Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]

http://dx.doi.org/10.10014-4827/& 2014 E

nCorresponding autE-mail addresses

Please cite this arHCT116 cells via p

Research Article

Special AT-rich sequence-binding protein 2 suppressesinvadopodia formation in HCT116 cells via palladin inhibition

Mohammed A. Mansoura,b,n, Eri Asanoa, Toshinori Hyodoa, K.A. Aktera, Masahide Takahashic,Michinari Hamaguchia, Takeshi Sengaa,n

aDivision of Cancer Biology, Nagoya University Graduate School of Medicine, 65 Tsurumai, Showa, Nagoya 466-8550 JapanbBiochemistry Section, Department of Chemistry, Faculty of Science, Tanta University, EgyptcDepartment of Pathology, Nagoya University Graduate School of Medicine, 65 Tsurumai, Showa, Nagoya 466-8550 Japan

a r t i c l e i n f o r m a t i o n

Article Chronology:

Received 3 October 2014Received in revised form28 November 2014Accepted 5 December 2014

Keywords:

Colorectal cancerSATB2InvadopodiaPalladin

016/j.yexcr.2014.12.003lsevier Inc. All rights reser

hor. Fax: þ81 52 744 2464.: biomansour@hotmail.com

ticle as: M.A. Mansour,alladin inhibition, Exp C

a b s t r a c t

Invadopodia are specialized actin-based microdomains of the plasma membrane that combineadhesive properties with matrix degrading activities. Proper functioning of the bone, immune,and vascular systems depend on these organelles, and their relevance in cancer cells is linked totumor metastasis. The elucidation of the mechanisms driving invadopodia formation is aprerequisite to understanding their role and ultimately to controlling their functions. SpecialAT-rich sequence-binding protein 2 (SATB2) was reported to suppress tumor cell migration andmetastasis. However, the mechanism of action of SATB2 is unknown. Here, we show that SATB2inhibits invadopodia formation in HCT116 cells and that the molecular scaffold palladin isinhibited by exogenous expression of SATB2. To confirm this association, we elucidated thefunction of palladin in HCT116 using a knock down strategy. Palladin knock down reduced cellmigration and invasion and inhibited invadopodia formation. This phenotype was confirmed by arescue experiment. We then demonstrated that palladin expression in SATB2-expressing cells

restored invasion and invadopodia formation. Our results showed that SATB2 action is mediatedby palladin inhibition and the SATB2/palladin pathway is associated with invadopodia formationin colorectal cancer cells.

& 2014 Elsevier Inc. All rights reserved.

Introduction

Colorectal cancer (CRC) is the second leading cause of cancerdeaths, with approximately 500,000 deaths per year worldwide.Over the past two decades, several treatments have been identi-fied for CRC patients. However, the rate of recurrence and failureafter these treatments remains high. Additionally, most CRCdeaths are associated with tumor invasion and metastasis [1].

ved.

(M.A. Mansour), tsenga@

et al., Special AT-rich seqell Res (2014), http://dx.d

Thus, there is a strong impetus to understand the function ofcancer genes involved in CRC invasiveness to develop newtherapeutic approaches.Special AT-rich sequence-binding protein 2 (SATB2) is a member

of the SATB family proteins, which share structural homologyconsisting of CUT and homeobox domains [2]. SATB2 is a transcrip-tional factor that specifically binds to the nuclear matrix attachmentregion (MAR) of AT-rich DNA sequences to regulate chromatin re-

med.nagoya-u.ac.jp (T. Senga).

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]2

modeling and transcription [3,4]. SATB2 has multiple roles indevelopmental processes including craniofacial patterning, braindevelopment, and osteoblast differentiation [5]. The chromosomaldeletions of 2q33.1 that cause SATB2 haploinsufficiency in humansare associated with a cleft or high palate, facial dysmorphism, andintellectual disability [6,7]. In addition to critical roles for develop-mental processes, previous studies have demonstrated that SATB2is associated with tumor suppression. An immunohistochemicalanalysis of laryngeal squamous cell carcinoma (LSCC) showed thatlower expression of SATB2 was correlated with advanced clinicalstaging, histological grade and tumor recurrence [8]. In colorectalcancer, high expression of SATB2 is associated with good prognosisand sensitivity to chemotherapy and radiation [9,10]. These studiesindicate that high expression of SATB2 is a promising marker for thegood prognosis of cancer patients; however, detailed analysis ofSATB2 function in cancer cells has not been fully performed.Palladin (PALLD) is an actin-associated protein with multiple

