mir-31 is a broad regulator of b1-integrin expression and ...647 tagged goat anti-mouse antibody was...

Signaling and Regulation

miR-31 Is a Broad Regulator of b1-Integrin Expression andFunction in Cancer Cells

Katarzyna Augoff, Mitali Das, Katarzyna Bialkowska, Brian McCue, Edward F. Plow, and Khalid Sossey-Alaoui

AbstractIntegrins are adhesion receptors involved in bidirectional signaling that are crucial for various cellular

responses during normal homeostasis and pathologic conditions such as cancer progression and metastasis.Aberrant expression of noncoding microRNAs (miRNA) has been implicated in the deregulation of integrinexpression and activity, leading to the development and progression of cancer tumors, including theiracquisition of the metastatic phenotype. miR-31 is a key regulator of several critical genes involved in theinvasion-metastasis cascade in cancer. Using diverse cell-based, genetic, biochemical, flow cytometry, andfunctional analyses, we report that miR-31 is a master regulator of integrins as it targets multiple a subunitpartners (a2, a5, and aV) of b1 integrins and also b3 integrins. We found that expression of miR-31 in cancercells resulted in a significant repression of these integrin subunits both at the mRNA and protein levels. Loss ofexpression of a2, a5, aV, and b3 was a direct consequence of miR-31 targeting conserved seed sequences inthe 30 untranslated region of these integrin subunits leading to their posttranscriptional repression, which wasreflected in their diminished surface expression in live cells. The biological consequence of decreased the cellsurface of these integrins was a significant inhibition of cell spreading in a ligand-dependent manner. Althoughdifferent reports have shown that a single integrin can be regulated by several miRNAs, here we show that asingle miRNA, miR-31, is able to specifically target several integrin subunits to regulate key aspects of cancercell invasion and metastasis. Mol Cancer Res; 9(11); 1500–8. �2011 AACR.

Introduction

As a large family of cell–cell and cell-extracellular matrix(ECM) receptors, integrin a/b heterodimers play promi-nent roles during embryonic development and postnatalphysiology and pathology (1). In addition to their adhesivefunctions, integrins mediate bidirectional signaling acrossthe cell membrane that regulates numerous cellularresponses, including motility, differentiation, growth, andgene expression (2–7). In this context, alterations in integrinexpression and function impart phenotypic properties totransformed cells that have significant impact on cancerprogression and metastasis (8, 9).MicroRNAs (miRNA), short noncoding, single-strand-

ed RNAs that regulate protein translation and mRNAdegradation of their target genes (10), have been shown to

be involved in a large number of biological processes. It isestimated that about one-third of all human mRNAs areregulated by miRNAs (11); and, thus far, more than 1,000miRNAs have been identified in humans with the poten-tial to control more than 30,000 target genes. The broadspectrum of genes that can be targeted by a single miRNAis attributed to the high level of conserved target motifs,seed sequences, within the 30untranslated regions (UTR)of the affected genes, thus making them powerful regu-lators of gene expression in numerous and complex cel-lular responses, including cancer cell invasion andmetastasis (12–15). miRNAs have also been found toregulate integrin activity in such complex biological pro-cesses, such as angiogenesis, cell migration, and invasion(reviewed in ref. 16). Recently, we and others haveidentified and characterized miR-31 as a master regulatorof different steps in the invasion-metastasis-cascade byvirtue of its capacity to modulate key metastasis-promotergenes, including WAVE3 and integrin a5 subunit (17,18). In this study, we show that in addition to a5, miR-31directly targets seed sequences in the 30UTRs of 2 otherintegrin alpha subunits, a2 and aV, and the b3 subunit.We also find that the miR-31–mediated suppression of theintegrin alpha subunits indirectly suppresses cell surfaceexpression of their b1 subunit partner and show thatsuppressed expression of these integrins reduces cellspreading on specific target ECM ligands of these affectedintegrins.

Authors' Affiliation: Department of Molecular Cardiology, Lerner Re-search Institute, Cleveland Clinic, Cleveland, Ohio

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

K. Augoff and M. Das contributed equally to this work

Corresponding Author: Khalid Sossey-Alaoui, Department of MolecularCardiology, Cleveland Clinic Lerner Institute, 9500 Euclid Ave., NB-50,Cleveland, OH 44195. Phone: 216-444-4314; Fax: 216-445-8204; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-11-0311

�2011 American Association for Cancer Research.

MolecularCancer

Research

Mol Cancer Res; 9(11) November 20111500

Research. on January 22, 2021. © 2011 American Association for Cancermcr.aacrjournals.org Downloaded from

Published OnlineFirst August 29, 2011; DOI: 10.1158/1541-7786.MCR-11-0311

http://mcr.aacrjournals.org/

Materials and Methods

ReagentsAll monoclonal antibodies (mAb) to integrins were from

Millipore except for anti-b5 integrin, which was fromeBioscience and b8 integrin, which was from R&D Sys-tems. UV Live Dead dye and all secondary antibodies werefrom Molecular Probes. Other reagents were from Sigma.

