trask loss enhances tumorigenic growth by liberating ... · for heregulin (ho786, sigma-aldrich)...

Molecular and Cellular Pathobiology

Trask Loss Enhances Tumorigenic Growth by LiberatingIntegrin Signaling and Growth Factor Receptor Cross-Talk inUnanchored Cells

Danislav S. Spassov1, Ching Hang Wong1, Sunny Y. Wong3, Jeremy F. Reiter2, and Mark M. Moasser1

AbstractThe cell surface glycoprotein Trask/CDCP1 is phosphorylated during anchorage loss in epithelial cells in which

it inhibits integrin clustering, outside-in signaling, and cell adhesion. Its role in cancer has been difficult tounderstand, because of the lack of a discernible pattern in its various alterations in cancer cells. To address thisissue, we generated mice lacking Trask function. Mammary tumors driven by the PyMT oncogene and skintumors driven by the SmoM2 oncogene arose with accelerated kinetics in Trask-deficient mice, establishing atumor suppressing function for this gene. Mechanistic investigations in mammary tumor cell lines derived fromwild-type or Trask-deficient mice revealed a derepression of integrin signaling and an enhancement of integrin-growth factor receptor cross-talk, specifically in unanchored cell states. A similar restrictive link betweenanchorage and growth in untransformed epithelial cells was observed and disrupted by elimination of Trask.Together our results establish a tumor-suppressing function in Trask that restricts epithelial cell growth to theanchored state. Cancer Res; 73(3); 1168–79. �2012 AACR.

IntroductionEpithelial tumor growth is dependent on a number of

cellular processes, one of which is cell adhesion. The relation-ship between tumor growth and cell adhesion is a complicatedone.While the normal epithelium is anchored on an underlyingbasement membrane, epithelial tumors grow without therequirement for anchorage and free of the boundaries andrestrictions associated with basement membranes. Althoughthe loss of anchorage can have a liberating effect, allowing theoften erratic and invasive growth of tumors, the absence of celladhesion can also have a restrictive effect, disabling signalingpathways important in promoting growth, and removingmechanisms that enable cell migration and invasion. In par-ticular, adhesion signaling through integrin receptors is knownto be important in promoting many aspects of tumor growthincluding proliferation, survival, migration, invasion, andmetastasis (1, 2). Integrins not only mediate the migratoryfunctions of tumor cells, but more specifically, cross-talkbetween integrins and growth factor receptors is thought to

be critically important for some tumors (3–5), particularlythose driven by oncogenic growth factor receptors, such asHER2/Neu, and in fact the development of HER2/Neu-driventumors is considerably delayed or suppressed in mice lackingb1- or b4-integrins (6, 7). Therefore, the relationship betweentumorigenesis and cell adhesion is a complex one that likelyinvolves reconciling the tumor promoting and suppressingconsequences of integrin engagement and activation. Whileintegrin signaling is generally linked with cell adhesion, theirrelationship is subject to layers of regulation, and potentiallycorruptible in tumors seeking freedom from anchorage butpersistency in integrin signaling. Here, we describe a Src-drivenantiadhesive pathway that negatively regulates integrin sig-naling and cell growth during anchorage deprivation, confer-ring a tumor-suppressive role and accounting for its observeddecrease or loss of expression in some tumors.

Trask (transmembrane and associated with Src kinases, akaCDCP1, SIMA135, gp140) was independently identified byseveral groups in the recent decade, studying different aspectsof cancer biology (8–11). The Trask protein is a 140 kDtransmembrane glycoproteinwith a large extracellular domaincontaining CUB domains, and a smaller intracellular domaincontaining 5 tyrosines and a proline-rich region. Trask isproteolytically cleaved within its extracellular domain by ser-ine proteases including MT-SP1 and is expressed in most cellsas a blend of 140 kD (uncleaved) and 80 kD (cleaved) forms invarying ratios (9, 12). Trask is widely and abundantly expressedthroughout epithelial tissues and specific hematopoietic cells,but its expression is not readily detected in mesenchymaltissues (13, 14).

In cultured cells, we found that Trask undergoes phosphor-ylation by Src kinases immediately upon the loss of anchorage,accounts for the majority of cellular phosphotyrosine content

Authors' Affiliations: 1Department of Medicine, Helen Diller Family Com-prehensive Cancer Center; 2Department of Biochemistry and Biophysics,Cardiovascular Research Institute, University of California, San Francisco,San Francisco, California; and 3Department of Dermatology, University ofMichigan, Ann Arbor, Michigan

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

CorrespondingAuthor:MarkM.Moasser, Department ofMedicine, HelenDiller Family Comprehensive Cancer Center, University of California, SanFrancisco, Box 1387, San Francisco, CA 94143. Phone: 415-476-0158;Fax: 415-502-5665; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-12-2496

�2012 American Association for Cancer Research.

CancerResearch

Cancer Res; 73(3) February 1, 20131168

Cancer Research. by guest on August 31, 2020. Copyright 2012 American Association forhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from

https://bloodcancerdiscov.aacrjournals.org

in the unanchored state in many cells, and remains phosphor-ylated until anchorage is restored (13, 15). When phosphory-lated, Trask functions to negatively regulate cell adhesion. Thisis accompanied by its association with b1-integrin complexes,interfering with integrin clustering, and preventing down-stream signaling including the phosphorylation of focal adhe-sion kinase (FAK), establishing pY-Trask as a negative regu-lator of FAK (16). The antiadhesive effects of phosphorylatedTrask are mediated specifically through the inhibition ofintegrin clustering without an effect on integrin affinity state(16). Consistent with its antiadhesive functions, Trask knock-down confers increased cell adhesiveness and constitutiveTrask phosphorylation blocks cell adhesion (9, 16, 17). P-Traskinterferes with most integrin complexes and adhesion to mostmatrix ligands, resulting in the inhibition of physical cell–matrix adhesion and consequently the disruption of focaladhesions and focal adhesion signaling through FAK, 130Cas,paxillin, and other focal adhesion proteins (16). Importantly,the phosphorylation of Trask and the phosphorylation of focaladhesion proteins, including FAK, occur in exclusion of eachother, defining a phosphotyrosine switch that underlies cellanchorage state (15, 16).In a survey of human epithelial tumors of different types, we

found that Trask is phosphorylated in many tumors, usually ina patchy distribution, including preinvasive, invasive, andmetastatic tumors (18). The phosphorylation of Trask is neverseen in normal epithelial tissues except for occasional mitoticor detached cells, consistent with the fully anchored state ofthe normal epithelium (18). The observed phosphorylation ofTrask in tumors may be either a physiologic event reflectingthe loss of cell anchorage in regions of tumors, or may be atumor-promoting pathologic event, perhaps due to the acti-vationof Src kinases. Experimental studiesusinghumancancercell lines fail to support a tumorpromoting function, and in factsuggest a tumor-suppressing function (19). Consistent with atumor-suppressing function, the expression of Trask is reduc-ed or lost in some tumors compared with the normal epitheliaand reduced or absent in some cancer cell lines (19).In the current study, we set out to more definitively deter-

mine the functions of Trask in tumorigenesis. We generatedmice lacking Trask function and studied its effects on thedevelopment and progression of 2 oncogene-induced tumors.We found that the loss of Trask significantly acceleratestumorigenesis, confirming a tumor-suppressing function.Mechanistic studies on these tumor cells and untransformedepithelial cells reveal that Trask phosphorylation is a mecha-nism that restricts growth in the unanchored state, at leastpartly through the inhibition of integrin signaling and itsassociated cross-talk with growth factor receptors.

