ICANCERRESEARCH54,2615-2621,May15,19941

ABSTRACT

Wit: gene expression was investigated by ribonuclease protection analysis in human breast cancer, nontumorous breast tissue, and a variety ofhuman breast cell lines. We report the expression of Wnt3, Wn14,andWnt7b In human breast cell lines and Wnt2, Wnt3, Wn14, and Wni7b inhuman breast tissues. Wnt3a and Wnt7a were absent in the cell lines andtissues tested. The level ofexpression of WntZand Wnt4 was 10. to 20-foldhigher In fibroadenomas than It was in normal or malignant breast tissue,and in 10% oftumors Wnt7b expression was 30-fold hiajier than in normalorbenign breast tissues. In contrast to the mouse, in which Wail and Wnt3are involved in tumorigenesis, our results suggest that Wn12,Wn14,andWnt7b may be assodated with abnormal proliferation in human breasttissue.

INTRODUCTION

Wnt genes (reviewed in Refs. 1 and 2) form a large family, the firstmember of which was discovered as a result of its role in mousemammary cancer (3). In murine breast tumors, mouse mammarytumor virus was found consistently at preferential loci in the tumorDNA. The gene activated at the commonest of these integration sites(initially called mt 1) was later shown to be homologous to theDrosophila “wingless―gene (4), and the combination of these twoterms gave rise to the name “Wnt―(5).

Numerous new Wnt genes have subsequently been identified inmanyorganismsincludingnot only Drosophila(6), Xenopus(7), mice(8), and humans (9) but also a variety of fish, birds, and reptiles (2).Of the above, the profile of Wnt expression and function has beenstudied in most detail in the mouse. Ten murine genes have beenidentified, which show homology ranging from 50 to 90% in aminoacid content (8).

The factors which control Wnt gene expression either as individualgenes or as an interacting group have not yet been identified. Wntproteins contain 350—380amino acids, including 24 highly conservedcysteines and a number of glycosylation sites which, with one exception, are poorly conserved. In Wntl, with the exception of sometemperature-sensitive mutants, mutations affecting cysteine residuesabolish biological activity, whereas those affecting glycosylation sitesdo not (10, 11). When in the cell, the protein is bound to thechaperonin-binding protein (12) and is associated with the endoplasmic reticulum and the Golgi apparatus. The N terminus of the proteincarries a signal peptide as well as an adjacent signal peptidase recognition site. This, together with the observations that mutants lackingthe signal peptide are inactive (10) and that there are no apparenttransmembrane domains, suggests that Wnt protein is secreted. Indeed, Papkoff and Schryver (13) have shown that Wntl is present on

Received 11/16/93; accepted 3/15/94.Thecostsof publicationof thisarticleweredefrayedin partby thepaymentof page

chargesThisarticlemustthereforebe herebymarkedadvertisementin accordancewith18 U.S.C. Section 1734 solely to indicate this fact.

I This work was funded by the Imperial Cancer Research Fund, the Oxford District

ResearchCommittee[E. L H., R. B., A. L H.], and HoffmanLa Roche[J. A. M.,A. P. M.J. E. L H. and J. A. M. contributed equally to this study.

2 To whom requests for reprints should be addressed, at Imperial Cancer Research

Fund, University of Oxford, John Radcliffe Hospital, Headington, Oxford 0X3 9DU,UnitedKingdom;ER,estrogenreceptor.

the cell surface. However, relatively little is known about the precisemolecular structure and function of the Wnts because the strongassociation of the proteins with the extracellular matrix has to dateprevented isolation of active protein.

At a cellular level, Wntgenes show specific spatiotemporalexpression in both embryonic and adult tissues of species in which they havebeen investigated. Studies in the mouse (14), Xenopus (15), and othersystems all suggest a role for Wnt genes as signal transducers functioning locally to produce biological effects in cells adjacent to thesource of secretion. Thus, Wnt genes play a part in tissue organizationevents such as central nervous system development in the mouse andsegmentation in Drosophila.

Interest in the role of Wnt genes in breast biology stems fromstudies in the mouse. Gavin and McMahon (16) observed thatmouse breast tissue showed differential expression of Wnt4,Wnt5a, Wnt5b, Wnt6, and Wnt7b which correlated with mammarychanges associated with pregnancy and lactation. This and thework of Buhler et a!. (17), demonstrating the expression of Wnt2in the mouse mammary gland, suggest a role for Wnt genes innormal breast development. In addition, a number of lines ofevidence show that some Wnt genes have growth-promoting properties in breast epithelium. Thus, Wntl and Wnt3 are two of thegenes, which, when activated by mouse mammary tumor virus,cause mouse mammary tumors. Also, Writi (18) and Wnt2 (19)transform some mouse mammary epithelial cell lines in vitro. Wntltransgenic mice (20) and mice bearing reconstituted mammaryglands transfected with Wntl (21) show hormone-independenthyperplasia and an increased incidence of breast tumors. Wntl mayproduce these effects in both autocrine and paracrine mechanisms(22). It is pertinent that those Wnt genes which are associated withgrowth deregulation are those which are absent in the normal adultbreast. Thus, an overall hypothesis consistent with the above datais that, in the mouse, Wnt genes are involved in the developmentalprocesses of normal breast but aberrant Wnt gene expression cancontribute to the development of mammary cancers.