isoforms derived from a single gene [11]. A recent review by Jin[12] described 7 isoforms and the Universal Protein Databasecurrently shows 9 isoforms of human palladin. Palladin is widelyexpressed in various tissues and cell lines as major 90 and140 kDa isoforms, whereas 200 kDa palladin is specificallydetected in heart and bone [11]. Palladin has an essential role inthe assembly and maintenance of multiple types of actin-richstructures, including stress fibers, dynamic dorsal ruffles andmatrix-degrading podosomes [13,14]. Accumulating evidencehas revealed that palladin is associated with malignant character-istics of cancer. Goicoechea et al. [15] reported that palladin waslocalized to the invadopodia of cancer cells and played importantroles for invasion of metastatic breast cancer cells. In pancreaticcancer, palladin is expressed in cancer associated fibroblast (CAF)and an animal model demonstrated that palladin expression inCAF enhanced invasion of pancreatic cancer cells [16,17]. Inaddition, a recent study showed that palladin interacts withmembrane-type 1 matrix metalloproteinase (MMP14) to promotedegradation of the extracellular matrix (ECM). This interactionlinks ECM degradation to cytoskeletal dynamics and migrationsignaling in mesenchymal breast cancer cells [18]. These studiesclearly indicate that palladin plays an important role for thepromotion of cancer cell invasion.In this study, we demonstrate that exogenous expression of

SATB2 inhibits invasion and invadopodia formation of colorectalcancer cells. SATB2 expression suppressed expression of palladin,and restoration of palladin expression rescued SATB2-mediatedinhibition of invasion and invadopodia formation. This studydefines a novel SATB2/palladin pathway for the suppression oftumor invasion.

Materials and methods

Cells and antibodies

HCT116, CW-2 and COLO 320 cells were obtained from ATCC andcultured in DMEM (HCT116) and RPMI (CW-2 & COLO 320),supplemented with 10% FBS and antibiotics. HEK293T cells usedfor retrovirus production were maintained in DMEM with 10%FBS. The antibodies were obtained from the following companies:anti-β-actin, Sigma-Aldrich (St. Louis, MO); anti-SATB2, Abcam(Cambridge, UK); anti-GFP, Neuro Mab (Davis, CA); anti-α-tubulin,

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

Sigma-Aldrich (St. Louis, MO). Anti-palladin antibody was gener-ated previously [19].

Generation of stable cell lines

Full-length SATB2 and palladin (90 kDa form) were PCR amplifiedfrom a cDNA library of HCT116 cells. SATB2 was cloned into thepQCXIP vector with an N-terminal GFP tag or FLAG tag andtransfected into 293T cells together with the pVPack-GP andpVPack-Ampho vectors using Lipofectamine 2000 (Invitrogen,Carlsbad, CA). Forty-eight hours after transfection, the super-natants were added to cells with 2 μg/ml polybrene (Sigma-Aldrich). The infected cells were selected by incubating with1 μg/ml puromycin for 2 days.

To establish HCT116 cell lines that constitutively expressed GFP,GFP–palladin (wt) (GFP-tagged wild-type palladin) and GFP–PALLD (Res) (GFP-tagged mutant palladin), each cDNA was clonedinto a pQCXIN retrovirus vector (Clontech, Mountain View, CA).Silent mutations were introduced in GFP–PALLD (Res) to conferresistance to shRNA targeting endogenous palladin. Each plasmidwas transfected into HEK 293T cells together with pVPack-GP andpVPack-Ampho vectors using Lipofectamine 2000 according tothe manufacturer's protocol. The culture supernatants werecollected 48 h after transfection and applied to HCT116 cells incombination with 2 mg/ml polybrene (Sigma). The cells werecultured for 24 h and infected cells were selected with 400 mg/ml of G418 (Nacalai Tesque, Tokyo, Japan). To establish shLuc(Ctrl), shPALLD #1, shPALLD #2/GFP and shPALLD #2/GFP–PALLD(Res) cells, oligonucleotides encoding shRNAs specific for humanpalladin (#1: 50-GTACTGGACGGCTAATGGT-30 & #2: 50-GCA-CAAAGGATGCTGTTAT-30) and luciferase (50-CTTACGCTGAG-TACTTCGA-30) were cloned into the pSIREN-RetroQ retroviralvector (Clontech). The cells were infected with recombinantretroviruses that encoded each shRNA and were selected with1 mg/ml puromycin.

Quantitative reverse transcription–polymerase chainreaction

RNA was extracted from HCT116 cells using the RNeasy Mini Kit(Qiagen, Venlo, Netherlands), and cDNA was generated using Prime-Script Reverse Transcriptase (TAKARA, Tokyo, Japan). PCR was per-formed using the SYBR Premix Ex TaqTM II (TAKARA), and a ThermalCycler DiceTM Real Time System TP800 (TAKARA) was used for theanalysis. The relative mRNA expression levels were normalized toGAPDH. The sequences of primers used to amplify each gene were 50-AGGTGGAGGAGTGGGTGTCGCTGTT-30 and 50-CCGGGAAACTGTGGC-GTGATGG-30 (GAPDH) and 50-GCAATTCAATGCTGCTGAGA-30 and 50-GTGGCTCCTTAGTGGGTGAA-30 (palladin).