Materials and MethodsTargeting vector construction and generation of Trask-deficient miceDNA was amplified from mouse 129 isogenic genomic DNA

with Pfu Ultra HF polymerase (Invitrogen). Three DNA frag-ments were amplified—A, B, and C. Fragment A was amplifiedwith primers ABF3 50-AGGCGCTGGACTCCCAGC and AR1 50-GGCAAGCCAGATTTCAGCCCAA. Fragment B was amplified

with primers BF1 50-ATCCACGAGGTCCCAGACCATC andABR1 50-CCTAAGGCTTGTTCCGCGCGC. Fragment C wasamplified with primers CF1 50-CGAGGACGTTCTCCAGGCAG-CAC and C20878R 50-CCTTGGCCAGCTTCCAGGTATG. Frag-ment B is 1,477 bp long and contains exon 1 of Trask andupstream promoter sequences. Loxp sites were engineered atthe flanks of fragments B and restriction sites EcoRV, SacI, andSwaI were introduced immediately adjacent to the loxp sitesfor diagnostic purposes. FragmentA is 3 kb long and serve as 50-arm of the vector, fragment C is 8.5 kb long and serve as a 30-arm. The fragment A, B, and C were cloned into pB6 vector. Inthis vector, the Neo cassette is flanked by flippase recognitiontarget sites, and the expression of Neo is driven by RNA-pol IIpromoter. The entire sequence of the targeting vector wasdetermined and found to be correct, and no additional muta-tions resulting from PCR amplification were found. The target-ing construct was linearized with ScaI and electroporated into129 embryonic stem cells. Genomic DNA isolated from 500G418 resistant embryonic stem cell colonies was analyzed bySouthern blotting. Several correctly targeted embryonic stemcell clones were identified and injected into blastocysts togenerate chimeric mice. Chimeras were mated with actin-FLPfemales on C57/BL6 background (Jackson Laboratory). Germ-line transmission andNeoexcisionwere confirmedbySouthernblot analysis. The floxedmice were bred for several generationswith normal C57/BL6 mice and later mated with EIIa-cre mice(Jackson Laboratory). Excision of fragment B and generation ofTrask heterozygous mice was confirmed by PCR. These micewere mated several times with normal C57BL/6 mice and theEIIa-cre transgenewas bred out. Trask heterozygousmice werebred to generate Traskþ/þ, Traskþ/�, and Trask�/� animals forsubsequent analysis. Trask heterozygous mice were bred for 6consequent generationswithnormal FVB/Nmice beforematedwith MMTV-PyMT mice. The experiments with K14-CreERT

SmoM2 mice were carried out on a mixed 129/SVJae-C57/BL6background as previously reported (20).

PCR and Southern blot analysesGenotyping of floxed mice was conducted with primer 2

(SeqNeo1) 50-CGCTCAGGCACGTGGGAGAG and primer 3(ABR3) 50-GTGTGGAGTGGGCCCGAAG.

Genotyping of Traskþ/þ, Traskþ/�, and Trask�/�mice weredone bymultiplex PCRwith 3 primers—primer 1 (FloxpEcoRV)50-TGACTGATACTGGGGCATGGGC, primer 2 (SeqNeo1) 50-CGCTCAGGCACGTGGGAGAG, and primer 3 (ABR3) 50-GTGTGGAGTGGGCCCGAAG. The 50-external probe wasamplified with primers 5F6472-50-ACTTGTGCCCATCCT-GCCTGTG and 5R7894-50-GCCAACTGCTCCTCCTACCTGCsubcloned and sequence confirmed.

The 30-external probe was amplified from mouse genomicDNA with primers F20923-50-AGCAGGTCCCTCCATTGCT-CAG and R21392-50-AGGAATCCAAGGGCCAGCCAG sub-cloned and sequence confirmed. A 556 bp BamHI SphI frag-mentwas excised from the coding sequence of theNeo cassetteand used as a Neo probe. Primers used for genotyping PYMTalleles were PYMT forward: 50-GGAAGCAAGTACTTCA-CAAGGG and PYMT reverse: 50-GGAAAGTCACTAGGAG-CAGGG. Primers used for genotyping SmoM2 were Smo

Trask Suppression of Integrin Cross-talk and Tumorigenesis

www.aacrjournals.org Cancer Res; 73(3) February 1, 2013 1169



forward: 50-TCCTTGAAGAAGATGGTGCG and Smo reverse:50-AAGTTCATCTGCACCACCG. Detecting K14-CreERT wasdone with primers CreF 50-ATACCGGAGATCATGCAAGCCreR 50-AGGTGGACCTGATCATGGAG.

Animal studiesAll animal work was conducted under the guidelines of the

University of California, San Francisco (UCSF; San Francisco,CA) Institutional Animal Care and Use Committee (IACUC)under an IACUC-approved protocol. MMTV-PyMT mice werepurchased from Jackson Laboratory. At 12 weeks of age, femalemice were sacrificed and all mammary tumors were carefullyexcised, weighted, and photographed. Portions of the tumorswere also frozen in liquid nitrogen or fixed in formalin. Forhistologic analysis, mouse tissues were fixed in 10% bufferedformalin for 24 hours and embedded in paraffin. Sections werestained with hematoxylin and eosin (H&E). For detection oflungmetastasis, lungs of Trask wt and null mice were analyzedinH&E sections. K14-CreERT and SmoM2 transgenicmice werefrom J. Reiter laboratory (UCSF). For induction of Cre recom-binase, 30-day-old mice were injected intraperitoneally withtamoxifen (T5648, Sigma-Aldrich) dissolved in corn oil (Sigma-Aldrich), at a dose 1 mg for 20 g body weight for 5 consequentdays. Skin samples were collected 12 weeks after tamoxifentreatment.

Wound healing was studied following the experimentalinduction of a skin wound as has been described previously(21, 22). Mice were anesthesized, the dorsal skin was shaved,and a thin 7 to 8 mm long full thickness skin cut was made andallowed to heal naturally. At the indicated early or late timepoints after wounding, mice were euthanized, the woundedskin region excised, fixed, embedded in paraffin, and sectionsstained as indicated.

Cell culture and immunoblottingFor derivation of cell lines, PYMT tumorswere cut into small

pieces and cultured in Dulbecco's Modified Eagle's Medium(DMEM)–10% FBS supplemented with 20 ng/mL EGF (Invi-trogen). To force cells into suspension, cells were washed inPBS, exposed to a 2 mmol/L solution of EDTA in Hank's buffer,and when fully detached cells were span down, resuspended infresh media and cultured in ultra-low cluster plates (Corning)for 2 hours. ULC plates are not permissive to cell adhesion.For heregulin (HO786, Sigma-Aldrich) stimulation cells wereserum- and EGF-starved overnight, and next day incubatedwith 10 ng/mL heregulin in serum-free media for 10 minutes.Cells were incubated with 1 mmol/L FAK inhibitor PF562271(Synmedchem) or 10 mg/mL anti-integrin-b1 antibody Ha2/5(BD Pharmingen) in suspension where indicated.

