DifferentiallyExpressedGenesRegulatingtheProgressionof ... · Gene expression proﬁles of DCIS and...

Tumor and Stem Cell Biology

Differentially ExpressedGenesRegulating theProgressionofDuctal Carcinoma In Situ to Invasive Breast Cancer

Sangjun Lee1, Sheila Stewart2,3,4, Iris Nagtegaal7, Jingqin Luo5, Yun Wu8, Graham Colditz6,Dan Medina9, and D. Craig Allred1

AbstractMolecular mechanisms mediating the progression of ductal carcinoma in situ (DCIS) to invasive breast cancer

remain largely unknown. We used gene expression profiling of human DCIS (n ¼ 53) and invasive breast cancer(n ¼ 51) to discover uniquely expressed genes that may also regulate progression. There were 470 totaldifferentially expressed genes (�2-fold; P < 0.05). Elevated expression of genes involved in synthesis andorganization of extracellular matrix was particularly prominent in the epithelium of invasive breast cancer.The degree of overlap of the geneswith nine similar studies in the literaturewas determined to help prioritize theirpotential importance, resulting in 74 showing overlap in�2 studies (average 3.6 studies/gene; range 2–8 studies).Using hierarchical clustering, the 74-gene profile correctly categorized 96% of samples in this study and 94% ofsamples from 3 similar independent studies. To study the progression of DCIS to invasive breast cancer in vivo, weintroduced human DCIS cell lines engineered to express specific genes into a "mammary intraductal DCIS"xenograft model. Progression of xenografts to invasive breast cancer was dramatically increased by suppressingfour genes that were usually elevated in clinical samples of DCIS, including a protease inhibitor (CSTA) and genesinvolved in cell adhesion and signaling (FAT1, DST, and TMEM45A), strongly suggesting that they normallyfunction to suppress progression. In summary, we have identified unique gene expression profiles of humanDCISand invasive breast cancer, which include novel genes regulating tumor progression. Targeting some of thesegenes may improve the detection, diagnosis, and therapy of DCIS. Cancer Res; 72(17); 4574–86. �2012 AACR.

IntroductionThe majority of human invasive breast cancer is the end

result of a usually decades long evolution of increasinglyabnormal premalignant stages (1, 2). Ductal carcinoma in situ(DCIS) is a late stage and the immediate precursor of mostinvasive breast cancers. Approximately 60,000 new cases ofDCIS were diagnosed in the United States in 2011 (3). Unde-tected, at least a third would progress to invasive breastcancers during an average lifespan (4–6). About 200,000 newcases of invasive breast cancer were also diagnosed (3), and alarge majority evolved from preexisting DCIS that were notdetected. Ten percent of DCIS which are detected still recurand/or progress to invasive breast cancer despite moderntherapy (4, 6). The incidence of invasive breast cancer can be

dramatically reduced by improving our abilities to detect,diagnose, and treat DCIS. Progress will be based on detailedunderstanding of molecular mechanisms responsible for thedevelopment and progression of DCIS to invasive breastcancer, which is currently quite limited (4, 7).

This study used microarray technology to identify differen-tially expressed genes in clinical samples of human DCIS andinvasive breast cancer, which may be involved in tumor pro-gression. The genes were compared with results from previousstudies comparing DCIS to invasive breast cancer to helpprioritize potential importance in tumor progression basedon the degree of overlap (8–16). The ability of selected genes toinfluence tumor progressionwas evaluated in a novel xenograftmodel involving injecting human DCIS cells into the primaryduct of intact mammary glands of immune-suppressed mice,referred to as the mammary intraductal DCIS (MIND) model(17). Cells were genetically modified before injection to mimicchanges in gene expression observed in clinical samples. Thecells grow within ducts in a manner histologically very similarto human DCIS, and may progress to invasive breast cancerdepending on the function of the modulated genes.

Materials and MethodsHuman tumor samples

The human samples used in this study are summarized inSupplementary Table S1. Samples were harvested anonymous-ly with appropriate IRB approvals (Baylor College of Medicine;H-10493 and H-12585).

Authors' Affiliations: Departments of 1Pathology and Immunology, 2CellMolecular Biology and Physiology, 3Medicine, 4BRIGHT Institute, 5Biosta-tistics, and 6Surgery, Washington University School of Medicine, St. Louis,Missouri; 7Department of Pathology, Radboud University Medical Center,Nijmengen, the Netherlands; 8Breast Center, and 9Department of Molec-ular Cellular Biology, Baylor College of Medicine, Houston, Texas

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: D. Craig Allred, Department of Pathology andImmunology, Washington University School of Medicine, 660 S. EuclidAvenue, Campus Box 8118, St. Louis, MO 63110. Phone: 314-362-6313;Fax: 314-362-5505; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-12-0636

�2012 American Association for Cancer Research.

CancerResearch

Cancer Res; 72(17) September 1, 20124574

on July 5, 2020. © 2012 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst July 2, 2012; DOI: 10.1158/0008-5472.CAN-12-0636

http://cancerres.aacrjournals.org/

Evaluation of gene expression by microarraysRNA from each sample was extracted, purified, amplified,

and evaluated for gene expression as previously described(2, 18). Affymetrix U95Av2 microarrays were utilized withsamples in groups 1 to 3, and U133-X3P microarrays withsamples in groups 4e and 4s.

Validation of gene expression by quantitative real-timePCRSelected genes (n ¼ 12) differentially expressed between

DCIS and invasive breast cancers by microarray analyses werevalidated by quantitative real-time PCR (qRT-PCR) as previ-ously described (2, 18), using new samples of total RNA fromgroup 4e. Sequences for the probes and primers were designedusing Primer Express Software v2.0 (Applied Biosystems), andsynthesized by a commercial vendor (Eurogentec).

Cell linesThis study used 3 DCIS-like human breast epithelial cell

lines. First, referred to as DCIS.COM, is widely available to theresearch community and has been used in nearly all previousfunctional studies of human DCIS (17, 19, 20). It was developedfrom ras-transformed MCF10AT cells originally cultured frombenign breast tissue (fibrocystic disease). Second, referred to asSUM225, was developed from a primary culture of a chest wallrecurrence in a patient with a history of DCIS treated bymastectomy (21). Third, referred to as h.DCIS.01, was devel-oped in our laboratory from a primary culture of hyperplasticbreast epithelial cells (columnar cell hyperplasia), which is animportant early precursor of breast cancer (22). Early passages(<25) developed into small hyperplastic lesions in MINDxenografts. The cells spontaneously transformed in cultureand later passages (>30) developed into DCIS in MIND xeno-grafts, similar to DCIS.COM. The related identity of early andlate passages was confirmed by array CGH. All cell lines weremaintained in humidified incubators at 37�C and 5% CO2 inDulbecco's modified Eagle's medium (DMEM)/F12 (50%:50%)media with 5% horse serum and 1% penicillin/streptomycin.293T cells were grown in DMEM media containing 10% fetalcalf serum and 1% penicillin/streptomycin.

Inhibiting gene expression with shRNAiTransductions of shRNAi against selected genes were car-

ried out as described by Stewart (23), utilizing an RNAi vector(pLKO.1puro) containing an RNAi library against the entirehuman genome. Briefly, 293T cells were infected with a lenti-viral vector containing the U6-driven stem-loop directedagainst candidate genes, and red fluorescent protein (RFP),using TransIT-LT1 (Mirius). Virus was harvested 48 hours laterand DCIS cells were infected with virus in the presence of 10mg/mL of protamine sulfate overnight, and reinfected 24 hourslater. Transduced cells were selected for in 2 mg/mL puromy-cin. The pLKO.1 bearing short hairpins targeting genes ofinterest and RFP, were produced by cotransfection withpCMVDR8.2 and pCMV-VSV-G (8:1 ratio). Knockdown effi-ciency was assessed by qRT-PCR and Western blot (whensuitable antibodies were available; Supplementary Fig. S2). Tocontrol for off-target effects, 3 or more separate hairpins were

transduced independently for each gene, and identical resultswere obtained with at least 2.

