pdlim2 is a marker of adhesion and b-catenin activity in ... · orla t. cox1, shelley j. edmunds1,...

Tumor Biology and Immunology

PDLIM2 Is a Marker of Adhesion and b-CateninActivity in Triple-Negative Breast CancerOrla T. Cox1, Shelley J. Edmunds1, Katja Simon-Keller2, Bo Li3, Bruce Moran3,NiamhE.Buckley4,MilanBustamante-Garrido1,NollaigHealy1,CiaraH.O'Flanagan1,William M. Gallagher3, Richard D. Kennedy5, Ren�e Bernards6, Carlos Caldas7,Suet-Feung Chin7, Alexander Marx2, and Rosemary O'Connor1

Abstract

The PDLIM2 protein regulates stability of tran-scription factors including NF-kB and STATs inepithelial and hemopoietic cells. PDLIM2 is strong-ly expressed in certain cancer cell lines that exhibitan epithelial-to-mesenchymal phenotype, and itssuppression is sufficient to reverse this phenotype.PDLIM2 supports the epithelial polarity of non-transformed breast cells, suggesting distinct roles intumor suppression and oncogenesis. To betterunderstand its overall function, we investigatedPDLIM2 expression and activity in breast cancer.PDLIM2 protein was present in 60% of tumorsdiagnosed as triple-negative breast cancer (TNBC),and only 20%of other breast cancer subtypes. HighPDLIM2 expression in TNBC was positively corre-lated with adhesion signaling and b-catenin activ-ity. Interestingly, PDLIM2 was restricted to thecytoplasm/membrane of TNBC cells and excludedfrom the nucleus. In breast cell lines, PDLIM2retention in the cytoplasm was controlled by celladhesion, and translocation to the nucleus wasstimulated by insulin-like growth factor-1 or TGFb.Cytoplasmic PDLIM2 was associated with activeb-catenin and ectopic expression of PDLIM2 wassufficient to increase b-catenin levels and its tran-scriptional activity in reporter assays. Suppressionof PDLIM2 inhibited tumor growth in vivo, whereasoverexpression of PDLIM2 disrupted growth in 3Dcultures. These results suggest that PDLIM2 mayserve as a predictive biomarker for a large subset ofTNBC whose phenotype depends on adhesion-reg-ulated b-catenin activity and which may be ame-nable to therapies that target these pathways.

Significance: This study shows that PDLIM2 expression defines a subset of triple-negative breast cancer thatmay benefit fromtargeting the b-catenin and adhesion signaling pathways.

Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/79/10/2619/F1.large.jpg.

β-catenin

β-catenin

β-cateninP

P

Membrane

Cytoplasm

PDLIM2

Nucleus

Outcome:Maintenance of tumorgrowth and migration

Loss of tumor growthand migration

β1-integrin

ECM

E-Cadherin

P

In TNBC, cytoplasmic PDLIM2 is associated with adhesion and growthfactor signaling (GFR), leading to β-catenin release from the membrane,phosphorylation, and increased activity.

PDLIM2suppressed

GFR

1Cell Biology Laboratory, School of Biochemistry and Cell Biology, UniversityCollege Cork, Cork, Ireland. 2Institute of Pathology, University Medical CentreMannheim, Heidelberg University, Germany. 3School of Biomolecular &Biomedical Science, Conway Institute, University College Dublin, Dublin,Ireland. 4School of Pharmacy, Queens University Belfast, Belfast, Northern

Ireland. 5Centre for Cancer Research and Cell Biology, Queens UniversityBelfast, Northern Ireland. 6Division of Molecular Carcinogenesis and CancerGenomics Netherlands, The Netherlands Cancer Institute, Amsterdam, TheNetherlands. 7Cancer Research UK Cambridge Research Institute, Li Ka ShingCentre, Cambridge, UK.

CancerResearch

www.aacrjournals.org 2619

on October 3, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 18, 2019; DOI: 10.1158/0008-5472.CAN-18-2787

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-18-2787&domain=pdf&date_stamp=2019-5-6

http://cancerres.aacrjournals.org/content/canres/79/10/2619/F1.large.jpg

http://cancerres.aacrjournals.org/

IntroductionTriple-negative breast cancers (TNBC) lack estrogen receptor,

progesterone receptor (ER�, PR�), and amplification of humanepidermal growth factor receptor 2 (HER2�), and constitute15% to 20% of all breast cancers. TNBC displays considerablegenetic heterogeneity, a high risk for distant metastasis, isrefractory to many therapies, and often has a poor progno-sis (1–5). Although many studies have helped define subtypeswithin TNBC, the underlying drivers of this cancer are stillunclear, and there is an urgent need for better biomarkers andtherapeutic targets (5–7).

Growth factor and adhesion signaling have been stronglyimplicated in TNBC (8–11). Our studies on insulin-likegrowth factor-1 (IGF1) signaling in cancer led us to identifythe PDLIM2 (Mystique) protein (12), as a feedback regulatorof IGF1 and adhesion signaling (12–14). PDLIM2 is a memberof the PDZ-LIM domain family encoded by a locus on chro-mosome 8p21 (12, 15, 16), a region that is associated withmetastasis and often disrupted in cancer (17). Located at thecytoskeleton and nucleus of epithelial cells and hemopoieticcells, PDLIM2 regulates the stability and activity of importanttranscription factors including STATs, NF-kB, and IRFs (12,18–21). PDLIM2 can suppress cellular transformation (12, 13,16), and it may be repressed by methylation in breast celllines (22, 23). PDLIM2 inactivation has also been reported inclassical Hodgkin and anaplastic large cell lymphoma (24).However, PDLIM2 is robustly expressed in breast and prostatecancer cells that exhibit a migratory phenotype (13, 16), andsustains a transcription programme for epithelial to mesen-chymal transition (EMT; ref. 21). In line with this, PDLIM2expression has been linked to highly aggressive ovarian can-cer (25), lymph node metastasis in low-grade breast can-cers (26), and was identified as a moderate dependency genein basal breast cancers (27). It has also been linked to thegrowth of neurofibromatosis 2 (NF2)-associated menginiomaand schwannoma (28).

In the nontumorigenic MCF10A breast epithelial cell line,PDLIM2 is required for maintaining cell polarization and 3Dacinar formation (14), and PDLIM2 expression increases uponretinoic-acid induced differentiation of breast cancer cells (23).However, although PDLIM2 expression has been linked withtumor suppression and with sustaining a migratory phenotype,it is still unclear what it contributes to the progression or phe-notype of any subgroup of breast cancer. Here, we investigatedthis by assessing PDLIM2 expression in different cohorts of breasttumors and manipulating its expression in cell lines. We foundthat PDLIM2 is expressed in a subset of TNBCs (50%–60%), andin approximately 20% of other breast cancer subtypes. In TNtumors and cell lines, cytoplasmic PDLIM2 promotes the accu-mulation of active b-catenin. Our study suggests that PDLIM2mediates adhesion and growth factor signals derived from thetumor microenvironment to enhance b-catenin activation in alarge proportion of TNBCs.

Materials and MethodsCell culture and assays

Cell lines were purchased from ATCC, except CAL-51 (DSMZ)and authentication established by PCR-single-locus-technology(Eurofins Medigenomix, Forensik GmbH, Ebersberg, Germany),up to November 2018. MDA-MB-231-LUC2 cells were purchasedfor in vivo experiments (Caliper Life Sciences). All cells were testedmonthly for mycoplasma by specific DNA staining and main-tained as previously described (14, 29, 30), up to 6 to 8 weeks foruse in experiments. Further details canbe found inSupplementaryMaterials and Methods: Tables 1, 3.

For serum starvation and growth factor stimulation, cells wereincubated in serum-freemedium for 4 hours prior to stimulation,or not, with 10 ng/mL TGFb1 or IGF1. For non-adherent condi-tions, cells were trypsinized, centrifuged, and washed prior to re-suspension in complete medium into 50 mL tubes. Cells wereincubated for 4 hours at 37�C, with gentle rotation. Alternatively,cells were seeded on petri dishes coated with 100 mL/cm2 ofPolyhema (6 mg/mL; Sigma).

For wound healing assays, confluent monolayer cultures werewounded by scoring with a sterile pipette tip, and imaged at 0 and24 hours post-wounding. Wound widths were measured asdescribed in Supplementary Materials and Methods.

Colony formation was measured by assessing plating efficien-cy. MDA-MB-231-LUC2 cells were seeded at 5 � 102 cells/wellin six-well plates, (minimum triplicates per clone). After 9 to10 days, cells were stained with 0.01% crystal violet and coloniesof more than 50 cells were counted.

For 3D on Top assays, cells were cultured on matrigel withmatrigel-supplemented medium, as described by Kenny andcolleagues (31) and monitored over 4 days. Entire 3D cultureswere extracted for Western blot analysis of protein expression asdescribed (31).

In vivo bioluminescence imaging of tumor spreadTwo separate clones of MDA-MB-231-LUC2 cells stably

expressing shRNA scramble control or shRNA targetingPDLIM2 (1 � 106 cells) were injected into the tail vein of 4-to 5-week-old female HsdOla:MF1-Foxn1nu mice (Harlan; n ¼6–7 per group, randomized; three separate cohorts). The Bio-luminescent signal (luminescence counts) per animal per cellclone and images were collected after 45 to 49 days by using theIVIS Spectrum in vivo imaging system (PerkinElmer). Theseexperiments were performed under licence to R. O'Connorfrom the Irish Department of Health and protocol approvalby University College Cork Animal Experimentation EthicsCommittee (#2012/008). Further details are in SupplementaryMaterials and Methods.