In view of this we have investigated whether a similar mechanism might operate in some human breast cancers. The only humanWrit genes so far described are Wntl, Wnt2 and Wnt3. Using PCR3techniques, we have generated fragments specific to several humanWnt genes for use as templates in ribonuclease protection analysis.We have examined Wntgene expression in benign and malignantdiseases of the breast and compared their expression profile to thatof normal breast tissues. An immortalized human breast epithelialcell line and human breast cancer cell lines were examined as apure epithelial population. Our results show that, as in the mouse(16), only a subset of Win gene are expressed in normal humanbreast tissue. Interestingly, quantitative differences in expressionare seen between normal tissue and benign and malignant tumorsof the human breast.

3 ‘l'heabbreviations used are: PCR, polymerase chain reaction; DMEM, Dulbecco's

modified Eagle's medium; FCS, fetal calf serum; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; EGFR, epidermal growth factor receptor; MCF7adr, Adryamicm-resistant MCF7; ER, estrogen receptor.

2615

Differential Expression of Human Wnt Genes 2, 3, 4, and 7B in Human Breast CellLines and Normal and Disease States of Human Breast Tissue'

Emmanuel L. Huguet, Jill A. McMahon, Andrew P. McMahon, Roy Bicknell, and Adrian L Harris2

Growth Factors Group, Imperial Cancer Research Fun4 University of Oxford, Institute ofMolecular Medicine, John Radcl(fte Hospita4 Headington, Oxford, 013 9DU, UnitedKingdom fE. L. H., R. B., A. L. H.], and Departments ofBiochemisny and Molecular Biology and Cellular and Developmental Biology, Harvard University, Massachusetts 02138(I. A. M., A. P. M.J

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

Wnt GENE EXPRESSION IN HUMAN BREAST TI@UE

breast tissue. Details of the samples of breast tissue from which RNA wasextracted are given in the legend to Fig. 2. Normal breast RNA wasobtained either from breast reduction tissue samples (n = 4) or from normalbreast tissue adjacent to diseased breast excised at operation (n = 4). Fornormal tissues the age range of the patients was 18—78years (mean, 42years; median, 46 years). The tissues representing benign breast diseaseconsisted of fibroadenomas (n = 5), benign phylloides tumors (n = 2), and

fibrocystic disease (n 2). For benign tissues the age range of the patients

was 25—52years (mean, 39 years; median, 38 years). The tumor tissueswere obtained postoperatively and were chosen to represent differentsubgroups (patient age range, 32—86years; mean, 56 years; median, 57years. By receptor status the tumors consisted of 40% ER positive, EGFRpositive; 30% ER positive, EGFR negative; 30% ER negative, EGFRpositive. The size of the tumors at the time of resection ranged from 8—40mm, 40% of the tumors were lymph node positive, 19 of the tumors wereductal and one was lobular in type). Receptor analysis and tumor tissuehandling was performed as previously described by Lejeune et al. (24).Tumors were considered ER positive if analysis showed at least 10 fmol