MIND xenograftsThe MIND xenograft studies were carried out as recently

described (17). Briefly, 20,000 cells were injected through thenipple into the primary duct of intact mammary glands (2 peranimal) of 6- to 8-week-old female SCID-Beige mice. Over aperiod of 10 weeks, variable proportions (depending on the cellline) develop intraductal tumors that are histologically verysimilar to human DCIS, and a proportion progress to invasivebreast cancers. The developing tumors were measured weekly,and all animals were euthanized by 10 weeks. Immediatelyafter being euthanized, mammary glands were excised andprocessed for further study (including whole-mounts, routinehistology, immunohistochemistry, and stored fresh frozen forlater use).

In vitro analyses of cell growth, migration, and invasionCell proliferation. Proliferation was measured as previ-

ously describedbyPazolli (24). Briefly, DCIS cell lines (�shRNAi)cells were seeded into 96-well plates in culture media. At 24,48, 72, and 96 hours, the cells were washed with PBS, andresuspended in cell-lysis buffer with luciferase substrate (Pro-mega). Relative luminescence was measured 5 minutes later.

3-Dimensional (3D) cell cultures. 3D growth was evalu-ated in 24-well culture plates coated with Matrigel (100%) orMatrigel:collagen-1 (200 mL at 1:1 ratio) seeded with 5000 DCIScells (�shRNAi) in regular culture media, and allowed to growfor 2 to 4 weeks (refed with fresh culture media every 2 days).These assays are not easily quantified, and the characteristicsand changes in cell growth were described in a qualitativemanner, including formation of multicellular spheres, andtubules; attachment, growth, and spreading on the gel surface;and invasion into the gels (subjective scale relative to control:"�" ¼ none; "þ"¼ low; "þþ"¼ intermediate; "þþþ"¼ high).Cultures were photographed at 1, 3, 5, 8, and 15 days, afterwhich gelswerefixed andprocessed for histologic evaluation ina routine manner.

In vitro cell migration and invasion were also measured withtheHTSFluoroBlok chamber assay (BDBiosciences), followingmanufacturer's instructions. Briefly, cells suspended in serum-free media were seeded onto the porous membrane of theupper chamber. The bottom chamber was filled with serum-containing media and the cells were incubated for 48 hours.Cells migrating and/or invading through the membrane intothe bottom chamber were stained with Calcein AM dye (Invi-trogen) and quantified using a bioluminescence reader. Tra-versing a baremembrane was defined asmigration. Traversinga membrane coated with 50 mL Matrigel or Matrigel þ COL1(1:1) was defined as invasion

Data analysesMicroarray results were evaluated using dChip software

(http://www.dchip.org) for estimating gene expression andclass comparisons as previously described (2, 18). Normaliza-tion of raw data at PM- and MM-probe level was by theinvariant set normalization method and expression was

Progression of DCIS to Invasive Breast Cancer

www.aacrjournals.org Cancer Res; 72(17) September 1, 2012 4575




estimated by the perfect match only model. Differentiallyexpressed genes were identified using the t test and theempirical false discovery rate (FDR) was estimated by permu-tation. Thefiltered genes have the "value of standard deviation/mean" of expression intensity between 0.5 and 1,000 (defaultsetting in dChip software). On average, about 2,500 genes ineach group of samples met these criteria.

Each of the 5 groups of samples was evaluated by unsuper-vised hierarchical clustering utilizing dChip software and largenumbers of genes (average ¼ 2,866) expressed in the samplesselected by identical normalization and filtering criteria (20%present calls; 0.5 < standard deviation/mean < 1,000 variationacross samples).

Gene ontologies (GO) associated with the progression ofDCIS to invasive breast cancer were inferred fromdifferentiallyexpressed genes utilizing Ontologizer 2.0 software (25). Thehierarchical tree of GO terms were ranked Term-for-Term, andby the Bonferonni method to correct to multiple comparisons.Details are available on the Ontologizer website http://comp-bio.charite.de/index.php/ontologizer2.html.

Differences in the size of MIND xenografts were determinedby averaging tumor size along the 7 time points (week 0 3 5 6 8 910), after square root transformation, was then fitted by amixed model for repeated measure, with group (gene knock-out), time (centered at the mid time point, week 6), and theirinteraction in the model. In the mixed model, the tumor sizeunder each experiment condition (gene knockout) over timewas each modeled as a linear line with its own intercept andslope. To test if a mean growth line under a gene knock outdiffers from that of the reference (RFP), we tested if the slope ofeach line was different from that of the reference group (RFP)and similarly on the intercept. As time is centered atweek 6, thecomparison of a group's intercept will indicate any differencebetween the group and the reference at week 6, whereas thedifference in slopes indicates different tumor growth rate.

ResultsGene expression profiles of DCIS and invasive breastcancer

There have been several hundred studies profiling geneexpression in fully-developed invasive breast cancers, provid-ing many insights into important molecular and clinical fea-tures of the disease. Our goal was to determine whetherexpression profiling can provide similar insights into theprogression of DCIS to invasive breast cancer, which repre-sents a critical step in the evolution of a nonlethal to potentiallylethal disease. Therefore, we carried out comprehensiveexpression profiling on clinical samples of DCIS and invasivebreast cancer, including separate samples of tumor epitheliumand adjacent stroma prepared by laser-capture microdissec-tion (LCM).

Briefly, samples were obtained from clinical cases of DCISand invasive breast cancer, distributed in 5 groups, which wereprofiled independently (Supplementary Table S1). Group 1consisted of 26 samples of DCIS and 24 samples of invasivebreast cancer enriched for tumor epithelial cells (>75% totalcells) bymanualmacrodissection. Groups 2, 3, and 4e consisted

of 7, 9, and 11 samples each of DCIS and invasive breast cancersenriched for tumor epithelial cells (>95%) by LCM (Fig. 1).Group 4s was composed of samples of near-pure (>95%)stroma immediately adjacent to tumor epithelium in 4e, alsoprepared by LCM (Fig. 1). The samples of DCIS in groups 1, 2,and 4 were harvested from breasts without invasive breastcancer. The samples of DCIS in group 3 were harvested frombreasts with adjacent invasive breast cancer.

The results from each groupwere evaluated by unsupervisedhierarchical clustering using variably expressed genes selectedby identical normalization and filtering criteria, averaging2,866 genes/group. As shown in Fig. 2 (top row), there was awide range in the degree that similar types of tumors clusteredtogether between groups. For example, in group 3, composedof DCIS and invasive breast cancer isolated from the samebreasts, analysis categorized patients, but not tumor type,suggesting that synchronous DCIS and invasive breast cancerare more similar than different, which may be expected,because the former evolved directly into the latter. In contrast,analysis of group 4e revealed that pure DCIS and invasivebreast cancer harvested from separate patients are very dif-ferent (100% and 91% clustering accuracy, respectively), indi-cating that there are also unique patterns of gene expression inthese 2 stages of breast cancer.