Western blotting

Cell lysates were loaded on SDS-polyacrylamide gels for electro-phoresis and transferred to polyvinylidene difluoride membranes(Millipore). The membranes were blocked with 1% skim milk for1 h and then incubated with primary antibodies for 1 h. Themembranes were then washed with TBS-T for 15 min andincubated with HRP-labeled secondary antibodies. The signalswere detected with the ECL system (GE Healthcare BioSciences).

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ] 3

The signal intensities were measured using Light Capture IIequipped with CS analyzer (ATTO Corp., Tokyo, Japan).

Wound-healing assay

Wound-healing assays were performed by scratching confluentmonolayers of cells with a pipette tip and incubating the cells at37 1C with 5% CO2. The distance of the leading edge of the

Fig. 1 – SATB2 expression inhibits cell invasion and invadopodia foFLAG or FLAG–SATB2 were established by retrovirus infection. Theimmunoblotting using anti-FLAG and anti-SATB2 antibodies. An arHCT116 cells were subjected to an invasion assay. Representative iaverage number of invaded cells per field (*Po0.05). (C) FLAG andimmunostained for F-actin (phalloidin; red) and cortactin (green).the graph shows the average percentage of cells with invadopodiaisothiocyanate (FITC)–gelatin coated coverslips, fixed with 4% paraRepresentative micrographs are shown (scale bar¼20 μm).

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

monolayer traveled was measured in five randomly selected fieldsevery 12 h. Three independent experiments were performed, andthe data are shown as the mean7SE.

Migration assay

To measure cell migration using Boyden chambers, a filter (8 mmpore size, 6.5 mm membrane diameter) was pre-coated with fibro-

rmation. (A) HCT116 cells that constitutively expressed eitherexpression of FLAG–SATB2 protein was examined byrow indicates the endogenous SATB2. (B) FLAG or FLAG–SATB2mages of invaded cells are shown, and the graph indicates theFLAG–SATB2 cells were fixed with 4% paraformaldehyde andRepresentative micrographs are shown (scale bar¼20 μm) and(*Po0.05). (D) HCT116 cells were seeded on fluoresceinformaldehyde and immunostained for cortactin (red).

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]4

nectin over night, and 2�105 cells were seeded onto the uppersurface of the chamber with DMEM and 0.1% BSA (serum-free).However, the lower chamber was filled with DMEM and 10% FBS(chemo-attractant). Eighteen hours after seeding, the cells were fixedwith 70% methanol and stained with 0.5% crystal violet. The cells thatmigrated through the lower surface of the filters were counted in fiverandomly selected fields. Three independent experiments wereperformed, and the data are shown as the mean7SE.

Fig. 2 – SATB2 expression inhibits palladin expression. (A) HCT116(GFP–SATB2) were generated by retroviral infection. The expressionby immunoblotting using anti-GFP and anti-SATB2 antibodies. An amRNA in GFP and GFP–SATB2 cells was determined by qRT-PCR (C)was determined by immunoblotting. The arrows indicate the 200,

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

Invasion assay

To measure cell invasion using Boyden chambers, a filter (8 mm poresize, 6.5 mm membrane diameter) was pre-coated with matrigel(mixture of matrix molecules including laminin, collagen type IVand entactin) overnight, and 2�105 cells were seeded onto theupper surface of the chamber with DMEM and 0.1% BSA (serum-free). However, the lower chamber was filled with DMEM and 10%FBS (chemo-attractant). Twenty-four hours after seeding, the cells

cells constitutively expressing GFP (GFP) or GFP-tagged SATB2of GFP or GFP–SATB2 proteins in each cell line was examined

rrow indicates the endogenous SATB2. (B) The level of palladinPalladin expression in either GFP or GFP–SATB2 expressing cells140, 115, 90 and 50 kDa palladin isoforms.

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

Fig. 3 – Palladin depletion reduces cancer cell migration and invasion. (A) Palladin expression was detected in seven colorectalcancer cell lines by immunoblotting. (B) HCT116 cells constitutively expressing shCtrl, shPALLD #1 and shPALLD #2 were generatedby retroviral infection. The palladin expression was detected in these cell lines by immunoblotting. (C) Confluent monolayers ofshCtrl, shPALLD #1 or shPALLD #2 HCT116 cells were scratched, and cell migration was examined every 12 h. Representative imagesof migrated cells are shown, and the graph shows the average migrated distance at the indicated time points (*Po0.05). (D) ShCtrl,shPALLD #1 or shPALLD #2 HCT116 cells were subjected to the migration assay. Representative images of migrated cells are shown,and the graph indicates the average number of migrated cells per field (*Po0.05). (E) ShCtrl, shPALLD #1 or shPALLD #2 HCT116cells were subjected to an invasion assay. Representative images of invaded cells are shown, and the graph indicates the averagenumber of invaded cells per field (*Po0.05).

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ] 5

Please cite this article as: M.A. Mansour, et al., Special AT-rich sequence-binding protein 2 suppresses invadopodia formation inHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]6

were fixed with 70% methanol and stained with 0.5% crystal violet.The cells that invaded the lower surface of the filters were countedin five randomly selected fields. Three independent experimentswere performed, and the data are shown as the mean7SE.