CWWSQPDFRACWWSQPDFRT

IGDFLKDKYDIGDFLXDKYD

SASEMVVEKHSASEMVVEXH

H Wnt-3a GLSGSCEVXTH Wnt-3a GLSGSCEVKT

H Wnt-3a RESRGWVETLM Wnt-3a RESRGWVETL

RPRYTYFKVP TERDLVYYEARPRYTYFXVP TERDLVYYEA

SPNFCEPNPESPNFCEPNPE

H Wnt-3a TGSFGTRDRTM Wnt-3a TGSFGTRDRT

CNVSSHGIDG CDLLCCGRGHCNVSSHGIDG CDLLCCGRGH

NARAERRREKNARTERRREK

H Wnt-3a CRCVM Wnt-3a CHCV

H Wnt-4 GVSGSCEVKTM Wnt-4 GVSGSCEVKT

CWRAVPPFRQ VGHALKEKPDCWRAVPPFRQ VGHALXEKFD

GATEVEPRRVGATE VEPRRV

H Wnt-4 GSSRALVPRNN Wnt-4 GSSRALVPRN

AQFXPHTDED LVYLEPSPDFAQFKPHTDED LVYLEPSPDF

CEQDMRSGVLCEQDIRSGVL

SXAIDGCELL CCGRGFHTAQSXAIDGCELL CCGRGFHTAQ

VELAERCSCKVELAERCGCR

H Wnt-4 GTRGRTCNXTM Wnt-4 GTRGRTCNXT

Fig. 1. Amino acid sequences of PCR-generatedfragments of human Wnt3a, Wnt4, Wnt7a, andWnt7b, aligned with the corresponding mouse sequences. H, human sequences; M, mouse sequences. H Wnt-7a GVSGSCTTXT

M Wnt-7a GVSGSCTTKTCWTTLPQFRECWTTLPQFRE

LGYVLKDKYNLGYVLKDKYN

EAVHVEPVRAEAVHVEPVRA

H Wnt-7a SRNXRPTFLKM Wnt-7a SRNKRPTFLK

IKKPLSYRKPIKXPLSYRKP

MDTDLVYIEKMDTDLVYIEL

SPNYCEGDPVSPNYCEEDPV

H Wnt-7a TGSVGTQGRAM Wnt-7a TGSVGTQGRA

CNKTAPQASG CDLMCCGRGYCNKTAPQASG CDLMCCGRGY

NTHQYARVWQNTHQYARVWQ

H Wnt-7a CNCKM Wnt—7a CNCX

H Wnt-7b GVSGSCTTKTM Wnt-7b GVSGSCTTKT

CWTTLPKFRE VGHLLXEKYNCWTTLPXFRE VGHLLKEKYN

AAVQVEVVRAAAVQVEVVRA

H Wnt-7b SRLRQPTFLRM Wnt-7b SRLRQPTFLR

IXQLRSYQKP METDLVYIEXIXQLRSYQKP METDLVYIEK

SPNYCEEDAASPNYCEEDAA

H Wnt—7b TGSVGTQGRIM Wnt-7b TGSVGTQGRI

CNRTSPGADG CDTMCCGRGYCNRTSPGADG CDTMCCGRGY

NTHQYTKVWQNTHQYTKVWQ

H Wnt-7b CNCKM Wnt-7b CNCK

2616

MATERIALS AND METHODS

CellCulture.Celllinesusedin thisstudywereobtainedfromthefollowing sources: MTSV1—7(23) from Dr Joyce Taylor, Imperial Cancer ResearchFund,London;T47D, MDA231, MDA361, BT2O,MDA415, MDA453, andMDA157 from the American Type Culture Collection, Bethesda, MD;ZR9BI1 and ZR75 from Dr. Eva Valvenus, University Hospital, 5-75185,Uppsala,Sweden;MCF7from Dr. BarbaraDurkacz,CancerResearchUnit,University of Newcastle upon Tyne; and MCF7adr from Dr. Ken Cowen,National Institutes of Health. MTSV17 were cultured in DMEM-10% FCS,with 5 g.@g/mlhydrocortisone (Sigma) and 20 gig/mIbovine insulin (Sigma).MDAMB157 was cultured in RPMI medium-10% FCS. All other cell lineswere cultured in DMEM-10% FfS. All cells were cultured on plastic cultureplates (Becton Dickinson) at 37°C,5% C02, 95% air, in a humidified incubator. All cultures were free of Mycoplasma. DMEM and RPMI were obtainedfrom the Imperial Cancer Research Fund Clare Hall laboratories. FCS wasobtained from Globepharm.

Tissue Selection. Human breast tissue samples were selected as follows.Wnt expression was examined in normal, benign disease, and malignant

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

Table 1 Summaryof Wnt expression profile in human breast celllineswnt3wnt4wnt7hMTSV1-7+@@(1)b+(50)MCF7——+(30)ZR75——+(60)T47D——+(50)MCF7adr+

(10)—+(1)MDA231——+(iOO)MDA361——+

(50)BT2O+

(1)—+(i60)MDA415—+(1)+(20)MDA457——+

(45)MDAMB157——+

(4)ZR9B11——+(50)

WNT2 EXPRESSIONPROFILE

Wag GENE EXPRES5ION IN HUMAN BREAST TISSUE

protocol. [a-32P]dCTP (Amersham) labeled probes were generated byrandom priming. Southern analysis of blotted DNA was performed usingstandard protocols (26).

