 85

Genes, Chromosomes & Cancer DOI 10.1002/gcc

Figure 2. Deletion of the BRCA1 gene as revealed by aCGH (A)and confirmed by LOH analysis (B). In the aCGH plot, individualprobes are represented by dots (red and green indicating gain andloss, respectively). The BRCA1 locus is highlighted in red, and thethree BRCA1 probes deleted in this sample are indicated. The upperpart of the LOH panel (B) shows the mutant (Mut) and wild-type(WT) alleles respectively, present in the blood sample from patientno. 7. The lower panel shows loss of the wild-type allele, which was

detected in samples from both breasts from the same individual.Panel C shows the results from the pyrosequencing analysis of meth-ylation sites 6–10 of the promoter of the BRCA1 gene. The upperand middle panels show the unmethylated and methylated controls,respectively. The lower panel shows the results of one sample frompatient no. 15. This sample, as all other samples tested (from allpatients) was unmethylated. AU, arbitrary units; Bp, base pair.

86 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

cell, giving it growth and survival advantages overthe surrounding normal cells. Our investigation oftissue samples from multiple breast locationsrevealed a high level of tissue heterogeneity; weidentified regions with genomic instability (i.e.,many genomic aberrations), as well as regions thatremained genomically stable (i.e., none or fewaberrations). Interestingly, the same genomic aber-rations were often discovered when genomicallyunstable tissue samples from different parts withinone breast were compared, and even upon com-parison with the contralateral breast. This mayindicate a common chain of events leading to trans-formation and initial tumor progression within anindividual, possibly influenced by hormonal levels,as well as exposure to environmental factors.

Using patient-matched peripheral blood as areference in each hybridization would haveenabled us to definitely exclude the possibilitythat normal CNVs or polymorphisms influencedour findings. However, if the detected aberra-tions were merely individual polymorphisms,these would have been present in all samples(from both breasts) from a single patient; thiswas not the case. In addition, we have excludedpreviously reported CNVs from our analysis.Taken together, we believe that the aberrationsreported herein represent true genomic changes,and are not merely a reflection of normal geneticvariation.

We were able to establish patterns of similarityacross samples from different individuals, includ-ing regions often involved in BRCA1-dependentinvasive breast tumors (this study and Jonssonet al., 2005). One-third of the significantly aber-rant regions in the histopathologically normal PMtissues have previously been described as specificto BRCA1-dependent cancers, and many morehave been implicated in breast cancer in general.Many of the changes found in the genomicallyunstable PM tissues were also found in the inva-sive tumors included in the present study (19losses and 12 gains, data not shown). Thesetumors were from patients other than those con-tributing to the PM data, and the fact that thesame genomic aberrations were observed in a dif-ferent cohort of patients strengthens our findings.

Interestingly, more than half of the most com-mon aberrations coincide with previouslyreported translocation break points in fibroade-noma of the breast (Petersson et al., 1997) orwith AI reported in atypical ductal hyperplasia(Allred et al., 2001). These lesions are consideredbenign alterations of the epithelium or indicatorsof breast cancer risk; discovering the same aber-rant regions in PM tissue as in these lesions maypossibly suggest that they in fact constitute earlyneoplastic genomic events that subsequently maygive rise to phenotypic changes and malignanttransformation. A few of these aberrations havealso been reported in premalignant or earlylesions in other tissue types such as squamousmetaplasia and dysplasia of the lung and Ta stagedisease of bladder cancer (Ma et al., 2006; Cor-don-Cardo, 2008). Taken together, the hereinreported findings of similarities to genomic altera-tions in BRCA1-dependent tumors or early neo-plastic events may indicate genomic regions thatare important for BRCA1-dependent tumorigene-sis, or are necessary for genomically aberrant cells

Figure 3. Expression of ERa in breast tissues. Immunohistochemi-cal staining of PM tissues (A), revealed moderate to strong positiveERa staining in !15% of the epithelial cells of all samples analyzed.The lower panel (B) shows the gene expression ratios of control RMand BRCA1 mutant PM tissues, and ERa positive and negative invasivetumors, respectively. ERa positive tumors displayed elevated expres-sion and ERa negative tumors displayed decreased expression com-pared to the reference. Non-malignant breast specimens fromcontrol RM as well as from BRCA1 mutant PM tissues displayed inter-mediate levels of ERa expression.

