EMT Transition Alters Interstitial Fluid Flow Induced ... · expressing cells, IFF-mediated invasion is chemokine receptor– independent and requires only p110a activation. To test

Oncogenes and Tumor Suppressors

EMT Transition Alters Interstitial FluidFlow–Induced Signaling in ERBB2-PositiveBreast Cancer CellsAlimatou M. Tchafa1, Mi Ta1, Mauricio J. Reginato2, and Adrian C. Shieh1

Abstract

A variety of biophysical forces are altered in the tumormicroenvironment (TME) and these forces can influence cancerprogression. One such force is interstitial fluid flow (IFF)—themovement of fluid through the tissue matrix. IFF was previ-ously shown to induce invasion of cancer cells, but the acti-vated signaling cascades remain poorly understood. Here, it isdemonstrated that IFF induces invasion of ERBB2/HER2-expressing breast cancer cells via activation of phosphoinosi-tide-3-kinase (PI3K). In constitutively activate ERBB2-expres-sing cells that have undergone epithelial-to-mesenchymal tran-sition (EMT), IFF-mediated invasion requires the chemokinereceptor CXCR4, a gradient of its ligand CXCL12, and activity ofthe PI3K catalytic subunits p110a and b. In wild-type ERBB2-

expressing cells, IFF-mediated invasion is chemokine receptor–independent and requires only p110a activation. To testwhether cells undergoing EMT alter their signaling responseto IFF, TGFb1 was used to induce EMT in wild-type ERBB2-expressing cells, resulting in IFF-induced invasion dependenton CXCR4 and p110b.

Implications: This study identifies a novel signaling mechanismfor interstitial flow–induced invasion of ERBB2-expressing breastcancer cells, one that depends on EMT and acts through aCXCR4–PI3K pathway. These findings suggest that the response of cancercells to interstitial flow depends on EMT status and malignancy.Mol Cancer Res; 13(4); 755–64. �2015 AACR.

IntroductionBreast cancer is the most commonly reported form of cancer in

women worldwide (excluding non-melanoma skin cancers) andis responsible for 13.7% of cancer-related deaths in women (1).The mortality rates associated with this and most other solidcancers is a consequence of cancer cell invasion into the surround-ing stroma and metastasis to distant organs. Although manyfactors responsible for invasion andmetastasis are still unknown,the influence of the breast tumor microenvironment (TME),which consists of stromal cells, extracellular matrix, and solublefactors, on cancer progression is well established (2).

While their importance in cancer progression has only beenrealized relatively recently, various biophysical factors are signif-icantly altered in the TME. Changes inmatrix density (3), stiffness(4), organization (5), interstitial fluid pressure (6), and flow (7)

are all biophysical consequences of tumor growth that affect geneexpression, proliferation, differentiation, and invasion. Intersti-tial fluid flow (IFF), the movement of fluid through the tissuematrix, is elevated in tumors compared with normal tissue (8).This elevated IFF is driven by steep fluid pressure gradients at thetumormargin (9).Measured levels of IFF range from0.1 to 1mm/sin normal tissue, and 0.1 to 55 mm/s in tumor tissue (10–13).Increased IFF is directly linked to lymph node metastasis inhuman cervical carcinoma, and increased cell motility and inva-sion in vitro in breast cancer, glioma, melanoma, and renalcarcinoma cells (11, 14–17). Although increased IFF appears todrive cancer cell invasion, the underlying mechanisms of IFFmechanotransduction are not fully understood.

To date, different studies have proposed mechanisms toexplain the strong tumor and stromal cell migration inducedby elevated IFF, though most have focused on highly invasivecells. In fibroblasts, IFF induced myofibroblast differentiationand secretion of matrix metalloproteinases (MMP), thus indi-rectly aiding tumor cell migration through matrix remodeling(16, 18). In metastatic cells, IFF altered extracellular chemokinegradients to accelerate tumor cell invasion via CXCR4- andCCR7-dependent chemotaxis (14, 15), and increased MMPactivity via shear stress sensing through the glycocalyx (17).More recently, IFF was shown to induce invasion against thedirection of flow due to fluid drag forces and activation ofintegrins on the upstream side of the cell (19). Although weknow some of its effects on invasive cells, the specific pathwaystriggered by IFF are still unclear, especially those that play acrucial role in the early stages of invasion.

At these early stages of invasion, the process of epithelial-to-mesenchymal transition (EMT) can play an important role. Dur-ing EMT, epithelial cells lose their distinctive morphology and

1School of Biomedical Engineering, Science and Health Systems,Drexel University, Philadelphia, Pennsylvania. 2Department of Bio-chemistry and Molecular Biology, Drexel University College of Medi-cine, Philadelphia, Pennsylvania.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Authors: Adrian C. Shieh, School of Biomedical Engineering,Science and Health Systems, Drexel University, 3141 Chestnut Street, Bossone710, Philadelphia, PA 19104. Phone: 215-895-0358; Fax: 215-895-4983; E-mail:[email protected]; and Mauricio J. Reginato, Department of Biochem-istry and Molecular Biology, Drexel University College of Medicine, 245 North15th Street, Philadelphia, PA 19102. Phone: 215-762-3554; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-14-0471

�2015 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 755

on December 5, 2020. © 2015 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst January 7, 2015; DOI: 10.1158/1541-7786.MCR-14-0471

http://mcr.aacrjournals.org/

molecular characteristics and acquire a mesenchymal phenotypethat is often associated with invasion. This phenomenon,although indispensable during embryogenesis and tissue remo-deling, also occurs during tumor development and likely plays arole in tumor progression (20). EMT is characterized by the loss ofepithelial basoapical polarity and cell–cell junctions, spindle-likecellular morphology, and activation of transcription factors thatlead to increased invasion (21, 22). The induction of EMT duringtumor development is a consequence of external cues in the TME,like changes in levels of cytokines and growth factors (22, 23). Todate, the relationship between EMT and how cancer cells respondto IFF has not been investigated.

Thus, the overall goal of this study was to elucidate the molec-ular pathways activated by IFF in invasive and noninvasive breastcancer cells and determine whether EMT alters IFF signaling andresponse. Here, we show for the first time that IFF causes anincrease in tumor cell invasion via activation of phosphoinosi-tide-3-kinase (PI3K) in ERBB2-positive breast cancer cells. Fur-thermore, we demonstrate that pre- and post-EMT cancer cellsrespond to IFF via different PI3K-dependent pathways.