Construction of Wnt Templates for RNase Protection Analysis. Using

reverse (5'-AAAATCFAGAACAACACCAATGAAAA-3') and forward (5'-

0000AATFCCAAGAATGTAAATGTCAT-3') primers, expected frag

ments of 406 and 424 base pairs were generated by PCR amplification fromplasmids pIRP (supplied by Dr. Robert Williamson, St. Mary's Hospital,

London) and pHPl (27) (supplied by Dr. Rod Nusse, Stanford University,Palo Alto, CA), which contain full-length human Wnt32 and Wnt3 cDNAs,respectively. The Wnt2 fragment was subcloned into the EcoRV site ofpBluescript 5K, and the Wnt3 fragment was subcloned into the EcoRI and

XbaI sites of pBluescipt KS. Both constructs were sequenced to confirm their

identity. The BamHI-linearized Wnt2 construct and the EcoRI-linearized Wnt3construct were used to generate [a-32P]CFP-labeled antisense probes with ‘11RNA polymerase. The full-length Wnt2 and Wnt3 probes are then 487 and 470base pairs, respectively, and can be distinguished from the respective protectedfragments of 389 and 400 base pairs.

Probes for human Wnt3a, Wnt4,Wnt7a, and Wnt7bwere isolated accordingto previously published procedures (8) by PCR amplification using 100 ng ofhuman genomic DNA and the degenerate forward (5@@GGGGAATfCCAA/

GGA1O'l'°/C@/GT0/CCAT3) and reverse (5?@AAAATCFAGA@@/0CAA/

0CACCAA/0TG@@/@AA@3l)primers@PCR products for Wnt4(expected size, 360base pairs) and Wnt3a, Wnt7a, and Wnt7b (expected size, 372 base pairs) weresubcloned into the plasmid pGem 3z (Promega) and their nucleotide sequencesdetermined by dideoxy chain termination sequencing. Antisense [a-32PJCTPlabeled RNA probes were generated from EcoRI-linearized templates by SP6polymerase. The amino acid sequences of the human Wnt3a, Wnt4, Wnt7a, and

Wnt7b fragmentsgeneratedby PCR are shown in Fig. 1 with the correspondingmouse sequences.

a @,expression detected; —,expression not detected.

b Numbers in parentheses, relative level of Wnt expression in each cell line.

specific binding sites per mg cytosolic protein and EGFR positive ifanalysis showed at least 20 fmol specific binding sites per mg membraneprotein.

RNA Preparation. RNA was prepared from cells in vitro and from tissuesby acid guanidium thiocyanate phenol chloroform extraction, as described byChomczinskiand Sacchi (25).

DNA Preparation. Genomic DNA was prepared from tissue samplesby lysis with proteinase K and sodium dodecyl sulfate, followed bypurification with phenol chloroform isoamyl alcohol using standardprotocols (26).

Southern Blot and Slot Blot Analysis. Genomic DNA was isolated fromtissues as above and digested with EcoRI. Digested DNA (20 pg) was sizeselected by electrophoresis on a 1% agarose gel and blotted onto a HybondN membrane. Serial dilutions of 1 ng of digested DNA were prepared andslot blotted onto a Hybond N membrane using a Bio-Rad apparatus and

WNT2 SIGNAL

normal tissue benign tissue malignant tissue

@.@.@

GAPDH SIGNAL

20 Wnt2: benign tissue Wnt2: tumor tissue

15

10

normal tissue20

15

10

5

:@@@@@ .

Fig. 2. RNase protection assay showing Wnt2and corresponding GAPDH signals in normal, benign, and malignant human breast tissue samples. Graphs show densitometric analysisofthedata.Normaltissues:A,normaltissueadjacentto tumor;8, normaltissueobtainedfrombreastreductions.Benigntissues:FA,fibroadenoma;FC,fibrocysticdisease;Ph,benignphylloides tumor. Malignant tissues: first ±,ER status, second ±,EGFR status. For all samples, 1, premenopausal tissue; 2, postmenopausal tissue. For each tissue sample the Wnisignal is shown above the GAPDH signal. The densitometric quantitation of the corrected Wnt signal is shown graphically in bar charts, and the order of the bars corresponds to theorder of the tissues above.

U :@:+:@:++ ++:1::@:++ + ++ + ::::@: :t:

2617

@..@ -S—-@@@@

.@ .@@ 0 @*

Wnt2:

10

5

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

WNT3 EXPRESSION PROFILE

_.-@@

Wnt GENE EXPRESSION IN HUMAN BREAST TISSUE

WNT3 SIGNAL

.

normal tissue benign tissue malignant tissue

_@.@ —-—— wGAPDH SIGNAL

Wnt3: normal tissue Wnt3: benign tissue Wnt3: tumortissue

A 120-base pair fragmentof the human GAPDH gene cloned into theplasmid P Bluescript SK+ (28) was used to generate antisense GAPDHprobes.