Cell lysates, subcellular fractionations, Western blotting, anddensitometry

Total cell protein was extracted using RIPA buffer and sub-cellular fractionations were performed as described previously

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

Corresponding Author: Rosemary O'Connor, Cell Biology Laboratory, Schoolof Biochemistry and Cell Biology, University College Cork, BioSciences

Institute, College Road, Cork, Ireland. Phone: 353-21-490-1312; Fax: 353-21-490-1382; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-18-2787

�2019 American Association for Cancer Research.

Cox et al.

Cancer Res; 79(10) May 15, 2019 Cancer Research2620




(20, 32). Western blot analyses were performed using theOdyssey Image scanner system and protein expression levelswere normalized to protein loading control by densitometryusing Licor Image Studio software as previously described(14, 20, 21, 32).

Breast cancer tissue sample collections and IHCAll tissues were stained using a PDLIM2 mouse monoclonal

antibody (Abcam). Details of IHC staining procedures used foreach sample collection are described in supplementary methods.

Northern Ireland Biobank cohort. Breast Cancer Tissue Micro-arrays were generated from Formalin-fixed paraffin-embeddedprimary tumor blocks as previously described (33). Eachtumor was represented by three independent cores. BreastCancer subtypes were determined from biomarker expression,and classified according to St Gallen International ExpertConsensus (34) as: luminal A (ER and/or PR positive, HER2negative); luminal B [ER and/or PR positive and HER2 neg-ative (HER2�) or HER2 overexpressed/amplified (HER2þ)];HER2 enriched (nonluminal, ER and PR negative and HER2overexpressed/amplified); basal-like/triple negative (ER, PR,and HER2�).

RATHER TNBC cohorts. Slides from the Netherlands CancerInstitute (NKI) and Cambridge University (CAM) containedtissue cores from individual patients within the cohorts (35), andwe analyzed IHC staining of tumor samples from 128 patientswith TNBC. Further details on RATHER cohorts analyses are inSupplementary Materials and Methods.

Heidelberg samples. Breast Cancer Tissue sections were preparedfrom TNBC biopsies at the University Medical Centre Mannheimas described previously (36).

Further details are in Supplementary Materials and Methods.Written informed consent and ethical approval was obtained ateach Institution (33, 35, 36).

RNA extraction, quantitative PCR, DNA, shRNA, and siRNAtransfections

RNA extraction, cDNA synthesis, and quantitative PCRs wereperformed as previously described (21), with details and primersequences in Supplementary Materials and Methods and Supple-mentary Table S2.

MDA-MB-231-LUC2 cells were transfected with pSUPERvectors encoding shRNAs targeting PDLIM2 (shPDLIM2) orcontrol shRNA (shScramble), and BT549 cells with pcDNA3-HA-Empty Vector (EV) or HA-PDLIM2 using Lipofectamine2000. Stable clones were generated by selection in G418(12, 21). HEK293T cells were transiently transfected withHA-EV or HA-PDLIM2 using calcium phosphate protocol (32).HCC1806 were transiently transfected with siRNA negativecontrol and 2 different siRNAs targeting PDLIM2 using oligo-fectamine (12). Further details are in Supplementary Materialsand Methods.

b-Catenin/Tcf transcriptional reporter activityHEK293T cells grown on 96-well OptiPlate microplates

(PerkinElmer) were transfectedwith the TOPflash plasmid, whichcontains three copies of the Tcf/Lef sites upstream of a thymidinekinase (TK) promoter and the firefly luciferase gene, and the

FOPflash, which contains mutated copies of Tcf/Lef sites usingLipofectamine 2000. Cells were cotransfected with 0.2 mg ofinternal control reporter Renilla reniformis luciferase construct(pRL-TK; Promega) to normalize transfection efficiency, andluciferase activity was measured using a dual luciferase AssaySystem Kit (Promega).

Immunofluorescence stainingCells were fixed with 4% paraformaldehyde, probed with

primary antibodies followed by Alexa 488- or Cy3-conjugatedsecondary antibodies and/or TRITC-phalloidin. Nuclei werevisualized with Hoechst dye and imaged as described previously(12, 14), and in Supplementary Materials and Methods.

RATHER cohorts RPPA, gene expression, and subtypesurvival analysis

RNA was purified, amplified, labeled, and hybridized to theAgendia custom-designed whole genome microarrays (AgilentTechnologies) and raw fluorescence intensities quantified, aspreviously described (35). GEO accessions are GSE66647 (RPPA)and GSE68057 (RNA microarray expression). Further details ofanalyses of the RATHER cohorts are in Supplementary Materialsand Methods.

Kaplan–Meier survival analysis was performed using the SPSSstatistics software (IBM). Univariate analysis was performed usingthe Cox regression model to illustrate the relationship betweenPDLIM2 protein expression and breast cancer-specific survival(BCSS), recurrence-free survival (RFS), and distant recurrence-freesurvival (DRFS), respectively. Hazard ratios (HR) and 95% con-fidence intervals (95% CI) were evaluated for each survivaloutcome using the Cox regression model.

Statistical analysisGraphs were prepared using GraphPad Prism. Data were ana-

lyzed for statistical significance using Fisher exact test, one-wayANOVAor Student t test as indicated in legends. AP value of <0.05was deemed significant and graded P values are denoted asfollows: �, P � 0.05; ��, P � 0.005; ��, P � 0.0005, unlessotherwise noted.

ResultsPDLIM2 expression is enriched in TNBC

To gain insight into how PDLIM2 contributes to human breastcancer, we first assessed its expression in tissuemicroarrays (TMA)using IHC scored as negative (0), low (1), moderate (2), or high(3). Images representing the relative intensities of these IHCstaining scores are shown in Supplementary Fig. S1A. In a TMArepresenting 248 breast tumors (Northern Ireland Biobank; NIB),PDLIM2 was expressed in 25.8% of all tumors (Fig. 1A). Byseparating these tumors using the Gallen scores of tumor classi-fication (34), it was clear that the majority of Luminal A (HER2-negative), Luminal B-HER2-negative or -positive and HER2-enriched tumors did not express PDLIM2 (Fig. 1B and C). How-ever, within the basal/TNBC samples more than half (53%) oftumors expressed PDLIM2 (Fig. 1B and C). Interestingly, PDLIM2expression in tumor-associated stroma or tumor-infiltrating leu-kocytes was evident in 30% to 40% of all non-TNBC types in theTMA, and again, expression was proportionately higher in TNBC(61.8%; Fig. 1D and E; Supplementary Fig. S1B and S1C).Overall,70% of PDLIM2-positive tumors (all subtypes) had PDLIM2-

PDLIM2 Enhances b-Catenin Activity in TNBC

www.aacrjournals.org Cancer Res; 79(10) May 15, 2019 2621




A

D

PDLIM2 expression in breast tumor cells

0

20

40

60

80

% o

f tum

ors

n = 64

n = 184

NegativePositive

E

B

% o

f tum

ors

PDLIM2 expression/tumor classificationNegative Positive

0

50

100

Luminal A (67)

Luminal B HER2- (55)

Luminal B HER2+ (26)

HER2 enriched (32)

Basal/triple negative (68)

****** ** *###

C PDLIM2 expression inbreast tumor stroma

0

20

40

60

NegativePositive

% o

f tum

ors

n = 109

n = 139

Luminal A

Luminal B HER2−

Luminal B HER2+

HER2 enriched

Basal/triple negative

Num

ber o

f cor

es

0

20

40

60

800 321IHC score:

PDLIM2 scores/tumor classification

***

***

***

###

PDLIM2 stroma expression/tumor classification

0

50

100

Negative Positive

Luminal A

Luminal B HER2−

Luminal B HER2+

HER2 enriched

Basal/triple negative

% o

f tum

ors

****###

F

100 μm

i ii

iii iv

20 μm

T S

T

S

T S

T

S

Figure 1.

PDLIM2 expression is enriched in TNBC. A, A breast cancer TMA of 248 samples from the Northern Ireland Biobank was stained for PDLIM2 expression byimmunohistochemistry. PDLIM2 expression was scored as negative (0), low (1), moderate (2), or high (3). Percentages of total negative or positive samples fromthe entire cohort are shown. B and C, The TMA samples were classified according to breast cancer subtype and percentage of PDLIM2 negative versus positive(B) and PDLIM2 expression scores (C) within each tumor type was assessed. Sample numbers for each breast cancer subtype are shown in brackets in B. D and E,The percentage of samples with PDLIM2 negative and positive expression in tumor stroma/infiltrating cells was assessed and quantified across the cohort (D)and within each breast cancer subtype (E). F, i–iv, Representative micrographs of PDLIM2 IHC staining in TNBCs showing tumor cell and stromal cell negativestaining (i); tumor positive/stroma negative (ii); tumor negative/stroma positive (iii), and tumor positive/stroma positive (iv). Insets show higher magnificationimages demonstrating PDLIM2 staining in tumor (T) or stroma (S) cells. �� , P < 0.0001; �� , P < 0.005; � , P < 0.05, Fisher exact test comparing PDLIM2 positiveand negative tumors in TNBC versus each non-TNBC subtype; ###, P < 0.0001 denotes TNBC versus all non-TNBC subtypes in B, C, and E. In C, only IHC score 1tumors were statistically significant.

Cox et al.





positive stroma, whereas within PDLIM2-negative tumors, 50%of TNBC and 30% of non-TNBC had PDLIM2-positive stroma(Supplementary Fig. S1C). PDLIM2 staining in TNBC and non-TNBC TMAs are shown in Fig. 1F and Supplementary Fig. S1D,respectively.

In summary, PDLIM2 is expressed in more than 50% of TNBCtumor cells and 60% have PDLIM2-positive stroma, whereas inother breast cancer subtypes, it is expressed in approximately 20%of tumors and 40% of stroma.