Materials and MethodsCell culture

MCF10A and MCF10A-ERBB2 (NeuN/NeuT; ref. 24) weremaintained in DMEM/F12 supplemented with 5% donor horseserum (Atlanta Biologicals), 20 ng/mL epidermal growth factor(EGF; Peprotech), 10 mg/mL insulin (Sigma-Aldrich), 100 ng/mLcholera toxin (Enzo Life Sciences), 500 ng/mL hydrocortisone(Sigma-Aldrich), and 1% penicillin/streptomycin (Mediatech).SKBR3, a breast cancer cell line that naturally expresses ERBB2,were grown in McCoy's 5A (Mediatech) supplemented with 10%FBS and 1% penicillin/streptomycin. BT474 and MDA-MB-453,other breast cancer cell lines that naturally express ERBB2, weregrown in DMEM (Mediatech) supplemented with 10% FBS and1%penicillin/streptomycin. All cellsweremaintained in ahumid-ified environment at 37�C and 5% CO2.

3D invasion assayIFF was applied to cells as previously described (14, 16, 25), in

serum-free conditions (tominimize the effect of serum factors onthe signaling pathways studied). Briefly, cells were embedded in amatrix composed of 1.3mg/mL rat tail tendon collagen type I (BDBiosciences) and 1 mg/mL Matrigel basement membrane matrix(BD Biosciences) at a final concentration of 5 � 105 cells/mLinside culture inserts with 8-mm diameter pores (Millipore; Sup-plementary Fig. S1). For static (control) conditions, serum-freemedia levels inside and outside the insert were kept approximate-ly equal, resulting in a minimal hydrostatic pressure differenceacross the gel and no measurable interstitial flow. For IFF condi-tions, serum-free media were added under the insert (100 mL)and above the gel (650 mL). The hydrostatic pressure differencegenerated was approximately 1 mm Hg. After 24 hours ofeither physiologic interstitial flow (approximate flow velocity,�0.1 mm/s) or static conditions, invasion was measured bycounting the number of cells that invaded through the matrixand migrated across the porous membrane. Migrated cells on theunderside of the culture insert membranes were fixed in 4%paraformaldehyde in PBS for 30 minutes and permeabilizedwith 0.5% Triton X-100 in PBS for 10 minutes. Cells were thenstained with Alexa Fluor 488–conjugated phalloidin (6 U/mL;

Life Technologies) and 40,6-diamidino-2-phenylindole (DAPI;2 mg/mL; MP Biomedicals). Labeled cells were visualized onan epifluorescence microscope (Leica Microsystems). DAPI-stained nuclei at five randomly selected locations of eachmembrane were counted. F-actin staining was used to confirmthat positive DAPI stain corresponded to cells. Percentageinvasion was calculated from the following equation:

%Invasion¼100 � Average cell count �Membrane surface areað ÞImage �Number of cells seededð Þ

In some instances, percentage invasion was normalized tothe static control condition to allow for comparison betweenindependent experiments. To determine the functional role ofspecific proteins, pharmacologic inhibitors were used to blockspecific signaling proteins (see Table 1 for a list of inhibitorsand corresponding concentrations employed in this study). Allexperimental concentrations were confirmed to be noncyto-toxic and effective on our cell lines (Supplementary Fig. S6).When used, inhibitors were added to the collagen matrix andexperimental medium at the appropriate concentrations.Experiments were repeated at least twice with a minimumsample size of n ¼ 3 for each experiment (total minimumsample size n ¼ 6).

TGFb1 treatmentEMT was induced in NeuN cells via treatment with recom-

binant human TGFb1 (Peprotech). An optimal concentrationof TGFb1 was first determined by treating NeuN cells at varyingconcentrations of TGFb1 (3, 5, 10, 20, and 50 ng/mL) for 6days. A concentration of 20 ng/mL was found to most effec-tively induce EMT (low E-cadherin expression, high vimentinexpression, elongated morphology). NeuN cells were treatedwith this concentration in full media for 6 days, with a mediachange after 3 days.

Western blot analysisFollowing the three-dimensional (3D) invasion assay, total and

phospho-protein levels were determined by Western blot analy-sis. Cells were isolated from the matrix using 2.5 mg/mL colla-genase D (Roche) for 30 minutes. The resulting solution wascentrifuged, and the pellet was washed with PBS and resuspendedin RIPA lysis buffer (150mmol/L NaCl, 1%NP40, 0.5%DOC, 50mmol/L Tris–HCl at pH 8, 0.1% SDS, 10% glycerol, 5 mmol/LEDTA, 20 mmol/L NaF, and 1 mmol/L Na3VO4). Lysates werecleared via centrifugation at 16,000 x g for 20 minutes at 4�C.Antibodies used were: rabbit polyclonal anti-phospho PI3K andtotal PI3K (1:500; Cell Signaling Technology), rabbit polyclonalanti-b-actin (1:3,000; Cell Signaling Technology), rabbit mono-clonal anti-vimentin (1:1,000; Abcam), rabbit monoclonal anti-E-cadherin (1:1,000; Abcam), rabbit polyclonal anti-phosphoand total CXCR4 (1:500; Abcam), rabbit polyclonal anti-p110a

Table 1. List of inhibitors, their targets, and concentrations used

Name Protein target Concentration (mmol/L)

LY294002 PI3K 10, 50PIK75 p110a 0.01TGX221 p110b 0.1AMD3100 CXCR4 12.6WZ811 CXCR4 0.1Pertussis toxin Gai 0.1

Tchafa et al.

Mol Cancer Res; 13(4) April 2015 Molecular Cancer Research756




and b (1:1,000; Cell Signaling Technology), and goat anti-rabbitHRP (1:10,000; Abcam).

ImmunofluorescenceFor visualization of EMTmarkers, cells were plated onto 8-well

chamber slides at 10,000 cells per well. After 16 hours, cells werefixed in 4% paraformaldehyde in PBS for 30 minutes and per-meabilized with 0.5% Triton X-100 in PBS for 10 minutes.This was followed by 1 hour in blocking solution (1% BSA,0.1% goat serum). Cells were stained with rabbit monoclonalanti-vimentin (1:100; Abcam) for 16 hours, then washed in PBSthree times. Alexa Fluor 488–conjugated phalloidin (6 U/mL),DAPI (2 mg/mL), and Alexa Fluor 555–conjugated anti-rabbit IgGwas incubated on the cells for 16 hours, then washed in PBSthree times. Fluorescence was visualized on an epifluorescencemicroscope (Leica Microsystems).

Microarray and quantitative RT-PCRChanges in gene expression in NeuN and NeuT cells were

identified by microarray analysis. Total RNA was extracted fromcells embedded in 3D gels (6 biological replicates per sample)using the RNeasy Mini Kit (Qiagen). RNA was submitted to theWistar Institute genomics core facility (Philadelphia, Pennsylva-nia) for amplification. Epicentre TargetAmpNano Labeling Kit forIllumina Expression BeadChip was used at 100-ng RNA andhybridized on an Illumina HT12v4 chip. Data were analyzedwith PARTEK Genomics Suite and Database for Annotation,Visualization and Integrated Discovery (DAVID) online tools.