BRCA1 AND GENOMIC INSTABILITY 87

Genes, Chromosomes & Cancer DOI 10.1002/gcc

to maintain their proliferative capacity or toevade apoptosis.Ontological enrichment analyses confirmed

these findings and implicated DNA-dependentregulation of transcription, multi-cellular organis-mal development, sequence-specific DNA bind-ing, and transcription factor activity as thebiological processes and molecular functionslinked to these common aberrations. Our data indi-cate that changes linked to transcriptional proc-esses constitute the first occurrence of genomicinstability in these cells with defective DNA repairmachinery, preceding aberrations of specific onco-genes or tumor suppressor genes, and even visuallydetectable phenotypical changes. This finding isnot surprising but in fact rather intuitive.We identified a large number of aberrant tran-

scription factors located in these regions, such asthe HOXA and HOXD gene clusters. These genesencode DNA-binding transcription factors whichregulate gene expression, morphogenesis, and cel-lular differentiation. They have previously beenassociated with expression of ERa, the TP53 pro-tein and the development of breast cancer (Ramanet al., 2000; Chen and Sukumar, 2003; Chu et al.,2004). A number of the aberrant regions also har-bor genes encoding zinc finger proteins which arethought to participate in the regulation of tran-scription. Other genes involved in the initiation oftranscription, DNA repair, cell cycle regulation,and apoptosis found in these regions are TAF9,MED18, CDK7, RAD17, FANCD2, CCNB1,CCNE1, SES2, CASP8AP2, and TP53INP1.The causative mechanism(s) underlying the

genomic imbalances described remain unclear,but could involve LOH or methylation of thewild-type allele of the BRCA1 promoter. Both ofthese mechanisms would result in BRCA1 mutantcells without functional BRCA1 protein, therebydriving the DNA repair machinery away from ho-mologous recombination and into more error-prone repair systems. It is also possible that thesechanges reflect a dominant-negative action of theproduct of the mutant BRCA1 gene over thewild-type BRCA1 protein. However, the patientsincluded in the present study harbor differentBRCA1 mutations (Table 1), and hence theresulting gene products would not all beexpected to have the same dominant-negativeaction. LOH of BRCA1 was detected in multiplesamples from only one of the individuals tested(simultaneously diagnosed with DCIS; Fig. 2B),suggesting that LOH is not the predominantmechanistic origin of the observed genomic insta-

bility. Moreover, BRCA1 promoter methylationwas not observed in any of the samples analyzed(Fig. 2C). Taken together, this suggests that thegenomic instability must be caused by an alterna-tive mechanism. Further to this notion, it haspreviously been proposed that BRCA1 haploin-sufficiency might be causal for the initiation ofbreast cancer, making cells vulnerable to mitoticrecombination (Staff et al., 2003; Cousineau andBelmaaza, 2007). Bartek et al., (2007) proposedthe theory of ‘‘conditional haploinsufficiency’’,arguing that DNA breakage and replication stressis low in adult tissue homeostasis, whereas in pre-cancerous and early cancer lesions the thresholdof endogenous DNA breakage is significantlyhigher. They suggest that there is a criticalthreshold of DNA damage that needs to be gen-erated for the DNA damage response (DDR) ma-chinery to become activated. Loss of the wild-type allele of BRCA1 as the first event in aBRCA1 mutation carrier would generate massiveDNA damage which would activate the DDRmachinery and most likely lead to cell death,whereas sub-threshold events such as the smallgenomic aberrations identified in the presentstudy may lead to clonal expansion. These aber-rations may lead to the evasion of cell cycle arrestand apoptosis, both oncogenic cellular processesimplicated by the aberrant genes in the histopa-thologically normal PM tissue in this study. Thisloss of cell cycle homeostasis would subsequentlyallow more genomic aberrations to accumulate,eventually leading to malignant transformation(Bartek et al., 2007).