The PDLIM2 protein is widely expressed in TNBCPDLIM2associationwith TNBCwas further tested in additional

TNBC cohorts. These were from the RATHER consortium (www.ratherproject.com), which includes cohorts from the NetherlandsCancer Institute (NKI, n ¼ 71) and the University of Cambridge,UK (Cam; n¼ 57). We also tested a smaller cohort (n¼ 15) fromHeidelberg University. As can be seen in Fig. 2A, PDLIM2 wasexpressed in 53%of TNBC from theNIB cohort (data fromFig. 1),72% of the NKI cohort, 47% of the CAM cohort and 70% of the

PDLIM2 expression: all cohorts

% o

f tum

ors,

n =

211

Negative Positive

C*P = 0.0268

D

B Negative (0) Low (1)Moderate-High (2-3)

NIB

CA

MN

KI

Hei

delb

erg

MES BLIA LAR BLIS

1.5

1

0.5

0

-0.5

-1

-1.5

E

TNBC subtype

0

20

40

60

80 PDLI

M2

expr

essi

on%

of t

umor

s, n

= 1

15

Negative Positive

Negative Positive

PDLIM2 expression: subtype (RATHER)

TNBC subtypeMES BLIA LAR BLIS

0

20

40

60

80

100

n = 79 n = 9 n = 14 n =1 3

TNBC subtype of tumors in RATHER cohortsA

0

20

40

60

80

100

% o

f tum

ors

PDLIM2 expression in TNBC tumors

NIB (68)

RATHER NKI (71)

RATHER CAM (57)

Heidelberg (15)

Figure 2.

PDLIM2 is expressed in a subset ofTNBCs. A, PDLIM2 expression wasexamined and quantified in threecohorts of TNBC samples; RATHERNetherlands Cancer Institute (NKI)and Cambridge University (CAM)TMAs and Heidelberg tissues(Heidelberg). The graph representsthe percentage of PDLIM2-positiveand -negative tumors within each ofthese three cohorts plus the NorthernIreland Biobank (NIB) cohortfrom Fig. 1. The number of tumors ineach cohort is shown in parentheses.B, Representative micrograph imagesillustrating PDLIM2-negative andPDLIM2-positive tumor stainingwithin each cohort, showing negative(0), low (1), or moderate-high (2 or 3)staining intensities. Scale bars, 50 mm(NIB, Heidelberg) or 25 mm (NKI,CAM). C,Graph shows total PDLIM2-negative and PDLIM2-positive TNBCsacross all four cohorts. P-valuedetermined using the Student t test.D, The Heatmap depicts TNBC tumorsubtypes from the RATHER NKI andCAM cohorts, determined from an80-gene signature according toBurstein and colleagues (6), asdescribed in the SupplementaryMaterials and Methods. E, Graphshowing the percentage ofPDLIM2-positive and -negativetumors within each TNBC subtype inthe RATHER cohorts, determinedfrom 80-gene signature as in D. NKI,n¼ 70; CAM, n¼ 45. In D and E,TNBC subtypes described by Bursteinand colleagues (6) are denoted as:MES, mesenchymal; BLIA, basal-likeimmune-activated; BLIS, basal-likeimmunosuppressed; LAR, luminalandrogen receptor.





www.ratherproject.com

www.ratherproject.com


Heidelberg cohort (Fig. 2A and B). The distribution of weightedIHC scores for PDLIM2 staining is shown in Supplementary Fig.S2A. Overall, PDLIM2 was expressed in 60% of TNBCs across thefour cohorts (Fig. 2C), and was low or absent in neighboringnormal tissue, as shown in Supplementary Fig. S2B (Heidelbergcohort). Interestingly, PDLIM2 is present in immune cells (mostlylymphocytes) surrounding epithelial ducts in this normal tissue(Supplementary Fig. S2B).

We next asked whether PDLIM2 expression could be correlatedwith survival and/or clinicopathological features. No significantcorrelation was found with PDLIM2 expression and overall sur-vival in patients with TNBC from the NIB cohorts, and in theRATHER cohorts, PDLIM2 was not significantly correlated withoutcome (breast cancer specific-, recurrence free-, or distant recur-rence-free survival; Supplementary Fig. S2C). Similar trends wereobserved when data were segregated into weighted scores, andthere was no significant association with histological grade,number of positive lymph nodes, or tumor size (SupplementaryFig. S2C).

To test whether PDLIM2 is associated with specific TNBCsubtypes, we applied the 80 gene signature from Burstein andcolleagues (6) to the RATHER cohorts (Fig. 2D). The majority oftumors were classified as mesenchymal (MES), but there was nosignificant correlation of PDLIM2 with any subtype (Fig. 2E;Supplementary Fig. S3A and S3B). However, trends indicatedhigher PDLIM2 expression in the Basal-like Immune Activated(BLIA) and Suppressed (BLIS) subtypes (the latter of which isassociated with poorest outcomes; ref. 6), compared with MESor luminal androgen receptor (LAR) subtypes (SupplementaryFig. S3B and S3C). Interestingly, in these analyses, we noted thatPDLIM2 mRNA levels were not strongly correlated with PDLIM2protein expression (Supplementary Fig. S3D).

Taken together, the data demonstrate that PDLIM2 is expressedin approximately 60%of tumor cells derived from four cohorts ofTNBC, but is not definitively linked to TNBC subtypes or clinicaloutcomes.

PDLIM2 is restricted to the cytoplasm in TNBC tumorsBecause PDLIM2 subcellular location has previously been

associated with integrating adhesion signals with transcriptionfactor stability/activity (12, 20, 21), we next analyzed its loca-tion in TNBC tissue. It was clear that PDLIM2 is predominantlyexpressed at the cytoplasm/membrane and excluded from thenucleus in tumor cells (Fig. 3A). By quantifying PDLIM2staining at the cytoplasm/membrane or also in the nucleus(nuclear-only expression was not observed in tumor cells), itwas evident that PDLIM2 was restricted to cytoplasmic/mem-branous areas (Fig. 3B and C). This is in contrast to high levelsof nuclear PDLIM2 in stromal cells (Fig. 1F insets; Supplemen-tary Figs. S2B and S3E).

To determine functional significance for cytoplasmic PDLIM2in TNBC, we used available reverse phase protein array (RPPA)and RNA profiling data for the RATHER cohorts, and askedwhether PDLIM2 expression was associated with signaling path-ways, or with transcription factor activity. Included in theseanalyses were 63 proteins and 139 genes with known or antici-pated links to PDLIM2 function (Supplementary Tables SI and SII;refs. 12, 14, 18, 20, 21). The Wilcoxon rank sum test was used todetermine significance. RPPA data showed that high levels ofphospho-b-catenin (serine 675; active) were significantly corre-lated with PDLIM2-positive tumors (Fig. 3D; Supplementary

Table SI). Phospho-MET (Tyrosine 1349) and the adhesion-asso-ciated proteins fibronectin and FAK were also high in PDLIM2-positive tumors. In contrast, mTOR and Ki67 were low inPDLIM2-positive tumors (Fig. 3D). Interestingly, although theexpression levels of several genes previously shown to be regu-lated by PDLIM2 (21), were significantly different in PDLIM2-negative and -positive tumors, mRNA levels corresponding to thedifferentially expressed proteins within the RPPA data were notsignificantly different (Supplementary Table SII and SIIb).

Overall, these data demonstrate that PDLIM2 expression inTNBC is restricted to the cytoplasm and this positively correlateswith b-catenin activity and adhesion signaling.

PDLIM2 protein expression in TNBC cell lines is associatedwith adhesion and b-catenin levels

To further test PDLIM2 association with specific signalingpathways in TNBC, we interrogated RNAseq and RPPA datasetsavailable for approximately 80 breast cell lines (30). PDLIM2 isnot represented in the RPPA analysis but is present in theRNAseq data. Given the observed lack of correlation betweenPDLIM2 mRNA and protein expression in the tumor samples(Supplementary Fig. S3D), we first tested whether PDLIM2mRNA could be correlated with protein expression in cell lines,using eight TNBC cell lines from the Marcotte study (30), andthe TN, nontumorigenic breast cell line, MCF10A (classifica-tions in Supplementary Fig. S4A). In agreement with observa-tions in tissues, we found that a subset of breast cell linesexpressed high levels of PDLIM2 protein, whereas others didnot (Fig. 4A and B).

Analysis of the RNAseq dataset from the TN cell lines (30)demonstrated that PDLIM2 RNA levels do not significantlycorrelate with protein expression (Supplementary Fig. S4B). ThePDLIM2 gene encodes several RNA variants including PDLIM2RNAvariants 1, 2, and 3 (NM_198042,NM_021630,NM_176871;also referred to as Mystique 1, 2, 3). We previously reported thatPDLIM2 protein is encoded by PDLIM2 variant 2 (12, 16), andin the nine TN cell lines, levels of PDLIM2 variant 2 mRNAcorrelated well with protein (Fig. 4C), whereas variant 3 mRNAwas generally inversely correlated with protein (Fig. 4D).PDLIM2 variant 1 was barely detectable. Thus, because RNAseqdata may include all PDLIM2 RNA transcripts, it cannot be usedto infer PDLIM2 protein expression. Moreover, becausePDLIM2 mRNA levels did not correlate with differentiallyexpressed proteins in the RATHER cohorts (Fig. 3; Supplemen-tary Tables SI and SII, SIIb), our analyses indicate that PDLIM2mRNA expression profiles in cancer cohorts do not reflectprotein expression or function. This disparity may also explainwhy total PDLIM2 mRNA does not correlate strongly withclinical outcomes.