Quantitative RT-PCR (qRT-PCR) was performed on cDNAreversed transcribed using QuantiTech Reverse Transcription kitand QuantiTech SYBR Green PCR kit (Qiagen) on three of themost altered genes within the microarray data to validate theresults. Primers for human interleukin 6, desmoplakin, and integ-rin subunit a6 (Quantitech; Qiagen) were used to validate themicroarray analysis and quantify mRNA expression differences.

Data and statistical analysisData are expressed as mean � standard error of the mean.

Differences among conditionswere tested using the Student t tests(for two groups) or two-way analysis of variance (ANOVA; forthree or more groups) using GraphPad Prism. When ANOVAidentified a significant difference, a Bonferroni post-test was usedformultiple comparisons.Differenceswere accepted as significantat P < 0.05.

ResultsIFF induces invasion of breast cancer cells through PI3Kactivation

To model noninvasive and invasive cells, we used a previouslydeveloped model of breast cancer based on MCF10A humanmammary epithelial cells engineered to overexpress ERBB2(24): MCF10A cells retrovirally transduced with either wild-typeERBB2 (NeuN)or a constitutively activemutant ofERBB2 (NeuT).When cultured in 3D conditions, NeuN cells exhibit behavior ofpreinvasive ERBB2-positive ductal carcinoma in situ (DCIS;ref. 24), while NeuT cells behave like ERBB2-positive invasiveductal carcinoma (IDC; ref. 26). Using a 3D invasion assay wheresingle cells are embedded in stroma-like matrix, we measuredinvasion by counting the number of transmigrated cells (Supple-mentary Fig. S1). We observed an increase in invasion of morethan 2-fold in response to 0.1 mm/s IFF in both NeuN and NeuTcells (Fig. 1A).

To determine the signaling pathways involved in the observedIFF-induced invasion,we examinedpathways known tobe impor-tant in invasion of HER2-positive breast cancer cells (HER2,MAPK, and PI3K). We isolated cells from gels after the 3Dinvasion assay and harvested protein from the resulting cells.Western blot analysis revealed an increase in phosphorylated p85(the regulatory subunit of class I PI3Ks) in both NeuN andNeuT cells treated with IFF (Fig. 1B). To confirm that our results

A B

C

StaticFlow

NeuN

4

3

2

1

0NeuT

NeuN NeuTFlow

p-p85

t-p85

% In

vasi

on

StaticFlow

NeuN

4

3

2

1

0LY294002

NeuT

% In

vasi

on

β-Actin

Figure 1.IFF-induced invasion occurs throughp85 activation. A, IFF increasesinvasion of NeuN and NeuT.Percentage invaded cells after 24hours in 3D IFF assay. All values aremean � SEM. The Student t test(�� , P < 0.01; �� , P < 0.001), n > 12. B,IFF induces activation of PI3K in NeuNand NeuT. Representative Westernblot analysis of phosphorylation ofPI3K regulatory subunit (p85) after 24hours of static or flow conditions in3D flow invasion assay; b-actin wasused as a loading control. C, PI3Kactivity is necessary for IFF-inducedinvasion in both cell lines. IFF-inducedinvasion is decreased in the presenceof 10 mmol/L LY294002, a pan PI3Kinhibitor. All values are mean � SEM.The Student t test and two-wayANOVA (�� , P < 0.01; �� , P < 0.001),n > 6.

EMT and Interstitial Flow in Breast Cancer Invasion

www.aacrjournals.org Mol Cancer Res; 13(4) April 2015 757




were not specific to the engineered MCF10A cell lines, exposureof other breast cancer cells overexpressing ERBB2, includingSKBR3, BT474, and MDA-MB-453, to IFF also increased invasionand PI3K phosphorylation compared with static conditions(Fig. 2A–C).

To test whether PI3K activity was required for IFF-inducedinvasion, we treated cells with the PI3K inhibitor LY294002(a pan-PI3K inhibitor). Treatment with LY294002 decreasedIFF-induced invasion in NeuN, NeuT (Fig. 1D), SKBR3, andMDA-MB-453 cells (Fig. 2D). These data suggest that IFFincreases invasion in ERBB2-overexpressing breast cancer cellsvia activation and subsequent downstream signaling throughPI3K.

p110b modulates response to IFF in NeuT but not NeuN cellsClass I PI3Ks are present in cells as heterodimers composed of a

regulatory (typically p85 or p55) and catalytic subunit (p110;ref. 27). The p110 catalytic subunit is responsible for propagatingdownstream signaling through its kinase activity (28). Usinginhibitors that specifically target the b (TGX-221) or a (PIK-75)p110 catalytic subunit of PI3K, we observed that in NeuNcells, only the p110a inhibitor blocked IFF-mediated invasion(Fig. 3A). However, in NeuT cells, inhibiting both p110a andp110b decreased IFF-induced invasion (Fig. 3B). NeuN andNeuTcells expressed similar protein levels of p110a and p110b (Fig.3C), thus ruling out altered levels of these isoforms as the reasonfor this difference.

CXCR4 regulates PI3K activity in response to IFF in NeuT cellsBecause previous studies have linked p110a to receptor

tyrosine kinase signaling and p110b to G protein–coupledreceptors (GPCR; refs. 29–31), we investigated the role ofchemokine receptors (a family of GPCRs) in the IFF-inducedinvasion of NeuT cells. CXCR4 is a known modulator of breastcancer invasion (32, 33) and IFF-induced glioma cell invasionwas previously shown to be dependent on CXCR4 (15). Treat-

ing NeuT cells with antagonists of the chemokine receptorCXCR4 (AMD3100 or WZ811) significantly reduced IFF-induced invasion (Fig. 4A). In contrast, AMD3100 and pertus-sis toxin (Gai subunit inhibitor of all GPCRs; ref. 34) had noeffect on IFF-mediated invasion of NeuN cells (Fig. 4B), sug-gesting that chemokine receptors, including CXCR4, are notinvolved in the molecular pathway activated by IFF in NeuNcells. CXCR4 inhibitors also had no effect on IFF-mediatedinvasion in the noninvasive HER2-positive cell line SKBR3(data not shown).