Total cellular lysates were harvested in modified radioim-munoprecipitation assay (RIPA) buffer [10 mmol/L Na phos-phate (pH 7.2), 150 mmol/L NaCl, 0.1% SDS, 1% NP-40, 1% Nadeoxycholate, protease inhibitors, 1 mmol/L sodium orthova-nadate). Tumor lysates were prepared after tumors werequickly frozen in liquid nitrogen, crushed with a mortar, andlysed in RIPA buffer. For Western blotting, 50 mg of each lysatewas separated by SDS-PAGE, transferred to membrane, andimmunoblotted using appropriate primary and secondary

antibodies and enhanced chemiluminescence visualization.For immunoprecipitation studies, 300 mg of lysate was incu-bated overnight with specific antibodies, immune complexeswere collected by protein G-Sepharose beads and washed, andthe denatured complexes were immunoblotted as describedearlier. Antiphosphotyrosine antibodies (PY99), anti-FAK, anti-HER2, and anti-HER3 antibodies were from Santa Cruz Bio-technology. Anti-phospho-[Y1248]HER2 antibodies antibodieswere from Cell Signaling. Anti-phospho-[T202/Y204]-MAPKantibodies and anti-p-[S473]-Akt antibodies were from CellSignaling.

Sphere formation assayFor sphere formation assay, MCF10A cells were seeded in

ultralow attachment plates (6-well clusters) at 10,000 cells perwell in 5 mL of full-growth medium. Traskþ/þ and Trask�/�

PyMT cell lines were similarly incubated in ultralow attach-ment plates, but in DMEM:F12 medium, containing 10% FBSand 20 ng/mL EGF. The number of spheres was enumeratedafter 6 days for MCF10A cells and after 7 days for PyMT celllines. Spheroids count consisted of approximately 50 cells ormore. The MTT assays were conducted by addition of 1.5 mL 1mg/mL solution of MTT (Sigma-Aldrich) in PBS to 5 mLmedium containing spheres directly into ultralow attachmentwells. Cells were incubated for 2 hours, spun down, andresuspended in 200 mL dimethyl sulfoxide (DMSO). Absor-bance was read in 96-well plates at 570 nm.

Matrigel growth assaysThree-dimensional cultures ofMCF-10A cells were prepared

as described (23). Briefly, 24-well culture plate was first coatedwith growth factor–reducedMatrigel matrix (BD Biosciences),followed by an overlay of MCF-10A cells (12.5 � 103/well) in a5% Matrigel/media suspension. Cells were grown in a 5% CO2

-humidified incubator at 37�C and fedwith fully supplementedMCF-10A media containing 2%Matrigel every 4 days. Cultureswere harvested on day 3, 7, and 12 and stained with antibodiesagainst Trask (1:1,000), phospho-Trask (1:160), and humanlaminin V (1:250, MAB19562X, Millipore). Staining was exam-ined and images were acquired by Zeiss LSM 510 NLO Metaconfocal microscope.

Generation of Trask knockdown cellsShort hairpin RNA (shRNA) sequences were cloned into

pSico-RNeo lentiviral vectors containing a neomycin resis-tance cassette. For the shTrask-1 construct, the oligonucleo-tide 50-TGAATGTTGCTTTCTCGTGGCAGTTCAAGAGACTG-CCA-CGAGAAAGCAACATTTTTTTTGGATCC-30 was annea-led to 50-TCGAGGATCCAAAAAAAATGTTGCTTTCTCGTGG-CAGTCTCTTGAACTGCCACG-AGAAAGCAACATTCA-30 andcloned into HpaI and XhoI sites of the pSicoRNeo vector. Tocreate the shTrask-2 construct, the oligonucleotide 50-TGA-TAGATGAGCGGTTTGCAATGCTGATTCAAG-AGATCAGCA-TTGCAAACCGCTCATCTATTTTTTTTGGCGCGCC-30 wasannealed to 50-TCGAGGCG-CGCCAAAAAAAATAGATGAGC-GGTTTGCAATGCTGATCTCTTGAATCAGCATTGCAAACCG-CTCATCTATCA-30 and cloned into HpaI and XhoI sites of thevector. For generation of the nonsilencing construct, the

Spassov et al.

Cancer Res; 73(3) February 1, 2013 Cancer Research1170



oligonucleotides 50-TGTCTCGCTTGGGCGAGAGTAAGTT-CAAGAGACTTACTCTC-GCCCAAGCGAGATTTTTTTGGCG-CGCC-30 and 50-TCGAGGCGCGCCAAAAAAATCTCGCTTGG-GC-GAGAGTAAGTCTCTTGAACTTACTCTCGCCCAAGCGA-GACA-30 were similarly annealed and cloned. The shRNAconstructs were transfected into 293T cells along with theappropriate packaging vectors. The resulting lentiviral parti-cles were used to infect MCF10A cells. Cells were selected with

G418 (400 mg/mL), and Trask knockdown was confirmed byWestern blotting.

ResultsGeneration of Trask-null mice

A 1.4 kb region including exon 1 and regulatory regions ofTrask/CDCP1 was flanked by loxP sites (Fig. 1A) and thegermline transmission of this floxed allele was confirmed by

Figure 1. Targeting of the murineTrask locus. A, diagrams of thetargeted locus and construct usedfor the generation of Trask-floxedand -null mice. The loxp sites areshown as big black arrowheads thatflank 1.4 kb fragment containingexon 1 (E) and core promotersequences (not shown). Neomycincassette (Neo) is flanked by flippaserecognition target sites (big whitearrows). The SacI (S), SwaI (Sw), ScaI(Sc) and AflII (A) restriction site usedfor Southern blot analysis areindicated. The probes used foranalysis include 50-external, internal(exon1), 30-external, and Neo aredepicted as small black boxes. Thediagnostic SacI and ScaI SwaIfragments are shown withpunctuated lines with doublearrowheads. The removal of the Neowas achieved by crossing with actin-FLP mice, and generation of the nullallele by crossing with EIIa-Cretransgenic mice. The PCR primersused for genotyping of floxed andnull mice are depicted as small grayarrows and numbered 1, 2, and 3. B,Southern blot analysis for germlinetransmission and Neo excision.Genomic DNA from the progeny ofTrask chimeric mice and actin-FLPmice were digested with theindicated restriction enzymes andanalyzed with 50-extrenal, internal,30-external, and Neo probes. DNAfrom wild-type (WT) mouse andtargeted embryonic stem (ES) clonewere used as controls. Mice 1, 3, 4,and 5 are floxedþ/� and used asfounders for subsequent crossing.C,genotyping with primers 2 and 3gives 187-bp product from the wttemplate and 314-bp product fromthe floxed allele. D, PCR analysis oftail DNAwith primers 1, 2, and 3wereused to genotype Traskþ/þ,Traskþ/�, and Trask�/� mice. E, theexpression of Trask protein wasassayedby immunoprecipitation andimmunoblotting of lysates in mouseskin (1 and 3) and intestine (2 and 4) inTraskþ/þ (1 and 2) and Trask�/�

(3 and 4) mice. Only the 140 kD Traskis expressed in these mouse tissues.

A

Sc

Sc A S S Sw

S Sw Sc

32 314bp

Cre

Sc S S Sc

1 3289bp

Mouse #

5′-external

3′-external

Internal

Neomycin

314 bp

WT

Flo

xe

d+/

−

WT

Flo

xe

d+/

−

Trask

187 bp

289 bp

187 bp

IP: TraskIB: Trask

1 2 3 4

Pro

be

Sacl

Fra

gm

en

ts

Sacl

Swal

Scal

Af/ II

1 2 3 4 5 6 7 8 9 1011WT ES

1

SSc

ES

6.7 kb 9.1 kb

FLP

SE

E

E

A Sc

S 1 2 187 bp

9.3 bp21 kb

S SwNEO

NEO

S

S Sc3

321

B

C D E

+/+ +/− +/−+/+





Southern blot analysis of tail genomic DNA (Fig. 1B) and a PCRassay established for rapid genotyping (Fig. 1C). Traskfl/þmiceon a C57/BL6 background were bred with EIIa-cre mice togenerate mice with a null Trask allele. Traskþ/� mice wereinterbred producing Traskþ/þ, Traskþ/�, and Trask�/� prog-eny in the predicted Mendelian ratio, suggesting that Trask isnot essential for mouse development. The generation of het-erozygous and homozygous Trask-null alleles was confirmedby PCR (Fig. 1D) and by Western blotting (Fig. 1E). Traskþ/�

and Trask�/� mice did not exhibit gross morphologic orbehavioral abnormalities compared with wild-type mice andhistologic examination of multiple organ systems found nosignificant pathology and no observed histologic differences.