Using the nine TN breast cell lines, we next asked whetherPDLIM2protein correlateswith adhesion signals andb-catenin, aswas observed in the RPPA data (Fig. 3D). b1-integrin was tested asa marker of extracellular matrix (ECM) adhesion, and we previ-ously reported that PDLIM2 is essential for feedback regulation ofb1-integrin signals (14). E-Cadherin and b-catenin are markers ofcell–cell adhesion. b1-integrin was observed to be generallyhigher in PDLIM2-positive than PDLIM2-negative breast cell lines(Fig. 4E andG; Supplementary Fig. S4C). Furthermore, E-cadherinand b-catenin were also expressed at higher levels in PDLIM2-positive than PDLIM2-negative cells (Fig. 4E–G), with the excep-tion of MDA-MB-231 cells where E-cadherin is suppressed by

Cox et al.





methylation (37). In addition, b-catenin phosphorylated onSerine 675 was higher in PDLIM2-positive TN cells (Fig. 4E).Furthermore, levels of phospho-EGFR and phospho-IGF1R were

high in PDLIM2-positive TN cells (Supplementary Fig. S4D andS4E), although there was no clear correlation between phospho-c-MET and PDLIM2 expression (Supplementary Fig. S4F).

NIB, n = 37 RATHER NKI, n = 51

PD

LIM

2 su

bcel

lula

r loc

aliz

atio

n(%

of p

ositi

ve tu

mor

s)

PD

LIM

2 su

bcel

lula

r loc

aliz

atio

n(%

of A

ll po

sitiv

e tu

mor

s)

B C

0

50

100

0

50

100

0

50

100Heidelberg, n = 11

A RATHER-NKI RATHER-CAM HeidelbergNIB

0

20

40

60

80

100

+NucleusCytoplasm/Membrane

*P = 0.0231

D




−1.5

−1.0

−0.5

0.0

0.5

1.0

Nor

m. R

FU

Phospho~β-catenin Ser675

−1.0

−0.5

0.0

0.5

1.0

Nor

m. R

FU

FAK

−1.0

0.0

1.0

Nor

m. R

FU

Ki67

Nor

m. R

FU

Phospho~Met Tyr1349

−2.0

−1.0

0.0

1.0

Nor

m. R

FU

Fibronectin

−1.0

0.0

1.0

2.0

NEG POS

Nor

m. R

FU

mTor

NEG POS NEG POS

NEG POS NEG POS

PDLIM2 statusNEG POS

−2.0

0.0

2.0

10 μm10 μm 10 μm10 μm

10 μm10 μm 10 μm10 μm

PROTEIN Neg vs. Pos IHC Score 1 IHC Score 2 IHC Score 3mTor 0.003 0.065 0.012 0.015Phospho-Beta Catenin Ser675 0.007 0.254 0.024 0.006Ki67 0.015 0.064 0.009 0.549Phospho-Met Tyr1349 0.031 0.480 0.030 0.098Fibronec�n 0.040 0.141 0.018 0.806Fak 0.043 0.732 0.123 0.005

* *** *** ** ** ** *

** *

*

Figure 3.

PDLIM2 retention at the cytoplasm inTNBC and correlation with adhesionsignaling. A, Representativemicrographs of a PDLIM2-positivetumor from each cohort examined,with higher magnification shown inbottom panels to highlight PDLIM2localization at the cytoplasm/membrane and lack of nuclearstaining. B, PDLIM2 subcellularlocalization within the TNBC cells wasscored for each PDLIM2-positivetumor, and scores are presented ingraphs as percentage of positivePDLIM2-tumors for each cohort. (TheRATHER-CAM samples were omittedfrom these analyses as diffusestaining in some samples confoundeda distinct localization score.) C,Overall localization of PDLIM2 isshown as the percentage of allPDLIM2-positive TNBCs across thethree cohorts, n¼ 99. Statisticalsignificance was analyzed using aStudent t test. D, Box and whiskersplots of RPPA data illustratingdistribution andmedian of expressionvalues of proteins differentiallyexpressed in PDLIM2-negative (NEG)versus PDLIM2-positive (POS) tumorsfrom the RATHER cohorts. Norm.RFU, normalized RPPA fluorescenceunits. Significant differences betweennegative and PDLIM2 expressionintensity from IHC analysis are alsoshown (IHC score 1–3); Wilcoxontest (rank sum). � , P < 0.05;�� , P < 0.005.






Overall, we conclude that PDLIM2expression is associatedwithhigh levels of cell adhesion markers, active growth factor recep-tors, and b-catenin.

PDLIM2 shuttling from cytoplasm to nucleus is stimulated byadhesion and growth factors

Although PDLIM2 is restricted to the cytoplasm of TN tumors(Figs. 1–3), it could be observed in the cytoplasm and nucleus ofcultured cell lines by both immunofluorescence and subcellularfractionation (Fig. 5A and B; Supplementary Fig. S5A and S5B).These observations suggest that cell adhesion and growth factorspromote PDLIM2 shuttling between the cytoplasm and nucleus,

so we tested this further. We found that under nonadherentculture conditions, PDLIM2 accumulates in the nucleus (Supple-mentary Fig. S5B and S5C). De-adhesion of MCF10A cells for24 hours was sufficient to reduce PDLIM2 in the cytoplasm andinduce accumulation in the nucleus, even under serum-starvedconditions (Fig. 5C and D; Supplementary Fig. S5B and S5C).Re-adhesion of cells for up to 24 hours was sufficient to restorecytoplasmic PDLIM2 with a concomitant decrease in nuclearlevels (Fig. 5C). Thus, cell adhesion is necessary and sufficientfor nuclear exclusion of PDLIM2.

Serum starvation caused markedly decreased nuclear andincreased cytoplasmic PDLIM2 (Fig. 5E–H; Supplementary

MCF10A

MDA231

HCC1806

HCC70

HS578T

MDA436

MDA157

CAL51BT549

PDLIM2

Tubulin3555

β-Catenin

β1-Integrin

90

100

G

E

35

55

PDLIM2

Tubulin

MCF10A

MDA231

MDA157

HS578T

MDA436

HCC70

HCC1806

BT549

CAL51

100 E-Cadherin

F

(Fol

d di

ffere

nce,

n =

5)

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

BT549

MDA436

MDA157

0

1

2

3

β-Catenin

Prot

ein

expr

esio

n(fo

ld d

iffer

ence

, n =

5)

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

BT549

MDA436

MDA157

0

1

2

3

β1-Integrin

(Fol

d di

ffere

nce,

n =

3)

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

BT549

MDA436

MDA157

0.0

0.5

1.0102030

E-Cadherin

BA

PD

LIM

2 pr

otei

n ex

pres

sion

(fold

diff

eren

ce, n

= 5

)

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

MDA436

MDA157

BT549

0.0

0.5

1.0

1.5

***P < 0.0005

DC PDLIM2 variant 2

mR

NA

expr

essi

on(fo

ld d

iffer

ence

, n =

4)

*

***

P < 0.0005

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

MDA436

MDA157

BT549

0

5

10

15

20

P~S675β-Catenin90

mR

NA

expr

essi

on(fo

ld d

iffer

ence

, n =

3)

PDLIM2 variant 3

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

MDA436

MDA157

BT549

0

2

4

6

MCF10A

MDA231

HCC1806

HCC70

HS578T

MDA436

MDA157

CAL51BT549

PDLIM2

Actin45

35

kDa

kDa

kDa

Figure 4.

PDLIM2 expression in TN cell linesthat also express adhesion receptorsand b-catenin. A,Whole cell lysateswere prepared from a panel of TN celllines cultured for 48 hours incomplete medium, and PDLIM2expression was examined byWesternblotting. Actin was used as loadingcontrol. The approximate proteinmolecular weights in kDa areindicated on the left of eachWesternblot analysis. B, PDLIM2 expression ineach of the cell lines was quantifiedby densitometry and data arepresented as fold difference inexpression compared with MCF10Acells, set at 1. Mean expression� SEMfrom five separate experiments areshown. C andD, PDLIM2 mRNAexpression was measured for PDLIM2variant 2 (C) and variant 3 (D) byqPCR. Primer details are listed inSupplementary Materials andMethods, Table 2. Graphs representmean� SEM from at least threeseparate experiments. Statisticalsignificance was analyzed using aStudent t test, comparing mRNAexpression in each cell line toMCF10A. � , P¼ 0.045 in C. E–G, TNbreast cell lines were grown undernormal conditions, lysed, andexamined for expression of adhesionsignaling proteins b1-integrin,E-cadherin, and b-catenin andphospho-ser675 b-catenin byWestern blotting. Data arerepresentative of at least threeexperiments. Protein expression wasquantified by densitometry asdescribed in Materials and Methods(G); quantification of phospho-ser675b-catenin is included inSupplementary Fig. S5D.

Cox et al.





A

G

PD

LIM

2 ex

pres

sion

(fold

cha

nge,

n =

4)

Completemedium

Serumfree

TGFβ10

1

2

**

**

Completemedium

Serumfree

IGF1

(Fol

d ch

ange

, n =

3)

CytoplasmNucleus

0

1

2

3 *

**

*

E

B Adherent Suspension

PDLI

M2

expr

essi

on(fo

ld c

hang

e, n

= 3

)

Cytoplasm Nucleus0

1

2

3

4

5*

C

HS578THCC1806MDA231 HCC70MCF10A

PDLIM2

PDLIM2/Actin

D

H PDLIM2

TGFβ124 hours

Serumfree

35

55

100

MCF10AMDA231

HCC70HCC1806

PDLIM2Cy NCy NCy NCy

kDa

kDa

kDa

kDaP~ P~

N

PARP

Tubulin35

55Cytoplasm Nucleus

Adherent

Suspensio

n

Re-attach

ed

Adherent

Suspensio

n

Re-attach

ed

PARPPDLIM2

Tub100

MCF-10A

Vinculin

PARP

PDLIM2

P~Smad2

35

50

100

100

TGFβ1, hr: 0 1 4 8 24

Cytoplasm Nucleus

0 1 4 8 24CM CM

F

Vinculin

PARP

PDLIM2

P~Akt

35

55

100

100

IGF1, hr: 0 1 4 8 24

Cytoplasm Nucleus

0 1 4 8 24CM CM

PDLIM2/Actin

Figure 5.