One of the unexplained features of BRCA1-de-pendent breast cancers is that up to 90% exhibitloss of ERa expression (Foulkes et al., 2004;Lakhani et al., 2005). Hosey et al., (2007)recently proposed an elegant link betweenBRCA1 mutations and ERa negativity. Theirresults suggest that the paucity of ERa expres-sion in BRCA1-dependent tumors is due to lossof BRCA1-mediated transcriptional activation ofESR1 and that this loss of ERa expression occursafter loss of the wild-type BRCA1 allele, eitherdue to LOH or epigenetic silencing. As ourresults indicate: (i) the presence of genomic insta-bility in normal breast tissue from BRCA1 muta-tion carriers prior to this secondary hit (LOH orpromoter methylation), (ii) the expression of ERaat the protein level in PM tissues is similar towhat has previously been reported for normalbreast tissue, and (iii) the mRNA levels are equalto those seen in normal control RM tissues and

88 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

intermediate between those seen in ERa positiveand ERa negative cancers (Fig. 3B), we proposethat the loss of ERa expression as well as LOHof BRCA1 occurs after genomic instability isestablished. This would imply that tumors arisingin BRCA1 mutation carriers may initially respondto estrogen stimulation. Furthermore, our findingsmay explain the suggested benefit of preventivetamoxifen treatment that has been reported forindividuals with BRCA1 mutations (Narod et al.,2000; Gronwald et al., 2006; Pierce et al., 2006),as well as the reduced risk of breast cancer inBRCA1 mutation carriers who have undergoneprophylactic removal of the ovaries (Rebbecket al., 1999), as these tumors may initially bedependant on ovarian hormone exposure. Alterna-tively, BRCA1-dependent tumors may arise fromthe ERa negative cells of the mammary epithe-lium, while our analysis has been performed onan unselected cell population, prior to the hypo-thetical expansion of an ERa negative compart-ment. These possibilities need to be furtherinvestigated, preferably by analyzing distinct cellpopulations, as does the potential benefit of pre-ventive anti-hormonal therapy for these women.Our findings indicate that the genomic instabil-

ity that can be detected in histopathologicallynormal breast tissue from women at high risk ofdeveloping breast cancer due to BRCA1 muta-tions cannot be explained solely by epigeneticsuppression or loss of the wild-type BRCA1 allele.This implies that molecular aberrations precedingmicroscopically detectable lesions may signifygenomic instability in BRCA1 mutation carriers.We propose that BRCA1 haploinsufficiency, caus-ing genomic instability, may be conductive tobreast cancer initiation. We have identifiedgenomic regions that may be of importance forthe initiation of tumorigenesis and regions thatmay sustain the progression of an establishedmalignancy. The origin of the genomic aberra-tions, specifically whether they originate fromstem cells or are the result of clonal evolution ofmature cells, and if these genomic aberrationsarise in all or part of either the epithelial or stro-mal cells, as well as their correlation to ERa sta-tus, may be further elucidated by microdissectionand subsequent individual analysis of putativeprecursor cells within the breast. Finding molecu-lar markers of tumor initiation and progression,for the potential use in early disease detection,may be of great clinical importance in terms ofimproved management of women at high risk ofdeveloping breast cancer.

ACKNOWLEDGMENTS

The authors are indebted to the women whoconsented to participate in this study, and wishto thank the Pathology Department at MalmoUniversity Hospital, Sweden for their assistancein providing tissue samples. They thank GoranJonsson for providing tumor DNAs, Janne Malinafor pathology review, and Elise Nilsson, Julie M.Chatigny, and Catherine M. Mancini for skillfultechnical assistance.

REFERENCES

Allred DC, Mohsin SK, Fuqua SA. 2001. Histological and biolog-ical evolution of human premalignant breast disease. EndocrRelat Cancer 8:47–61.

Antoniou A, Pharoah PD, Narod S, Risch HA, Eyfjord JE, HopperJL, Loman N, Olsson H, Johannsson O, Borg A, Pasini B, Rad-ice P, Manoukian S, Eccles DM, Tang N, Olah E, Anton-Cul-ver H, Warner E, Lubinski J, Gronwald J, Gorski B, TuliniusH, Thorlacius S, Eerola H, Nevanlinna H, Syrjakoski K, Kallio-niemi OP, Thompson D, Evans C, Peto J, Lalloo F, EvansDG, Easton DF. 2003. Average risks of breast and ovarian can-cer associated with BRCA1 or BRCA2 mutations detected incase series unselected for family history: A combined analysisof 22 studies. Am J Hum Genet 72:1117–1130.