RNase Protection Analysis. RNase protection analysis was performedusing the described constructs by standard protocols (26). In all assays 10 p@gof RNA was hybridized to a cocktail containing 10@cpm of each of Wnt andGAPDH probes. Protected fragment signals for both Wnt and GAPDH werequantified by laser densitometry using a Bio Image analyzer (Millipore). In allassays the GAPDH signal was used as a loading control.

Statistical Analysis. P values were derived from Mann-Whitney U testcalculations, using the “statview―version 4 program (Abacus Concepts Inc.)

RESULTS

Expression of Wnt RNA was examined by RNase protection usingprobes for Wnt2, Wnt3, Wnt3a, Wnt4, Wnt7a, and Wnt7b. Expressionwas assayed in human cell lines, as well as in normal, benign, andmalignant human breast tissue.

Wnt Expression in Breast Cell Lines. The panel of cell lines(shown in Table 1 with their Wnt expression profile) consisted ofMTSV1—7(a human immortalized benign mammary epithelial cellline) (23), three ER-positive breast cancer cell lines (MCF7, ZR75,T47D), six ER-negative breast cancer cell lines (Adriamycin-resistantMCF7, MDA231, MDA361, MDA415, MDA457, MDAMB157), andthe EGFR-transfected breast cancer cell line ZR9B1 1. Wnt3, Wnt4,and Wnt7b were expressed in some of the cell lines, but Wnt2, Wnt3a,and Wnt7a were not.

Wnt3 was expressed in 3 cell lines. MTSV1—7and BT2O showed asimilar level of expression, which was 10-fold lower than that foundin MCF7adr. Wnt4 was present only in MDA415. Wnt7b was presentin all cell lines examined. There was a 100-fold range in the level of

expression of Wnt7b between the lowest and highest expressing celllines. No consistent relationship was found between levels of expression and ER or EGFR status, although the 3 Wnt3-expressing cell lineswere all ER negative, EGFR positive.

Wnt Expression in Human Breast Tissue. Figs. 2—5show thesignals obtained for Wnt2, Wnt3, Wnt4, and Wnt7b with the conesponding GAPDH signals and densitometric quantitation of the data.

Wnt2 expression was 10- to 20-fold higher (P = 0.03) in fibroadenomas (n = 5) than it was in normal, malignant, or benign tissues(taken together n = 30). Wnt2 expression was also found in fibroblastsfreshly isolated from a breast reduction sample (data not shown).

Wnt3 was expressed at similar levels in normal, benign, and malignant breast tissue. There was a 20-fold range in expression betweenthe lowest and highest expressing samples.

Like Wnt2, Wnt4 expression was higher (P = 0.001) in fibroadenomas (n = 5) than it was in other tissues (taken together n = 31).The increase in expression was approximately 10-fold, with onesample showing a 40-fold elevation.

Wnt7b was expressed at similar levels in normal and benign tissuesbut showed an elevation of approximately 30-fold in 10% of tumors.Southern analysis of genomic DNA extracted from normal tissue withlow Wnt7b expression and from the tumor expressing the highest levelof Wnt7b showed no obvious rearrangement of the gene, and slot blotanalysis of the same DNA samples showed no genomic amplification(slot blot and Southern data not shown).

DISCUSSION

Our results show that several Wnt genes are expressed in normalhuman breast tissue. The specific role of Wnt genes as developmental

2618

—--...@

3 3

2 21

111LJ@ic@@@@'-,-csJv- 0@-C@@.1-C%iC%4@- -+++•._,_,•

@ @O@O.c .tfftf@@

Fig. 3. RNase protection assay showing Wnt3and corresponding GAPDH signals in normal, benign, and malignant human breast tissue samples. Graphs show densitometric analysisof the data. See Fig. 2 for additional details.

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

WNT4 EXPRESSIONPROFILE

@ -ff@T:j@@

WagGENE EXPRESSIONIN HUMAN BREAST TISSUE

WNT4 SIGNAL

normal tissue benign tissue malignant tissue

___-@—‘I@@@ .-. —@.@@GAPDH SIGNAL

Wnt4: normal tissue Wnt4: tumortissue50 50

40 40

30 30

20 20

10 10

Wnt4: benign tissue

1@

-,-CsJ@-@ _-esJ.,.-c'JI-,-.@ @o@d.c

u_ U. U@Li@U@0 u u@Q. C@-'-C@'-C@JC%JC%J

Fig. 4. RNase protection assay showing Wnt4 and corresponding GAPDH signals in normal, benign, and malignant human breast tissue samples. Graphs show densitometric analysisof the data. See Fig. 2 for additional details.

regulators in all systems in which they have been studied supports thehypothesis that they may play a similar role in human breast tissue.The fact that not all Wnt genes examined are expressed suggests thatthey have different functions. It is interesting that even the mosthomologous Wnt genes, such as pairs 3 and 3a and 7a and 7b, showdifferent expression profiles (3 and 7b are expressed, 3a and 7a arenot), suggesting a lack of redundancy in Wnt function.