Cytoplasmic and nuclear shuttling of PDLIM2 in response to adhesion and growth factor signals. A, Representative immunofluorescence micrographs from atleast three separate experiments, showing PDLIM2 expression and localization in TN cell lines. The remainder of the cell panel is shown in Supplementary Fig.S5A. Cells were costained with TRITC-Phalloidin (actin) and Hoechst (Nuclei). Scale bar, 20 mm. B, Subcellular fractions were prepared from TN breast cellscultured for 48 hours in complete medium and analyzed byWestern blotting for PDLIM2 expression. Fraction markers are tubulin (cytoplasm, Cy) and PARP(nuclei, N); representative of at least three separate experiments. C,MCF10A were plated on polyhema-treated petri dishes to prevent adhesion for 24 hours(suspension), followed by re-plating on tissue-culture treated plates to allow reattachment for 24 hours. PDLIM2 expression and localization was compared withadherent cells by subcellular fractionation andWestern blotting. A representative experiment of three is shown. D, Graph shows densitometric quantification ofPDLIM2 expression in the cytoplasm or nuclei of adherent and suspension cells from three separate experiments. Graphs represent mean� SEM; statisticalsignificance was analyzed using Student t test. E and F,MCF10A cells were serum starved, followed by stimulation with either 10 ng/mL TGFb (TGFb1; E) or IGF1(F) for up to 24 hours. Cells were harvested at each time-point indicated, and processed for subcellular fractionation of cytoplasmic and nuclear fractions, whichwere probed for PDLIM2, signaling pathwaymarkers phospho-Smad2 (TGFb1) or phospho-Akt serine 473 (IGF1), and fraction markers vinculin (cytoplasm) andPARP (nuclear) byWestern blotting. Arrows, phosphorylated PDLIM2. G,Quantification of PDLIM2 expression and localization following serum starvation inpresence or absence of growth factor stimulation for 24 hours (TGFb1, left; IGF1, right) from at least three separate experiments. Expression levels weremeasured by densitometry. Graphs represent mean� SEM, and statistical significance was analyzed using Student t test. H, PDLIM2 expression and localization(gray/green) in MCF10A cells following serum starvation alone, or in presence of 10 ng/mL TGFb1. Representative images taken at�100 magnification of eachcondition from at least three separate experiments are shown. Costaining with TRITC-Phalloidin (actin, red) and Hoechst (nuclei, blue). Scale bars, 20 mm.� , P� 0.05; �� , P� 0.005 in D andG.






D

E

A

β-C

aten

in a

ctiv

ity(T

OP

flash

ass

ay,

fold

cha

nge,

n =

3)

HEK 293T

HA-PDLIM2HA-EV

***

0

2

4

6

Act

ive

β-ca

teni

n dx

pres

sion

(fold

cha

nge,

n =

3)

HA-PDLIM2HA-EV

**

0

1

2

3

4C

Act

ive

β-ca

teni

n ex

pres

sion

(fold

cha

nge,

n =

3)

HA-PDLIM2HA-EV0

10

20

30 **

β-Catenin

MCF10A

MDA231

MDA157

HS578T

MDA436

HCC70

HCC1806

BT549

CAL51

PDLIM2

Activeβ-Catenin

Tubulin3555

90kDa

kDa kDa

kDa

kDa

90

***

MCF10A

MDA231

HCC1806

HCC70

HS578T

CAL51

BT549

MDA436

MDA157

Act

ive

β-ca

teni

nex

pres

sion

(fold

diff

eren

ce, n

= 4

)

0

1

2

3

4

5

*****

*

***

Cy N Cy N

SusAdhCy N Cy N

SusAdh

Active β-Cateninβ-Catenin

PDLIM2

Tubulin

PARP

MCF10A HCC1806

e

35

55

100

100

100

Cy N Cy N

SusAdh

HS578T

F

Active β-Cateninβ-CateninPDLIM2

TubulinPARP

HA-PDLIM

2

HA-EV

NCy N CyNCyNCyHA-P

DLIM2

HA-EV

BT549Adhesion Suspension

35

55

110

100

100

G

B

HA-PDLIM

2

HA-EV

BT549

PDLIM2

Tubulin

Activeβ-Catenin

35

55

90

90 β-Catenin

HEK 293T

HA-PDLIM

2

HA-EV

HA-PDLIM2

Tubulin

Activeβ-Catenin

35

55

90

90β-Catenin

MDA 231MCF10A HCC1806 HCC70 HS578TCAL51 BT549

PD

LIM

2M

erge

MDA 157

Merge: PDLIM2/β-catenin P~S675/Nuclei

β-C

aten

inP

~S67

5

Cox et al.





Fig. S5C). Stimulation of MCF10A with either TGFb (Fig. 5E, G,and H) or IGF1 (Fig. 5F and G) promoted translocation ofPDLIM2 into the nucleus over time. Interestingly, the substan-tial retardation of cytoplasmic PDLIM2 protein mobility,which we previously confirmed as serine phosphorylation (20),was less evident in the nuclear fractions (arrows; Fig. 5E and F),indicating that PDLIM2 serine phosphorylation facilitates itcytoplasmic sequestration. Overall, these data show that adhe-sion and growth factor signaling in TN breast cells controlPDLIM2 phosphorylation and subcellular localization.

PDLIM2 expression is sufficient to activate b-catenin in TNBCBecause PDLIM2 expression correlates with active b-catenin

(Fig. 3D, Fig. 4E), we next asked whether PDLIM2 enhancesb-catenin activity by using two approaches. First, we assessedactive b-catenin in cell lines using an antibody that detectsb-catenin only when it is not phosphorylated on serine 45 andtherefore active (38). This showed that PDLIM2-expressingcells generally express active b-catenin (Fig. 6A). Exceptionswere MDA-MB-231 cells, which express PDLIM2 in the pres-ence of low b-catenin, and HS578T, which do not expressPDLIM2 but have high active b-catenin expression. Similarresults were observed using the antibody that detects activeb-catenin phosphorylated on serine 675 (Fig. 4E; Supplemen-tary Fig. S5D; refs. 38, 39). Immunofluorescence stainingillustrated that active b-catenin is present at the plasma mem-brane, throughout the cytoplasm and in the nucleus ofPDLIM2-positive cells (Fig. 6B). Interestingly, in cells withlow PDLIM2, active b-catenin is predominantly at the plasmamembrane (Fig. 6B).

The second approach was to test the effects of ectopicHA-PDLIM2 expression on b-catenin activity. In BT549 cells,HA-PDLIM2 was mostly evident in the cytoplasm and at theactin cytoskeleton (Supplementary Fig. S5E). Higher total andactive b-catenin levels (Fig. 6C; Supplementary Fig. S6A) wereobserved in the cytoplasm and nucleus (Supplementary Fig.S6A and S6B), and the migratory potential was enhancedcompared with controls (Supplementary Fig. S6C). Further-more, ectopic expression of PDLIM2 was sufficient to activateb-catenin in HEK293T cells in luciferase reporter assays(Fig. 6D and E). PDLIM2 suppression had minor effects onthe levels of b-catenin and adhesion proteins, in HCC1806

cells and MDA-MB-231-LUC2 cells (Supplementary Fig. S6Eand S6F).

Because active b-catenin is associated with PDLIM2 in TNBCtissues and cell lines (Figs. 3 and 6A and B), and PDLIM2 isrestricted to the cytoplasm of TNBC tumor cells (Fig. 3), wenext asked whether the subcellular localization of PDLIM2 isimportant for b-catenin expression and activation. To test this,we used cell detachment to induce nuclear accumulation ofPDLIM2 and serum starvation to induce nuclear exclusion,and then measured b-catenin levels and phosphorylation(activity) status in PDLIM2-positive and -negative cells. Asexpected (Fig. 5; Supplementary Fig. S5; refs. 12, 20), cells insuspension exhibited less cytoplasmic PDLIM2 than adherentcells, with a concomitant accumulation of nuclear PDLIM2(Fig. 6F; Supplementary Fig. S6G and S6H). These cells showeda clear reduction in active b-catenin levels in both the cyto-plasm and nucleus that was not evident in PDLIM2-negativecells. Serum starvation had little effect on levels of activeb-catenin, indicating that nuclear PDLIM2 is not directlyinvolved in regulating b-catenin levels or activity (Supplemen-tary Fig. S6H and S6I). We also confirmed that overexpressedHA-PDLIM2 accumulated more in the nucleus than in thecytoplasm in non-adherent BT549 cells, and this was accom-panied by a marked reduction in active b-catenin, especially inthe nucleus (Fig. 6G).

Overall, we conclude that PDLIM2 accumulation in the cyto-plasm (and out of the nucleus) enhances nuclear levels of activeb-catenin.