Differentially expressed genes between DCIS andinvasive breast cancer

To identify differentially expressed genes that may be asso-ciated with and/or regulate the progression of DCIS to invasivebreast cancer, the microarray results were evaluated by super-vised hierarchical clustering at a threshold of�2-fold/p� 0.05.The numbers of genes meeting these criteria were 117, 76,30,165, and 184 in groups 1, 2, 3, 4e, and 4s, respectively(Supplementary Table S2). These results were distilled to470 genes differentially expressed in at least one group, whichwas further refined to 74 genes showing similar profiles in 2 ormore groups (Table 1). The 74 genes were then compared withresults of 9 similar studies in the literature to help prioritizepotential importance based on degree of overlap (8–16, 26).Forty-nine of the 74 genes (66%) overlapped with one or more

Figure 1. LCM of samples in Group 4. Tumor epithelial cells andimmediately adjacent stroma were independently isolated by LCM to apurity of about 95%. IBC, invasive breast cancer.

Lee et al.

Cancer Res; 72(17) September 1, 2012 Cancer Research4576




previous study (range 2–8 studies/gene; average¼ 3.7 studies/gene). The relative expression levels of selected genes (n¼ 12)were evaluated by qRT-PCR in the same study samples,validating the original microarray results for all of the genestested (Supplementary Fig. S1). The majority (78%) of the 74genes have not been studied in the progression of human DCISto invasive breast cancer beyond the original microarraystudies mentioned earlier. The exceptions include genes pre-viously shown to be expressed at different levels in DCIS andinvasive breast cancer by other techniques (COL1A2, MMP11,THBS2, VCAN, CTSK, FN1, FAP, GJB2, CCL5, MYH11, TP63,KRT14, CASP8, NGFR, CAV1, and FAT1), and/or functionallyinvolved in the progression of DCIS to invasive breast cancer(MMP11, CTSK, TP63, and CAV1)—the latter providing somereassurance that this vetting strategy is helpful in identifyingpotentially important genes.

Classification of DCIS and invasive breast cancer basedon differentially expressed genes

Each group of tumor sampleswas reanalyzed by hierarchicalclustering using the 74-gene profile to determine its ability toclassify DCIS and invasive breast cancer (Fig. 2, middle row).There was considerable improvement over initial analysesbased on all filtered genes expressed within each group (Fig.2, top row). The 74-gene profile correctly categorized 97% of allDCIS and 95% of all invasive breast cancers. A high degree ofaccuracy was expected because the 74 genes were derived fromanalysis of the same samples considered collectively. However,each group contributed only a subset of the genes, rangingfrom 28% (21/74) in group 3 to 62% (46/74) in group 4e (average¼ 47%; 34/74).

The ability of the 74-gene profile to classify DCIS versusinvasive breast cancer was further evaluated in cohorts from 3

Figure 2. Hierarchical clustering. The top row shows dendrograms from unsupervised hierarchical clustering analyses based on genes expressed withineach group of samples in this study (average ¼ 2,866). The center row shows results from hierarchical clustering of the same samples using therefined profile of 74 differentially expressed genes determined in this study. The bottom row shows results (dendrograms, heat maps, and genes)using the 74-gene profile in 3 independent cohorts from previous gene expression profiling studies comparing human DCIS and invasive breast cancer (IBC).Note: analysis of the Scheutz study was restricted to 7 of 9 samples each of DCIS and invasive breast cancer that were all evaluated using the sameplatform (U133A Plus 2.0 microarray chips).






Tab

le1.

Gen

es(N

¼74

)sho

wingincrea

sed(n¼42

)and

dec

reas

ed(n¼32

)exp

ressionininva

sive

breas

tcan

cersve

rsus

duc

talcarcino

mainsitu

(DCIS)inat

leas

t2grou

psof

samplesin

this

stud

y,an

dov

erlapwith

previou

sstud

iesco

mparinghu

man

inva

sive

breas

tca

ncersan

dDCIS.

Differen

tially

express

edgen

esGroup

sofsa

mplesin

this

stud

y(fold-cha

nge)

Previous

stud

ies(direc

tionsignifica

ntdifferen

ces)

Sym

bol

Locu

slin

kGroup

1Group

2Group

3Group

4ep

ithe

lium

Group

4stroma

Ave

rage

P-value

Porter

2001

Ma

2003

Porter

2003

Abba

2004

Alline

n20

04Han

neman

2006

Sch

eutz

2006

Ma

2009

Knu

dse

n20

11Studiesin

Common

Increa

sedin

inva

sive

breas

tca

ncer

vs.D

CIS

COL1

A2a

1278

2.27

5.23

3.2

4.66

0.00

24"

""

"8

MMP11

a,b

4320

3.33

3.19

4.17

0.00

90"

""

""

8COL1

1A1c

1301

5.43

11.54

8.95

5.49

0.00

09"

"6

THBS2a

7058

2.05

7.25

4.17

5.67

0.01

37"

"6

LRRC15

c13

1578

2.68

5.09

2.06

2.43

0.00

55"

"6

AEBP1c

165

2.02

4.19

2.35

0.01

37"

""

6COL6

A3c

1293

3.56

4.1

5.55

0.00

86"

""

6COL1

0A1c

1300

2.77

5.13

3.34

0.00

95"

""

6VCANa

1462

3.59

2.75

3.12

0.01

52"

""

6POSTN

d10

631

3.86

3.35

2.99

0.01

86"

""

6MXRA5c

2587

83.57

2.27

2.64

0.01

01"

""

6COL3

A1c

1281

3.58

3.29

4.87

0.01

90"

""

6SPARCd

6678

3.04

2.22

0.00

76"

""

5COL5

A2c

1290

2.41

3.92

3.65

0.00

86"

"6

COL8

A1c

1295

2.28

2.47

2.16

0.01

54"

"6

CILPc

8483

2.05

16.09

4.75

0.00

91"

"6

CTS

Ka,b

1513

3.8

3.08

0.02

64"

""

6FB

N1c

2200

4.82

2.89

0.01

57"

""

6HLA

-DPA1d

3113

2.02

2.02

0.03

00"

""

5LU

Md

4060

4.57

30.01

60"

""

5HLA

-DRAc

3122

2.17

2.07

0.02

94"

"4

FN1a

2335

2.89

2.75

0.01

01"

"4

CDH11

d10

093.8

2.01

0.02

90"

"4

COL5

A1c

1289

3.18

2.26

0.02

98"

"4

ASPNd

5482

97.18

2.7

0.00

26"

"4

FAPa

2191

6.61

4.58

0.00

82"

"4

HTR

A1d

5654

2.39

2.14

0.02

33"

"4

NID2c

2279

52.05

2.89

0.03

07"

"4

STA

T1d

6772

2.11

2.66

0.00

76"

3LY

96c

2364

33.07

2.14

0.01

12"

3COMPd

1311

2.38

3.11

0.02

54"

3INHBAd

3624

2.13

2.59

0.00

92"

3FC

GR3A

c22

142.1

30.00

52"

3FN

DC1c

8462

44.54

3.27

0.01

082

GJB

2a27

064.16

4.25

0.00

912

(Con

tinue

don

thefollo

wingpag

e)

Lee et al.