Tumorigenesis in Trask-null miceMMTV-PyMT mice develop mammary tumors character-

ized by activation of Src kinases (24) and fairly extensivephosphorylation of Trask (18). Traskþ/� mice, backcrossedfor 6 generations in the FVB/N background, were bred withMMTV-PyMT transgenic mice in the FVB/N background andsubsequent matings undertaken to generate sufficient num-bers of female mice carrying the MMTV-PyMT transgene ineither of 3 Trask genotypic backgrounds (Traskþ/þ, Traskþ/�,and Trask�/�). PyMT Trask�/� mice showed significantlyaccelerated tumorigenesis in comparison with PyMTTraskþ/þ mice (Fig. 2A). The PyMT Trask�/� mice had to beeuthanized at 12 weeks of age because of the presence ofnumerous large tumors (Fig. 2B andC). In comparisonwith thePyMT Traskþ/þ of the same age, the 12-week-old PyMTTrask�/� mice had on average 2.6-fold increase in the numberof tumors per mouse (P < 0.001) with an average of 5.7-foldincrease in the total tumor burden (P < 0.001; Fig. 2A). ThePyMT Traskþ/� mice showed a slight, but statistically signif-icant increase in tumorigenesis in comparison with PyMTTraskþ/þ mice. Lung metastases were present in both arms(Supplementary Fig. S1), but a comparative analysis was notpossible as the frequency was low due to the 12-week studytermination compelled by the rapid primary tumor growth.

We examined the expression and phosphorylation of Traskin the normal mammary gland and mammary tumors derivedfrom MMTV-PyMT mice. Trask is not phosphorylated in thenormal mammary gland consistent with the fully anchoredstate of the normal epithelium, but its phosphorylation isinduced in PyMT tumors (Fig. 2D, lane 2), consistent with theanchorage-free state in at least some regions of these tumors.The normal mammary gland expresses predominantly the 140kD uncleaved form of Trask, whereas there is considerablecleavage of Trask and expression of both 140 and 80 kD formsof Trask in the PyMT tumors (Fig. 2D), likely reflectingenhanced protease activity in the tumor microenvironment.Tumor lysates derived fromTrask-nullmice have nodetectableexpression of Trask protein (Fig. 2D, lane 3).

In a second model of tumorigenesis, we studied the devel-opment of hyperplasia and tumors in mouse skin driven by aconditional oncogenic Smoothened allele. In SmoM2cond, K14-CreERT bitransgenic mice, treatment with tamoxifen inducesCre recombinase activity in the epidermis, leading to activationof the conditional SmoM2 allele and the development of

widespread epidermal hyperplasia and neoplasia with com-plete penetrance (20). These neoplastic lesions also haveelevated activity of Src kinases (Supplementary Fig. S2). Tostudy the function of Trask in this type of tumor, we bredTraskþ/� mice (3 generations of backcrossing in the C57/BL6background since gene targeting) with K14-CreERT andSmoM2cond mice in the C57/BL6 background to generateTraskþ/� SmoM2cond K14-CreERT mice and bred them withTraskþ/� mice. From among the progeny, we selectedTraskþ/þ SmoM2cond K14-CreERT and Trask�/� SmoM2cond

Trask+/+ Trask−/−

+/+ +/- -/- +/+ +/- -/-

n = 26 23 21 n = 26

Tu

mo

r w

eig

ht

(g)

Nu

mb

er

of

tum

or

pe

r m

ou

se 6

5

4

3

2

1

0

6

5

4

3

2

1

0

23 21


A

B

C

D

Figure 2. Inactivation of Trask accelerates PyMT-induced tumorigenesis.A, mammary tumorigenesis in mice carrying an MMTV-PyMT transgenewas compared among mice carrying the indicated Trask genotypes.Trask�/� mice have both an increase in the number of detectabletumors and in the weight of the tumors. The difference betweenþ/þ and�/� genotypes is highly significant in both tumor number (P < 0.001) andtumorweight (P<0.001). Thedifferencebetweenþ/þandþ/�genotypesis more modest but also significant in both tumor number (P < 0.005) andtumorweight (P<0.04).B, picturesof representativeage-matched tumor-bearing mice are shown for the Traskþ/þ (left) and Trask�/� (right)genotypes. C, representative ex vivo images of PyMT mammary tumorsfrom Traskþ/þ and Trask�/� mice. Each dish contains the entire tumorburden of a single mouse. Trask-null mice have both an increase in thenumber of tumors and in the size of the tumors. D, lysates from mousetissue were assayed by immunoprecipitation. Lanes correspond tonormal mammary gland from Traskþ/þ mouse (lane 1), mammary tumorfromPyMT, Traskþ/þmouse (lane 2),mammary tumor fromMMTV-PymT,Trask�/�mouse (lane 3). Trask is phosphorylated in PyMT tumors but notin the normal mouse mammary gland. IgG, immunoglobulin G.

Spassov et al.




K14-CreERT mice for tamoxifen treatment. Tamoxifen treat-ment of thesemice resulted in the development of hyperplasticskin disease of the ear and tail as previously reported (20). Thephenotype of the epidermal disease was significantly moresevere in Trask�/� mice compared with Traskþ/þ mice. InTrask�/� mice the skin became much more thick and morewidespread, with gross changes including hair loss and scaling,particularly evident on the abdominal skin (Fig. 3A). Thesemice developed hunched postures and had to be euthanized 12weeks after tamoxifen treatment, significantly earlier thanpreviously reported with this model (20). The Traskþ/þ controlmice were euthanized at the same 12-week post-tamoxifeninduction time point to enable a synchronous examination ofhistologic sections in comparison with the Trask�/� mice.

Tumor burden was quantitatively studied by the microscopicmeasurement of epidermal thickness as previously done (20).The epidermal thickness was on average 3.5-fold higher inTrask�/� SmoM2cond K14-CreERT mice compared with theirTraskþ/þ counterparts (P < 0.0001; Fig. 3B and C; Supplemen-tary Fig. S3), indicating accelerated tumorigenesis in theabsence of Trask function. Control mice from bothTraskþ/þ and Trask�/� genotypes that were not treated withtamoxifen did not develop any gross skin disease or evidence ofskin pathology on histologic sections (Fig. 3B and C).