PDLIM2 suppression impairs TNBC tumor spread in vivo andalters spheroid growth

In a previous study, PDLIM2 suppression in DU145 pro-state cells was shown to reverse the EMT phenotype characterizedby re-expression of epithelial markers and loss of directionalmigration (21). In that study we also showed MDA-MB-231-LUC2 cells with stably suppressed PDLIM2 exhibited enhancedproliferative rates and reduced growth in soft agarose comparedwith controls. As can be seen in Fig. 7A and B, two clones ofMDA-MB-231-LUC2 cells stably expressing shRNA targeting PDLIM2(shPDLIM2) exhibited greatly reduced colony formation (platingefficiency) compared with controls (shScr). To test the in vivosignificance of PDLIM2 suppression in TNBC, we assessed in vivo

Figure 6.PDLIM2 expression is sufficient to activate b-catenin in TNBC. A, Immunoblots of whole cell lysates from TN breast cell lines grown under normalconditions for 48 hours were probed with active b-catenin (nonphosphorylated serine 45), total b-catenin, PDLIM2, and tubulin antibodies. Graphshows quantification of active b-catenin expression, measured by densitometry; mean expression relative to that of MCF10A cells � SEM is shown.Student t test analyses of significant differences in expression compared with those of MCF10A cells, which are set at 1. � , P � 0.05; �� , P � 0.005;�� , P � 0.0001. B, Immunofluorescence micrographs of TN breast cell lines grown on coverslips in complete medium for 24 hours. Cells were fixedand stained for antibodies against active b-catenin (phospho-Ser675; green) and PDLIM2 (red). Nuclei were stained with Hoechst (blue). Scale bars,20 mm. C and D, BT549 cells (C) or HEK293T cells (D) were stably transfected with HA-Empty Vector (EV) or HA-PDLIM2 as described inSupplementary Materials and Methods. Cells were cultured under normal conditions for 48 hours and assessed for protein expression as in A. Graphsshow quantification of active b-catenin expression (non-phospho-Ser 45), measured by densitometry; mean expression � SEM, relative to that ofHA-EV control cells. Data are from three separate stable clones of HA-EV and HA-PDLIM2- expressing cells. �� , P � 0.001, Student t test. E,HEK293T cells stably transfected with HA-EV or HA-PDLIM2 were assessed for b-catenin transcriptional activity using a b-catenin/TCF TOPflashluciferase assay as described in Materials and Methods. Data from three separate experiments are presented as fold change � SEM of transcriptionalactivity of HA-PDLIM2-expressing cells compared with HA-EV controls. �� , P � 0.0001, Student t test. F and G, Subcellular fractions of adherent orsuspension cells were prepared for PDLIM2-positive or -negative TN cell lines (F) or BT549 stably expressing HA-EV or HA-PDLIM2 (G). Westernblots of cytoplasmic and nuclear fractions were probed for active b-catenin [non-phospho-Ser 45 (F) or phospho-Ser 675 (G)], b-catenin, PDLIM2,and fraction markers tubulin (cytoplasm, Cy) and PARP (nuclear, N), e, empty lane. Counterparts with additional cell lines for parts F and G areshown in Supplementary Fig. S6G and S6H.






tumor spread of these MDA-MB-231-LUC2 clones in nude mice.Each of the shPDLIM2 clones exhibited little if any tumor burdenafter 45 to 49 days compared with controls, noting that controlshScr 1 cells exhibited better growth in vivo than shScr 2 (Fig. 7C).This result is consistent with reduced clonogenic growth and

migratory potential previously observed with PDLIM2 suppres-sion (21), and suggests that high PDLIM2 may facilitate breastcancer progression.

To further test PDLIM2 function, we assessed the effects ofPDLIM2 overexpression on BT549 3D cell growth using 3D "on

ED

3D assay BT549

HA-PDLIM

2(1)

HA-EV(1)

β-Catenin

Activeβ-Catenin

PDLIM2Tubulin

HA-PDLIM

2(2)

HA-EV(2)

HA-PDLIM

2(3)

HA-EV(3)

kDa

Lum

ines

cenc

e (m

ean

coun

ts x

10⁴)

shScr-

1

shScr-

2

shPDLIM

2-1

shPDLIM

2-20

10

20

30

40

90

****

shScr-

1

shScr-

2

shPDLIM

2-1

shPDLIM

2-2

30,000

20,000

10,000

ImageMin = -41336Max = 75672p/sec/cm^2/sr

C

MDA-MB-231-LUC2

HA-EV HA-PDLIM2 HA-EV HA-PDLIM2

3D assay- Day 2 3D assay- Day 4

Clo

ne 1

Clo

ne 3

Clo

ne 2

BT549

A BColony formation

Num

ber

of c

olon

ies

(n =

3)

shScr-

1

shScr-

2

shPDLIM

2-1

shPDLIM

2-20

20

40

60

80

******

% P

DLI

M2

expr

essi

on(n

= 5

)

shScr-

1

shScr-

2

shPDLIM

2-1

shPDLIM

2-2

*** ***

020406080

100120

PDLIM2

shScr-

2

shPDLIM

2-1

shPDLIM

2-2

shScr-

1

Actin

kDa

3545

100

100

3535

Figure 7.

PDLIM2 suppression inhibits 3D andcolony formation in vitro and growthof TNBC cells in vivo.A,Western blotanalyses of suppression of PDLIM2expression in MDA-MB-231-LUC2cells stably expressing shScramble(shScr) or shPDLIM2 (clones 1 and 2),used for in vivo studies. PDLIM2expression was quantified bydensitometry, normalized to shScr1levels, n¼ 5. �� , P < 0.0005,Student t test. B, Colony formationby MDA-MB-231-LUC2 clones wasassessed by plating efficiency assay,n¼ 3. �� , P < 0.0005, Student t test.C, Scatter plot showing decreasedtumor burden of tumor cells withPDLIM2 suppressed. Data are wholebody bioluminescence counts 45 to49 days post-tail vein injection, withtwo clones each of MDA-MB-231-LUC2 cells stably expressing shScr orshPDLIM2 (clones 1 and 2). Examplesof IVIS images are also shown. Thecolor scale depicts the photon flux(photons per second) emitted fromanimals. �� , P < 0.0005; �� , P <0.005; one-way ANOVA,Bonferroni's posttest in B and C. D,Micrographs of 3D cultures formedby BT549 clones stably expressingHA-EV or HA-PDLIM2, taken on Day2 and 4 following plating in a 3D "ontop" assay as described in Materialsand Methods. Original magnification,�10. E, 3D cultures extracted fromMatrigel were lysed and analyzed byWestern blotting for expression ofphospho-ser675 and total b-catenin,HA-PDLIM2 and tubulin loadingcontrol.

Cox et al.





top" cell culture assays (31). In control BT549 cultures, cellsformed stellate 3D structures that have previously been asso-ciated with an invasive phenotype (31). However, BT549 cellsstably expressing HA-PDLIM2 with higher active b-catenin,initially formed more rounded, small clustered structures thancontrols at day 2, and failed to fully adopt the stellate mor-phology by day 4 (Fig. 7D and E). These results indicate thatPDLIM2 regulates the pathway controlling the stellate pheno-type and also indicates that b-catenin activity is regulatedduring this reversible process.

DiscussionAlthough several studies have characterized distinct types of

TNBC (1–3, 6, 7), there remains an urgent need for biomarkerswith prognostic value or to predict TNBC subgroups that wouldbe amenable to novel therapies. Here, we found that thePDLIM2 protein is enriched in TNBCs (overall approximately60%), whereas it is less frequently expressed in other breastcancers. Furthermore, the correlation of PDLIM2 expressionwith b-catenin activity and adhesion signaling suggests itdefines a large cohort of TNBC with similar features. A directlink between PDLIM2 expression in TNBC and clinicopatho-logical parameters was not established. However, this may bedue to TNBC heterogeneity and the fact that PDLIM2 mRNAprofiles in available TNBC databases actually represent severalPDLIM2 mRNA variants, whereas only one mRNA variant (2)encodes protein. Our conclusion that the PDLIM2 protein is amarker for a large subset of TNBC is further supported bypublished genomic studies, including a gene dependencyscreen, that identified PDLIM2 as a moderate dependency genein basal TNBC and functionally linked this to cell adhe-sion (27). Interestingly, this study also reported that basalTNBC is addicted to proteasomal activity, whereas genes encod-ing components of the proteasome and the COP9 signalosome(CSN) have been described as essential in TNBC basal A celllines (30). These observations are all consistent with thereported actions of PDLIM2 in shuttling between the cytoskel-eton and nucleus to regulate transcription factor stability(19, 21, 40), via association with the COP9 signalosome (21).

Suppression of PDLIM2 in MDA-MB-231 (and DU145) cells issufficient to impair migratory potential and clonogenicgrowth (21), and as shown here, to impair in vivo spread ofMDA-MB-231-LUC2 cells, whereas re-introduction of PDLIM2into BT549 TNBC cells alters their invasive growth in 3D culturesand enhances directional migration. The exclusion of PDLIM2from the nucleus of tumor cells, and the effects of adhesion andgrowth factors on promoting PDLIM2 translocation from thecytoplasm to nucleus in cell cultures, indicate that adhesionsignaling promotes cytoplasmic retention of PDLIM2 in TNBCto regulate its function. Moreover, TN cell lines expressing cyto-plasmic PDLIM2 generally also expressed high levels of b1-integ-rin, E-cadherin, b-catenin, and phosphorylated IGF1 or EGFreceptors. However, although PDLIM2 suppression enhancedexpression of integrins and ECM proteins in a MCF10A 3Dmodel (14), suppression of PDLIM2 in TNBC cell lines did nothave a significant effect on the levels of b1-integrin, for example.This may be expected considering the high levels of adhesionproteins in PDLIM2-expressing TNBC cells, and that PDLIM2 isitself regulated by, and a component of adhesion signaling inthese cells. Overall, we propose that PDLIM2 functions to main-

tain a polarized phenotype associated with cell–cell adhesion innormal epithelial cells, but in cancer cells, it is a component ofadhesion signaling that promotes loss of cell polarization, tumorgrowth, and motility.