Immunofluorescence microscopy

Cells were grown on glass coverslips, fixed with 4% paraformal-dehyde for 20 min and then permeabilized with 0.5% Triton X-100for 3 min. The cells were blocked with phosphate-buffered saline(PBS) containing 7% fetal bovine serum for 30 min. The cells wereincubated with primary antibody in PBS for 1 h and washed threetimes with PBS. After washing, the cells were incubated withFITC-conjugated anti-rabbit antibody (Invitrogen) or Rhodamine-conjugated anti-mouse antibody (Invitrogen). The images wereacquired using a laser scanning confocal microscope FV1000(OLYMPUS, Tokyo, Japan). To visualize the actin cytoskeleton thefixed cells were incubated with rhodamine-conjugated phalloidinfor 30 min and washed.

Gelatin degradation assay

Coverslips were coated with 50 μg/ml (H2O) Poly-L(D)-lysine (Sigma-Aldrich), fixed with 0.5% glutaraldehyde (Sigma-Aldrich) and coatedwith 0.25 mg/ml fluorescein isothiocyanate (FITC) conjugated gelatin(Elastin Products Co.). Then, cells were seeded on the coverslips andafter 22 h, cells were fixed and immunostained.

Statistical analysis

The statistical analysis is indicated in each corresponding figurelegend. All data are the mean7SE from 3 independent assays. Allanalyses were examined using the SigmaPlot program version10.0 (Systat Software, Inc., San Jose, CA). The P-values werecalculated from two-tailed statistical tests. A difference wasconsidered statistically significant when Po0.05.

Results

SATB2 expression inhibits invasion and invadopodiaformation

SATB2 has been shown to have tumor suppression function anddownregulation of SATB2 is associated with metastasis and poorprognosis in CRC [10]. We generated HCT116 cells that constitutivelyexpressed FLAG–SATB2 or FLAG tag and examined cell invasion.Expression of FLAG–SATB2 was significantly higher than theendogenous level of SATB2 (Fig. 1A). We assessed cell invasionusing a matrigel-coated Boyden chamber. We seeded either FLAG orFLAG–SATB2 expressing HCT116 cells in the upper chamber (serum-free) and allowed the cells to invade through a matrigel coated filtertoward serum containing medium. After 24 h, cells were fixed andthe number of invaded cells was counted. As shown in Fig. 1B,invasion of FLAG–SATB2-expressing HCT116 cells was clearlyreduced compared to that of FLAG-expressing cells.Invadopodia are highly motile membrane extensions that promote

degradation of extracellular matrix for cell invasion. We examinedwhether SATB2 expression in HCT116 cells suppressed invadopodia

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

formation. Invadopodia are usually visualized by co-staining cellswith cortactin and F-actin. Additionally, it is important to determinethat the co-staining is on the ventral surface of the cell using eitherconfocal or TIRF microscopy. We used the confocal microscopy afterco-staining with phalloidin (red) and cortactin (green) to detectinvadopodia. As shown in Fig. 1C, the invadopodia can be identifiedby their dot-like concentrations of cortactin. The invadopodia werefound scattered over the ventral membrane of the FLAG-expressingcells. However, FLAG–SATB2-expressing cells showed reduced dot-like concentrations of cortactin and a decreased percentage of cellswith invadopodia with a 1.6-fold decrease compared to FLAG-expressing cells (Fig. 1C). To confirm that the cortactin-positive dotsare functional invadopodia, we used glass coverslip coated withFITC-labeled gelatin. Cells were cultured on the FITC-labeled gelatinand then degradation of gelatin was observed by confocal micro-scopy. Degradation of FITC-labeled gelatin was observed wherecortactin was prominently concentrated (Fig. 1D), suggesting thatcortactin-rich spots have the ability to degrade extracellular matrix.

Expression of SATB2 inhibits palladin

Palladin was reported to promote formation of podosomes andinvadopodia in normal and cancer cells, respectively [20]. Toelucidate the molecular mechanism by which SATB2 exerts itsaction, we examined palladin expression in SATB2-expressingcells. We generated either GFP- or GFP–SATB2-expressing HCT116cells by retrovirus infection (Fig. 2A) and examined palladinmRNA level by RT-PCR. As shown in Fig. 2B, palladin mRNA wassignificantly reduced by GFP–SATB2 expression. To further con-firm the suppression of palladin expression by exogenous SATB2,we generated CW-2 and COLO320 cells that constitutivelyexpressed GFP–SATB2. Immunoblot analysis showed that expres-sion of multiple palladin isoforms (200, 140, 115, 90 and 50 kDa)was significantly decreased by GFP–SATB2 (Fig. 2C).