Flow cytometryMDA-MB-231 cells expressing GFP-pBABE vector or

GFP-miR-31, a kind gift from Dr. Robert Weinberg (18),were propagated in Dulbecco's modified Eagle's medium(DMEM) culture medium supplemented with 10% FBSand puromycin (5 mg/mL). Subconfluent cells were de-tached by trypsinization and washed with Hank's BalancedSalt Solution with Ca2þ, Mg2þ, and 0.1% BSA. Cells wereprocessed for flow cytometry as described (19). Alexa Fluor647 tagged goat anti-mouse antibody was used to detect allintegrin mouse mAbs, including the b1 activation–specificHUTS-4, except for rat anti-a6 integrin mAb, which wasdetected with Alexa Fluor 647 tagged anti-rat antibody. Allprimary and secondary antibodies were used at 10 mg/mL.As a positive control for integrin activation, 3.5 mmol/LMnCl2 was used to treat the cells for 5 minutes at 37�C in ahumidified tissue culture incubator. All data were acquiredin a BD LSRII instrument and analyzed with FlowJo 7.6.3(Treestar) software. Binding of only live, GFPþ cells todifferent antibodies was used in the analyses. Nonspecificbinding due to secondary antibody alone to empty vectortransfected cells was subtracted from all flow cytometricdata. The resultant fluorescence intensity values were thenplotted graphically.

Cell culture, transfections, and immunoblottingThe established human cancer cell lines were obtained

from American Type Culture Collection and cultured aspreviously described (20). Growth and morphology of eachcell line were routinely monitored and key features werefound to be consistent with prior description of each cellline. No further genetic analyses or fingerprinting weredone. Transient transfections and immunoblotting analyseswere done as described previously (20–23).

Plasmid construction, site-directed mutagenesis, and30UTR luciferase reporter analysisThe nucleotide sequence (�300 bp) flanking the miR-31

seed sequence within the 30UTR of each integrin subunitunder study was PCR-amplified from MDA-MB-231 ge-nomic DNA and subcloned into the pmiRGlo vectordownstream of the firefly luciferase expression cassette(Promega). The correct sequence and orientation wereverified by sequencing. QuikChange site-directed mutagen-esis kit (Stratagene) was used to generate the 30UTRmutants where the seed sequence recognized by miR-31was deleted. pmirGlo reporter plasmids (1 mg total plasmid)were transfected with Lipofectamine 2000 (Invitrogen) intothe specified cancer cells and then seeded in 12-well plates

(3 � 104 cells per well). Cells were collected after 48 hoursfor assay using the Dual-Luciferase reporter assay system(Promega), as described previously (17). For cotransfectionexperiments, synthetic miRNA mimics or miR NegativeControl (5 nmol/L), were added to the transfection mix. Allexperiments were done in triplicate with data averaged fromat least 3 independent experiments.

Semiquantitative and real-time quantitative RT-PCRTotal RNA was extracted from established cancer cell

lines using TRIzol reagent (Invitrogen), following themanufacturer's instructions. cDNA was generated andused as a template for semiquantitative RT-PCR as pre-viously described (20, 22–24). Expression levels of miR-NAs were quantified by real-time qRT-PCR using thehuman TaqMan MicroRNA Assays Kits (Applied Biosys-tems). The reverse transcription reaction was carriedout with TaqMan MicroRNA Reverse TranscriptionKit (Applied Biosystems) following the manufacturer'sinstructions. For each gene, quantification of expressionlevels was done using the respective gene-specific primers(Supplementary Table S1) and the RT2 SYBR Green/Fluorescein qPCR Master Mix (SABiosciences) followingthe manufacturer's instructions. qRT-PCR was done onthe Bio-Rad iCycler PCR system where the reactionmixtures were incubated at 95�C for 10 minutes, followedby 40 cycles of 95�C for 15 seconds and 60�C for 1minute. The cycle threshold (Ct) values were calculatedwith SDS 1.4 software (Bio-Rad). The expression levels ofeach transcript were normalized using the 2�DDCt method(25) relative to glyceraldehyde-3-phosphate dehydroge-nase (GAPDH). The DCt was calculated by subtractingthe Ct values of GAPDH from the Ct values of thetranscript of interest. The DDCt was then calculated bysubtracting DCt of the parental cells from the controlvector- or miR-31–expressing cells. Fold change in thegene expression was calculated according to the equation2�DDCt.

ImmunofluorescenceMDA-MB-231 cells expressing miR-31 or the control

vector were seeded onto 6-well plates coated with 10 mg/mLlaminin, 10 mg/mL fibronectin, 10 mg/mL collagen, or 5times diluted growth factor-reduced Matrigel for 1 hour,fixed, permeabilized with 0.1% Triton X-100, blocked inhorse serum, and stained for 30 minutes with Alexa 568-phalloidin to visualize actin filaments (26).

Cell spreading and integrin blockingCell spreading assays were carried out on round coverslips

(Fisher Scientific) in 12-well plates (Falcon, Becton Dick-inson & Co.). Coverslips were coated 1 hour with MatrigelMatrix (BD Biosciences) diluted 1:6 in PBS, 10 mg/mLcollagen I, 10 mg/mL laminin or 2 mg/mL vitronectin (allfrom Sigma) at 37�C. Following 3 washes with PBS cover-slips were blocked for 1 hour with 1% BSA in PBS at 37�C.BSA (1 mg/mL) coated coverslips were used as negativecontrols.