Bartek J, Lukas J, Bartkova J. 2007. DNA damage response as ananti-cancer barrier: Damage threshold and the concept of ‘‘con-ditional haploinsufficiency.’’ Cell Cycle 6:2344–2347.

Campbell LL, Polyak K. 2007. Breast tumor heterogeneity: Can-cer stem cells or clonal evolution? Cell Cycle 6:2332–2338.

Cavalli LR, Singh B, Isaacs C, Dickson RB, Haddad BR. 2004.Loss of heterozygosity in normal breast epithelial tissue and be-nign breast lesions in BRCA1/2 carriers with breast cancer. Can-cer Genet Cytogenet 149:38–43.

Chen H, Sukumar S. 2003. Role of homeobox genes in normalmammary gland development and breast tumorigenesis. J Mam-mary Gland Biol Neoplasia 8:159–175.

Chu MC, Selam FB, Taylor HS. 2004. HOXA10 regulates p53expression and matrigel invasion in human breast cancer cells.Cancer Biol Ther 3:568–572.

Clarke CL, Sandle J, Jones AA, Sofronis A, Patani NR, LakhaniSR. 2006. Mapping loss of heterozygosity in normal humanbreast cells from BRCA1/2 carriers. Br J Cancer 95:515–519.

Cordon-Cardo C. 2008. Molecular alterations associated with blad-der cancer initiation and progression. Scand J Urol NephrolSuppl 218:154–165.

Cousineau I, Belmaaza A. 2007. BRCA1 haploinsufficiency, butnot heterozygosity for a BRCA1-truncating mutation, deregu-lates homologous recombination. Cell Cycle 6:962–971.

Deng G, Lu Y, Zlotnikov G, Thor AD, Smith HS. 1996. Loss ofheterozygosity in normal tissue adjacent to breast carcinomas.Science 274:2057–2059.

Domchek SM, Weber BL. 2006. Clinical management of BRCA1and BRCA2 mutation carriers. Oncogene 25:5825–5831.

Foulkes WD, Metcalfe K, Sun P, Hanna WM, Lynch HT, Gha-dirian P, Tung N, Olopade OI, Weber BL, McLennan J, Oli-votto IA, Begin LR, Narod SA. 2004. Estrogen receptor statusin BRCA1- and BRCA2-related breast cancer: The influence ofage, grade, and histological type. Clin Cancer Res 10:2029–2034.

Goldberg JI, Borgen PI. 2006. Breast cancer susceptibility testing:Past, present and future. Expert Rev Anticancer Ther 6:1205–1214.

Gronwald J, Tung N, Foulkes WD, Offit K, Gershoni R, Daly M,Kim-Sing C, Olsson H, Ainsworth P, Eisen A, Saal H, Fried-man E, Olopade O, Osborne M, Weitzel J, Lynch H, GhadirianP, Lubinski J, Sun P, Narod SA,Hereditary Breast Cancer Clini-cal Study Group. 2006. Tamoxifen and contralateral breast can-cer in BRCA1 and BRCA2 carriers: An update. Int J Cancer118:2281–2284.

Guillem JG, Wood WC, Moley JF, Berchuck A, Karlan BY, MutchDG, Gagel RF, Weitzel J, Morrow M, Weber BL, Giardiello F,Rodriguez-Bigas MA, Church J, Gruber S, Offit K. 2006. ASCO/

BRCA1 AND GENOMIC INSTABILITY 89

Genes, Chromosomes & Cancer DOI 10.1002/gcc

SSO review of current role of risk-reducing surgery in commonhereditary cancer syndromes. J Clin Oncol 24:4642–4660.

Hagen AI, Kvistad KA, Maehle L, Holmen MM, Aase H, Styr B,Vabo A, Apold J, Skaane P, Moller P. 2007. Sensitivity of MRIversus conventional screening in the diagnosis of BRCA-associ-ated breast cancer in a national prospective series. Breast16:367–374.

Hedenfalk I, Ringner M, Ben-Dor A, Yakhini Z, Chen Y, Chebil G,Ach R, Loman N, Olsson H, Meltzer P, Borg A, Trent J. 2003.Molecular classification of familial non-BRCA1/BRCA2 breastcancer. Proc Natl Acad Sci USA 100:2532–2537.