Palladin depletion in HCT116 suppresses cell migration andinvasion

We speculated that SATB2-mediated suppression of invasion andinvadopodia formationwere mediated by decreased palladin expres-sion. To evaluate the function of palladin, we used two differentshRNAs to deplete palladin expression. Palladin was highlyexpressed in HCT116, CW-2 and COLO320 cells, but not in DLD-1,CACO-2, SW620 and COLO205 cells (Fig. 3A). We used HCT116 cellsfor further experiment and established HCT116 cells that constitu-tively expressed control shRNA (shCtrl) or palldin shRNA(shPALLD#1 and shPALLD#2). Expression of palladin isoforms wasclearly suppressed in shPALLD#1 and shPALLD#2 cells (Fig. 3B). Wefirst examined cell migration in the absence of palladin. Confluentmonolayers of cells were scratched and distance of the migratedleading edges was measured every 12 h. The migration of palladin-depleted cells was clearly delayed compared with that of the shCtrlcells (Fig. 3C). To further confirm this result, we used a modifiedBoyden chamber assay. The cells (serum-starved) were placed onthe upper surface of a filter coated with fibronectin, and the cellswere then allowed to migrate to the bottom surface filled withserum-containing medium. We counted cells that migrated to thebottom surface 18 h after the seeding. The migration of both

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

Fig. 4 – Palladin depletion inhibits invadopodia formation. (A) HCT116 cells that constitutively expressed shCtrl, shPALLD #1 orshPALLD #2 were fixed with 4% paraformaldehyde and immunostained for F-actin (phalloidin; red) and cortactin (green).Representative micrographs are shown (scale bar¼20 μm). The arrows indicate the invadopodia concentrations. (B) The graphshows the average percentage of cells with invadopodia (*Po0.05) in shCtrl, shPALLD #1 or shPALLD #2 cells. (C) HCT116 cells werefixed with 4% paraformaldehyde and immunostained for cortactin (green) and palladin (red). Representative micrographs areshown (scale bar¼20 μm).

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ] 7

Please cite this article as: M.A. Mansour, et al., Special AT-rich sequence-binding protein 2 suppresses invadopodia formation inHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]8

shPALLD#1 and shPALLD#2 cells was clearly suppressed comparedwith that of the shCtrl cells (Fig. 3D). We next examined cell invasionusing matrigel-coated Boyden chambers. As shown in Fig. 3E, cellinvasion was significantly suppressed by palladin knockdown.

Palladin localizes to invadopodia and enhances itsformation

Cancer cell invasion involves degradation of ECM via formation ofinvasive morphological features including invadopodia [21]. Weimmunostained cells for F-actin and cortactin, and then observedusing a confocal microscopy to evaluate invadopodia formation. InshCtrl HCT116 cells, invadopodia appeared as isolated punctalocated behind the leading edge of the cell or under the nucleus(Fig. 4A). Consistent with the reduced cell invasion in the absenceof palladin, invadopodia formation was suppressed in bothshPALLD#1 and shPALLD#2 cells (Fig. 4A). Nearly 60% of shCtrlcells had cortactin concentrated area, whereas less than 30% ofpalladin-depleted cells showed cortactin accumulation (Fig. 4B).These results suggest that palladin plays an important role inpromoting invadopodia formation in HCT116 cells. We usedimmunostaining to visualize endogenous palladin in HCT116 cellsto determine whether palladin is recruited to these actin-basedstructures. As shown in Fig. 4C, partial co-localization of palladinand cortactin was observed in HCT116 cells.

Rescue of palladin in palladin depleted cells restoresinvadopodia formation

We next performed a rescue experiment using HCT116 cells toexclude the possibility that the observed phenotype was induced bythe off-target effect of the shRNAs. To exogenously express palladin,we used 90 kDa isoform because the isoform is highly expressed inmultiple tumor tissues and is known to promote tumor cell invasion[16,18]. Importantly, C-terminal Ig domains of the 90 kDa palladinwere found to promote tumor cell invasion by linking ECMdegradation to cell cytoskeleton [18]. These results strongly suggesta unique function for the 90 kDa palladin in the regulation of ECMdegradation and cancer cell invasion. We generated HCT116 cellsthat constitutively expressed GFP or GFP–palladin (Res) by retrovirusinfection. GFP–palladin (Res) has silent mutations to be resistant toshPALLD #2-mediated knockdown. Both cell lines were theninfected with shPALLD #2-encoding retrovirus to generate shPALLD#2/GFP and shPALLD #2/GFP PALLD (Res) cells. Expression ofendogenous palladin was reduced by shPALLD #2, but the expres-sion of GFP–palladin (Res) was not affected (Fig. 5A). We performedan immunofluorescence analysis to observe invadopodia formation.The GFP-expressing cells showed a reduced frequency of invadopo-dia formation by shPALLD#2 expression. In contrast, the introduc-tion of shPALLD#2 did not affect invadopodia formation in GFPPALLD (Res) cells (Fig. 5B). These results show that invadopodiaformation was specifically promoted by palladin.