miR-31 Regulates Several Integrins

www.aacrjournals.org Mol Cancer Res; 9(11) November 2011 1501




MDA-MB-231-vector and MDA-MB-231-miR-31breast cancer cells grown to approximately 80% confluencywere detached with trypsin, washed with serum-free-DMEM, resuspended in the same medium and then platedin triplicate at a cell density of 6� 104 cells/well. For a2b1integrin blocking experiments, cells were resuspended inmedia containing 10 mg/mL of the a2b1 integrin blockingmAb, LS-C247646, (LS-Bio) and incubated 30 minutes at37�C before they were seeded on coverslips coated withcollagen I or Matrigel. Control cells were resuspended withnonblocking IgG control under the same conditions. Cellswere allowed to adhere and spread for 30 minutes and 1hour in a 37�CCO2 incubator. At the end of the incubationperiod, unattached cells were removed by careful aspirationand washing 2 times with PBS. Adherent cells were fixedwith 4% PFA for 10 minutes and washed twice with PBS.Fixed cells were permeabilized with 0.1% Triton X-100 inPBS for 5 minutes and blocked with 4% bovine serumovernight at 4�C and then stained for 1 hour with AlexaFluor 568-conjugated phalloidin (Invitrogen, MolecularProbes). After washing, the coverslips were transferred onslides and mounted in immune-fluor mounting medium(MP Biomedicals). Images were acquired with a 20�/0.7 or40�/1.25 numerical aperture oil objective on a fluorescentmicroscope (Leica) equipped with digital camera usingImage-Pro Plus version 5.1.2.59. Cells were quantifiedusing the ImageJ software version 1.43u and the resultsare presented as the total surface area covered by cells, insquare micrometer, averaged to 300 cells.

Oligonucleotide sequencesSequences of the oligonucleotide primers used for geno-

mic PCR, RT-PCR, to amplify the entire WAVE3 30UTRor the individual seed sequences, and the primers used formutagenesis were from IDT and are listed in Supplemen-tary Table S1.

Statistical analysesThe data are presented as the means � SE of at least 3

independent experiments. The results were tested for sig-nificance using an unpaired Student's t test and P values of <0.05 were considered statistically significant.

Results

miR-31 targets the subunits of several integrinsPrevious reports have shown that miRNA miR-31 targets

the integrin a5 subunit (ITGA5) by interacting with aperfectly matched seed sequence in the 30UTR of theITGA5 transcript (17, 18). We sought to determine wheth-er miR-31 could also target other integrin subunits. By insilico analyses, we found that the seed sequence of miR-31was conserved in the 30UTR of 3 other integrin subunits;a2, aV, and b3 (ITGA2, ITGAV, and ITGB3, respectively;Fig. 1A). Because binding of miRNAs to their specific seedsequences often represses gene expression, we examined theeffect of miR-31 on the expression levels of these candidatetarget integrin subunits. The analysis was carried out inMDA-MB-231 cells because these breast cancer cells

0

0.5

1

1.5

ITGB5ITGB3ITGB1ITGAVITGA5ITGA2

Fo

ld c

han

ge t

o M

DA

-MB

-231 c

ells

MDA-MB-231

Vector

miR-31

* * * **

Actin

ITGA2

ITGB3

ITGB1

ITGA5

ITGAV

MD

A-M

B-2

31

Vecto

r

miR-3

1

miR-31 seed sequenc 3’-CATCTTGC-5’eITGA2 nt 4314 to 5’-CATCTTGC-3’nt 4321ITGAV nt 3446 to 5’-CATCTTGA-3’nt 3452ITGA5 nt 3620 to 5’-AGCCTTGC-3’nt 3627ITGB3 nt 4848 to 5’-GATCTTGT-3’nt 4855

A

MD

A-2

31

Vecto

r

miR-3

1

MD

A-2

31

Vecto

r

miR-3

1

MD

A-2

31

Vecto

r

miR-3

1

MD

A-2

31

Vecto

r

miR-3

1

MD

A-2

31

Vecto

r

miR-3

1

MD

A-2

31

Vecto

r

miR-3

1B

C D

Figure 1. miR-31 inhibitsexpression of specific a and bintegrin subunits. A, alignment ofmiR-31 seed sequence withcandidate target sequences withinthe 30UTRs of the indicatedintegrin subunits. B,semiquantitative RT-PCR, (C) real-time qRT-PCR, and (D)immunoblot of the indicatedintegrin subunits from MDA-MB-231 cells or MDA-MB-231 cellswith stable expression of eithercontrol or miR-31 vector.Densitometry and statisticalanalyses of the Western blot datais shown in Supplementary FigureS2. *, P < 0.05.

Augoff et al.

Mol Cancer Res; 9(11) November 2011 Molecular Cancer Research1502




have very low levels of endogenous miR-31 (17, 18). Stableexpression of miR-31 in these cells had no effect onproliferation compared with control vector transfected cells(Supplementary Fig. S1). Overexpression of miR-31 inMDA-MB-231 cells resulted in approximately 50% de-crease in the levels of a2, a5, aV, and b3. This suppressionwas found both at the mRNA as observed by semiquanti-tative RT-PCR (Fig. 1B) and qRT-PCR (Fig. 1C) and thecellular protein levels as determined by Western blots(Fig. 1D and Supplementary Fig. S2). All these 4 integrinsubunits contain target sites for miR-31 in their respective30UTRs (Fig. 1A), suggesting that the silencing of theirexpression is consequent to miR-31 targeting of the 30UTRin their transcripts. Integrin subunit b5, which lacks a miR-31 seed sequence, was not affected by miR-31 expression inthese cells. We also found that miR-31 overexpressionresulted in approximately 2-fold decrease in b1 subunit(ITGB1) mRNA and protein expression levels comparedwith the parental cells or the vector-transfected cells(Fig. 1, all panels), even though b1 does not contain atarget sequence for miR-31 in its 30UTR. This result couldbe a secondary to effect of miR-31 on the expression levelsof a2, a5, and aV, alpha subunits partners of b1; whichcould lead to degradation of surplus b1 subunit. The effectof miR-31 on the expression of these specific integrinsubunits, a2, a5 and aV, and b1, was also confirmed inLNCaP prostate cancer cells, and b5, which lacks the miR-31 target sequence, was again unaffected (SupplementaryFig. S3).