Our findings indicate that b-catenin activity is a key output ofcytoplasmic PDLIM2 in TNBC. Levels of active b-catenin weresignificantly elevated in PDLIM2-positive tumors and ectopicexpression of PDLIM2 was sufficient to induce b-catenin expres-sion andactivity in cell lines. This is consistentwith suppression ofPDLIM2 in EMT-like cells causing reduced b-catenin nuclearactivity (21). Several studies have linked canonical and atypicalWNT/b-catenin signaling to breast cancer, although the role ofb-catenin is complex (41–44). Canonical WNT signaling in TNBCmay promote stabilization of b-catenin in the cytoplasm andsubsequent transcription of genes involved in EMT (44). How-ever, elevated b-catenin at the plasmamembrane has been linkedwith a good prognosis in TNBC (42),whereas low b-catenin at cellmembranes in EGFR-positive TNBC has been associated withpoor survival (45). Low levels of E-cadherin at cell membranesmay also correlate with elevated b-catenin activity and poorsurvival in TNBC (46).

How PDLIM2 integrates adhesion and growth factor signalingto enhance b-catenin phosphorylation and activation is not fullyestablished. It is likely that PDLIM2 mediates cell adhesion-dependent phosphorylation and relocation of b-catenin to facil-itate its subsequent nuclear activity. This is supported by ourobservations that cell de-adhesion reduces cytoplasmic PDLIM2levels and phosphorylation/nuclear accumulation of activeb-catenin, and that ectopic expression of PDLIM2 enhancesb-catenin phosphorylation on serine 675 and its nuclear accu-mulation. Importantly, we also observed that although activeb-catenin is low in PDLIM2-negative cell lines, it was predomi-nantly at the plasma membrane or cell–cell contacts, whereas, inPDLIM2-positive cells, it was localized in the cytoplasm andnucleus. These suggest that PDLIM2 is not implicated in basalb-catenin phosphorylation, but rather, it facilitates the release ofb-catenin from the membrane, thereby enabling its stabilization,phosphorylation, and subsequent activation. Interestingly,growth factor and adhesion signals have recently been reportedto act through cellular kinases such as SRC and PAK to enhanceb-catenin phosphorylation on serine 675, and its subsequentactivation (39, 47).

In summary, PDLIM2 is a marker for a large subgroup ofTNBC that is driven by adhesion, growth factor signals andb-catenin activity. The identification of a potential new proteinmarker to classify functional subsets of TNBCs across previ-ously defined subtypes is an important development. AlthoughPDLIM2 per se may not have a prognostic role, it could be apredictive biomarker to stratify TNBC for therapies that targetthe signaling pathways regulated by PDLIM2 or that directlytarget active b-catenin.

Disclosure of Potential Conflicts of InterestB. Moran has Marie Curie Industry Fellowship at Oncomark Ltd. W.M.

Gallagher is a Chief Scientific Officer at OncoMark Limited; reports receivingcommercial research grant from Carrick Therapeutics; has ownership interest(including stock, patents, etc.) in OncoMark Limited; and is a member ofconsultant/advisory board at Carrick Therapeutics. R.D. Kennedy is a MedicalDirector at Almac Diagnostics. C. Caldas reports receiving commercial researchgrant fromAstraZeneca, Genentech, Roche, Servier; and is a consultant/advisoryboard member at AstraZeneca. No potential conflicts of interest were disclosedby the other authors.






Authors' ContributionsConception and design: O.T. Cox, R. O'ConnorDevelopment of methodology: O.T. Cox, S.J. Edmunds, M. Bustamante-Garrido, R.D. Kennedy, A. Marx, R. O'ConnorAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): O.T. Cox, S.J. Edmunds, K. Simon-Keller,N.E. Buckley, M. Bustamante-Garrido, N. Healy, C.H. O'Flanagan,W.M. Gallagher, R. Bernards, S.-F. Chin, A. Marx, R. O'ConnorAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): O.T. Cox, S.J. Edmunds, K. Simon-Keller, B. Li,B. Moran, N.E. Buckley, M. Bustamante-Garrido, C. Caldas, A. Marx,R. O'ConnorWriting, review, and/or revision of the manuscript: O.T. Cox, S.J. Edmunds,K. Simon-Keller, B. Moran, N.E. Buckley, M. Bustamante-Garrido, N. Healy,C.H. O'Flanagan, W.M. Gallagher, R.D. Kennedy, R. Bernards, A. Marx,R. O'ConnorAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): O.T. Cox, K. Simon-Keller, B. Li, C. Caldas,A. Marx, R. O'ConnorStudy supervision: O.T. Cox, R. O'Connor

AcknowledgmentsThis work was funded by the Irish Cancer Society Collaborative Cancer

Centre Breast Predict CCRC13GAL (toW.M.Gallagher, R. O'Connor, supported

O.T. Cox, B. Li, B. Moran), Science Foundation Ireland Principal Investigatorawards 11/PI/1139 and 16/IA/4505 (to R. O'Connor, supported O.T. Cox,M. Bustamente-Garrido), and the European Union FP7 Marie Curie Industry-Academia Partnerships and Pathways (IAPP) Programme 251480: Biomarker-IGF (to R. O'Connor, supported S. E. Edmunds, O. T. Cox, C. H. O'Flannagan).The Northern Ireland Biobank (NIB) received funding from the HSC Researchand Development Division of the Public Health Agency in Northern Ireland,Cancer Research UK through the Belfast CR- UK Centre, the Northern IrelandExperimentalCancerMedicineCentre, and the Friends of theCancerCentre. TheNorthern Ireland Molecular Pathology Laboratory, which creates resources forthe NIB has received funding from Cancer Research UK, the Friends of theCancer Centre and the Sean Crummey Foundation. The Rational Therapy forBreast Cancer (RATHER), a Collaborative Project is funded under the EuropeanUnion 7th Framework Programme (grant agreement no. 258967). We thankDr. Gary Loughran, University College Cork, for helpful discussions on PDLIM2RNA isoforms and RNA-seq analysis.

The costs of publication of this articlewere defrayed inpart 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 September 6, 2018; revised January 9, 2019; accepted March 12,2019; published first March 18, 2019.

References1. Shah SP, Roth A, Goya R, Oloumi A, Ha G, Zhao Y, et al. The clonal and

mutational evolution spectrum of primary triple-negative breast cancers.Nature 2012;486:395–9.

2. Curtis C, Shah SP, Chin SF, Turashvili G, RuedaOM,DunningMJ, et al. Thegenomic and transcriptomic architecture of 2,000 breast tumours revealsnovel subgroups. Nature 2012;486:346–52.

3. Liedtke C, Bernemann C, Kiesel L, Rody A. Genomic profiling in triple-negative breast cancer. Breast Care (Basel) 2013;8:408–13.

4. Ali HR, Rueda OM, Chin SF, Curtis C, Dunning MJ, Aparicio SA, et al.Genome-driven integrated classification of breast cancer validated in over7,500 samples. Genome Biol 2014;15:431. doi: 10.1186/s13059-014-0431-1.

5. Denkert C, Liedtke C, Tutt A, von Minckwitz G. Molecular alterations intriple-negative breast cancer-the road to new treatment strategies. Lancet2017;389:2430–42.

6. Burstein MD, Tsimelzon A, Poage GM, Covington KR, Contreras A,Fuqua SA, et al. Comprehensive genomic analysis identifies novelsubtypes and targets of triple-negative breast cancer. Clin Cancer Res2015;21:1688–98.

7. LehmannBD, Bauer JA, ChenX, SandersME, Chakravarthy AB, Shyr Y, et al.Identification of human triple-negative breast cancer subtypes and pre-clinical models for selection of targeted therapies. J Clin Invest 2011;121:2750–67.

8. Bahhnassy A, Mohanad M, Shaarawy S, Ismail MF, El-Bastawisy A, Ash-mawy AM, et al. Transforming growth factor-beta, insulin-like growthfactor I/insulin-like growth factor I receptor and vascular endothelialgrowth factor-A: prognostic and predictive markers in triple-negative andnon-triple-negative breast cancer. Mol Med Rep 2015;12:851–64.

9. Taliaferro-Smith L,Oberlick E, Liu T,McGlothen T, Alcaide T, Tobin R, et al.FAK activation is required for IGF1R-mediated regulation of EMT, migra-tion, and invasion in mesenchymal triple negative breast cancer cells.Oncotarget 2015;6:4757–72.

10. Yin HL, Wu CC, Lin CH, Chai CY, HouMF, Chang SJ, et al. b1-Integrin as aprognostic and predictive marker in triple-negative breast cancer. Int J MolSci 2016;17. doi: 10.3390/ijms17091432.

11. Hamurcu Z, Kahraman N, Ashour A, Ozpolat B. FOXM1 transcriptionallyregulates expression of integrin beta1 in triple-negative breast cancer.Breast Cancer Res Treat 2017;163:485–93.

12. Loughran G, Healy NC, Kiely PA, Huigsloot M, Kedersha NL, O'Connor R.Mystique is a new insulin-like growth factor-I-regulated PDZ-LIM domainprotein that promotes cell attachment and migration and suppressesanchorage-independent growth. Mol Biol Cell 2005;16:1811–22.

13. Cox OT, O'Shea S, Tresse E, Bustamante-Garrido M, Kiran-Deevi R,O'Connor R. IGF-1 receptor and adhesion signaling: an important axis indetermining cancer cell phenotype and therapy resistance. Front Endocrinol(Lausanne) 2015;6:106. doi: 10.3389/fendo.2015.00106.

14. Deevi RK, Cox OT, O'Connor R. Essential function for PDLIM2 in cellpolarization in three-dimensional cultures by feedback regulation of thebeta1-integrin-RhoA signaling axis. Neoplasia 2014;16:422–31.

15. TorradoM, Senatorov VV, Trivedi R, Fariss RN, Tomarev SI. Pdlim2, a novelPDZ-LIM domain protein, interacts with alpha-actinins and filamin A.Invest Ophthalmol Vis Sci 2004;45:3955–63.