SATB2 represses cell invasion and invadopodia formationvia palladin inhibition

We next tested whether SATB2-induced suppression of cell invasionand invadopodia formation was mediated by palladin inhibition.We infected FLAG.SATB2 HCT116 cells with GFP or GFP–palladin(90 kDa isoform) and then assessed cell invasion. Although

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

expression of endogenous palladin in FLAG–SATB2/GFP–palladincells was reduced compared to that in FLAG cells, GFP–palladin wasexpressed to the level compatible with the endogenous palladin inFLAG cells (Fig. 6A). Cell invasion assay demonstrated that expres-sion of palladin in FLAG–SATB2 cells clearly restored invasiveactivity (Fig. 6B) similar to the level of FLAG cells (Fig. 1B). We nextexamined whether invadopodia formation was restored by palladinexpression. Cells were immunostained for cortactin and observedunder the confocal microscopy.

As shown in Fig. 6C, FLAG–SATB2/GFP cells showed GFP fluor-escence distributed over the cells and low dot-like concentrationsof cortactin. However, FLAG–SATB2/GFP–palladin cells showed anaccumulation of GFP–palladin and cortactin at the edge of cells(Fig. 6D). We quantified these results and found that the percen-tage of cells with invadopodia in FLAG–SATB2/GFP–palladin cellsincreased 1.6-fold compared to FLAG-SABT2/GFP cells (Fig. 6D).Collectively, these results indicate that the suppression of cellinvasion and invadopodia formation by SATB2 was partly mediatedby inhibition of palladin.

Discussion

In this report, we demonstrate that exogenous expression ofSATB2 in HCT116 reduced invasion and invadopodia formationthrough the inhibition of palladin. Expression of SATB2 in HCT116,CW-2, COLO 320 clearly suppressed expression of multiple iso-forms of palladin. In addition, palladin mRNA in HCT116 wasreduced by SATB2 expression. These results clearly indicate thatSATB2 expression can directly or indirectly suppress palladinexpression. SATB2 is composed of two CUT domains and ahomeobox domain at the C-terminus [2]. The homeobox domainbinds to conserved DNA sequences to promote the transcriptionof target genes [22]. However, the CUT domains trigger binding tothe AT-rich nuclear matrix attachment region (MAR) sequencesfor chromosomal remodeling [3,4]. SATB2-mediated chromosomerearrangement may induce structural changes in the genomicregion of palladin and suppress its expression. Recently, SATB2was also found to mediate differential expression of microRNAs(miRNAs) in murine bone marrow stromal cells for osteogenicdifferentiation [23]. MicroRNAs are small RNAs that bind to mRNAto inhibit translation or promote mRNA degradation. Most of themicroRNA target regions are located to the 3'UTR of mRNAs.Therefore, we cannot exclude the possibility that miRNAs inducedby SATB2 may have caused reduction of palladin isoforms know-ing that they have the same C-terminal region.

The role of palladin in the metastasis-promoting behavior ofcancer-associated fibroblasts (CAF) of pancreatic tumors and otherinvasive tumor types has been well investigated [15–18]. Palladincan bind to multiple actin-associated proteins, such as actinin,lasp1 and CLP36, and regulate rearrangement of actin cytoskele-ton for cell migration [24]. Recently, palladin was found toregulate podosome and invadopodia formation through RhoGTPases, which are known regulators of cellular responsesrequired for cell migration and metastasis [17]. Nevertheless,other reports [25,26] including our previous study [19] supportedthe antimigratory or antitumorigenic functions of palladin. In ourprevious study, we reported that extracellular signal-regulatedkinase (ERK) phosphorylated palladin to inhibit the palladinassociation with Abl tyrosine kinase and reduce cell migration.

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

Fig. 5 – Rescue of palladin restores invadopodia formation. (A) HCT116 constitutively expressing shCtrl, shPALLD #2/GFP andshPALLD #2/GFP PALLD (Res) were generated by retroviral infection. The endogenous and exogenous palladin expression wasdetected in these cell lines by immunoblotting using anti-palladin and anti-GFP. An arrow indicates exogenous palladin level.(B) HCT116 cells that constitutively expressed shPALLD #2/GFP and shPALLD #2/GFP PALLD (Res) were fixed with 4%paraformaldehyde and immunostained for GFP (green) and cortactin (red). Representative micrographs are shown (scalebar¼20 μm). The graph indicates the average percentage of cells with invadopodia in the cells (*Po0.05).

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ] 9

However, expression of non-phosphorylated palladin enhancedcell migration [19]. Chin and Toker [26] reported that Akt1phosphorylates palladin to modulate its activity in bundling actinstress fibers and to inhibit cell migration. However, phosphoryla-tion levels of palladin were uneven among breast cancer cells andcorrelated inversely with cancer cell invasiveness [26,27]. There-fore, depending on the cellular context, expression of palladinmay promote or suppress cell migration according to phosphor-ylation status in each cell.