Surface expression of a-subunits partners of b1integrins are reduced due to miR-31Because integrin-mediated cellular responses depend

upon the expression at the cell surface, as a next step, wemonitored the surface expression patterns of the differ-

ent a-subunits that commonly form heterodimers withb1 subunit in cells stably expressing either the emptycontrol vector or miR-31 by flow cytometry. In accordwith our qRT-PCR, semiquantitative RT-PCR, andWestern blot results (Fig. 1), flow cytometry analysesrevealed that miR-31 significantly reduced the surfaceexpression of a2 (P < 0.05), a5 (P < 0.05), and aV (P <0.001) subunits (Fig. 2). Consistent with the observedreduction in expression of b1 integrin subunit detectedat the mRNA and protein levels, there was also areduction in surface expression of the b1 subunit partnerof these a subunits on the surface of the miR-31expressing cells.

b3 integrin is a target of miR-31aV integrin subunit is known to complex with several b

integrins, including b1, b3, b5, and b6 integrins. Becauseour results show miR-31 can affect b1 indirectly, weevaluated whether other common partners of aV, b3,b5, b6, and b8, were also targeted by miR-31. Only oneof these 4 integrin b subunits, the 30UTR of b3, was foundto harbor the seed sequence that could be targeted by miR-31 (Fig. 1). Consistent with this prediction, we found b3surface levels were significantly reduced (P < 0.05) in cellsexpressing high levels of miR-31 compared with vectorcontrol (Fig. 3A), which is concordant with the RT-PCRandWestern blotting analyses (Fig. 1). Surface expression ofb5 subunit was also reduced in the presence of miR-31,probably as an indirect effect because there is no seedsequence of miR-31 in the 30UTR of b5 (Fig. 3A). Reducedexpression levels of either aV or b3 subunits resulted ina significant reduction of cell surface expression of theaVb3 and the aVb5 heterodimers using antibodies thatspecifically recognize these integrins heterodimers (Fig. 3B).Thus, miR-31 can simultaneously regulate the expression

Figure 2. Surface expression of integrina-subunits and their b1 subunit partner issuppressed by miR-31. Expression profiles ofdifferent integrin a subunits in MDA-MB-231cells overexpressing GFP-pBABE vectorcontrol or GFP-miR-31. Surface expression ofdifferent a subunit partners of the b1 integrinsubunit is quantitated as the relativefluorescence intensity (RFI) with subunit-specific mAbs by flow cytometry. Values aremeans � S.E. and are representative of 3independent experiments. Representativehistograms are provided below in which theempty histograms are for the vector control andthe filled histograms are for miR-31overexpressing cells.

0

3,000

6,000

9,000

12,000

15,000

α1 α2 α3 α4 α5 α6 αV β1

Vector

mir31

Inte

grin

exp

ressio

n (R

FI)

P < 0.05

P < 0.001P < 0.05

P < 0.05

miR

31

Vecto

r

miR

31

miR

31

miR

31

miR

31

miR

31

miR

31

miR

31

Vecto

r

Vecto

r

Vecto

r

Vecto

r

Vecto

r

Vecto

r

Vecto

r






of multiple a and b integrin subunits in a cellular environ-ment.

Reduced expression of a2, a5, and aV subunits bymiR-31 results in decreased expression of activated b1integrin on the cell surfaceOn the basis of the above results, we assessed whether

miR-31–mediated suppression also affected the activationstate of integrins expressed on the surface of the MDA-MB-231 cells. This analysis was done using HUTS-4, a mAbthat recognizes the active conformation of b1 integrinsupon dimerization with one of its a subunit partner(27). miR-31 transfectants showed a significantly reductionin their capacity to spontaneously bind HUTS-4 comparedwith vector control (P < 0.001), and this differenceremained significant in the presence of Mn2þ (P < 0.05),which activates integrins upon binding to their extracellulardomain (Fig 4A). However, this decrease in HUTS-4binding was proportional to the decrease in the surfaceexpression of b1 integrins in these cells (Fig. 4B). Thus,when the activation status of these integrins is expressed as aratio of the total b1 surface expression, measured in HUTS-4 binding, no significant differences were found betweenthe vector or miR-31 expressants in either unstimulated(resting) cells or cells in which integrins were activated withMn2þ (Fig. 4B).

miR-31 directly targets the 30UTR of ITGA2, ITGA5,ITGAV, and ITGB3 integrins and represses theirexpressionWe used the Firefly-Renilla dual luciferase reporter gene

assay to show that the posttranscriptional repression of thea2, a5, aV, and b3 integrin subunits is a consequence ofdirect binding of miR-31 to target seed sequences within the30UTR of the transcripts for these integrin subunits. ADNA fragment, approximately 350 bp-long from the30UTR sequence of these integrins encompassing themiR-31 seed sequence, was subcloned within the 30UTRof the firefly luciferase gene into the pmirGlo vector, andluciferase activity was measured in MDA-MB-231 breastcancer cells, which express very low or no levels of miR-31.For each construct, firefly luciferase activity was normalizedto Renilla luciferase activity and the ratio of miR-31 com-pared with control miR treatment was calculated.We foundno difference in luciferase activity between MDA-MB-231cells transfected with control vector (Glo) alone or withcontrol vector cotransfected with either miR-31 or controlmiRNA (Supplementary Fig. S4), showing the lack of amiR-31 effect in the absence of its specific target sequence.In contrast, in cells transfected with the individual integrin30UTR-containing vector, overexpression of miR-31resulted in a significant reduction in luciferase activity,ranging from 30% to 55% depending on the integrin