To test whether CXCR4 signaling was associated with PI3Kactivation, we examined the effects of treating NeuT and NeuNcells with AMD3100 on p85 phosphorylation during IFF-medi-ated invasion. IFF-induced PI3K phosphorylation wasdecreased in NeuTs when the cells were treated with AMD3100but phosphorylation was only weakly inhibited in NeuN (Fig.4C). Interestingly, CXCR4 and CXCL12 protein levels weresimilar in NeuN and NeuT cells (Supplementary Fig. S2).CXCR4 and its ligand CXCL12 have been previously implicatedin IFF-mediated glioma invasion by autologous chemotaxis(15). In autologous chemotaxis, a transcellular gradient isgenerated by the combination of autocrine chemokine secre-tion and interstitial flow (35). To determine whether a CXCL12gradient was necessary for IFF-induced invasion of NeuT cells,exogenous CXCL12 was added to the surrounding media andmatrix at a uniform concentration of 80 ng/mL (10 nmol/L). Atthis concentration, the cells' CXCR4 receptors should beapproaching saturation (KD, �14 nmol/L; ref. 36), effectivelyblocking gradient sensing. Under these conditions, NeuT cellsno longer responded to IFF (CXCL12 condition) comparedwith a 2-fold increase in invasion in the control flow condition(Fig. 4D). When NeuN cells were subjected to IFF in thepresence of exogenous CXCL12, the uniform chemokine con-centration did reduce invasion, but, the cells still exhibited asignificant IFF-induced response (Supplementary Fig. S3), incontrast to NeuT cells. These data suggest that NeuT cells, unlike

A B

C D

SKBR3 BT474

Static

4

3

2

1

0

% In

vasi

on

3

2

1

0

% In

vasi

on

Flow Static Flow

Flow

StaticFlow

Flow

p-p85

β-Actin

MDA-MD-453

Static

2

1

0

% In

vasi

on

4

6

2

0LY294002

SKBR3 MDA-MB-453

% In

vasi

on

Flow

Flow

p-p85

β-Actin

p-p85

β-Actin Figure 2.IFF-induced invasion through PI3K isalso observed in other HER2-positivecell lines. IFF induces invasion ofSKBR3 (A), BT474 (B), MDA-MB-453(C) coupled with PI3K activation.Percentage invaded cells after24 hours in 3D IFF assay andrepresentative Western blot analysisof phosphorylation of PI3K. D, SKBR3and MDA-MB-453 show decreasedIFF-induced invasion in the presenceof 50 mmol/L LY294002, a pan PI3Kinhibitor. All values are mean � SEM.The Student t test and two-wayANOVA (�� , P < 0.01; �� , P < 0.001),15 > n > 6.

Tchafa et al.





NeuN cells, require a gradient of CXCL12 for IFF-inducedactivation of PI3K and invasion, consistent with CXCL12-dependent autologous chemotaxis.

TGFb1-induced EMT alters the mechanism of interstitial flow–induced invasion in NeuN cells

Having identified separate IFF-induced pathways in NeuN andNeuT cells, our next goal was to determine some of the funda-mental differences between these cells that are responsible for theobserved differences in IFF-induced invasion. NeuN cells adoptedmore epithelial-like shapes on 2D cell culture plates and invadedin clumps of cells similar to their parental MCF10A counterparts(Supplementary Fig. S4). However, NeuT invaded as single cellsand were characterized by a spindle-like morphology. Thus, wehypothesized that NeuT cells may have undergone EMT and thismay explain the difference in IFF response. We probed for proteinexpression of epithelial andmesenchymal markers, and observedloss of E-cadherin and increased levels of vimentin in NeuT cellswhen compared with NeuN cells (Fig. 5A and B). Both of theseproteins are known markers of EMT (37). To examine genedifferences between NeuN and NeuT cells further, we profiledgene expression using microarray analysis. Table 2 lists a numberof known EMT-associated genes that were significantly differentwhen comparing NeuT and NeuN mRNA expression (log2-trans-formed fold change is represented). These data have been depos-ited in NCBI's Gene Expression Omnibus (GEO; ref. 38) and areaccessible through GEO Series accession number GSE64670(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc¼GSE64670).The expressions of all these genes agree with published EMT-associated data (39, 40). We validated the microarray resultwith qRT-PCR of three of the identified genes (SupplementaryFig. S5). These results supported the hypothesis that NeuT cellsdisplay a more mesenchymal-like phenotype when comparedwith NeuN cells.

To determine whether EMT contributes to the altered IFFresponse, EMT was induced in NeuN cells using the known

A

C

B

NeuN NeuT2.5

2.0

1.5

1.0

0.5

0.0

StaticControl

No

rmal

ized

inva

sio

nPIK75 TGX221

2.5

2.0

1.5

1.0

0.5

0.0Control

No

rmal

ized

inva

sio

n

PIK75 TGX221

Flow

NeuN NeuT

p110 α

p110 β

β-Actin

Figure 3.Different p110 isoforms are necessaryfor IFF-induced invasion in NeuN andNeuT. A and B, both p110a and p110bare necessary for flow response inNeuT only. Cellular response to P110isoforms specific inhibitors PIK75 andTGX221 (p110a and p110b inhibitors,respectively) in NeuN (A) and NeuT(B). All values are mean � SEM. TheStudent t test (� , P < 0.05), n ¼ 6.C, both cells express similar levels ofp110 isoforms. RepresentativeWestern blot analysis of p110a and b inNeuN and NeuT; b-actin was used as aloading control.

NeuT

NeuTAMD3100 AMD3100 ControlControl

NeuN

NeuT

Control

4

3

2

1

0AMD3100 WZ811

StaticFlow

Flow

p-p85

β-Actin

A

C D

B

% In

vasi

on

Control

Control CXCL12

4

3

2

1

0

0

1

2

3

AMD3100 PTX

% In

vasi

on

No

rmal

izw

d in

vasi

on

NeuN

Figure 4.CXCR4 activity is required for IFF-induced invasion in NeuT but not NeuN.A, changes in invasion in the presence or absence of CXCR4 inhibitors(AMD3100 and WZ811) in NeuT. B, AMD3100 and pertussis toxin are notnecessary for IFF-induced invasion in NeuN. C, AMD3100 inhibits IFF-inducedPI3K activation inNeuTbut not inNeuN. RepresentativeWestern blot analysisof phospho-p85 after 24 hours in 3D invasion assay with and without theinhibitor in NeuN and NeuT; b-actin was used as a loading control. D, changesto flow-induced invasion when exogenous CXCL12 is added to surroundingmatrix andmedia. NeuT respond to fluid flowonly in the presence of a CXCL12gradient. All values are mean � SEM. The Student t test (� , P < 0.05;�� , P < 0.01); n � 6.