Tab

le1.Gen

es(N

¼74

)sho

wingincrea

sed(n¼42

)and

dec

reas

ed(n¼32

)exp

ressionininva

sive

breas

tcan

cers

versus

duc

talcarcino

mainsitu

(DCIS)in

atleas

t2grou

psof

samplesin

this

stud

y,an

dov

erlap

with

previou

sstud

iesco

mparing

human

inva

sive

breas

tca

ncersan

dDCIS.(Con

t'd)

Differen

tially

express

edgen

esGroup

sofsa

mplesin

this

stud

y(fold-cha

nge)

Previous

stud

ies(direc

tionsignifica

ntdifferen

ces)

Sym

bol

Locu

slin

kGroup

1Group

2Group

3Group

4ep

ithe

lium

Group

4stroma

Ave

rage

P-value

Porter

2001

Ma

2003

Porter

2003

Abba

2004

Alline

n20

04Han

neman

2006

Sch

eutz

2006

Ma

2009

Knu

dse

n20

11Studiesin

Common

HLA

-Gd

3135

2.28

2.23

0.00

782

CCL5

a63

522.04

2.25

0.02

732

CXCL1

1d63

733.52

4.52

0.01

572

IFI30c

1043

72.21

3.08

0.01

342

BNC2c

5479

62.22

2.7

0.00

262

CITED4d

1637

324.05

3.3

0.02

072

KANK4c

1637

823.27

2.46

0.01

752

Dec

reas

edin

inva

sive

breas

tca

ncer

vs.D

CIS

MYH11

a46

29�4

.2�3

.1�2

�2.7

0.00

97#

#6

TP63

a,b

8626

�2.5

�2.2

0.02

63#

##

5DSTc

667

�5.8

�4.6

�5.3

0.02

33#

4SYNM

d23

336

�2.7

�2�2

.50.02

06#

4C7c

730

�2.8

�3.9

0.01

44#

#4

KRT1

4a38

61�2

.7�4

.30.01

87#

#4

EFE

MP1d

2202

�2�2

.40.00

32#

3DLE

U2d

8847

�2.6

�2.2

0.03

71#

3TINAGL1

d64

129

�2.2

�2.4

0.01

33#

3IL33

d90

865

�2.4

�4.8

0.00

92#

3SERHL2

c25

3190

�4.2

�3�2

.50.00

17#

3DUSP6d

1848

�2.6

�20.00

99#

3INSIG

1d36

38�2

.1�2

.50.02

80#

3COL1

7A1c

1308

�2.1

�2.3

0.01

31#

3FA

BP4d

2167

�2.2

�4.5

0.02

43#

3ENPP2d

5168

�2.1

�2.4

0.01

10#

3ATP

2A2c

488

�2.1

�2.2

0.01

462

CSTA

d14

75�9

.4�3

.80.03

242

CYP51

A1d

1595

�2.2

�2.2

0.02

572

CASP8a

841

�2.1

�2.1

0.00

332

GHRd

2690

�2.6

�2.8

0.02

752

IDH1d

3417

�2.1

�2.3

0.02

612

LEPRd

3953

�2.1

�2.4

0.04

162

MYLK

d46

38�2

.2�2

.20.01

272

NGFR

a48

04�2

.2�3

.70.01

272

CAV1a

,b85

7�2

.2�2

.50.00

902

STB

D1c

8987

�2.6

�2.1

0.03

282

(Con

tinue

don

thefollo

wingpag

e)






similar independent studies (14, 16, 26) (Fig. 2, bottom row). Inthe study by Scheutz (14), which involved 7 paired samples ofDCIS and invasive breast cancer epithelium (purified by LCM)from the same breasts, 100% and 86% of the samples wereclassified accurately, respectively—which is promising consid-ering how similar gene expression profiles are in samples ofthis nature (8–10). In the study by Knudsen (16), involving 10samples of DCIS and invasive breast cancer from differentbreasts, accuracy was 100% and 90% in the epithelial samples,and 90% and 70% in the stromal samples (both independentlypurified by LCM). Classification of DCIS and invasive breastcancer in the study by Kretchmer (26) was 100% accurate forboth types of samples. However, the sample (9 DCIS and 5invasive breast cancer) in the latter study were not micro-dissected and, therefore, each contained varying unknownamounts of tumor epithelium and stroma. The expressionprofile (heat map) of upregulated genes from the 74-geneprofile in this analysis showed some inconsistencies with theother cohorts (e.g., elevation and suppression of several col-lagens in 3 DCIS and 2 invasive breast cancer samples, respec-tively), although the majority were consistent (especiallyinvolving downregulated genes). Overall, the combined resultsstrongly suggest that some of these genes may be usefulclinically for molecular classification and diagnostic targetedimaging of DCIS, which will require additional studies.

Ontologies and pathways associated with differentiallyexpressed genes

To gain a general sense of biological processes, molecularfunctions, and cellular structures involved in the progression ofDCIS to invasive breast cancer, ontologies and pathwaysassociated with differentially expressed genes were evaluatedusing Ontologizer 2.0 software (Table 2 and SupplementaryTable S3; ref. 25). Analysis of the 470 unique differentiallyexpressed genes revealed many highly significant associations(adjusted P-value <0.01), including migration, proliferation,cell–cell adhesion, cell–substrate adhesion, angiogenesis,response to wounding, synthesis of extracellularmatrix (ECM),organization of ECM, skin development, and response toglucocorticoids. Analysis of the 74-gene signature reenforcedmany of the same functions. Based on analyses of groups 4eand 4s, ECM-related ontologies were primarily restricted totumor epithelial cells, and a majority of genes involved weresignificantly elevated in invasive breast cancers versus DCIS. Incontrast, genes involved in skin development were suppressedin the epithelium of invasive breast cancers. Changes in genesinvolved in angiogenesis and response to gluccocorticoidswere primarily restricted to tumor stroma, and the majorityof genes were suppressed in invasive breast cancers.

In vivo functional studies of differentially expressedgenes

The high degree of overlap between studies, impressiveclassification accuracy, and types of ontologies associated withthe 74-gene profile suggest that some of the genes may befunctionally important in the progression of DCIS to invasivebreast cancer. We investigated the function of selected genesusing human DCIS-like cell lines and the MIND xenograft

Tab

le1.Gen

es(N

¼74

)sho

wingincrea

sed(n¼42

)and

dec

reas

ed(n¼32

)exp

ressionininva

sive

breas

tcan

cers

versus

duc

talcarcino

mainsitu

(DCIS)in

atleas

t2grou

psof

samplesin

this

stud

y,an

dov

erlap

with

previou

sstud

iesco

mparing

human

inva

sive

breas

tca

ncersan

dDCIS.(Con

t'd)

Differen

tially

express

edgen

esGroup

sofsa

mplesin

this

stud

y(fold-cha

nge)

Previous

stud

ies(direc

tionsignifica

ntdifferen

ces)

Sym

bol

Locu

slin

kGroup

1Group

2Group

3Group

4ep

ithe

lium

Group

4stroma

Ave

rage

P-value

Porter

2001

Ma

2003

Porter

2003

Abba

2004

Alline

n20

04Han

neman

2006

Sch

eutz

2006

Ma

2009

Knu

dse

n20

11Studiesin

Common

SLC

16A4d

9122

�2.1

�2.9

0.02

282

ATP

6V1G

1c95

50�4

.1�2

.50.02

882

MAFF

d23

764

�2.2

�3.2

0.01

972

ADIPOQ

d93

70�3

.2�7

.00.01

752

FAT1

a21

95�4

.9�2

.80.00

802

Num

ber

in74

3235

2146

37Perce

ntin

7443

.2%

47.3%

28.4%

62.2%

50%

aPreviou

slystud

iedin

DCIS,p

rimarily

comparingex

pressionleve

lswith

inva

sive

breas

tca

ncers.

bPreviou

slystud

iedin

DCIS,inc

ludingfunc

tiona

linv

olve

men

tin

progres

sion

toinva

sive

breas

tca

ncer.

cNot

previou

slystud

iedin

breas

tca

ncer.

dPreviou

slystud

iedin

inva

sive

breas

tca

ncer.