β3 β4 β5 β6 β8αvβ3 αvβ5 αvβ6 αv

P < 0.05

A B

P < 0.05

P < 0.05

P < 0.05

P < 0.05

P < 0.05

Vector

mir31

800

600

400

200

0

1,600

1,200

800

400

0Inte

grin

exp

ressio

n (R

FI)

Inte

grin

exp

ressio

n (R

FI)

Vector

mir31

Figure 3. Expression profile of different integrin b subunits in MDA-MB-231 cells expressing and GFP-miR-31 or control GFP-pBABE. A, surface expressionof different b integrin subunits (other than b1) as determined by flow cytometry with representative histograms given below. Empty histogram is forvector and filled histogram is for miR-31 overexpressing cells. Values are means � S.E. and are representative of 3 independent experiments. B, surfacelevels of aVb3, aVb5, and aVb6 reveal differences between vector and miR-31 cells for aVb3 and aVb5 integrins. The integrin aV level has been usedas the control. Empty histogram is for vector and filled histogram is for miR-31 overexpressing cells. Values are means � S.E. and are representativeof 3 independent experiments.

Augoff et al.





subunit compared with the cells transfected with the samevector in the absence of miR-31 or cotransfected withcontrol miRNA in MDA-MB-231 (Fig. 5 and Supplemen-tary Fig. S4). As a control to confirm that miR-31 isspecifically targeting and binding seed sequences withinthese 30UTR, we deleted the seed sequence of miR-31 in the30UTR of a2, a5, aV, and b3 integrin subunits. Thismaneuver abrogated the effect of the exogenous miR-31 onthese integrin subunit reporters in the MDA-MB-231 cells(Fig. 5 and Supplementary Fig. S4). Our in silico analysesalso found a second potential seed sequence of miR-31 inthe 30UTR of ITGAV (ITGAV-2, Fig. 5), which onlypartially aligned to the miR-31 target sequence. However,this second seed sequence was insufficient to be regulated bymiR-31; its presence had no effect on luciferase activity. As anegative control for the effect of miR-31 on the luciferase

activity, the 30UTR of ITGB5 which does not contain aseed sequence for miR-31, showed no different in Renillaluciferase activity in the low or high miR-31 expressingcells. The luciferase assay for ITGB1 could not be donebecause of the large size (more than 1,200 bp) of its30UTR. Similar results were obtained when these experi-ments were repeated in the LNCaP prostate cancer cells(Supplementary Fig. S5); namely, miR-31 repressed ex-pression of the genes encoding for a2, a5, aV, and b3integrin subunits but did not affect b5 expression. Theseresults show that the effect of miR-31 is not restricted to asingle cell line but seems to extend to other cancers andshow that miR-31 specifically targets and represses severalintegrin subunits simultaneously.

miR-31 affects cell spreading in a ligand-dependentmannerBecause miR-31 overexpression leads to diminished ex-

pression of b1 integrin and several of its a subunits partnerson the cell surface (Fig. 2), we anticipated that specificintegrin-mediated cellular responses might be suppressed.Accordingly, we plated control or miR-31-MDA-MB-231cells on different ECM ligands that are recognized byvarious b1-contaning integrins and examined cell spreading.The cells were allowed to spread for 30 or 60 minutes oncollagen (an a2b1 ligand), fibronectin (an a5b1 ligand) orlaminin (an a6b1 ligand) or Matrigel [containing mostlylaminin and collagen IV, ligands for a1b1, a2b1, a3b1,and a6b1 integrins (28)]. At the 2 selected time points,cells were fixed and stained with phalloidin to visualizeactin (Fig. 6A). We found a striking difference in cell sizeand actin organization between control cells and miR-31–expressing cells on collagen and Matrigel after 1 hour(Fig. 6, top 2 panels). The effect of miR-31 on the abilityof cells to spread on these 2 substrates was also observedafter 30 minutes (Supplementary Fig. S6) but becamemore pronounced after 1 hour. While the control cellsbecame large and fully spread, and developed extensivestress fibers and lamellipodia at the cell periphery, themiR-31 expressing cells remained small and round onthese substrata. In control experiments, a blocking mAb(LS-C247646) to a2b1 completely inhibited spreading ofboth the parental and the miR-31 expressing MDA-MB-231 cells on collagen (Fig. 7), but not on Matrigel (notshown), suggesting the roles of not only a2b1 but alsoother integrins that were suppressed by miR-31 expres-sion. The difference in cell spreading between the controlcells and the miR-31 expressing cells was also evident onlaminin and vitronectin, but no difference was observedwhen cells were plated on fibronectin (Fig. 6, bottompanels).