EMT-inducing growth factor TGFb1 (22). The cellswere treated for6 days with 20 ng/mL TGFb1 and EMT induction was confirmedusing a series of protein, morphological, and functional markers.NeuNEMT cells exhibited amoremarked spindle-likemorphologyconsistent with EMT (Supplementary Fig. S7A). NeuNEMT cellsshow lower levels of E-cadherin and higher levels of vimentinwhen compared with their control NeuN counterparts (Fig. 5Cand D), as well as increased levels of fibronectin (SupplementaryFig. S7B). These are all well-established EMT markers (22). Asexpected, the NeuN cells induced to undergo EMT had higherlevels of basal invasion (Fig. 6B, Supplementary Fig. S7C). In linewith our previous findings in NeuT cells, we observe that IFF-induced invasion in NeuNEMT cells was inhibited by both per-tussis toxin and CXCR4 (Fig. 6A and B). In addition, IFF-inducedPI3K phosphorylation in the presence of AMD3100 was reduced(Fig. 6C) in NeuNEMT cells, similar to NeuT cells (Fig. 4C).Consistent with this result, treatment of NeuNEMT cells withp110b inhibitor (TGX-221) reduced IFF-mediated invasion, in

contrast to regular NeuN (Fig. 6D), but similar to NeuT cells(Fig. 3B). Thus, these data support the hypothesis that ERBB2-expressing breast cancer cells that have undergone EMT invadein response to IFF through a CXCR4- and p110b-dependentmechanism.

DiscussionIn this study, we examined the role of key intracellular signaling

pathways thatmediate IFF-induced invasion in ERBB2-expressingbreast cancer cells. These experiments were performed underserum-free conditions to decrease confounding factors associatedwith serum components and to reduce basal cellular activities. Inall breast cancer cells tested, we found that IFF induced activationof PI3K as measured by p85 phosphorylation. In addition, thesebreast cancer cells required PI3K activity for IFF-mediated inva-sion as inhibition of PI3K and the p110a catalytic subunitblocked IFF-mediated invasion. However, cells that have under-gone EMT (NeuT andNeuNEMT cells) required additional signals,including activation of chemokine receptors. Indeed, we showedthat inhibitors of CXCR4 and p110b, a p110 isoform that isGPCR-regulated (29–31), specifically blocked IFF-mediated inva-sion in breast cancer cells that have undergone EMT. Our datasuggest a model where, as cells undergo EMT, the signalingpathways activated by IFF to induce invasion change significantly(Fig. 7).

The PI3K pathway plays a central role in numerous cellularprocesses crucial for cancer progression. This pathway has beenimplicated in invasion, proliferation, transformation, and cellsurvival (41, 42). Mutations or alterations of this pathway are themost frequent in all human cancers,making it an important targetof cancer therapeutics (41). This work supports the importance oftargeting the PI3K pathway to curtail breast cancer invasion,especially in those with prominent IFF. PI3K is present in cells

A B

C D

E-cadherin

NeuN NeuT

NeuN NeuT

NeuN

NeuN

Vimentin

Vimentin

NeuT

NeuN NeuNEMTNeuNEMT

Vimentin

β-Actin

E-cadherin

β-Actin

β-Actin

Figure 5.NeuT cells have undergone EMTsimilar to TGFb1-dependent EMTinduction in NeuN. A and B, comparedwith NeuN, NeuT cells express lowerlevels of epithelial markers and higherlevels of mesenchymal markers.Representative Western blot analysisof E-cadherin and vimentin (A);b-actin was used as a loadingcontrol. B, representativeimmunofluorescence of vimentin fromcells plated on collagen I–coatedcoverslip. C, compared with control,NeuNEMT cells express lower levels ofepithelial markers. RepresentativeWestern blot analysis of E-cadherin;b-actin was used as a loading control.D, compared with control, NeuNEMT

cells express higher levels of vimentin.Representative immunofluorescenceof vimentin from cells attached to theunderside of membrane after IFFinvasion assay. White bars represent50 mm.

Table 2. List of EMT genes differentially expressed between NeuT and NeuN

EMT upregulated genesGene symbol (name) NeuT/NeuN P

CDH2 (N-cadherin) þ2.7 6.8 � 10�8

GNG11 (guanine nucleotide-binding protein) þ2.9 7.2 � 10�11

IGFBP4 (insulin-like growth factor-binding protein 4) þ1.9 1.2 � 10�7

STEAP1 (metalloreductase STEAP1) þ3.4 1.1 � 10�9

VCAN (versican) þ1.5 1.6 � 10�8

WNT5A (Protein Wnt-5a) þ1.6 2.0 � 10�8

EMT downregulated genesGene symbol (name) NeuT/NeuN PPPPDE2 (desumoylating isopeptidase 1) �1.4 4.0 � 10�7

CDH1 (E-cadherin) �5.4 8.4 � 10�9

NOTE: The genes below are published EMT genes whose expression profileagrees with our dataset. They were identified in our microarray dataset with asignificant log2 transformed expression ratio (NeuT/NeuN) of at least �1.4.

Tchafa et al.





as a heterodimer complex made up of a regulatory and catalyticsubunit. Cancer-specificmutations havebeenobserved in the fourcatalytic p110 isoforms (a, b, d, and g) of the class I kinase (42).IFF led to phosphorylation of p85 in all cells tested, suggestingthat class I PI3K is central to signals activated by IFF. However, ascancer cells underwent EMT, we found changes in p110 subunit

dependency. Only breast cancer cells expressing a mesenchymalphenotype (determined by E-cadherin and vimentin expression)required p110b activity. As mentioned above, p110a has beenlinked to downstream activity of receptor tyrosine kinases, whilep110b has been linked to GPCRs (29–31). This is consistent withour data that EMT cells required the GPCR and chemokine

A 10

NeuN NeuNNeuNEMT

NeuN NeuNEMT

NeuN NeuNEMT

NeuN

NeuNEMT

NeuNEMT

8

6

4

% In

vasi

on

1.5

1.0

0.5

0.0

No

rmal

ized

p-P

I3K

% In

vasi

on

0Control

10

8

6

4

22

0Control AMD3100

Control ControlAMD3100

Control AMD3100

10

8

6

4

2

0

% In

vasi

on

Control TGX221 TGX221Control

AMD3100

Control AMD3100PTX Control

n.s.n.s.

n.s.