16. Loughran G, Huigsloot M, Kiely PA, Smith LM, Floyd S, Ayllon V, et al.Gene expression profiles in cells transformedbyoverexpressionof the IGF-Ireceptor. Oncogene 2005;24:6185–93.

17. Macartney-Coxson DP, Hood KA, Shi HJ, Ward T, Wiles A, O'Connor R,et al.Metastatic susceptibility locus, an 8p hot-spot for tumour progressiondisrupted in colorectal livermetastases: 13 candidate genes examinedat theDNA, mRNA and protein level. BMC Cancer 2008;8:187. doi: 10.1186/1471-2407-8-187.

18. Tanaka T, Soriano MA, Grusby MJ. SLIM is a nuclear ubiquitin E3ligase that negatively regulates STAT signaling. Immunity 2005;22:729–36.

19. Tanaka T, Grusby MJ, Kaisho T. PDLIM2-mediated termination of tran-scription factor NF-kappaB activation by intranuclear sequestration anddegradation of the p65 subunit. Nat Immunol 2007;8:584–91.

20. Healy NC, O'Connor R. Sequestration of PDLIM2 in the cytoplasm ofmonocytic/macrophage cells is associated with adhesion and increasednuclear activity of NF-kappaB. J Leukoc Biol 2009;85:481–90.

21. Bowe RA,CoxOT, AyllonV, Tresse E,HealyNC, Edmunds SJ, et al. PDLIM2regulates transcription factor activity in epithelial-to-mesenchymal transi-tion via the COP9 signalosome. Mol Biol Cell 2014;25:184–95.

22. QuZ, Fu J, YanP,Hu J,Cheng SY, XiaoG. Epigenetic repressionof PDZ-LIMdomain-containing protein 2: implications for the biology and treatmentof breast cancer. J Biol Chem 2010;285:11786–92.

23. Vanoirbeek E, Eelen G, Verlinden L, Carmeliet G, Mathieu C, Bouillon R,et al. PDLIM2 expression is driven by vitamin D and is involved in thepro-adhesion, and anti-migration and -invasion activity of vitamin D.Oncogene 2014;33:1904–11.

24. Wurster KD, Hummel F, Richter J, Giefing M, Hartmann S, HansmannML, et al. Inactivation of the putative ubiquitin-E3 ligase PDLIM2 inclassical Hodgkin and anaplastic large cell lymphoma. Leukemia 2017;31:602–13.

Cox et al.





25. Du F, Li Y, Zhang W, Kale SP, McFerrin H, Davenport I, et al. Highly andmoderately aggressive mouse ovarian cancer cell lines exhibit differentialgene expression. Tumour Biol 2016;37:11147–62.

26. Bouchal P,DvorakovaM,RoumeliotisT,BortlicekZ, Ihnatova I, ProchazkovaI, et al. Combined proteomics and transcriptomics identifies carboxypepti-dase B1 and nuclear factor kB (NF-kB) associated proteins as putativebiomarkers of metastasis in low grade breast cancer. Mol Cell Proteomics2015;14:1814–30.

27. Petrocca F, Altschuler G, Tan SM, Mendillo ML, Yan H, Jerry DJ, et al. Agenome-wide siRNA screen identifies proteasome addiction as a vulner-ability of basal-like triple-negative breast cancer cells. Cancer Cell 2013;24:182–96.

28. Bassiri K, Ferluga S, Sharma V, Syed N, Adams CL, Lasonder E, et al. Globalproteome and phospho-proteome analysis of Merlin-deficient meningio-ma and schwannoma identifies PDLIM2 as a novel therapeutic target.EBioMedicine 2017;16:76–86.

29. Neve RM,Chin K, Fridlyand J, Yeh J, Baehner FL, Fevr T, et al. A collection ofbreast cancer cell lines for the study of functionally distinct cancer subtypes.Cancer Cell 2006;10:515–27.

30. Marcotte R, Sayad A, Brown KR, Sanchez-Garcia F, Reimand J, Haider M,et al. Functional genomic landscape of human breast cancer drivers,vulnerabilities, and resistance. Cell 2016;164:293–309.

31. Kenny PA, Lee GY,Myers CA, Neve RM, Semeiks JR, Spellman PT, et al. Themorphologies of breast cancer cell lines in three-dimensional assayscorrelate with their profiles of gene expression. Mol Oncol 2007;1:84–96.

32. Stanicka J, Rieger L, O'Shea S, CoxO, ColemanM,O'Flanagan C, et al. FES-related tyrosine kinase activates the insulin-like growth factor-1 receptor atsites of cell adhesion. Oncogene 2018;37:3131–50.

33. Buckley NE, Haddock P, De Matos Simoes R, Parkes E, Irwin G, Emmert-Streib F, et al. A BRCA1 deficient, NF-kB driven immune signal predictsgood outcome in triple negative breast cancer. Oncotarget 2016;7:19884–96.

34. Goldhirsch A, Wood WC, Coates AS, Gelber RD, Thurlimann B, Senn HJ,et al. Strategies for subtypes—dealing with the diversity of breast cancer:highlights of the St. Gallen International Expert Consensus on the PrimaryTherapy of Early Breast Cancer 2011. Ann Oncol 2011;22:1736–47.

35. Michaut M, Chin SF, Majewski I, Severson TM, Bismeijer T, de Koning L,et al. Integration of genomic, transcriptomic and proteomic data identifiestwo biologically distinct subtypes of invasive lobular breast cancer. Sci Rep2016;6:18517.

36. Riabov V, Yin S, Song B, Avdic A, Schledzewski K, Ovsiy I, et al.Stabilin-1 is expressed in human breast cancer and supports tumorgrowth in mammary adenocarcinoma mouse model. Oncotarget 2016;7:31097–110.

37. Graff JR, Herman JG, Lapidus RG, Chopra H, Xu R, Jarrard DF, et al.E-cadherin expression is silenced by DNA hypermethylation in humanbreast and prostate carcinomas. Cancer Res 1995;55:5195–9.

38. Taurin S, Sandbo N, Qin Y, Browning D, Dulin NO. Phosphorylation ofbeta-catenin by cyclic AMP-dependent protein kinase. J Biol Chem 2006;281:9971–6.

39. Zhu G,Wang Y, Huang B, Liang J, Ding Y, Xu A, et al. A Rac1/PAK1 cascadecontrols beta-catenin activation in colon cancer cells. Oncogene 2012;31:1001–12.

40. Yan P, Fu J, Qu Z, Li S, Tanaka T, Grusby MJ, et al. PDLIM2 suppresseshuman T-cell leukemia virus type I Tax-mediated tumorigenesis by target-ing Tax into the nuclear matrix for proteasomal degradation. Blood 2009;113:4370–80.

41. Khramtsov AI, Khramtsova GF, Tretiakova M, Huo D, Olopade OI, GossKH. Wnt/beta-catenin pathway activation is enriched in basal-likebreast cancers and predicts poor outcome. Am J Pathol 2010;176:2911–20.

42. Geyer FC, Lacroix-Triki M, Savage K, Arnedos M, Lambros MB, MacKay A,et al. beta-Catenin pathway activation in breast cancer is associatedwith triple-negative phenotype but not with CTNNB1 mutation.Mod Pathol 2011;24:209–31.

43. DeyN, Barwick BG,MorenoCS,Ordanic-KodaniM, Chen Z,Oprea-Ilies G,et al. Wnt signaling in triple negative breast cancer is associated withmetastasis. BMC Cancer 2013;13:537.

44. Pohl SG, Brook N, Agostino M, Arfuso F, Kumar AP, Dharmarajan A. Wntsignaling in triple-negative breast cancer. Oncogenesis 2017;6:e310.

45. Lakis S, Dimoudis S, Kotoula V, Alexopoulou Z, Kostopoulos I, Koletsa T,et al. Interaction between b-catenin and EGFR expression by immunohis-tochemistry identifies prognostic subgroups in early high-risk triple-neg-ative breast cancer. Anticancer Res 2016;36:2365–78.

46. Shen T, Zhang K, Siegal GP,Wei S. Prognostic value of E-cadherin and beta-catenin in triple-negative breast cancer. Am J Clin Pathol 2016;146:603–10.

47. Gayrard C, Bernaudin C, Dejardin T, Seiler C, Borghi N. Src- and confine-ment-dependent FAK activation causes E-cadherin relaxation and beta-catenin activity. J Cell Biol 2018;217:1063–77.






2019;79:2619-2633. Published OnlineFirst March 18, 2019.Cancer Res Orla T. Cox, Shelley J. Edmunds, Katja Simon-Keller, et al. Triple-Negative Breast Cancer

-Catenin Activity inβPDLIM2 Is a Marker of Adhesion and

Updated version

10.1158/0008-5472.CAN-18-2787doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2019/03/16/0008-5472.CAN-18-2787.DC1

Access the most recent supplemental material at:

Overview

Visual

http://cancerres.aacrjournals.org/content/79/10/2619/F1.large.jpgA diagrammatic summary of the major findings and biological implications:

Cited articles

http://cancerres.aacrjournals.org/content/79/10/2619.full#ref-list-1

This article cites 46 articles, 11 of which you can access for free 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/79/10/2619To 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-18-2787

http://cancerres.aacrjournals.org/content/suppl/2019/03/16/0008-5472.CAN-18-2787.DC1

http://cancerres.aacrjournals.org/content/79/10/2619/F1.large.jpg

http://cancerres.aacrjournals.org/content/79/10/2619.full#ref-list-1

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/79/10/2619


pdlim2 is a marker of adhesion and b-catenin activity in ... · orla t. cox1, shelley j. edmunds1,...

Documents