Lee et al.





model (17). In this model, the cells grow as authentic DCIS (i.e.,growing within ducts of intact mammary glands), whereasmost previous studies utilized subcutaneous xenografts todraw conclusions about DCIS (12, 19).Our initial studies focused on genes expressed at relatively

high levels in samples of DCIS versus invasive breast cancers,hypothesizing that they normally suppress tumor progression.Ten total candidate suppressor genes from chosen for thesestudies, taken from the 74-gene profile (CHD1, CSTA, FAT1,DST, SERHL2, and SYNM) or 470 gene list (KRT10,PHLDA2, SFN,and TMEM45A). All have known functions that can reasonablybe involved in the progression of DCIS (proteases, adhesionmolecules, etc.) and none except SFN and CDH1 have beenstudied in DCIS. The latter have both been shown to besuppressed in clinical invasive breast cancer versus DCIS(27, 28). The DCIS-like cell lines we utilized included DCIS.COM and SUM225, which are the only widely available tothe research community (20, 29), and a third, referred to ash.DCIS.01, which we recently developed in a manner similar tothat used to develop DCIS.COM. DCIS.COM has been usedalmost exclusively in previous studies addressing gene func-tion in human DCIS (20).DCIS.COM cells expressed relatively high levels of all 10

candidate suppressor genes, andwe predicted that knockdownwould promote the progression of DCIS.COM to invasivebreast cancer MIND xenografts. Therefore, the cells wereindependently transduced with shRNAi corresponding to eachgene (3–5 hairpins/gene), which decreased expression averag-ing about 80% for all genes (showed with �2 hairpins/gene)compared with control cells (Supplementary Fig. S2). Controls

were transducedwith shRNAi to RFP, which was not expressedin the cells, to partially control for off-target effects, and tofacilitate in vitro bioluminescence growth assays. As shownin Fig. 3 and Supplementary Table S4, all transduced cell lineswere able to grow and progress from DCIS to invasive breastcancer in MIND xenografts, although at varying rates com-pared with control cells (parental and shRFP): 2 (shSFN andshCDH1) showed similar rates as controls; 4 (shKRT10,shSYNM, shPHLDA2, and shSERHL2) showed moderatelyincreased rates; and 4 (shCSTA, shDST, shFAT1, andshTMEM45A) showed highly increased (P < 0.001) rates at10 weeks. Xenografts �5 mm were about equally dividedbetween DCIS alone and DCIS þ invasive breast cancer. Allxenografts >5 mm were composed of DCIS þ invasive breastcancer.

Similar experiments were conducted using h.DCIS.01 andSUM225 DCIS cells lines (Fig. 4). h.DCIS.01 expressed highlevels of CSTA, FAT1, and DST. It was transduced with thesingle most suppressive hairpin observed for each gene inDCIS.COM cells, which significantly reduced expression(average about 80%), and significantly (P < 0.001) increasedthe growth and progression of DCIS to invasive breastcancer xenografts, nearly identical to the results obtainedwith DCIS.COM. SUM225 cells expressed very low levels ofCSTA, FAT1, and DST, and we predicted that further sup-pression would not influence tumor progression. Trans-ducing shFAT1 and shDST was lethal in SUM225, so theycould not be further evaluated. However, shCSTA resultedin about 50% suppression and, as predicted, growth andprogression were not increased, although we cannot

Table 2. Selected ontologies associated with the progression of DCIS to invasive breast cancer based ondifferentially expressed genes in this study.

470-gene profile 74-gene profile Group 4 epithelium Group 4 stroma

GO ID GO term Adjusted P-value

GO:0005578 Proteinaceous extracellular matrix 1.8E�15 7.6E�18 7.3E�13 nsGO:0030198 Extracellular matrix organization ns 1.1E�06 7.7E�03 nsGO:0007155 Cell adhesion 2.1E�09 1.3E�06 ns nsGO:0016477 Cell migration 1.0E�04 ns ns nsGO:0042127 Regulation of cell proliferation 2.9E�03 ns ns nsGO:0019838 Growth factor binding 3.1E�05 1.7E�02 ns nsGO:0001568 Blood vessel development 1.4E�05 ns ns 1.3E�02GO:0009611 Response to wounding 1.4E�05 1.7E�02 ns nsGO:0043588 Skin development 7.1E�03 9.5E�06 8.4E�04 nsGO:0051384 Response to glucocorticoid stimulus 2.1E�03 ns ns 7.1E�02

Group 4 epithelium Differentially Expressed Genes in Invasive Breast Cancer vs. DCIS (up and down)Proteinaceousextracellular matrix

ASPN, CILP, COCH, COL10A1, COL11A1, COL1A2, COL3A1, COL5A1, COL6A3, COL8A1, CTHRC1, DCN, FBLN1,FBN1, LUM, MMP11, MMP12, OMD, POSTN, SPARC, SPOCK1, SPON1, VCAN, VEGFA

Extracellular matrixorganization

ASPN, CILP, COCH, COL10A1, COL11A1, COL1A2, COL3A1, COL5A1, COL6A3, COL8A1, CTHRC1, DCN, FBLN1,FBN1, LUM, MMP11, MMP12, OMD, POSTN, SPARC, SPOCK1, SPON1, VCAN, VEGFA

Skin development COL1A2, COL3A1, COL5A1, NGFR, SFN, TP63Group 4 stromaBlood vessel development C1GALT1, CASP8, CAV1, CDH5, COL5A1, COL8A1, FLT1, LMO2, MAPK1, MEOX2, PDGFA, ROBO4, SEMA5A,

TGFBR2Response to glucocorticoidstimulus

ADIPOQ, ADM, CAV1, CCL2, FABP4, GHR, SDC1






exclude the possibility that the level of suppression wasinadequate.

Collectively, our microarray results in clinical samples, andin vivo results with human DCIS cell lines, provide compellingevidence that CSTA, FAT1, and DST normally suppress theprogression of DCIS to invasive breast cancer.

In vitro functional studies of suppressor genesThe progression of DCIS to invasive breast cancer is a

complex process involving many types of cell behaviors,including growth, migration, and invasion. To further char-acterize the functions of CSTA, FAT1, and DST observed inMIND xenografts, DCIS.COM.sh cells were evaluated in 3-dimensional cultures composed of Matrigel � COL1 (Fig. 5).All cells on Matrigel alone grew as small spheres. Controlcells on Matrigel þ COL1 formed small spheres, which wereoccasionally connected by branching tubules. In contrast,shCSTA, shFAT1, and shDST cells spread and piled up on the

gel surface, and invaded into the gels. Interestingly, COL1,which was necessary to reveal these altered behaviors, wasexpressed at highly elevated levels in clinical samples ofhuman invasive breast cancer versus DCIS, suggesting that itmay be particularly important in enabling or promotingtumor invasion.

Other more quantifiable methods were also utilized toassess the growth (2-dimension culture), migration and inva-sion (both in 3-dimensional Boydon chamber assays) of DCIS.COM and h.DCIS.01 cells transduced with shCSTA, shFAT1,and shDST (Supplementary Fig. S3). Migration and invasionwere significantly increased in both cell lines transduced withshFAT1. All other results were similar to controls, or highlyvariable and difficult to interpret. However, despite the neg-ative or ambiguous results of some of these in vitro studies,especially involving shCSTA and shDST cells, tumor cellgrowth, migration, and invasion were unequivocally elevatedin all corresponding MIND xenografts.