StaticFlow

PTX

B

DC

p-p85

Flow

β-Actin

Figure 6.TGFb1-induced EMT in NeuN leads toCXCR4- and p110b-dependent IFF-induced invasion. A, changes ininvasion in the presence or absence ofpertussis toxin. Inhibiting GPCRactivity in NeuNEMT blocks IFF-induced invasion but not in controlNeuN. B, changes in invasion in thepresence or absence of CXCR4inhibitors, AMD3100. InhibitingCXCR4activity in NeuNEMT blocks IFF-induced invasion but not in controlNeuN. C, AMD3100 decreases IFF-induced PI3K activation in NeuNEMT

but not in control. RepresentativeWestern blot analysis of phospho-p85after 24 hours in 3D invasion assaywith and without the inhibitor; b-actinwas used as a loading control. Theband densities were measured andnormalized to their respective b-actin.D, changes in invasion in the presenceor absence of p110b inhibitors, TGX221.Inhibiting CXCR4 activity in NeuNEMT

blocks IFF-induced invasion but not incontrol NeuN. All values are mean �SEM. The Student t test (� , P < 0.05;�� , P < 0.001); n � 6.

NeuN(preinvasive cells)

INVASION

EMT

Flow

Flow

AMD3100WZ811

PIK75

LY294002

TGX221

LY294002

PIK75

??? CXCR4

p85 p85 p85

p110α p110α p110β

CXCL12gradient

INVASION

NeuT, NeuNEMT

(invasive cells)

Figure 7.Proposedmechanism of IFF in ERBB2-positive breast cancer. Separatesignaling pathways are activated inresponse to IFF in our cell model butthey all converge at class I PI3K. Inpreinvasive cells (NeuN), IFF activatesPI3K through an unknown receptorleading to increased invasion via thep110a catalytic subunit. In invasivecells, similar to those that haveundergone EMT (NeuT, NeuNEMT), IFFactivates PI3K through CXCR4, whichleads to increased invasion via bothp110 catalytic subunits. This occursbecause the combination of autocrineCXCL12 secretion and IFF creates achemokine gradient around the cells,driving chemoinvasion in the directionof IFF via autologous chemotaxis (14).






receptor CXCR4 for IFF-induced invasion. Our findings suggestthat breast cancer cells undergoing EMT require both PI3K cata-lytic subunits p110a and p110b for IFF-mediated invasion andthat this pathway may be targeted to block invasion of aggressivebreast cancers. Because we could decrease IFF-induced invasion inEMT-like cells by specifically blocking either p110a or p110b,drugs targeting these p110 isoforms may be used to block inva-sionof aggressive breast cancers and thismayhave less toxic effectscompared with using pan-PI3K inhibitors.

We demonstrated that IFF-induced PI3K activation occursthrough separate upstream receptors, leading to downstreamsignaling in a p110 isoform-specific fashion (Fig. 7). IFF-spe-cific activation of PI3K in cells that have undergone EMT (NeuTand NeuNEMT) occurred via CXCR4. CXCR4 is a known prog-nosis marker for breast cancer metastasis (32) and has beenimplicated in breast cancer invasion (43). The role CXCR4 playsin IFF-induced glioma invasion was previously described as aresult of autologous chemotaxis through its ligand CXCL12(15). Here, as the cells secrete CXCL12, IFF creates a CXCL12gradient around the cells in the direction of flow, leading to ahigher chemokine concentration on the downstream side of thecells. We did not observe an increase in total or phosphorylatedCXCR4 due to IFF or a difference in CXCL12 levels in these cells(Supplementary Fig. S2). This suggests that IFF does not alterthe levels of these proteins but may alter how the cells migrateby creating a small but biologically relevant transcellular gra-dient (35). The fact that this phenomenon is only observed incells that have undergone EMT suggests that in addition toinducing invasion in preinvasive cells, IFF may play a role insustaining invasion as cells move through the stroma individ-ually. Invasive cells, like those which have undergone EMT,have been shown to be more sensitive to chemokine gradients(44, 45). This may allow them to home to specific organsproducing a chemoattractant (46, 47), but it may also facilitateIFF-induced invasion, as we have outlined here.

Although static invasion levels between NeuN and NeuT aresimilar, the mechanism by which these cells invade appears tobe different. NeuN cells have a tendency to invade collectivelyand NeuTs invade as single cells (Supplementary Fig. S4). Toidentify possible phenotypic differences responsible for theirdifferential responses to IFF, we compared NeuN and NeuTgene expression through microarray. More than 2,500 geneswere significantly differentially expressed between NeuN andNeuT. Among them, genes known to suppress the PI3K path-way, such as PTEN and PIK3IP1, were downregulated inNeuT cells. Genes associated with the ERBB2 pathway, suchas GRB2 and BCL2, were upregulated in NeuT cells. We furtherprofiled these cells' mRNA and protein levels and observedthat NeuT lacked key epithelial cell–associated proteins/genes(E-cadherin and integrin a6) and expressed higher levels ofmesenchymal markers (N-cadherin and vimentin; Fig. 5A andB), suggesting that EMT alters the mode of invasion of cells,and this leads to changes in the response of cells to IFF. To testthis idea specifically, we showed that induction of EMT inNeuN converted the IFF invasion response from chemokine-independent to chemokine receptor–dependent. Previous stud-ies have implicated ERBB2 expression and activation in EMT(31, 48, 49), therefore our findings could suggest that theobserved EMT-specific response to IFF was related to expressionof HER2. Although the HER2 pathway is constitutively active inNeuT, both cell lines overexpress similar levels of HER2 so this

alone could not explain the differential effects of IFF on thesecells.

One limitation of our study is the difference in pressuregradients between our model and that of advanced tumors. Thehydrostatic pressure difference generated by our 3D system wasapproximately 1 mm Hg, comparable with physiologic fluidpressure levels in normal tissue and early lesions (11) but thehydrostatic pressure differences present in advanced tumors aremuch larger (10, 11). Nevertheless, the in vitro collagen gelsused in our experiments are much more permeable than nor-mal and cancerous tissues, resulting in IFF velocities that arewithin a tumor-relevant range (12, 13). We cannot, however,discount the possibility that the larger absolute pressure mag-nitudes or pressure gradients associated in more advancedtumors may have an effect on cell behavior. Future studiesshould therefore consider how activation of the signaling path-ways described here is affected by fluid pressures and pressuregradients closer to those levels observed in vivo. In addition,although our data suggest potential relevance of our findings totreatment of HER2-positive DCIS, the study focuses on inva-sion of single cells embedded in a matrix. DCIS cells in vivo areorganized in 3D complex structures that are highly dependenton cell-to-cell interactions and basement membrane restric-tions (50). Their invasion through the stroma is also dependenton stroma-associated cells, which are well known to have bothpositive and negative influences on cancer progression (2). Inthe future, investigating the role of IFF on 3D DCIS structures ina more clinically relevant matrix will lead to better understand-ing of the role of IFF on breast cancer invasion. We must alsoconsider the contribution of other proposed mechanisms of IFFmechanotransduction, including mechanosensors such asintegrins and the tumor cell glycocalyx, which have both beenimplicated in IFF-induced cell invasion (17, 19). Examiningthese other mechanisms may help identify the upstream medi-ator of PI3K phosphorylation in NeuN cells. Further studiesshould also focus on identifying the implications of EMT inaltering IFF response in other cancers (e.g., melanoma, gliomaand renal cell carcinoma).