Discussion

Cell adhesion, spreading, and migration on ECM pro-teins are critical processes for both physiologic responses,such as organ development, cell-mediated immunity, andwound healing, and pathologic responses, such as cancer

0

200

400

600

800

1,000

1,200

mir31 + MnVector + Mnmir31Vector

HU

TS

-4 b

indin

g (

RF

I)

P < 0.001

P < 0.05

A

0

1

2

3

Mn-activatedResting

Vector

miR-31

HU

TS

-4 b

indin

g n

orm

aliz

ed

for

β1 inte

grin e

xpre

ssio

n (

RF

I)

B

Figure 4. Targeting of a2, a5, and aV subunits by miR-31 decreases b1integrin activation. MDA-MB-231 cells stably overexpressing GFP-miR-31show reduced b1 integrin activation due to a quantitative reduction in b1integrin surface expression. A, activation of b1 integrins was monitoredwith HUTS-4 antibody, specific for activated b1 integrins, by flowcytometry. Spontaneous HUTS-4 binding or that induced by MnCl2 areshown. Cells were either left untreated or treated with 3.5 mmol/L MnCl2and incubated with HUTS-4 antibody followed by anti-mouse Alexa Fluor647 IgG to monitor b1 integrin activation status. B, HUTS-4 bindingnormalized to total b1 integrin expression measured with mAb MAB2000(Millipore) on control vector- or the miR-31–expressing cells in theresting and Mn2þ-activated states. HUTS-4 binding was done as in (A)and integrin expression was monitored using the same protocol as forHUTS-4 staining, using the expression-specific b1 antibody. Values inboth panels are means � S.E. and representative of 3 independentexperiments.






progression and metastasis. Cell adhesion and spreading aretriggered by cell contact with ECM proteins such ascollagen, laminin, fibronectin, vitronectin, and fibrinogen.Physiologic regulation of cell-matrix interactions is medi-ated by cell-to-ECM (inside out) and ECM-to-cell (outsidein) signals that are themselves transmitted by various ad-hesion receptors on the cell surface, including integrins (28–31). The b1 integrin subunit, in association with severaldifferent a integrin subunits, serve as the principal receptorsfor type I collagen, the most abundant extracellular matrixprotein in mammals (28). Binding of b1 integrin to collagenresults in downstream rearrangements of the actin cytoskel-eton by providing a scaffold for cytoskeletal proteins andmultiple signaling molecules that are involved in the reg-ulation of cell adhesion and spreading (32). In this report,we identify the regulation of expression and activity ofmultiple integrins by miR-31 as a mechanism for themodulation of cancer cell spreading. We and others havepreviously reported the targeting of integrin a5 subunit bymiR-31 (17, 18). In this report, we show that, in addition toa5, miR-31 directly targets and represses the mRNAs andproteins of 2 other integrin alpha subunits (a2 and aV).This repression is mediated by recognition of a conservedseed sequence of miR-31 in the 30UTRs of the respectivetarget transcripts. The suppression of these integrin alphasubunits by miR-31 was observed in a series of in vitromolecular and cellular assays, as well as independent experi-ments to verify that miR-31 regulates the expression andactivity of these integrins subunits. First, we showed thatoverexpression of miR-31 in MDA-MB-231 breast cancercells, which has low endogenous levels of miR-31, resultedin a significant decrease in the expression levels of multiplealpha integrins subunits along with the b1 subunit, usingboth real-time qRT-PCR and Western blotting analyses.Second, we used a luciferase reporter gene assay to show that

the miR-31–mediated suppression of these integrins sub-units arises from direct and specific targeting of a conservedmiR-31 seed sequence in their 30UTRs, and that mutationof the seed sequence abrogates the effect of miR-31.Targeting of these a subunits by miR-31 is consistent withindirect repression of the b1 integrin subunit which associ-ates with a2, a5, and aV partners to form functionalintegrin heterodimers. The expression of a1, a3, a4, anda6 integrins remained similar between the vector and miR-31 expressing cells, showing that the effect of miR-31 isselective for specific a integrins. Interestingly, althoughanalysis of b1 integrin 30UTR revealed it lacks a seedsequence for miR-31, it was significantly (P < 0.05) reducedat the mRNA and protein levels, which further resulted in asignificant reduction in its cell surface expression. Becausethere was no compensatory increase in the levels of the a1,a3, a4, and a6 integrins in the miR-31 transfectants, it canbe expected that there is a net decrease in the number ofavailable a integrin subunits for b1 to complex within thesecells. Third, we showed that loss of mRNA and proteinexpression of these integrins downstream of miR-31resulted in loss of both surface expression and activationas shown by flow cytometric analyses with integrin subunitspecific mAbs and the b1 activation-specific mAb, HUTS4.Fourth, and importantly, we provided evidence that themiR-31-mediated loss of a2, a5, and aV integrins, andtheir b1 partner, and b3 integrins resulted in a significantinhibition of cell spreading in a ligand-dependent manner.The inhibition in cell spreading could reflect reduction insurface expression of total or activated integrins.This study, together with the previously published

reports from both our group (17) and others (16, 18)establish the important role of miR-31 in regulating thedifferent aspect of cell adhesion, spreading and cancer cellinvasion, and metastasis; and solidifies its role a master

0.000

0.500

1.000

1.500

ITGA2 d31ITGA2Vector ITGA5 d31ITGA5 ITGAV-1 ITGAV-1 d31 ITGB3 d31ITGB3 ITGAV-2 ITGB5

Fo

ld c

han

ge o

f lu

cif

era

se a

cti

vit

y

P = 0.019P = 0.005

P = 0.005 P < 0.001

Figure 5.miR-31 directly targets the 30UTR of ITGA2, ITGA5, ITGAV, and ITGB3 integrins and represses their expression in MDA-MB-231 cells. MDA-MB-231cells were transfected with the luciferase reporter plasmids containing the 30UTR of the indicated integrin subunit or with deletion of the miR31 seedsequences, and cotransfected with or without miR31. Luciferase activities were measured after 24 hours. For all luciferase activity assays, Renillaluciferase activity was used for normalization. The normalized firefly luciferase activity of each construct was plotted as a ratio of the control miRNAover the miR-31 treatment. The data are the mean � S.D. of at least 3 independent transfections.