Figure 3. Effect of modulating geneexpression in DCIS.COM cells onMIND xenografts. DCIS.COM cellswere transduced with shRNAicorresponding to 10 candidateinvasion-suppressor genes (thecells originally expressed highlevels of all genes). Four of themost efficiently suppressed genes(CSTA, DST, FAT1, andTMEM45A) resulted in a highlysignificant increase in theprogression of DCIS to invasivebreast cancer, consistent with thehypothesis that these genesfunction normally to suppresstumor invasion. The pie chartssummarize xenograft growth(tumor size). The bar graphssummarize knockdown efficiencyof the gene in DCIS.COM cells.

Lee et al.





DiscussionGene expression profiling has taught us a great deal about

the progression of fully-developed invasive breast cancer,and our study used this approach as a starting point to learnmore about the progression of DCIS to invasive breastcancer. We are aware of at least 9 previous studies compar-ing expression between DCIS and invasive breast cancer,publishing results that can be compared with ours (8–16, 30).Collectively, these studies included 130 cases of DCIS and 126of invasive breast cancer, which pales in comparison toprevious expression profiling studies of invasive breast can-cer alone involving thousands of cases. Our study increasesthe number of samples comparing DCIS to invasive breastcancer by about 50%, which is a helpful contribution giventhe relatively small numbers of cases overall addressing thisimportant question.A proportion of samples from this (17%) and the previous

(37%) studies were paired DCIS and invasive breast cancerfrom the same breasts, which may not be the ideal setting (atleast exclusively) to identify invasion-regulating genes becausesomeportion of these DCIS have already progressed to invasivebreast cancer, and may already possess some alterationsassociated with progression. Consistent with this, unsuper-vised hierarchical clustering based on expressed filtered genesof Group 3 in our study, composed of paired samples of DCISand invasive breast cancer from the same breasts, was able todistinguish the pairs, but not tumor type. Amajority of samplesfrom this (83%) and previous (63%) studies were invasivebreast cancer and DCIS from different breasts, and the DCISwere not associated with invasive breast cancer in the samebreast. Similar analyses of the samples in our study fromdifferent breasts (groups 1, 2, and 4) were more accuratelycategorized, presumably because of biological differences

that may be related to tumor progression. The 74-gene profilewas able to correctly classify 96% of all samples in this study,and 94% of samples from 3 independent studies. Thus, some ofthese genes are promising targets for the development ofimproved methods of detection and diagnosis of DCIS (e.g.,targeted imaging).

Identifying genes important in tumor progression fromstatic measurements of expression by microarrays is challeng-ing. The main criterion defining "important" in this study wasthe degree of overlap of specific genes betweenmultiple studiescomparing DCIS and invasive breast cancer. Similarly, infer-ring ontologies and pathways important in tumor progressionfrom differentially expressed genes is also challenging. Anal-yses of the 470 unique differentially expressed genes across allgroups in this study, and the refined subset of 74 overlappinggenes, converged on essentially the same ontologies and path-ways. Nearly all have been implicated in the progression ofmany types of cancers over the years, including growth,adhesion, andmotility. Our results support the obvious impor-tance of these behaviors in the progression of DCIS to invasivebreast cancer, and point to many new genes that may beinvolved.

Group 4 in this study enabled independent analyses of tumorepithelium and adjacent stroma, and both tissue compart-ments were implicated inmany of the same ontologies—with afew interesting exceptions. For example, changes in genesassociated with synthesis and organization of extracellularmatrix (especially fibrillar collagens) were particularly prom-inent in invasive breast cancer epithelial cells, consistent withepithelial-to-mesenchymal transformation (EMT), and addingto the molecular definition of EMT. Previous similar studies byEmory (31) and Knudsen (16) also observed elevated expres-sion of ECM-related genes in invasive breast cancer epithelium

Figure 4. Effect of modulating geneexpression in h.DCIS.01 andSUM225 cells on MIND xenografts.h.DCIS.01 cells expressed highlevels of CSTA, DST, and FAT1, andwere transduced with thecorresponding shRNAi, which weremost effective in suppressingexpression in DCIS.COM cells. All 3genes were highly suppressed,which resulted in significantlyincreased progression of DCIS toinvasive breast cancer (IBC)xenografts, confirming resultsobtained with DCIS.COM. SUM225cells expressed very low levels of all 3genes. Further suppression using thesame hairpins for CSTA and FAT1were lethal, and further analysis wasnot possible. Further suppression ofDST was not lethal and did notincrease in the progression of DCISto invasive breast cancer xenografts.The pie charts summarize xenograftgrowth (tumor size). The bar graphssummarize knockdown efficiency ofthe gene in DCIS.COM cells.






(31). In particular, the Knudsen study, identified and priori-tized important ontologies in amanner almost identical to ourstudy, which is scientifically reassuring, although the specificgenes utilized for identification overlapped with our group 4only about 20% and 10% in the epithelium and stroma,respectively.

Skin development was another prominent ontology restrict-ed to epithelium, and genes regulating terminal differentia-tion and polarity of epidermal cells (e.g., SFN and TP63) weresuppressed in invasive breast cancers. Breasts arise frombudding skin appendages during embryonic development(32), and invasion of many types of solid cancers involvesreactivation of embryonic pathways. Changes in genes respon-sive to glucocorticoids were prominent in stroma, and nearlyall were suppressed in invasive breast cancer versus DCIS,although the consequence of these changes on tumor pro-gression is unclear. Angiogenesis as a GO term was restrictedto stroma, but many genes regulating growth of endothelialcells were suppressed in invasive breast cancers (e.g., VEGFA,FLT1, LMO2, PDGFA, and ROBO4), which runs counter toprevious studies demonstrating increased angiogenesis ininvasive breast cancers (33). However, most previous studies

observed increased vascular density predominately at theleading edge of invasive breast cancers (34). The stroma inthis study was adjacent to clusters of invasive epitheliumprimarily from the interior of tumors (each case was a singlepiece of tissue harvested before the study, so only a singleedge, at best, could represent the leading edge)—and per-haps the regulation of angiogenesis is different in theselocations. Regardless, this interesting issue requires furtherstudy.

Group 4s in this study, and the studies by MA (15) andKnudsen (16), are the only directly comparing transcriptomesof stroma immediately adjacent to tumor epithelial cells ofDCIS and invasive breast cancer. The samples in 4s were fromdifferent breasts, although those of Ma were pairs from thesame breasts, and there was relatively little overlap of differ-entially expressed genes (12%), emphasizing the difference instudy design. A study by Allinen (12) evaluated stromal geneexpression by SAGE in samples from different breasts, andthere was also little overlap with 4s (5%). However, this study isdifficult to compare because expression was evaluated inpurified subtypes of stromal cells rather than intact stroma.Like our study, Knudsen compared purified samples of stromafrom invasive breast cancer and pure DCIS in different breasts.However, overlap of specific differentially expressed genes wasonly about 10%, and only 5 were in common with our 74-geneprofile (COL11A1, TP63, MYH11, FABP4, and ENPP2). Thereasons for the low level of overlap overall are unknown atthe moment.