Trask is not phosphorylated in the normal skin, but it isphosphorylated in the SmoM2-induced tumors (Fig. 3D). Boththe normal skin and the skin tumors express only theuncleaved 140 kD Trask form. We have previously described



+ T

am

oxif

en

-T

am

oxif

en

A

B

*

––

SmoM2cond;K14-CreERT

2 20172=n

TAM TAM


C

D

Ep

iderm

al th

ickn

ess (

μm)

Figure 3. Inactivation of Trask accelerates SmoM2-induced neoplasia. A, conditional SmoM2 mice in the indicated Trask genotypes were treated withtamoxifen to induce the oncogenic SmoM2 allele. Dorsal and abdominal view of representative Traskþ/þ (left) and Trask�/� (right) mice are shown12weeks after tamoxifen treatment. Notemore severe hair loss, scaling, and skin thickening in Trask-nullmice. B, representativeH&E staining of skin samplesfrom K14-CreERT SmoM2 Traskþ/þ and K14-CreERT SmoM2 Trask�/� mice treated with or without tamoxifen. Skin samples were taken 12 weeks aftertamoxifen injection. C, quantification of epidermal thickness (reflecting tumor burden) from histopathology sections showing much thicker tumors in theTrask�/� mice (P < 0.0001). This analysis was conducted on 17 K14-CreERT SmoM2 Traskþ/þ and 20 K14-CreERT SmoM2 Trask�/� mice 12 weeks aftertamoxifen treatment. Control mice were not injected with tamoxifen. D, the expression and phosphorylation of Trask were assayed as indicated frommouseskin by immunoprecipitation. Lanes correspond to normalmouse skin (lane 1), skin from 2Traskþ/þmice bearing SmoM2-induced skin tumors (lanes 2 and 3),skin from a Trask�/� mouse bearing an SmoM2-induced skin tumor (lane 4), IgG immunprecipitate from a tumor-bearing Traskþ/þ mouse skin (lane 5). Themouse skin and SmoM2-induced tumors express only the full-length 140 kD Trask form. Trask is phosphorylated in SmoM2-induced tumors but not in thenormal mouse skin.





that the degree of cleavage of Trask varies considerably amongcell types and its cleavage does not induce its phosphorylation(17–19).

Signaling in Trask-null tumor cellsImmunoblot analysis of tumor lysates from several PyMT-

driven mammary tumors revealed overall higher levels ofmitogen-activated protein kinase (MAPK) signaling in theTrask�/� compared with the Traskþ/þ tumors (Fig. 4A),suggesting enhanced growth factor signaling in Trask�/�

tumors as well as an overall increased level of FAK phos-phorylation, consistent with the derepression of integrinsignaling in the absence of Trask. Receptor tyrosine kinase(RTK) profiling revealed that the predominant active RTK inthese tumors is the HER family, specifically HER2 and to alesser extent platelet-derived growth factor receptor(PDGFR; Fig. 4B) without much difference between thewild-type and Trask-null tumors (data not shown). Primarytumor cell lines were generated from excised mammarytumors obtained from MMTV-PyMT, Traskþ/þ andMMTV-PyMT, Trask�/� mice. The in vitro proliferativegrowth of these tumors is dependent on growth factors.When grown in monolayer, there is no significant differencein the growth rates of the Traskþ/þ or Trask�/� tumor cells(Fig. 5A). However, when grown in nonadhering plates, thereis much higher spheroid formation in the Trask�/� com-pared with the Traskþ/þ tumor cells (Fig. 5B and C). In theTraskþ/þ genotype, these tumor cells form large spheroidswith low efficiency, however, in the Trask�/� genotype,spheroid formation is significantly enhanced with manymore spheroids of all sizes evident. The phenotypic differ-

ence is specifically related to cell growth and not to cell–cellaggregation (Supplementary Fig. S4).

We then studied differences in growth factor receptor signaltransduction in the unanchored state. Our previous work hadshown that Trask, when phosphorylated, inhibits integrinclustering and integrin outside-in signaling (9, 15–17). Becauseintegrin signaling is known to cross-talk with RTK signaling, itis a logical next hypothesis that the phosphorylation of Trask,through its inhibition of integrin signaling, can inhibit integrin-RTK cross-talk, and negatively impact RTK signaling (3–5). Totest this hypothesis, we looked at RTK signaling in theTraskþ/þ and Trask�/� genotypes, and in the presence orabsence of Trask phosphorylation. Trask phosphorylation bySrc kinases is tightly linked with cell anchorage and is instantlyinduced upon loss of anchorage as previously described(11, 13, 15). Trask phosphorylation in PyMT-driven mammarytumor cells is similarly determined by their state of anchorage.In PyMT, Traskþ/þ tumor cells Trask is not phosphorylatedwhen cultured in the adherent state (Fig. 5D, lane 1), but isphosphorylated if cultured in suspension (Fig. 5D, lane 2).Similarly, Trask is phosphorylated in tumor lysates from invivo growing mammary tumors from MMTV-PyMT Traskþ/þ

mice (Fig. 5D, lane 3). Therefore, when cultured in vitro, it is theanchorage-deprived state of these cells that mimicks their invivo state, not theirmonolayer adherent state.When looking atgrowth factor signal transduction, we found that heregulin-driven signaling through MAPK is suppressed in suspendedcells (Fig. 5E, compare lane 2 with 6). Trask is required for thisdetachment-specific suppression of MAPK signaling, as thissuppression is not seen when Trask�/� tumor cells aredeprived of anchorage (Fig. 5E, compare lane 4 with 8). Thereis a similar, although less pronounced suppression of Aktsignaling in Traskþ/þ tumor cells, which is also not seen inTrask�/� tumor cells.

The negative regulation of growth factor signaling does notseem to occur directly through the inhibition of RTK phos-phorylation, as there is no difference in phosphorylation ofHER2 or HER3 between Traskþ/þ and Trask�/� tumor cells(Fig. 6A). The negative effect is seen downstream of RTKs, suchas in MAPK signaling. To interrogate whether this negativeeffect of Trask is specifically mediated through its inhibition ofintegrin signaling, we undertook experiments to manipulateintegrin outside-in signaling both positively and negativelywhile in the unanchored state. To negatively impact integrinsignaling, we used the FAK inhibitor PF562271. As alreadyshown earlier, heregulin-drivenMAPK signaling is enhanced inTrask-null tumor cells (Fig. 6, compare lanes 2 with 4). Thisenhancement is mediated through FAK because it is not seenin cells treated with the FAK inhibitor PF562271 (Fig. 6,compare lanes 6 with 8). For positively impacting integrinsignaling, we used the integrin-b1 antibody Ha2/5. This anti-body interferes with the adhesion properties of integrin-b1, butat the same time activates integrin-b1outside-in signaling (25,26). Cell treated with the integrin-b1 antibody Ha2/5 haveenhancement of heregulin-driven MAPK signaling in Traskþ/þ

tumor cells without any additional enhancement seen inTrask�/� cells (Fig. 6A, compare lanes 10 with 12). Thus, theenhancement of RTK signaling seen in unanchored Trask-null

Figure 4. Integrin and growth factor signaling in PyMT tumors in vivo. A,lysates fromPyMTTraskþ/þandPyMTTrask�/� tumorswere analyzedbyimmunoblotting. Phosphorylation of MAPK (T202/Y204), tyrosinephosphorylation of FAK, and their total protein expressions are shown.Each lane corresponds to a tumor derived from a different mouse. B, theTraskþ/þ tumor lysates were used in the R&D Systems p-RTK profiling.This analysis revealed that the predominant active RTK in these tumors isHER2 (labeled A) and to a lesser extent PDGFR (labeled B).