Our data allow a comparison to be made between the Writgenesexpressed in normal human breast and those Wnt genes that areexpressed at some stage of development in the mouse breast (Table 2).Mouse and human breast resemble each other in their expression ofWnt2, Wnt4, and Wnt7b and in the absence of Wnil, Wnt3a, andWnt7a. However, they differ in that Wnt3 (absent in normal mousebreast, by Northern analysis at least), is present in the human breast.When studying the mouse, Gavin and McMahon (16) were able toshow a clear correlation between breast dcye1op@nt during pregnancy and lactation and differential Wnt expression. In our experiments the profile of Wnt expression of postmenopausal breast was nodifferent from that of premenopausal breast. Clearly, the fact thatbreast tissue from pregnant and lactating women is not readily available will make it difficult to repeat the experiments of Gavin andMcMahon in the human. However, the question of Wnt gene regulalion may be partially answered by investigating the effect of pregnancy-related and other growth factors on the Wnt expression profile ofhuman mammary epithelial cells in vitro.

Cell line expression of Wrngenes mirroredthe tissue expressionwith the exception of Wnt2, which was absent in cell lines but presentin tissues. It is possible that this discrepancy is due to the expressionof Wnt2by stromalcells, which arepresentin tissuesbut not in thecell line cultures. This hypothesis is supported by the finding that

Wnt2 mRNA is clearly present in fibroblasts freshly isolated from abreast tissue sample, and it is interesting that Wnt2 is elevated intissues from benign breast disease, where there is a large stromalelement. In addition, mesenchymal expression of Wnt2 has beendemonstrated in other systems (29, 30). If the in vivo expression ofWnt2 is indeed stromal, this would suggest that the protein can act inparacrine fashion in the human breast and would parallel the in vitrodemonstration of Wntl paracrine action in coculture assays (22).

If Wnt genes have a role in normal developmental processes of thehuman breast, then disease states of the breast in which tissue morphology is altered may be associated with changes in Writexpression.Our results show an increased level of Wnt2 and Wnt4 in fibroadenomas and an elevated level of Wnt7b in approximately 10% of breasttumors. The finding that Wnt2 and Wnt4 are overexpressed in fibroadenomas does not establish an etiological relationship. However, thefacts that Wntl can cause mammary hyperplasia in the mouse and thatfibroadenomas are hyperplastic rather than neoplastic (31) support acausal role for Wnt genes in the pathology of these lesions. It isgenerally accepted that the term “fibrocysticdisease―has no prognostic significance per se but that some of the histological featuressometimes exhibited do correlate with an increased risk of breastcancer. These include the different types of epithelial hyperplasia(32). In the mouse, aberrant Wnt expression causes mammary epithehal cell hyperplasia, and it would, therefore, be of great interest toinvestigate Wnt expression in human breast epithelial hyperplasia,particularly “atypicalepithelial hyperplasia― in which a 5-fold increased risk of breast cancer is recognized (32).

Wnt7b was expressed in all tissues examined but was elevated in10% of tumors. These tumors could not be distinguished from othersin the panel on the basis of clinical criteria such as estrogen receptor

2619

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

WJ1T7bEXPRESSIONPROFILE

Wnt233a47a7bMousebreast÷a+—+(Refs.

16 and 17)Human celllines—+—+—+Humantissue++—+—+NormalBenign

Malignant+ ++ +— -+ +— —++a

@ expression detected;—, expression not detected.

WaS GENE ExPRESSION IN HUMAN BREAST TISSUE

WNT7b SIGNAL

normal tissue benign tissue

•.e..I-I• •1..@L1

GAPDH SIGNAL

Wnt7b: benign tissues Wntlb: tumortissue50

40

30

20

10I—

Wntlb: normal tissue50 50

40 40

30 30

20 20

10 10

!@@

@@<crcca:<crcrFig. 5. RNase protection assay showing Wnt7b and corresponding GAPDH signals in normal, benign,

analysis of the data. See Fig. 2 for additional details.

V

,@:C%JCSJCsJCsJ(‘JCsJC@C%JCsJ,- CsJCsJ1@I@OJ@-C@JCSJC'J

and malignanthumanbreasttissuesamples.Graphsshowdensitometric

levels, epidermal growth factor receptor levels, differentiation grade,histological type, or nodal status. Our analysis of the tumor expressingthe highest levels of Wnt7b would not demonstrate a minor translocation of the gene but makes a major rearrangement of Wnt7b in thistumor unlikely. Furthermore, the gene shows no genomic amplification; therefore, the increased RNA level may be due to increasedtranscription or decreased degradation.