Overall, there are relatively few in vivo functional studiesaddressing the progression of human DCIS to invasive breastcancers, and all utilized DCIS.COM cells (20), primarily assubcutaneous xenografts in mice (12, 19). Our study alsoutilized DCIS.COM to study tumor progression, but in MINDxenografts, which more closely resemble human DCIS (17).In particular, the microenvironment in the mouse mammaryducts (myoepithelial cells, intact basement membrane, vas-culature, etc.) is structurally similar to ducts in humanbreasts. The basement membrane may be particularlyimportant. Interestingly, all 3 cell human DCIS cell linesused in this study somehow displaced the mouse myoe-pithelial cells, and made their own (Supplementary Fig. S4).It is important to keep in mind, however, that all currentlyutilized human xenografts models have limitations in termsof relevance to human disease. Within these limitations, ourstudies based on 3 human DCIS-like cell lines stronglysuggest that CSTA, FAT1, and DST normally function tosuppress the progression of DCIS to invasive breast cancer,which underscores the complexity of progression of humanDCIS, given the diverse functions of these genes.

CSTA is a protease inhibitor that is thought to primarilyinhibit cathepsin B activity (35). A previous study by Kleer (36)showed that in vitro invasion of DCIS.COM cells was signifi-cantly increased by coculturing with fibroblasts expressinghigh levels of cathepsin K, which is closely related to cathepsinB. Low levels of CSTA, and elevated levels of cathepsin B, havebeen associatedwith poor prognosis in prostate (37) and head/neck cancers (35). Zajc (38) observed direct and indirectassociations with cathepsin B and CSTA levels, respectively,

Figure 5. Effects of modulating gene expression in DCIS.COM cells in3-dimensional culture. To further characterize the functions of CSTA,FAT1, and DST observed in MIND xenografts, DCIS.COM.sh cells wereevaluated in 3-dimensional cultures composed of Matrigel � COL. Allcells onMatrigel alone grew as small spheres. Control cells onMatrigelþCOL1 formed small spheres that were occasionally connected bybranching tubules. In contrast, shCSTA, shFAT1, and shDST cells spreadand piled up on the gel surface, and invaded into the gels.

Lee et al.





with the degree of invasiveness of breast cancer cells linesin vitro. Our findings suggest that unchecked cathepsin activity(through loss of CSTA) is important in the progression ofclinical DCIS.DST is a cytoplasmic protein involved in cell adhesion and

cytoskeletal organization (39). It regulates the polarity ofsquamous epithelium and auto antibodies against DST resultin bullous pemphigoid, a blistering skin disease (40). DSTexpression is directly regulated by p63 in normal keratinocytes(41). Myoepithelial cells, which express high levels of p63, playan important role inmaintaining differentiation and polarity ofluminal epithelial cells in the breast. Hu (42) showed thatinhibiting the differentiation of myoepithelium in subcutane-ous xenografts of DCIS.COM promoted progression to invasivecancer. Together with our results, this suggests that there maybe an important link between p63 and DST in the progressionof DCIS.FAT1 encodes a surface membrane protein involved in cell

adhesion, polarity, and signaling (43). It regulates cell prolif-eration and has tumor suppressor activity in Drosophila (43).Homozygous deletions of the FAT1 gene, and loss of proteinexpression, are common in human squamous cell carcinomas(44), in line with our results suggesting that suppression ofFAT1 promotes the progression of DCIS.

In conclusion, new more effective methods of detecting,diagnosing, and treating DCIS can be developed based ontargeting some of the genes identified in this study.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: S. Lee, S. Stewart, D.C. AllredDevelopment of methodology: S. Lee, S. Stewart, D.C. AllredAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Lee, S. Stewart, Y. Wu, D.C. AllredAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): S. Lee, S. Stewart, I. Nagtegaal, J. Luo, D. Medina,D.C. AllredWriting, review, and/or revision of the manuscript: S. Lee, S. Stewart,I. Nagtegaal, J. Luo, G. Colditz, D. Medina, D.C. AllredAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): Y. Wu, D. Medina, D.C. AllredStudy supervision: S. Stewart, D.C. Allred

Grant SupportThis study was supported by the Breast Cancer Research Foundation through

donation by Harry and Lillian Glassman, and NIH P50CA58183.The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received February 20, 2012; revised June 14, 2012; accepted June 15, 2012;published OnlineFirst July 2, 2012.

References1. Allred DC. Biological features of human premalignant breast disease

and the progression to cancer. In:Harris JR, Lippman ME, Mlorrow M,Hellman S, Osborne CK, editors. Diseases of the breast. 4th ed.Philadelphia: Lippincott Williams and Wilkins; 2009. p. 323–34.

2. Allred DC, Wu Y, Mao S, Nagtegaal ID, Lee S, Perou CM, et al. Ductalcarcinoma in situ and the emergence of diversity during breast cancerevolution. Clin Cancer Res 2008;14:370–8.

3. American Cancer Society. Breast cancer facts and figures 2011-2012;p. 2 (Table 1).

4. Kuerer HM, Albarracin CT, Yang WT, Cardiff RD, Brewster AM, Sym-mans WF, et al. Ductal carcinoma in situ: state of the science androadmap to advance the field. J Clin Oncol 2009;27:279–88.

5. PageDL,DupontWD, Rogers LW, JensenRA,Schuyler PA.Continuedlocal recurrence of carcinoma 15–25 years after a diagnosis of lowgrade ductal carcinoma in situ of the breast treated only by biopsy.Cancer 1995;76:1197–200.

6. Virnig BA, Shamliyan T, Tuttle TM, Kane RL,Wilt TJ, editors. Diagnosisand management of ductal carcinoma in situ (DCIS). Rockville, MD:Agency for Healthcare Research and Quality, U.S. Department ofHealth and Human Services; 2009.

7. Allegra CJ, Aberle DR, Ganschow P, Hahn SM, Lee CN, Millon-Underwood S, et al. State-of-the-Science Conference: diagnosis andmanagement of ductal carcinoma in situ (DCIS). J Natl Cancer Inst2010;102:161–9.

8. Porter D, Lahti-Domenici J, Keshaviah A, Bae YK, Argani P, Marks J,et al. Molecular markers in ductal carcinoma in situ of the breast. MolCancer Res 2003;1:362–75.

9. Porter DA, Krop IE, Nasser S, Sgroi D, Kaelin CM, Marks JR, et al. ASAGE (serial analysis of gene expression) view of breast tumor pro-gression. Cancer Res 2001;61:5697–702.

10. Ma XJ, Salunga R, Tuggle JT, Gaudet J, Enright E, McQuary P, et al.Gene expression profiles of human breast cancer progression. ProcNatl Acad Sci U S A 2003;100:5974–9.

11. Abba MC, Drake JA, Hawkins KA, Hu Y, Sun H, Notcovich C, et al.Transcriptomic changes in human breast cancer progression asdetermined by serial analysis of gene expression. Breast Cancer Res2004;6:R499–513.

12. AllinenM, Beroukhim R, Cai L, Brennan C, Lahti-Domenici J, Huang H,et al. Molecular characterization of the tumor microenvironment inbreast cancer. Cancer Cell 2004;6:17–32.

13. Hannemann J, Velds A,Halfwerk JB, KreikeB, Peterse JL, van deVijverMJ. Classification of ductal carcinoma in situ by gene expressionprofiling. Breast Cancer Res 2006;8:R61.

14. Schuetz CS, Bonin M, Clare SE, Nieselt K, Sotlar K, Walter M, et al.Progression-specific genes identified by expression profiling ofmatched ductal carcinomas in situ and invasive breast tumors, com-bining laser capture microdissection and oligonucleotide microarrayanalysis. Cancer Res 2006;66:5278–86.