Spassov et al.




tumor cells is due to loss of the integrin-suppressing functionsof Trask manifested in the unanchored state, as it is recapit-ulated when integrin signaling is experimentally induced (byHa2/5 antibody treatment), and it is overcome when integrinsignaling is experimentally inhibited (by PF562271).The activity of FAK, known to be linked with tumorigenic

growth, is notably affected by the loss of Trask. This is mostevident in the analysis of FAK inactivation during anchoragedeprivation. When PyMT tumor cells are detached, FAK isdephosphorylated as expected (Fig. 6B, lanes 1 and 2). How-ever, this dephosphorylation is impaired in Trask�/� tumorcells leading to inappropriate FAK signaling despite the loss ofanchorage (Fig. 6B, lanes 3 and 4). A similar effect can also beseen in human tumor cells through the shRNA knockdown of

Trask (Fig. 6C). Therefore, the loss of Trask function leads toconstitutive and inappropriate FAK signaling, associated withenhancement of tumorigenic growth (27, 28).

The growth inhibitory functions of p-Trask seem to be ageneral mechanism that manifests during anchorage loss. Todetermine this, Trask was silenced in MCF10A nontrans-formed epithelial cells (Fig. 7A).When cultured in nonadherentplates, MCF10A cells form small spheroids with only very lowefficiency, and mostly remain as single cells, or 2 to 3 cellaggregates. However, the Trask knockdown cells show muchmore efficient spheroid formation with significantly increasednumbers of spheroids, increased size of spheroids, andincreased total cell counts (determined through the analysisof viable cells in an MTT assay) when grown in anchorage-free

Figure 5. Loss of Trask augmentsgrowth factor signaling in PyMTtumor-derived cell lines. A, tumor celllines from MMTV-PyMT mice in theindicated Trask genotypes weregrown in the presence of either EGFor heregulin for 6 days and counted.B, tumor cell lines from PyMT-drivenmouse mammary tumors originatingfrom Traskþ/þ or Trask�/� mice werecultured in nonadherent plates for 7days and the number of spheroidscounted manually undermicroscopic examination. C, phasecontrast microscope images of thespheroids are shown here. D, celllysates were assayed for theexpression and phosphorylation ofTraskby immunoprecipitation. Lanescorrespond to PyMT tumor cell linecultured in the adherent state (lane 1),PyMT tumor cell line cultured in thesuspended state (lane 2), in vivoPyMT tumor from a mouse (lane 3).Trask is phosphorylated duringtumor growth in vivo, resembling thesuspended state when grown inculture. E, serum-starved PyMT celllines from the indicated genotypeswere grown in adherent orsuspended formandwere stimulatedwith heregulin. The growth factor–induced activation of MAPK issuppressed in the suspendedTraskþ/þ cells. This suppressiondoes not occur in Trask�/� cells.There is a more subtle enhancementof Akt signaling in the Trask�/� cells.The arrows indicate the 140 and 80kD forms of Trask. The phospho-immunoblots reflect the T202/Y204phosphorylation of MAPK and theS473 phosphorylationof Akt.





conditions (Fig. 7B and C). MCF10A cells growing in Matrigeltypically grow into spherical forms of a limited size defined by asecreted basement membrane. Trask knockdown cells showenhanced growth with crowding of the central lumen andincreased cell density and cell content of the spheroids, con-sistent with increased proliferative growth in the unanchoredstate (Fig. 7D and E). Anchorage loss was also studied in vivousing an experimental model of skin wounding. In skin wound-ing experiments, there is damage to tissue integrity surround-ing the wound, leading to the phosphorylation of Trask in

epithelial cells at the wound edge (Supplementary Fig. S5),suggesting at least a partial impairment of cell anchorage andintegrin signaling in these cells. When the wound repairprocess is histologically examined in the subsequent days, itseems that the epithelial layer at thewounded edge is thicker inTrask�/�mice compared with Traskþ/þmice (SupplementaryFig. S6). This is consistent with the in vitro experiments above,which show a derepression of growth in anchorage-impairedepithelial cells lacking Trask function.

DiscussionA major attribute of epithelial tumor cells is their ability to

grow in an anchorage-independent manner. However, themechanisms that ordinarily restrict epithelial cell growth tothe anchored state are not well defined. Here, we define thephosphorylation of Trask as one suchmechanism. Although inuntransformed cells Trask is only phosphorylated whenanchorage is lost (13), it is commonly phosphorylated in manyareas of tumor tissues in vivo, consistent with a state of growththat is either fully or partially unanchored or at least defectivein anchorage (18). While the fully or partially unanchored stateof growth has advantages for tumorigenic growth, the phos-phorylation of Trask in this state is a disadvantage, andtherefore the loss of Trask enables tumors to grow muchbetter in an anchorage-independent manner. The fact thatthe negative effects of Trask are much more significant in ourin vivomodels compared with themonolayer growthmodels isconsistent with the fact that its tumor suppressing functionsare manifest predominantly during the unanchored state.Consistent with this, we see a reduction or loss of Traskexpression compared with the normal epithelium in manyhuman cancers, including a survey of breast, colon, andprostate cancers (19, 29). The loss of Trask function in tumorsmay occur through allelic loss, as Trask resides on chromo-some 3 at 3p21.3, an area associatedwith high-frequency allelicloss in human cancers. The location of Trask between markersD3S1029 and D3S1478 place it within an area known to haveloss of heterozygosity in a large fraction of lung cancer in a highresolutionmapping effort (30, 31). Alternatively, the expressionof Trask is lost in some tumors due to methylation silencing(19, 32). In a different mode of functional loss, Trask fails toundergo detachment-induced phosphorylation in some cancercell lines (19). In our analysis of a panel of approximately 50cancer cell lines, the cancer cell lines that expressed Trask butfailed to phosphorylate it were all of the HER2-amplifiedsubtype. This is entirely consistent with our findings thatphosphorylated Trask interferes with integrin-growth factorreceptor cross-talk, and previous studies showing that integrinsignaling is particularly important in HER2-driven tumorigen-esis (6, 7). In fact, HER2-driven tumor cells have been shown tobe particularly dependent on integrin signaling during theirunanchored state (33), creating unique pressure in this tumortype to eliminate the expression or phosphorylation of Trask.

In comparing PyMT tumor cell lines from the Trask wild-type or null background, it is apparent that growth factorreceptor signaling is significantly enhanced in Trask-nulltumors. The enhancement is specifically seen in the

Figure 6. Trask inhibits integrin signaling and disrupts integrin-growthfactor receptor cross-talk. A, PyMT cell lines were serum-starved for 24hours, detached, and incubated in suspended conditions. Cells wereincubated with FAK inhibitor PF562271 or b1-integrin-activatingantibody Ha2/5 for 2 hours, followed by stimulation with herugulin for 10minutes as indicated. Phosphorylation of MAPK (T202/Y204), FAK (alltyrosines), HER2 (Y1248), and HER3 (Y1289) and their total expressionlevels was determined by immunoblotting. B, PyMT tumor cell lines fromTraskþ/þ and Trask�/� mice were assayed while adherent (A) or after 2hours of culture in suspension in nonadhering plates (S). Thephosphorylation state of FAK was assayed by IP/pTyr analysis. FAK isnormally dephosphorylated when anchorage is lost (lane 2). But inTrask�/� cells, there is abnormal persistence of FAK phosphorylationin the unanchored state (lane 4). C, the same analysis was done in MDA-468 breast cancer cells looking at p-FAK in suspension using either of 2Trask shRNA knockdown pools compared with control cells (parental ornonsilencing shRNA transfectant).

Spassov et al.




unanchored state when Trask is phosphorylated and seems tosuppress growth factor receptor signaling. The suppression ofgrowth factor signaling in the unanchored state is mediatedthrough the inhibition of integrin signaling, as this suppressioncan be made redundant if FAK is directly inhibited, and it canbe overcome if integrin-b1 signaling is directly activated.The fact that phosphorylated Trask inhibits integrin signal-

ing is entirely consistent with the fact that it also inhibitsgrowth factor signaling, as integrin signaling is well known tocross-talk with and enhance signaling from growth factorreceptors (3, 4, 34, 35). Consistent with this, knockdown ofb1-integrin diminishes EGF receptor (EGFR) signaling activity(36). How integrins facilitate growth factor receptor signaling isnot yet clear. Integrins can be found in association with severalgrowth factor receptors, including EGFR and HER2 (37, 38).