Palladin has multiple isoforms and most of them haveC-terminal immunoglobulin (Ig) domains [12]. Although thedifference in function between palladin isoforms remainsobscure, a previous study [11] reported that whereas the90 kDa isoform binds to three actin-associated proteins (ezrin,VASP and alpha-actinin), the 140 kDa palladin has a conservedbinding site for SH3 domains, suggesting that it binds directly

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

to the SH3-domain protein, Lasp-1. Overexpression of the90 kDa palladin induced bundling of stress fibers, whereas the140 kDa isoform produced star-like arrays of actin fibers,indicating that these isoforms have different functions in actincytoskeletal rearrangement [11]. In the current study, we usedthe 90 kDa isoform for the rescue and overexpression experi-ments. Although HCT116 cells expressed multiple isoforms ofpalladin and both shRNAs suppressed these isoforms, expres-sion of the 90 kDa palladin in SATB2-expressing cells wassufficient to restore invasion and invadopodia formation.Although other isoforms may be associated with actin cytoske-letal rearrangement and cellular functions, our results showthat the 90 kDa palladin is a major isoform that controlsinvasion and invadopodia formation.In conclusion, we provide insights into the role of SATB2 in

suppressing invadopodia formation and invasion of CRC cells, and

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003

Fig. 6 – Palladin expression restores invadopodia formation in SATB2-expressing cells. (A) HCT116 cells constitutively expressingFLAG, FLAG.SATB2/GFP and FLAG.SATB2/GFP–palladin were generated by retroviral infection. Endogenous and exogenous palladinexpression were detected by immunoblotting using anti-palladin and confirmed by anti-GFP. An arrow indicates the exogenouspalladin expression (B) HCT116 cells constitutively expressing FLAG.SATB2 or FLAG.SATB2/GFP–palladin were subjected to aninvasion assay. Representative images of invaded cells are shown, and the graph indicates the average number of invaded cells perfield (*Po0.05). (C) HCT116 cells that constitutively expressed FLAG.SATB2/GFP were fixed with 4% paraformaldehyde andimmunostained for GFP (green) and cortactin (red). Representative micrographs are shown (scale bar¼20 μm) (D) HCT116 cellsthat constitutively expressed FLAG.SATB2/GFP–palladin were fixed with 4% paraformaldehyde and immunostained for GFP (green)and cortactin (red). Representative micrographs are shown (scale bar¼20 μm). The graph indicates the average percentage of cellswith invadopodia in FLAG.SATB2/GFP and FLAG.SATB2/GFP–palladin cells (*Po0.05).

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ]10

Please cite this article as: M.A. Mansour, et al., Special AT-rich sequence-binding protein 2 suppresses invadopodia formation inHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.12.003

E X P E R I M E N T A L C E L L R E S E A R C H ] ( ] ] ] ] ) ] ] ] – ] ] ] 11

demonstrate its action is mediated by palladin inhibition. Palladinwas found to be essential for cancer cell migration and invasion.These results suggest that the SATB2/palladin pathway will likelyprovide new targets for molecular therapies that aim to controlthe aggressive metastasis of colorectal cancer.

Disclosure

The authors have no conflicts of interest.

Acknowledgments

We would like to thank the members of the Division of CancerBiology for their helpful discussions and Dr. Yoshihiko Yamakita(Nagoya University, Japan) for gelatin degradation assay. This researchwas funded by a grant from the Ministry of Education, Culture, Sports,Science and Technology of Japan (23107010 to T. Senga).

r e f e r e n c e s

[1] J. Weitz, M. Koch, J. Debus, T. Höhler, P.R. Galle, M.W. Büchler,Colorectal cancer, Lancet 365 (9454) (2005) 153–165.

[2] G. Dobreva, J. Dambacher, R. Grosschedl, SUMO modification of anovel MAR-binding protein, SATB2, modulates immunoglobulinmu gene expression, Genes Dev. 17 (24) (2003) 3048–3061.

[3] D.R. FitzPatrick, I.M. Carr, L. McLaren, J.P. Leek, P. Wightman,K. Williamson, et al., Identification of SATB2 as the cleft palategene on 2q32-q33, Hum. Mol. Genet. 12 (19) (2003) 2491–2501.

[4] O. Britanova, S. Akopov, S. Lukyanov, P. Gruss, V. Tarabykin, Noveltranscription factor Satb2 interacts with matrix attachmentregion DNA elements in a tissue-specific manner and demon-strates cell-type-dependent expression in the developing mouseCNS, Eur. J. Neurosci. 21 (3) (2005) 658–668.

[5] G. Dobreva, M. Chahrour, M. Dautzenberg, et al., SATB2 is amultifunctional determinant of craniofacial patterning andosteoblast differentiation, Cell 125 (5) (2006) 971–986.

[6] O. Britanova, M.J. Depew, M. Schwark, et al., Satb2 haploinsuffi-ciency phenocopies 2q32-q33 deletions, whereas loss suggests afundamental role in the coordination of jaw development, Am.J. Hum. Genet. 79 (4) (2006) 668–678.

[7] J.A. Rosenfeld, B.C. Ballif, A. Lucas, et al., Small deletions of SATB2cause some of the clinical features of the 2q33.1 microdeletionsyndrome, PLoS One 4 (8) (2009) e6568.