Hosey AM, Gorski JJ, Murray MM, Quinn JE, Chung WY, Stew-art GE, James CR, Farragher SM, Mulligan JM, Scott AN, Der-van PA, Johnston PG, Couch FJ, Daly PA, Kay E, McCann A,Mullan PB, Harkin DP. 2007. Molecular basis for estrogen re-ceptor alpha deficiency in BRCA1-linked breast cancer. J NatlCancer Inst 99:1683–1694.

Jonsson G, Naylor TL, Vallon-Christersson J, Staaf J, Huang J,Ward MR, Greshock JD, Luts L, Olsson H, Rahman N, Strat-ton M, Ringner M, Borg A, Weber BL. 2005. Distinct genomicprofiles in hereditary breast tumors identified by array-basedcomparative genomic hybridization. Cancer Res 65:7612–7621.

Lakhani SR, Chaggar R, Davies S, Jones C, Collins N, Odel C,Stratton MR, O’Hare MJ. 1999. Genetic alterations in ‘‘normal’’luminal and myoepithelial cells of the breast. J Pathol 189:496–503.

Lakhani SR, Reis-Filho JS, Fulford L, Penault-Llorca F, van derVijver M, Parry S, Bishop T, Benitez J, Rivas C, Bignon YJ,Chang-Claude J, Hamann U, Cornelisse CJ, Devilee P, Beck-mann MW, Nestle-Kramling C, Daly PA, Haites N, Varley J,Lalloo F, Evans G, Maugard C, Meijers-Heijboer H, Klijn JG,Olah E, Gusterson BA, Pilotti S, Radice P, Scherneck S, SobolH, Jacquemier J, Wagner T, Peto J, Stratton MR, McGuffog L,Easton DF. 2005. Prediction of BRCA1 status in patients withbreast cancer using estrogen receptor and basal phenotype. ClinCancer Res 11:5175–5180.

Larson PS, Schlechter BL, de las Morenas A, Garber JE, CupplesLA, Rosenberg CL. 2005. Allele imbalance, or loss of hetero-zygosity, in normal breast epithelium of sporadic breast cancercases and BRCA1 gene mutation carriers is increased comparedwith reduction mammoplasty tissues. J Clin Oncol 23:8613–8619.

Liebens FP, Carly B, Pastijn A, Rozenberg S. 2007. Managementof BRCA1/2 associated breast cancer: A systematic qualitativereview of the state of knowledge in 2006. Eur J Cancer 43:238–257.

Ma XJ, Wang Z, Ryan PD, Isakoff SJ, Barmettler A, Fuller A,Muir B, Mohapatra G, Salunga R, Tuggle JT, Tran Y, Tran D,Tassin A, Amon P, Wang W, Enright E, Stecker K, Estepa-Sabal E, Smith B, Younger J, Balis U, Michaelson J, Bhan A,Habin K, Baer TM, Brugge J, Haber DA, Erlander MG, SgroiDC. 2004. A two-gene expression ratio predicts clinical outcomein breast cancer patients treated with tamoxifen. Cancer Cell5:607–616.

Ma J, Gao M, Lu Y, Feng X, Zhang J, Lin D, Xiao T, Hu Z,Yuan J, Su K, Shipley J, Xue J, Gao Y. 2006. Gain of 1q25–32,12q23–24.3, and 17q12–22 facilitates tumorigenesis and pro-gression of human squamous cell lung cancer. J Pathol 210:205–213.

Mullan PB, Quinn JE, Harkin DP. 2006. The role of BRCA1 intranscriptional regulation and cell cycle control. Oncogene25:5854–5863.

Narod SA, Brunet JS, Ghadirian P, Robson M, Heimdal K, Neu-hausen SL, Stoppa-Lyonnet D, Lerman C, Pasini B, de los

Rios P, Weber B, Lynch H. 2000. Tamoxifen and risk of con-tralateral breast cancer in BRCA1 and BRCA2 mutation carriers:A case-control study. Hereditary Breast Cancer Clinical StudyGroup. Lancet 356:1876–1881.

Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z,Hernandez-Boussard T, Livasy C, Cowan D, Dressler L,Akslen LA, Ragaz J, Gown AM, Gilks CB, van de Rijn M,Perou CM. 2004. Immunohistochemical and clinical characteri-zation of the basal-like subtype of invasive breast carcinoma.Clin Cancer Res 10:5367–5374.

Petersson C, Pandis N, Mertens F, Adeyinka A, Ingvar C, Ring-berg A, Idvall I, Bondeson L, Borg A, Olsson H, KristofferssonU, Mitelman F. 1996. Chromosome aberrations in prophylacticmastectomies from women belonging to breast cancer families.Genes Chromosomes Cancer 16:185–188.

Petersson C, Pandis N, Rizou H, Mertens F, Dietrich CU,Adeyinka A, Idvall I, Bondeson L, Georgiou G, Ingvar C, HeimS, Mitelman F. 1997. Karyotypic abnormalities in fibroadeno-mas of the breast. Int J Cancer 70:282–286.

Pierce LJ, Levin AM, Rebbeck TR, Ben-David MA, Friedman E,Solin LJ, Harris EE, Gaffney DK, Haffty BG, Dawson LA,Narod SA, Olivotto IA, Eisen A, Whelan TJ, Olopade OI,Isaacs C, Merajver SD, Wong JS, Garber JE, Weber BL. 2006.Ten-year multi-institutional results of breast-conserving surgeryand radiotherapy in BRCA1/2-associated stage I/II breast can-cer. J Clin Oncol 24:2437–2443.

Raman V, Martensen SA, Reisman D, Evron E, Odenwald WF,Jaffee E, Marks J, Sukumar S. 2000. Compromised HOXA5function can limit p53 expression in human breast tumours. Na-ture 405:974–978.

Rebbeck TR, Levin AM, Eisen A, Snyder C, Watson P, Cannon-Albright L, Isaacs C, Olopade O, Garber JE, Godwin AK, DalyMB, Narod SA, Neuhausen SL, Lynch HT, Weber BL. 1999.Breast cancer risk after bilateral prophylactic oophorectomy inBRCA1 mutation carriers. J Natl Cancer Inst 91:1475–1479.

Staff S, Isola J, Tanner M. 2003. Haplo-insufficiency of BRCA1 insporadic breast cancer. Cancer Res 63:4978–4983.

Teixeira MR, Pandis N, Gerdes AM, Dietrich CU, Bardi G,Andersen JA, Graversen HP, Mitelman F, Heim S. 1996. Cyto-genetic abnormalities in an in situ ductal carcinoma and fiveprophylactically removed breasts from members of a familywith hereditary breast cancer. Breast Cancer Res Treat 38:177–182.

Tirkkonen M, Johannsson O, Agnarsson BA, Olsson H, Ingvars-son S, Karhu R, Tanner M, Isola J, Barkardottir RB, Borg A,Kallioniemi OP. 1997. Distinct somatic genetic changes asso-ciated with tumor progression in carriers of BRCA1 andBRCA2 germ-line mutations. Cancer Res 57:1222–1227.

van Beers EH, van Welsem T, Wessels LF, Li Y, OldenburgRA, Devilee P, Cornelisse CJ, Verhoef S, Hogervorst FB,van’t Veer LJ, Nederlof PM. 2005. Comparative genomichybridization profiles in human BRCA1 and BRCA2 breasttumors highlight differential sets of genomic aberrations. Can-cer Res 65:822–827.

Venkitaraman AR. 2002. Cancer susceptibility and the functionsof BRCA1 and BRCA2. Cell 108:171–182.

Warner E, Plewes DB, Hill KA, Causer PA, Zubovits JT, JongRA, Cutrara MR, DeBoer G, Yaffe MJ, Messner SJ, MeschinoWS, Piron CA, Narod SA. 2004. Surveillance of BRCA1 andBRCA2 mutation carriers with magnetic resonance imaging,ultrasound, mammography, and clinical breast examination.JAMA 292:1317–1325.

Wooster R, Weber BL. 2003. Breast and ovarian cancer. N Engl JMed 348:2339–2347.

90 RENNSTAM ETAL.

Genes, Chromosomes & Cancer DOI 10.1002/gcc