Integrins can induce ligand-independent activation of theEGFR or can synergistically enhance the growth factor–induced activation of downstream MAPK without affectingreceptor phosphorylation directly (35, 37). How integrinsenhance growth factor receptor signaling is a subject ofcontinuing debate. One proposed mechanism is that coclus-tering with integrins creates increased local receptor density,promoting receptor dimerization and signaling. Another pro-posed mechanism is that colocalization with integrin clustersbrings receptors into a subcellular microenvironment rich insignaling molecules, including various adaptor proteins, orsecond messengers including Src or FAK (reviewed in refs. 3–5). In our study of PyMT-driven tumors, we do not see anenhancement of HER2 or HER3 (or EGFR) phosphorylation inTrask-null tumor cells, but we do see an enhancement in the

Figure 7. Anchorage-independentgrowth in the presence or absence ofTrask. A, MCF10A cells wereengineered to stably express eitherof 2 Trask-targeting shRNAs or anonsilencing shRNA and the loss ofTrask expression in these cells wasconfirmed by immunoblotting.MCF10A/shTrask1 cells have acomplete loss of Trask, whereasMCF10A/shTrask2 cells have apartial loss of Trask expression.B, the indicated cell types werecultured in nonadherent plates for 6days and the number of spheroidscounted manually by microscopicexamination in triplicate wells (left).The total number of cells in each wellwas also compared using the MTTassay (right). C, the appearance ofthe spheroids was captured byphase contrast microscopy at theindicatedmagnifications for the 3 celltypes. D, MCF10A/shControl andMCF10A/shTrask1 cells were grownon Matrigel for the indicated numberof days and the cellular architecture,density, and expression of lamininwere determined by immunofluore-scence staining and confocalmicroscopy. Laminin was stainedusing green-fluorescent reagentsand 40,6-diamidino-2-phenylindole(DAPI) staining was used to stain thecell nuclei. E, additional sectionswere stained with anti-Traskantibodies in green as shown.

MCF10A

MC

F1

0A

MC

F1

0A

/sh

Co

ntr

ol

MCF10A/shControl

MC

F1

0A

/sh

Tra

sk

1

MCF10A/shTrask1

Day 3

Day 7

Day 12

Day 3

Day 7

Day 12

Laminin DAPI Laminin DAPITrask

MCF10A/shControl

MCF10A/shTrask1

MC

F1

0A

/sh

Tra

sk

2

IB: Trask

Nu

mb

er

of

sp

he

res

140

120

100

80

60

40

20

0

Ab

so

rba

nc

e,

57

0 n

m 0.16

0.14

0.12

0.1

0.08

0.06

0.04

0.02

0IB: actin

shControl

MCF10

A

shCon

trol

MCF10A

shTrask1

MCF10

A

shTr

ask1

MCF10A

shTrask2

MCF10

A

shTr

ask2

MCF10

A

shCon

trol

MCF10

A

shTr

ask1

MCF10

A

shTr

ask2

×100 ×200

A B

C

ED





activation of MAPK and to a lesser extent the activation of Akt.This is consistent with an effect of integrin signaling down-stream of receptor phosphorylation in this tumor context. FAKitself has been proposed to mediate integrin-growth factorreceptor cross-talk (39). Our data are also consistent with thisas we see an attenuation of growth factor receptor signalingwhen cells are treated with FAK inhibitors. As seen in thePyMT-driven tumors in vivo and in culture, the loss of Traskleads to constitutive signaling by FAK, uncoupled from thestate of anchorage. This by itself can enhance tumorigenicgrowth, as experimentally induced constitutive FAK signalingenhances tumorigenesis (28, 40).

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

Authors' ContributionsConception and design: D.S. Spassov, C.H. Wong, M.M. MoasserDevelopment of methodology: D.S. Spassov, C.H. Wong, S.Y. Wong, M.M.MoasserAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D.S. Spassov, C.H. Wong, S.Y. Wong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.S. Spassov, C.H. Wong, M.M. MoasserWriting, review, and/or revision of themanuscript:D.S. Spassov, J.F. Reiter,M.M. MoasserAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): M.M. MoasserStudy supervision: J.F. Reiter, M.M. Moasser

AcknowledgmentsThe authors thank the use of the Core Facilities of the Helen Diller Family

Comprehensive Cancer Center, including the Gene Targeting Core, the MousePathology Core, the Laboratory for Cell Analysis Core, and the UCSF SandlerLentiviral Core.

Grant SupportThis work was funded by grant CA113952 (M.M. Moasser) from the NIH. D.S.

Spassovwas supported by a SusanG.Komen for theCure Postdoctoral fellowshipand C.H. Wong was supported by a California Breast Cancer Research ProgramPostdoctoral fellowship. S.Y. Wong was supported by NIH grant 1K99AR059796.

The costs of publication of this article were defrayed in part by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received June 27, 2012; revisedNovember 9, 2012; acceptedNovember 12, 2012;published OnlineFirst December 12, 2012.

References1. Goel HL, Languino LR. Integrin signaling in cancer. Cancer Treat Res

2004;119:15–31.2. Rathinam R, Alahari SK. Important role of integrins in the cancer

biology. Cancer Metastasis Rev 2010;29:223–37.3. Alam N, Goel HL, Zarif MJ, Butterfield JE, Perkins HM, Sansoucy BG,

et al. The integrin-growth factor receptor duet. J Cell Physiol 2007;213:649–53.

4. Streuli CH, Akhtar N. Signal co-operation between integrins and otherreceptor systems. Biochem J 2009;418:491–506.

5. Soung YH, Clifford JL, Chung J. Crosstalk between integrin andreceptor tyrosine kinase signaling in breast carcinoma progression.BMB Rep 2010;43:311–8.

6. GuoW, Pylayeva Y, Pepe A, Yoshioka T, Muller WJ, Inghirami G, et al.Beta 4 integrin amplifies ErbB2 signaling to promotemammary tumor-igenesis. Cell 2006;126:489–502.

7. Huck L, Pontier SM, Zuo DM,Muller WJ. Beta1-integrin is dispensablefor the induction of ErbB2 mammary tumors but plays a critical role inthe metastatic phase of tumor progression. Proc Natl Acad Sci U S A2010;107:15559–64.

8. Scherl-Mostageer M, Sommergruber W, Abseher R, Hauptmann R,Ambros P, Schweifer N. Identification of a novel gene, CDCP1, over-expressed in human colorectal cancer. Oncogene 2001;20:4402–8.

9. Bhatt AS, Erdjument-Bromage H, Tempst P, Craik CS, Moasser MM.Adhesion signaling by a novel mitotic substrate of src kinases. Onco-gene 2005;24:5333–43.

10. Hooper JD, Zijlstra A, Aimes RT, Liang H, Claassen GF, Tarin D, et al.Subtractive immunization using highly metastatic human tumor cellsidentifies SIMA135/CDCP1, a 135 kDa cell surface phosphorylatedglycoprotein antigen. Oncogene 2003;22:1783–94.