Taken together, our results demonstrate for the first time thatIFF increases invasion of cells via separatemechanisms dependingon the stage of the cancer cells. This is thefirst study to examine theeffect of IFF on separate stages of breast cancer, especially in thecontext of how it may influence the early steps of tumor progres-sion. We showed that IFF increases the invasion of different typesof human mammary breast cancer cells through activation ofPI3K. We further demonstrated that EMT alters how cells respondto IFF, engaging a CXCR4/CXCL12-dependent autologous che-motaxis mechanism that signals through p110b. Understandingthese cellular responses to IFF will increase our understanding ofhow biophysical forces interact with molecular factors to drivebreast cancer progression andmay help identify novel therapeuticregimens.

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

Authors' ContributionsConception and design: A.M. Tchafa, M.J. Reginato, A.C. ShiehDevelopment of methodology: A.M. TchafaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A.M. Tchafa, M. Ta

Tchafa et al.





Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A.M. Tchafa, M. Ta, M.J. Reginato, A.C. ShiehWriting, review, and/or revision of themanuscript:A.M. Tchafa,M.J. Reginato,A.C. ShiehAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): A.M. TchafaStudy supervision: A.C. Shieh

AcknowledgmentsThe authors thank Arpit Shah and Rawan Shraim for their technical support.

Grant SupportThis work was funded, in part, by a CURE grant from Drexel University

College of Medicine and the Pennsylvania Department of Health (to M.J.Reginato).

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 August 21, 2014; revisedNovember 20, 2014; acceptedDecember 7,2014; published OnlineFirst January 7, 2015.

References1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of

worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer2010;127:2893–917.

2. Cichon MA, Degnim AC, Visscher DW, Radisky DC. Microenvironmentalinfluences that drive progression from benign breast disease to invasivebreast cancer. J Mammary Gland Biol Neoplasia 2010;15:389–97.

3. Provenzano PP, Inman DR, Eliceiri KW, Knittel JG, Yan L, Rueden CT, et al.Collagen density promotes mammary tumor initiation and progression.BMC Med 2008;6:11.

4. Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, et al.Tensional homeostasis and the malignant phenotype. Cancer Cell2005;8:241–54.

5. Levental KR, Yu H, Kass L, Lakins JN, Egeblad M, Erler JT, et al. Matrixcrosslinking forces tumor progression by enhancing integrin signaling. Cell2009;139:891–906.

6. Ariffin AB, Forde PF, Jahangeer S, Soden DM, Hinchion J. Releasingpressure in tumors: what do we know so far and where do we go fromhere? A review. Cancer Res 2014;74:2655–62.

7. Butler TP, Grantham FH, Gullino PM. Bulk transfer of fluid in the inter-stitial compartment of mammary tumors. Cancer Res 1975;35:3084–8.

8. Shieh AC. Biomechanical forces shape the tumor microenvironment. AnnBiomed Eng 2011;39:1379–89.

9. WuM, Frieboes HB, McDougall SR, Chaplain MA, Cristini V, LowengrubJ. The effect of interstitial pressure on tumor growth: coupling withthe blood and lymphatic vascular systems. J Theor Biol 2013;320:131–51.

10. Dafni H, Landsman L, Schechter B, Kohen F, Neeman M. MRI andfluorescence microscopy of the acute vascular response to VEGF165:vasodilation, hyper-permeability and lymphatic uptake, followed by rapidinactivation of the growth factor. NMR Biomed 2002;15:120–31.

11. Hompland T, Ellingsen C, Ovrebo KM, Rofstad EK. Interstitial fluidpressure and associated lymph node metastasis revealed in tumors bydynamic contrast-enhanced MRI. Cancer Res 2012;72:4899–908.

12. Chary SR, Jain RK. Direct measurement of interstitial convection anddiffusion of albumin in normal and neoplastic tissues by fluorescencephotobleaching. Proc Natl Acad Sci U S A 1989;86:5385–9.

13. Dafni H, Israely T, Bhujwalla ZM, Benjamin LE, Neeman M. Overexpres-sion of vascular endothelial growth factor 165 drives peritumor interstitialconvection and induces lymphatic drain: magnetic resonance imaging,confocal microscopy, and histological tracking of triple-labeled albumin.Cancer Res 2002;62:6731–9.

14. Shields JD, Fleury ME, Yong C, Tomei AA, Randolph GJ, Swartz MA.Autologous chemotaxis as a mechanism of tumor cell homing to lympha-tics via interstitial flow and autocrine CCR7 signaling. Cancer Cell2007;11:526–38.

15. Munson JM, Bellamkonda RV, Swartz MA. Interstitial flow in a 3D micro-environment increases glioma invasion by a CXCR4-dependent mecha-nism. Cancer Res 2013;73:1536–46.

16. Shieh AC, Rozansky HA, Hinz B, Swartz MA. Tumor cell invasion ispromoted by interstitial flow-induced matrix priming by stromal fibro-blasts. Cancer Res 2011;71:790–800.

17. Qazi H, Palomino R, Shi ZD, Munn LL, Tarbell JM. Cancer cell glycocalyxmediates mechanotransduction and flow-regulated invasion. Integr Biol2013;5:1334–43.

18. Ng CP, Hinz B, Swartz MA. Interstitial fluid flow induces myofibroblastdifferentiation and collagen alignment in vitro. J Cell Sci 2005;118:4731–9.

19. Polacheck WJ, German AE, Mammoto A, Ingber DE, Kamm RD. Mechan-otransductionoffluid stresses governs 3D cellmigration. ProcNatl AcadSciU S A 2014;111:2447–52.

20. Hay ED. The mesenchymal cell, its role in the embryo, and the remarkablesignaling mechanisms that create it. Dev Dyn 2005;233:706–20.

21. Sleeman JP, Thiery JP. SnapShot: the epithelial–mesenchymal transition.Cell 2011;145:162 e1.

22. Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial–mes-enchymal transition. Nat Rev Mol Cell Biol 2014;15:178–96.

23. Said NA, Williams ED. Growth factors in induction of epithelial–mesenchymal transition and metastasis. Cells Tissues Organs 2011;193:85–97.

24. Haenssen KK, Caldwell SA, Shahriari KS, Jackson SR, Whelan KA, Klein-Szanto AJ, et al. ErbB2 requires integrin alpha5 for anoikis resistance via Srcregulationof receptor activity in humanmammary epithelial cells. J Cell Sci2010;123:1373–82.