In all cases in which Wnt signal was detected, the level of expression was nevertheless generally low, in that it was usually necessaryto expose the gels for 1 week to obtain signals intense enough forreliable quantitation. This suggests that low quantities of Writmay besufficient to produce biological effects and would fit with the observation that, in the appropriate context, low-level Wnt expression isenough to achieve transformation (33).

An association between disease states and differential Wnt expression in the breast does not establish a causal relation between the two.It would be of interest to transfect human Wnt genes into humanmammary epithelial cells to demonstrate whether or not altered levels

of Writ expression can be transforming, or at least alter normalmammary epithelial cell biology.

In summary,ourresultsshow thatseveralhumanWntgenes areexpressedin normal breast tissue. Diseased breast shows levels of expression of some

Wnt genes different from that found in normal breast These data establish arole for Writgenes in human breast biology andjustify a further examinationof apossiblerolein breastpethologj@.

REFERENCES

1. Nusse, R., and Varmus, H. E. Wnt genes. Cell, 69: 1073-1087, 1992.2. McMahon, A. P. The wnt family of developmental regulators. Trends Genet., 8:

236—242,1992.3. Nusse, R., and Varmus, H. E. Many tumours induced by the mouse mammary tumour

virus contain a provirus integrated in the same region of the host genome. Cell, 31:99—109,1982.

4. Rijsewijk, F., Schuermann, M., wagcnasr, E., Parren, P., Weigel, D., and Nusse, R.The Drosophila homolog of the mouse mammary oncogene int-1 is identical to thesegment polarity gene wingless. Cell, 50: 649—657,1987.

5. Nusse, R., Brown, A., Papkoff, J., Scambler, P., Shackleford, G., McMahon, A,Moon, R., and Varmus, H. A new nomenclature for mt-i and related genes: the Wntgene family (letter). Cell, 64: 231, 1991.

6. Russell, J., Gennissen, A., and Nusse, R. Isolation and expression of two novelWnt/wingless gene homologues in Drosophila. Development (Camb.), 115:475—485,1992.

7. Wolda, S. L., and Moon, R. 1. Cloning and developmental expression in Xenopuslaevis of seven additional members of the wnt family. Oncogene, 7: 1941—1947,1992.

8. Gavin, B. 1., McMahon, J. A., and McMahon, A. P. Expression of multiple novelWnt-1/int-1-related genes during fetal and adult mouse development. Genes Dcv., 4:2319—2332, 1990.

9. Wainwright, B. 1., Scambler, P. J., Stanier, P., Watson, E. K., Bell, 0., Wicking, C.,Estivill, X., Courtney, M., Boue, A., and Pedersen, P. S. Isolation of a human genewith proteinsequencesimilarityto humanand murinemt-i and the Drosophilasegment polarity mutant wingless. EMBO J., 7: 1743-1748, 1988.

10. McMahon, A. P., and Moon, R. T. Ectopic expression of the proto-oncogene int-1 inXenopus embryos leads to duplication of the embryonic axis. Cell, 58: 1075-1084,1989.

Table2 Wntgeneexpressionin mousemammaryglandand in humanbreastcell linesand breast tissues

2620

malignant tissue

@.. 0SO.@ .@ I. •..O@

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

Wnt GENE EXPRESSIONIN HUMAN BREAST TISSUE

11. Mason, J. 0., Kitajewski, J., and Varmus, H. E. Mutational analysis of mouse Wnt-1identifies two temperature-sensitive alleles and attributes of Wnt-1 protein essentialfor transformation of a mammary cell line. MoL Biol. Cell, 3: 521—533,1992.

12. Kitajewski, J., Mason, J. 0., and Varmus, H. E. Interaction ofWnt-1 proteins with thebinding protein B1P.Mol Cell. Biol., 12: 784—790,1992.

13. Papkoff, J., and Schryver, B. Secreted int-1 protein is associated with the cell surface.MoL Cell. Biol., 10: 2723—2730,1990.

14. Dickinson,M. E., and McMahon,A P. The role of Wnt genes in vertebratedevelopment. Curr. Opin. Genet. Dev., 2: 562—566,1992.

15. Moon, R. T. In pursuit of the functions of the Wnt family of developmentalregulators:insightsfromXenopuslaevis.Bioessays,15:91-97,1993.