11. Brown TA, Yang TM, Zaitsevskaia T, Xia Y, Dunn CA, Sigle RO, et al.Adhesion or plasmin regulates tyrosine phosphorylation of a novelmembrane glycoprotein p80/gp140/CUBdomain-containing protein 1in epithelia. J Biol Chem 2004;279:14772–83.

12. He Y, Wortmann A, Burke LJ, Reid JC, Adams MN, Abdul-Jabbar I,et al. Proteolysis induced amino terminal ectodomain shedding of theintegral membrane glycoprotein CUB domain containing protein 1(CDCP1) is accompanied by tyrosine phosphorylation of its carboxyterminal domain and recruitment of Src and PKC{delta}. J Biol Chem2010;285:26162–73.

13. SpassovDS,Baehner FL,WongCH,McDonoughS,MoasserMM. Thetransmembrane src substrate Trask is an epithelial protein that signalsduring anchorage deprivation. Am J Pathol 2009;174:1756–65.

14. Conze T, Lammers R, Kuci S, Scherl-Mostageer M, Schweifer N, KanzL, et al. CDCP1 is a novel marker for hematopoietic stem cells. AnnNYAcad Sci 2003;996:222–6.

15. Spassov DS, Wong CH, Moasser MM. Trask phosphorylation definesthe reverse mode of a phosphotyrosine signaling switch that underliescell anchorage state. Cell Cycle 2011;10:1225–32.

16. SpassovDS,WongCH, SerginaN, Ahuja D, FriedM, Sheppard D, et al.Phosphorylation of Trask bySrc kinases inhibits integrin clustering andfunctions in exclusion with focal adhesion signaling. Mol Cell Biol2011;31:766–82.

17. Spassov DS, Ahuja D,WongCH,Moasser MM. The structural featuresof Trask that mediate its anti-adhesive functions. PLoS ONE 2011;6:e19154.

18. Wong CH, Baehner FL, Spassov DS, Ahuja D, Wang D, Hann B, et al.Phosphorylation of the SRC epithelial substrate Trask is tightly reg-ulated in normal epithelia but widespread in many human epithelialcancers. Clin Cancer Res 2009;15:2311–22.

19. Spassov DS, Wong CH, Harris G, McDonough S, Phojanakong P,Wang D, et al. A tumor-suppressing function in the epithelial adhesionprotein Trask. Oncogene. 2012;31:419–31.

20. Wong SY, Seol AD, So P-L, Ermilov AN, Bichakjian CK, Epstein EH,et al. Primary cilia can bothmediate and suppressHedgehog pathway-dependent tumorigenesis. Nat Med 2009;15:1055–61.

21. Chmielowiec J, Borowiak M, Morkel M, Stradal T, Munz B, Werner S,et al. c-Met is essential for wound healing in the skin. J Cell Biol2007;177:151–62.

22. Romer J, Bugge TH, Pyke C, Lund LR, Flick MJ, Degen JL, et al.Impaired wound healing in mice with a disrupted plasminogen gene.Nat Med 1996;2:287–92.

23. Debnath J, Muthuswamy SK, Brugge JS. Morphogenesis andoncogenesis of MCF-10A mammary epithelial acini grown inthree-dimensional basement membrane cultures. Methods2003;30:256–68.

24. Guy CT, Cardiff RD, Muller WJ. Induction of mammary tumors byexpression of polyomavirus middle T oncogene: a transgenic mousemodel for metastatic disease. Mol Cell Biol 1992;12:954–61.

25. Armulik A, Velling T, Johansson S. The integrin beta1 subunit trans-membrane domain regulates phosphatidylinositol 3-kinase-depen-dent tyrosine phosphorylation of Crk-associated substrate. Mol BiolCell 2004;15:2558–67.

26. Wennerberg K, Fassler R, Warmegard B, Johansson S. Mutationalanalysis of the potential phosphorylation sites in the cytoplasmic

Spassov et al.




domain of integrin beta1A. Requirement for threonines 788–789 inreceptor activation. J Cell Sci 1998;111 (Pt 8):1117–26.

27. LahlouH,Sanguin-GendreauV, ZuoD,Cardiff RD,McLeanGW,FrameMC, et al. Mammary epithelial-specific disruption of the focal adhesionkinase blocks mammary tumor progression. Proc Natl Acad Sci U S A2007;104:20302–7.

28. Pylayeva Y, Gillen KM, Gerald W, Beggs HE, Reichardt LF, GiancottiFG. Ras- and PI3K-dependent breast tumorigenesis in mice andhumans requires focal adhesion kinase signaling. J Clin Invest2009;119:252–66.

29. Siva AC, Wild MA, Kirkland RE, Nolan MJ, Lin B, Maruyama T, et al.Targeting CUB domain-containing protein 1 with a monoclonal anti-body inhibits metastasis in a prostate cancer model. Cancer Res2008;68:3759–66.

30. Ji L, Minna JD, Roth JA. 3p21.3 tumor suppressor cluster: prospectsfor translational applications. Future Oncol 2005;1:79–92.

31. Wistuba II, Behrens C, Virmani AK, Mele G,Milchgrub S, Girard L, et al.High resolution chromosome3pallelotyping of human lung cancer andpreneoplastic/preinvasive bronchial epithelium reveals multiple, dis-continuous sites of 3p allele loss and three regions of frequent break-points. Cancer Res 2000;60:1949–60.

32. Ikeda JI, Morii E, Kimura H, Tomita Y, Takakuwa T, Hasegawa JI, et al.Epigenetic regulation of the expression of the novel stem cell markerCDCP1 in cancer cells. J Pathol 2006;210:75–84.

33. Jenndahl LE, Taylor-Papadimitriou J, Baeckstrom D. Characterizationof integrin and anchorage dependence in mammary epithelial cells

followingc-erbB2-inducedepithelial-mesenchymal transition. TumourBiol 2006;27:50–8.

34. Aplin AE, Juliano RL. Integrin and cytoskeletal regulation of growthfactor signaling to the MAP kinase pathway. J Cell Sci 1999;112 (Pt5):695–706.

35. Miyamoto S, Teramoto H, Gutkind JS, Yamada KM. Integrins cancollaborate with growth factors for phosphorylation of receptor tyro-sine kinases and MAP kinase activation: roles of integrin aggregationand occupancy of receptors. J Cell Biol 1996;135:1633–42.

36. Morello V, Cabodi S, Sigismund S, Camacho-Leal MP, Repetto D,Volante M, et al. [beta]1 integrin controls EGFR signaling andtumorigenic properties of lung cancer cells. Oncogene 2011;30:4087–96.

37. Moro L, Dolce L, Cabodi S, Bergatto E, Boeri Erba E, Smeriglio M, et al.Integrin-induced epidermal growth factor (EGF) receptor activationrequires c-Src and p130Cas and leads to phosphorylation of specificEGF receptor tyrosines. J Biol Chem 2002;277:9405–14.

38. Falcioni R, Antonini A, Nistico P, Di Stefano S, Crescenzi M, Natali PG,et al. Alpha 6 beta 4 and alpha 6 beta 1 integrins associate with ErbB-2in human carcinoma cell lines. Exp Cell Res 1997;236:76–85.

39. Renshaw MW, Price LS, Schwartz MA. Focal adhesion kinase med-iates the integrin signaling requirement for growth factor activation ofMAP kinase. J Cell Biol 1999;147:611–8.

40. Frisch SM, Vuori K, Ruoslahti E, Chan-Hui PY. Control of adhesion-dependent cell survival by focal adhesion kinase. J Cell Biol1996;134:793–9.





trask loss enhances tumorigenic growth by liberating ... · for heregulin (ho786, sigma-aldrich)...

Documents