[8] T.R. Liu, L.H. Xu, A.K. Yang, Q. Zhong, M. Song, M.Z. Li, et al.,Decreased expression of SATB2: a novel independent prognosticmarker of worse outcome in laryngeal carcinoma patients, PloSOne 7 (7) (2012) e40704.

[9] S. Wang, J. Zhou, X.Y. Wang, J.M. Hao, J.Z. Chen, X.M. Zhang, et al.,Down-regulated expression of SATB2 is associated with metas-tasis and poor prognosis in colorectal cancer, J. Pathol. 219 (1)(2009) 114–122.

Please cite this article as: M.A. Mansour, et al., Special AT-rich seqHCT116 cells via palladin inhibition, Exp Cell Res (2014), http://dx.d

[10] J. Eberhard, A. Gaber, S. Wangefjord, B. Nodin, M. Uhlen,K. Ericson Lindquist, et al., A cohort study of the prognostic andtreatment predictive value of SATB2 expression in colorectalcancer, Br. J. Cancer 106 (5) (2012) 931–938.

[11] A.S. Rachlin, C.A. Otey, Identification of palladin isoforms andcharacterization of an isoform-specific interaction between Lasp-1 and palladin, J. Cell Sci. 119 (Pt 6) (2006) 995–1004.

[12] L. Jin, The actin associated protein palladin in smooth muscle andin the development of diseases of the cardiovasculature and incancer, J. Muscle Res. Cell Motil. 32 (1) (2011) 7–17.

[13] M.M. Parast, C.A. Otey, Characterization of palladin, a novelprotein localized to stress fibers and cell adhesions, J. Cell Biol.150 (3) (2000) 643–656.

[14] S. Goicoechea, D. Arneman, A. Disanza, R. Garcia-mata, G. Scita,C.A. Otey, Palladin binds to Eps8 and enhances the formation ofdorsal ruffles and podosomes in vascular smooth muscle cells,J. Cell Sci. 119 (Pt 16) (2006) 3316–3324.

[15] S.M. Goicoechea, B. Bednarski, R. García-mata, H. Prentice-dunn,H.J. Kim, C.A. Otey, Palladin contributes to invasive motility inhuman breast cancer cells, Oncogene 28 (4) (2009) 587–598.

[16] S.M. Goicoechea, B. Bednarski, C. Stack, et al., Isoform-specificupregulation of palladin in human and murine pancreas tumors,PLoS One 5 (4) (2010) e10347.

[17] S.M. Goicoechea, R. García-mata, J. Staub, et al., Palladin promotesinvasion of pancreatic cancer cells by enhancing invadopodiaformation in cancer-associated fibroblasts, Oncogene 33 (10)(2014) 1265–1273.

[18] P. Von nandelstadh, E. Gucciardo, J. Lohi, et al., Actin-associatedprotein palladin promotes tumor cell invasion by linking extra-cellular matrix degradation to cell cytoskeleton, Mol. Biol. Cell 25(17) (2014) 2556–2570.

[19] E. Asano, M. Maeda, H. Hasegawa, et al., Role of palladinphosphorylation by extracellular signal-regulated kinase in cellmigration, PLoS One 6 (12) (2011) e29338.

[20] P. Najm, M. El-sibai, Palladin regulation of the actin structuresneeded for cancer invasion, Cell Adhes. Migr. 8 (1) (2014) 29–35.

[21] H.S. Leong, A.E. Robertson, K. Stoletov, et al., Invadopodia arerequired for cancer cell extravasation and are a therapeutic targetfor metastasis, Cell Rep 8 (5) (2014) 1558–1570.

[22] N. Shah, S. Sukumar, The Hox genes and their roles in onco-genesis, Nat. Rev. Cancer. 10 (5) (2010) 361–371.

[23] Y. Gong, F. Xu, L. Zhang, et al., MicroRNA expression signature forSatb2-induced osteogenic differentiation in bone marrow stro-mal cells, Mol. Cell Biochem. 387 (1–2) (2014) 227–239.

[24] M. Maeda, E. Asano, D. Ito, et al., Characterization of interactionbetween CLP36 and palladin, FEBS J. 276 (10) (2009) 2775–2785.

[25] P.N. Tay, P. Tan, Y. Lan, et al., Palladin, an actin-associated protein,is required for adherens junction formation and intercellularadhesion in HCT116 colorectal cancer cells, Int. J. Oncol. 37 (4)(2010) 909–926.

[26] Y.R. Chin, A. Toker, The actin-bundling protein palladin is anAkt1-specific substrate that regulates breast cancer cell migra-tion, Mol. Cell 38 (3) (2010) 333–344.

[27] Y.R. Chin, A. Toker, Akt isoform-specific signaling in breastcancer: uncovering an anti-migratory role for palladin, CellAdhes. Migr. 5 (3) (2011) 211–214.

uence-binding protein 2 suppresses invadopodia formation inoi.org/10.1016/j.yexcr.2014.12.003