25. Tchafa AM, Shah AD, Wang S, Duong MT, Shieh AC. Three-dimensionalcell culture model for measuring the effects of interstitial fluid flow ontumor cell invasion. J Vis Exp 2012;11:2236–46.

26. Caldwell SA, Jackson SR, Shahriari KS, Lynch TP, Sethi G, Walker S, et al.Nutrient sensor O-GlcNAc transferase regulates breast cancer tumorigen-esis through targeting of the oncogenic transcription factor FoxM1. Onco-gene 2010;29:2831–42.

27. Geering B, Cutillas PR, Nock G, Gharbi SI, Vanhaesebroeck B. Class IAphosphoinositide 3-kinases are obligate p85-p110heterodimers. ProcNatlAcad Sci U S A 2007;104:7809–14.

28. Vogt PK, Hart JR, Gymnopoulos M, Jiang H, Kang S, Bader AG, et al.Phosphatidylinositol 3-kinase: the oncoprotein. Curr Top MicrobiolImmunol 2010;347:79–104.

29. Fritsch R, de Krijger I, Fritsch K, George R, Reason B, Kumar MS, et al. RASandRHO families ofGTPases directly regulate distinct phosphoinositide 3-kinase isoforms. Cell 2013;153:1050–63.

30. Dbouk HA, Vadas O, Shymanets A, Burke JE, Salamon RS, Khalil BD, et al.G protein-coupled receptor-mediated activation of p110beta by Gbeta-gamma is required for cellular transformation and invasiveness. Sci Signal2012;5:ra89.

31. Jia S, Roberts TM, Zhao JJ. Should individual PI3 kinase isoforms betargeted in cancer? Curr Opin Cell Biol 2009;21:199–208.

32. Sobolik T, Su YJ, Wells S, Ayers GD, Cook RS, Richmond A. CXCR4 drivesthemetastatic phenotype in breast cancer through induction ofCXCR2 andactivation of MEK and PI3K pathways. Mol Biol Cell 2014;25:566–82.

33. Cojoc M, Peitzsch C, Trautmann F, Polishchuk L, Telegeev GD, DubrovskaA. Emerging targets in cancer management: role of the CXCL12/CXCR4axis. Onco Targets Ther 2013;6:1347–61.

34. Oppermann M. Chemokine receptor CCR5: insights into structure, func-tion, and regulation. Cell Signal 2004;16:1201–10.

35. Fleury ME, Boardman KC, Swartz MA. Autologous morphogen gradientsby subtle interstitial flow and matrix interactions. Biophys J 2006;91:113–21.

36. Xu H, Heilshorn SC. Microfluidic investigation of BDNF-enhanced neuralstem cell chemotaxis in CXCL12 gradients. Small 2013;9:585–95.

37. Yilmaz M, Christofori G. EMT, the cytoskeleton, and cancer cell invasion.Cancer Metastasis Rev 2009;28:15–33.

38. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI geneexpression and hybridization array data repository. Nucleic Acids Res2002;30:207–10.






39. GrogerCJ,GrubingerM,Waldhor T, Vierlinger K,MikulitsW.Meta-analysisof gene expression signatures defining the epithelial to mesenchymaltransition during cancer progression. PLoS ONE 2012;7:e51136.

40. Sims AH, Larionov AA, Harrison DJ, Katz E. Use of microarray analysis toinvestigate EMT gene signatures. Methods Mol Biol 2013;1046:85–95.

41. Baselga J. Targeting the phosphoinositide-3 (PI3) kinase pathway in breastcancer. Oncologist 2011;16(Suppl 1):12–9.

42. Zhao L, Vogt PK. Class I PI3K in oncogenic cellular transformation.Oncogene 2008;27:5486–96.

43. ZhangZ,NiC, ChenW,WuP,WangZ, Yin J, et al. Expression ofCXCR4 andbreast cancer prognosis: a systematic review and meta-analysis. BMCCancer 2014;14:49.

44. IannoneM, Ventre M, Pagano G, Giannoni P, Quarto R, Netti PA. Definingan optimal Stromal Derived Factor-1 presentation for effective recruitmentof mesenchymal stem cells in 3D. Biotechnol Bioeng 2014;111:2303–16.

45. Roussos ET, Condeelis JS, Patsialou A. Chemotaxis in cancer. Nat RevCancer 2011;11:573–87.

46. Mosadegh B, Saadi W, Wang SJ, Jeon NL. Epidermal growth factor pro-motes breast cancer cell chemotaxis in CXCL12 gradients. BiotechnolBioeng 2008;100:1205–13.

47. Swamydas M, Ricci K, Rego SL, Dreau D. Mesenchymal stem cell-derivedCCL-9 and CCL-5 promote mammary tumor cell invasion and theactivation of matrix metalloproteinases. Cell Adh Migr 2013;7:315–24.

48. Carpenter RL, Paw I, Dewhirst MW, Lo HW. Akt phosphorylates andactivates HSF-1 independent of heat shock, leading to Slug overexpressionand epithelial–mesenchymal transition (EMT) of HER2-overexpressingbreast cancer cells. Oncogene 2015;34:546–57.

49. AleskandaranyMA, NegmOH,Green AR, AhmedMA, Nolan CC, Tighe PJ,et al. Epithelial mesenchymal transition in early invasive breast cancer: animmunohistochemical and reverse phase protein array study. Breast Can-cer Res Treat 2014;145:339–48.

50. Burstein HJ, Polyak K, Wong JS, Lester SC, Kaelin CM. Ductal carcinoma insitu of the breast. N Engl J Med 2004;350:1430–41.


Tchafa et al.




2015;13:755-764. Published OnlineFirst January 7, 2015.Mol Cancer Res Alimatou M. Tchafa, Mi Ta, Mauricio J. Reginato, et al.

-Positive Breast Cancer CellsERBB2Induced Signaling in −EMT Transition Alters Interstitial Fluid Flow

Updated version

10.1158/1541-7786.MCR-14-0471doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2015/01/09/1541-7786.MCR-14-0471.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/13/4/755.full#ref-list-1

This article cites 50 articles, 14 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://mcr.aacrjournals.org/content/13/4/755To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-14-0471

http://mcr.aacrjournals.org/content/suppl/2015/01/09/1541-7786.MCR-14-0471.DC1

http://mcr.aacrjournals.org/content/13/4/755.full#ref-list-1

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/13/4/755


Documents

EMT Transition Alters Interstitial Fluid Flow Induced ... · expressing cells, IFF-mediated invasion is chemokine receptor– independent and requires only p110a activation. To test