Augoff et al.





metastasis suppressor gene (17, 18). Natural regulation ofcancer metastasis is mediated by cancer metastasis sup-pressor genes which either prevent the progression oftumor cells to a metastatic phenotype or reverse themetastatic phenotype. miR-31 is one of these metastasissuppressor genes, whose pleiotropic functions regulatemultiple steps of the invasion-metastasis cascade by tar-

geting a cohort of prometastatic targets genes all of whichare known to play an important role in metastasis, in-cluding integrins. The pleiotropic effect of miR-31 is ofpotential interest from a translational perspective becausemodulation of a single miRNA can affect the function ofseveral genes involved in cancer metastasis. Thus, thesedata reinforce the important role of miR-31 in the reg-ulation of the cancer invasion-metastasis cascade andidentify miR-31 as an attractive therapeutic target forblunting cancer metastasis.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Tim Burke for proofreading the manuscript. We also thank members ofthe flow cytometry core with assistance with flow cytometry analyses.

0

10

20

30

40

FibronectinVitronectinLamininCollagenMatrigelCell s

urf

ace a

rea x

103 (

mm

2) Spreading assay-1 h

MDA-MB-231

MDA-MB-231-miR-31

P < 0.05P < 0.05

P < 0.05

P < 0.05

A

B

Collagen

Matrigel

Laminin

Fibronectin

Vitronectin

Control cells miR-31-transfected cells

Figure 6. miR-31 affects cell spreading in a ligand-dependent manner.Wild-type and miR-31-expressing MDA-MB-231 cells spread on collagen,Matrigel, fibronectin, and laminin for 1 hour, were fixed, permeabilized andstained for 30 minutes with Alexa 568-phalloidin to visualize actinfilaments. A, in comparison with vector control cells, miR-31–transfectedcells showed significant inhibition of spreading and actin organizationon collagen and Matrigel; lesser perturbation on laminin, and no differenceon fibronectin. The images shown are representative micrographs of12 to 15 independent scans for each integrin ligand. Bar ¼ 40 mm. B,quantitation of cell spreading presented as the total surface area coveredby cells, in square micrometer, averaged to 300 cells.

IgG

α2β1

0

20

40

60

80

100

120

MDA-MB-231-miR-31MDA-MB-231

Su

rface c

ell a

rea (

%)

Spreading on collagen for 1 hFunctional blocking with α2β1 Abs

MDA-MB-231 MDA-MB-231-miR-31

Co

ntr

ol Ig

Gα2

β1 b

lockin

g A

bs

B

A

Figure 7. A function blocking mAb to a2b1 completely inhibits cellspreading in both the control cells and the miR31-expressing MDA-MB-231 cells. Wild-type and miR-31–expressing MDA-MB-231 cells whereincubated with either an a2b1 monoclonal blocking antibody (LS-C247646) or control IgG and spread on collagen for 1 hour. Cells were thenfixed, permeabilized and stained for 30 minutes with Alexa 568-phalloidinto visualize actin filaments. A, the images shown are representativemicrographs of 12 to 15 independent scans for each treatment. Bar ¼ 40mm. B, quantitation of cell spreading presented as the total surface areacovered by cells, in square micrometer, from 300 cells.






Grant Support

DOD Breast Cancer Idea Award grant W81XWH0810236 (K. Sossey-Alaoui),NIH grants P01HL073311 and P50HL077107 (E.F. Plow). K. Augoff was funded inpart by a research fellowship "Development Program of Wroclaw MedicalUniversity from the European Social Fund, Human Capital, National CohesionStrategy" (contract # UDA-POKL.04.01.01-00-010/08-00).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received June 29, 2011; revised August 10, 2011; accepted August 23,2011; published OnlineFirst August 29, 2011.

References1. Johnson MS, Lu N, Denessiouk K, Heino J, Gullberg D. Integrins

during evolution: evolutionary trees and model organisms. BiochimBiophys Acta 2009;1788:779–89.

2. Humphries JD, Byron A, Humphries MJ. Integrin ligands at a glance. JCell Sci 2006;119:3901–3.

3. Cavallaro U, Dejana E. Adhesion molecule signalling: not always asticky business. Nat Rev Mol Cell Biol 2011;12:189–97.

4. Schmidt S, Friedl P. Interstitial cell migration: integrin-dependent andalternative adhesion mechanisms. Cell Tissue Res 2010;339:83–92.

5. Desgrosellier JS, Cheresh DA. Integrins in cancer: biological implica-tions and therapeutic opportunities. Nat Rev Cancer 2010;10:9–22.

6. Danen EH, van RJ, Franken W, Huveneers S, Sonneveld P, Jalink K,et al. Integrins control motile strategy through a Rho-cofilin pathway. JCell Biol 2005;169:515–26.

7. Mierke CT, Frey B, Fellner M, Herrmann M, Fabry B. Integrin alpha5-beta1 facilitates cancer cell invasion through enhanced contractileforces. J Cell Sci 2011;124:369–83.

8. Cabodi S, del PC-L, Di SP, Defilippi P. Integrin signalling adaptors: notonly figurants in the cancer story. Nat Rev Cancer 2010;10:858–70.

9. Tabatabai G, Weller M, Nabors B, Picard M, Reardon D, Mikkelsen T,et al. Targeting integrins in malignant glioma. Target Oncol 2010;5:175–81.

10. Arsanious A, Bjarnason GA, Yousef GM. From bench to bedside:current and future applications of molecular profiling in renal cellcarcinoma. Mol Cancer 2009;8:20.

11. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, et al.MicroRNA expression profiles classify human cancers. Nature 2005;435:834–8.

12. Esquela-Kerscher A, Slack FJ. Oncomirs - microRNAs with a role incancer. Nat Rev Cancer 2006;6:259–69.

13. Korpal M, Kang Y. The emerging role of miR-200 family of microRNAsin epithelial-mesenchymal transition and cancer metastasis. RNA Biol2008;5:115–9.

14. Peter ME. Let-7 and miR-200 microRNAs: guardians against pluripo-tency and cancer progression. Cell Cycle 2009;8:843–52.

15. Spaderna S, Brabletz T, Opitz OG. The miR-200 family: central playerfor gain and loss of the epithelial phenotype. Gastroenterology2009;136:1835–7.

16. Valastyan S, Weinberg RA. Roles for microRNAs in the regulation ofcell adhesion molecules. J Cell Sci 2011;124:999–1006.

17. Sossey-Alaoui K, Downs-Kelly E, Das M, Izem L, Tubbs R, Plow EF.WAVE3, an actin remodeling protein, is regulated by the metastasissuppressor microRNA, miR-31, during the invasion-metastasis cas-cade. Int J Cancer 2011;129:1331–43.

18. Valastyan S, Reinhardt F, Benaich N, Calogrias D, Sz�asz AM, WangZC, et al. A pleiotropically acting microRNA, miR-31, inhibits breastcancer metastasis. Cell 2009;137:1032–46.

19. Ithychanda SS, DasM,MaYQ, Ding K,Wang X, Gupta S, et al. Migfilin,a molecular switch in regulation of integrin activation. J Biol Chem2009;284:4713–22.

20. Sossey-Alaoui K, Bialkowska K, Plow EF. The miR200 family ofmicroRNAs regulates WAVE3-dependent cancer cell invasion. J BiolChem 2009;284:33019–29.

21. Sossey-Alaoui K, Li X, Ranalli TA, Cowell JK. WAVE3-mediated cellmigration and lamellipodia formation are regulated downstream ofphosphatidylinositol 3-kinase. J Biol Chem 2005;280:21748–55.

22. Sossey-Alaoui K, Ranalli TA, Li X, Bakin AV, Cowell JK. WAVE3promotes cell motility and invasion through the regulation of MMP-1, MMP-3, and MMP-9 expression. Exp Cell Res 2005;308:135–45.

23. Sossey-Alaoui K, Safina A, Li X, Vaughan MM, Hicks DG, Bakin AV,et al. Down-regulation ofWAVE3, ametastasis promoter gene, inhibitsinvasion andmetastasis of breast cancer cells. Am J Pathol 2007;170:2112–21.

24. Sossey-Alaoui K, Li X, Cowell JK. c-Abl-mediated phosphorylation ofWAVE3 is required for lamellipodia formation and cell migration. J BiolChem 2007;282:26257–65.

25. Schmittgen TD, Livak KJ. Analyzing real-time PCR data by thecomparative C(T) method. Nat Protoc 2008;3:1101–8.

26. Bialkowska K, Ma YQ, Bledzka K, Sossey-Alaoui K, Izem L, Zhang X,et al. The integrin co-activator Kindlin-3 is expressed and functional ina non-hematopoietic cell, the endothelial cell. J Biol Chem 2010;285:18640–9.

27. Luque A, Gomez M, Puzon W, Takada Y, Sanchez-Madrid F,Cabanas C. Activated conformations of very late activation integrinsdetected by a group of antibodies (HUTS) specific for a novelregulatory region (355-425) of the common beta 1 chain. J BiolChem 1996;271:11067–75.

28. Barczyk M, Carracedo S, Gullberg D. Integrins. Cell Tissue Res 2010;339:269–80.

29. LuoBH, Springer TA. Integrin structures and conformational signaling.Curr Opin Cell Biol 2006;18:579–86.

30. Plow EF, Haas TA, Zhang L, Loftus J, Smith JW. Ligand binding tointegrins. J Biol Chem 2000;275:21785–8.

31. Qin J, Vinogradova O, Plow EF. Integrin bidirectional signaling: amolecular view. PLoS Biol 2004;2:e169.

32. Stossel TP, Condeelis J, Cooley L, Hartwig JH, Noegel A, SchleicherM, et al. Filamins as integrators of cell mechanics and signalling. NatRev Mol Cell Biol 2001;2:138–45.

Augoff et al.





2011;9:1500-1508. Published OnlineFirst August 29, 2011.Mol Cancer Res Katarzyna Augoff, Mitali Das, Katarzyna Bialkowska, et al. Function in Cancer Cells

1-Integrin Expression andβmiR-31 Is a Broad Regulator of

Updated version

10.1158/1541-7786.MCR-11-0311doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2011/08/26/1541-7786.MCR-11-0311.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/9/11/1500.full#ref-list-1

This article cites 32 articles, 11 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/9/11/1500.full#related-urls

This article has been cited by 7 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

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

Permissions

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

.http://mcr.aacrjournals.org/content/9/11/1500To 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-11-0311

http://mcr.aacrjournals.org/content/suppl/2011/08/26/1541-7786.MCR-11-0311.DC1

http://mcr.aacrjournals.org/content/9/11/1500.full#ref-list-1

http://mcr.aacrjournals.org/content/9/11/1500.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/9/11/1500


mir-31 is a broad regulator of b1-integrin expression and ...647 tagged goat anti-mouse antibody was...

Documents