15. MaXJ,DahiyaS,RichardsonE, ErlanderM,SgroiDC.Geneexpressionprofiling of the tumormicroenvironment during breast cancer progres-sion. Breast Cancer Res 2009;11:R7.

16. Knudsen ES, Ertel A, Davicioni E, Kline J, Schwartz GF,Witkiewicz AK.Progression of ductal carcinoma in situ to invasive breast cancer isassociated with gene expression programs of EMT and myoepithelia.Breast Cancer Res Treat 2012;133:1009-24.

17. Behbod F, Kittrell FS, LaMarca H, Edwards D, Kerbawy S, HeestandJC, et al. An intraductal human-in-mouse transplantation modelmimics the subtypes of ductal carcinoma in situ. Breast Cancer Res2009;11:R66.

18. Lee S, Medina D, Tsimelzon A, Mohsin SK, Mao S, Wu Y, et al.Alterationsof gene expression in the development of early hyperplasticprecursors of breast cancer. Am J Pathol 2007;171:252–62.

19. Hu M, Yao J, Carroll DK, Weremowicz S, Chen H, Carrasco D, et al.Regulation of in situ to invasive breast carcinoma transition. CancerCell 2008;13:394–406.

20. AllredDC,MedinaD. The relevance ofmousemodels to understandingthe development and progression of human breast cancer. J Mam-mary Gland Biol Neoplasia 2008;13:279–88.

21. Ignatoski Km Fau - Lapointe AJ, Lapointe Aj Fau - Radany EH, RadanyEh Fau - Ethier SP, Ethier SP. erbB-2 overexpression in humanmammary epithelial cells confers growth factor independence. Endo-crinology 1999;140:3615–22.

22. Feeley L, Quinn CM. Columnar cell lesions of the breast. Histopathol-ogy 2008;52:11–9.






23. Stewart SA, Dykxhoorn DM, Palliser D, Mizuno H, Yu EY, An DS, et al.Lentivirus-delivered stable gene silencing by RNAi in primary cells.RNA 2003;9:493–501.

24. Pazolli E, Luo X, Brehm S, Carbery K, Chung JJ, Prior JL, et al.Senescent stromal-derived osteopontin promotes preneoplastic cellgrowth. Cancer Res 2009;69:1230–9.

25. Bauer S, Grossmann S, Vingron M, Robinson PN. Ontologizer 2.0—amultifunctional tool for GO term enrichment analysis and data explo-ration. Bioinformatics 2008;24:1650–1.

26. Kretschmer C, Sterner-Kock A, Siedentopf F, Schoenegg W, SchlagPM, Kemmner W. Identification of early molecular markers for breastcancer. Mol Cancer 2011;10:15.

27. Hoque MO, Prencipe M, Poeta ML, Barbano R, Valori VM, Copetti M,et al. Changes in CpG islands promoter methylation patterns duringductal breast carcinoma progression. Cancer Epidemiol BiomarkersPrev 2009;18:2694–700.

28. Yoon NK, Seligson DB, Chia D, Elshimali Y, Sulur G, Li A, et al. Higherexpression levels of 14-3-3sigma in ductal carcinoma in situ of thebreast predict poorer outcome. Cancer Biomark 2009;5:215–24.

29. Gupta PB, Kuperwasser C. Disease models of breast cancer. DrugDiscov Today 2004;1:9–16.

30. Castro NP, Os�orio CA, Torres C, Bastos EP, Mour~ao-Neto M, SoaresFA, et al. Evidence that molecular changes in cells occur beforemorphological alterations during the progression of breast ductalcarcinoma. Breast Cancer Res 2008;10:R87.

31. Emery LA, Tripathi A, King C, Kavanah M, Mendez J, Stone MD, et al.Early dysregulation of cell adhesion and extracellular matrix pathwaysin breast cancer progression. Am J Pathol 2009;175:1292–302.

32. Cowin P, Wysolmerski J. Molecular mechanisms guiding embryonicmammarygland development. ColdSpringHarbPerspect Biol 2010;2:a003251.

33. Kerbel RS. Tumor angiogenesis. N Engl J Med 2008;358:2039–49.34. Uzzan B, Nicolas P, Cucherat M, Perret GY. Microvessel density as

a prognostic factor in women with breast cancer: a systematic

review of the literature and meta-analysis. Cancer Res 2004;64:2941–55.

35. Strojan P, Budihna M, Smid L, Svetic B, Vrhovec I, Kos J, et al.Prognostic significance of cysteine proteinases cathepsins B and Land their endogenous inhibitors stefins A and B in patients withsquamous cell carcinoma of the head and neck. Clin Cancer Res2000;6:1052–62.

36. Kleer CG, Bloushtain-Qimron N, Chen YH, Carrasco D, Hu M, Yao J,et al. Epithelial and stromal cathepsin K and CXCL14 expression inbreast tumor progression. Clin Cancer Res 2008;14:5357–67.

37. Sinha AA, Quast BJ, Wilson MJ, Fernandes ET, Reddy PK, Ewing SL,et al. Ratio of cathepsin B to stefin A identifies heterogeneity withinGleason histologic scores for human prostate cancer. Prostate2001;48:274–84.

38. Zajc I, Sever N, Bervar A, Lah TT. Expression of cysteine peptidasecathepsin L and its inhibitors stefins A and B in relation to tumor-igenicity of breast cancer cell lines. Cancer Lett 2002;187:185–90.

39. Sonnenberg A, LiemRK. Plakins in development and disease. ExpCellRes 2007;313:2189–203.

40. Hamill KJ, Hopkinson SB, DeBiase P, Jones JC. BPAG1e maintainskeratinocyte polarity through beta4 integrin-mediated modulation ofRac1 and cofilin activities. Mol Biol Cell 2009;20:2954–62.

41. Osada M, Nagakawa Y, Park HL, Yamashita K, Wu G, Kim MS, et al.p63-specific activation of the BPAG-1e promoter. J Invest Dermatol2005;125:52–60.

42. Hu M, Polyak K. Molecular characterisation of the tumour microenvi-ronment in breast cancer. Eur J Cancer 2008;44:2760–5.

43. KatohY,KatohM.Comparative integromics onFAT1, FAT2, FAT3, andFAT4. Int J Mol Med 2006;18:523–8.

44. Nakaya K, Yamagata HD, Arita N, Nakashiro KI, Nose M, Miki T,et al. Identification of homozygous deletions of tumor suppressorgene FAT in oral cancer using CGH-array. Oncogene 2007;26:5300–8.

Lee et al.





2012;72:4574-4586. Published OnlineFirst July 2, 2012.Cancer Res Sangjun Lee, Sheila Stewart, Iris Nagtegaal, et al.

to Invasive Breast CancerIn SituDuctal Carcinoma Differentially Expressed Genes Regulating the Progression of

Updated version

10.1158/0008-5472.CAN-12-0636doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2012/07/02/0008-5472.CAN-12-0636.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/72/17/4574.full#ref-list-1

This article cites 40 articles, 13 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/72/17/4574.full#related-urls

This article has been cited by 6 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/72/17/4574To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-12-0636

http://cancerres.aacrjournals.org/content/suppl/2012/07/02/0008-5472.CAN-12-0636.DC1

http://cancerres.aacrjournals.org/content/72/17/4574.full#ref-list-1

http://cancerres.aacrjournals.org/content/72/17/4574.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/72/17/4574


DifferentiallyExpressedGenesRegulatingtheProgressionof ... · Gene expression proﬁles of DCIS and...

Documents

Transcript of DifferentiallyExpressedGenesRegulatingtheProgressionof ... · Gene expression proﬁles of DCIS and...