16. Gavin, B. J., and McMahon, A. P. Differential regulation of the Wnt gene familyduringpregnancyand lactationsuggestsa role in postnataldevelopmentof themammary gland. Mo! Cell. Biol., 12: 2418-2423, 1992.

17. Buhler, T. A@,Dale, T. C., Kieback, C., Humphreys, R. C., and Rosen, J. M.Localization and quantification of Wnt-2 gene expression in mouse mammary development. Dev. Biol., 155: 87—96,1993.

18. Bmwn, A@M., Wildin, R. S., Prendergast, T. J., and Varmus, H. E. A retrovirus vectorexpressing the putative mammary oncogene mt-i causes partial transformation of amammary epitheial cell line. Cell, 46: 1001-1009, 1986.

19. Blasband,A., Schryver,B., and Papkoff,J. The biochemicalpropertiesand testisformingpotentialofhumanWnt-2aresimilartoWnt-1.Oncogene,7: 153—161,1992.

20. Un, T. P., Guzman, R. C., Osborn, R. C., Thordarson, G., and Nandi, S. Role ofendocrine, autocrine, and passerine interactions in the development of mammaryhyperplasia in Wnt-1 transgemc mice. Cancer Res., 52:4413-4419,1992.

21. Edwards, P. A, Hiby, S. E., Papkoff, J., and Bradbury, J. M. Hyperplasia of mousemammary epithelium induced by expression of the Wnt-1 (int-1) oncogene in reconstituted mammary gland. Oncogene, 7: 2041—2051,1992.

22. Jue, S. F., Bradley, R. S., Rudnicki, J. A.. Varmus, H. E., and Brown, A. M. Themouse Wnt-1 gene can act via a passerine mechanism in transformation of mammaryepithelial cells. Mol Cell. Biol, 12: 321—328,1992.

23. Bartek, J., Bartkova, J., Kiprianou, N., Lalani, E., Staskova, S., Shearer, M., Chang,S., and Taylor-Papadimitriou, J. Efficient immortalization of luminal epitheial cellsfrom human mammary gland by introduction of simian virus 40 large tumor antigenwith a recombinant retrovirus. Proc. Nati. Acad. Sci. USA, 88: 3520—3524, 1991.

24. Leieune, S., Leek, R., Horak, E., Plowman, G., Greenall, M., and Harris, A. LAmphiregulin, EGF receptor and estrogen receptor in human breast cancer. CancerRes., 53: 3597—3602,1993.

25. Chomczinski, P., and Sacchi, N. Single step isolation of RNA by acid guanidiumthiocyanate-phenol-chloroform extraction. Anal. Biochem., 162: 321—328,1987.

26. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith,J. A., and StruM,K. CurrentProtocolsin MolecularBiology.Vol. 1. New York:Green Publishing Associates & Wiley lnterscience, 1990.

27. Roelink, H., Wang, J., Black, D. M., Solomon, E., and Nusse, R. Molecular cloningand chromosomal localisation to 17q21 of the human wnt3 gene. Genomics, 17:790—792,1993.

28. McCarthy, S. A, and Bicknell, R. Responses ofpertussis toxin-treated microvascularendothelial cells to transforming growth factor fll. J. Biol. Chem., 267: 21617—21622,1992.

29. Levay-Young, B. K., and Navre, M. Growth and developmental regulation of wnt-2(irp) gene in mesenchymal cells of fetal lung. Am. J. Physiol., 262: 672—683,1992.

30. McMahon, J. A., and McMahon, A. P. Nucleotide sequence, chromosomal localization and developmental expression of the mouse mt-i-related gene. Development(Camb.), 107: 643—650,1989.

31. Noguchi, S., Motomura, K., Inaji, H., lmaoka, S., and Koyama, H. Clonal analysis offibroadenoma and phyllodes tumor of the breast. Cancer Rca., 53: 4071—4074,1993.

32. Connolly, J. L, and Schnitt, S. i. Benign breast disease. Resolved and unresolvedissues. Cancer (Phila.), 71: 1187—1189,1993.

33. Nusse, R., van-Ooyen, A@,Cox, D., Fung, Y. K., and Varmus, H. Mode of proviralactivation of a putative mammary oncogene (mt-i) on mouse chromosome 15. Nature(Lond.), 307: 131—136,1984.

2621

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

1994;54:2615-2621. Cancer Res   Emmanuel L. Huguet, Jill A. McMahon, Andrew P. McMahon, et al.   Human Breast TissueHuman Breast Cell Lines and Normal and Disease States of

Genes 2, 3, 4, and 7B inWntDifferential Expression of Human

  Updated version

  http://cancerres.aacrjournals.org/content/54/10/2615

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/54/10/2615To request permission to re-use all or part of this article, use this link

Research. on September 12, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from