Exosomes Promote Ovarian Cancer Cell Invasion through ... · Exosomes Promote Ovarian Cancer Cell Invasion through Transfer of CD44 to Peritoneal Mesothelial Cells Koji Nakamura1,

Signal Transduction

Exosomes Promote Ovarian Cancer Cell Invasionthrough Transfer of CD44 to PeritonealMesothelial CellsKoji Nakamura1, Kenjiro Sawada1, Yasuto Kinose1, Akihiko Yoshimura1, Aska Toda1,Erika Nakatsuka1, Kae Hashimoto1, Seiji Mabuchi1, Ken-ichirou Morishige2,Hirohisa Kurachi3, Ernst Lengyel4, and Tadashi Kimura1

Abstract

Epithelial ovarian cancer (EOC) cells metastasize within theperitoneal cavity and directly encounter human peritonealmesothelial cells (HPMC) as the initial step of metastasis. Thecontact between ovarian cancer cells and the single layer ofmesothelial cells involves direct communications that modulatecancer progression but the mechanisms are unclear. One can-didate mediating cell–cell communications is exosomes, 30–100 nm membrane vesicles of endocytic origin, through thecell–cell transfer of proteins, mRNAs, or microRNAs. Therefore,the goal was to mechanistically characterize how EOC-derivedexosomes modulate metastasis. Exosomes from ovarian cancercells were fluorescently labeled and cocultured with HPMCswhich internalized the exosomes. Upon exosome uptake,HPMCs underwent a change in cellular morphology to a mes-enchymal, spindle phenotype. CD44, a cell surface glycoprotein,was found to be enriched in the cancer cell–derived exosomes,transferred, and internalized to HPMCs, leading to high levels of

CD44 in HPMCs. This increased CD44 expression in HPMCspromoted cancer invasion by inducing the HPMCs to secreteMMP9 and by cleaning the mesothelial barrier for improvedcancer cell invasion. When CD44 expression was knocked downin cancer cells, exosomes had fewer effects on HPMCs. Theinhibition of exosome release from cancer cells blocked CD44internalization in HPMCs and suppressed ovarian cancer inva-sion. In ovarian cancer omental metastasis, positive CD44expression was observed in those mesothelial cells that directlyinteracted with cancer cells, whereas CD44 expression wasnegative in the mesothelial cells remote from the invading edge.This study indicates that ovarian cancer–derived exosomestransfer CD44 to HPMCs, facilitating cancer invasion.

Implications:Mechanistic insight from the current study suggeststhat therapeutic targeting of exosomes may be beneficial intreating ovarian cancer. Mol Cancer Res; 15(1); 78–92. �2016 AACR.

IntroductionOvarian cancer, which has a cure rate of only 30%, is

generally characterized by widespread peritoneal dissemina-tion and ascites (1). During ovarian cancer dissemination, cellsdetach from the primary site of origin, which was initiallythought to be the ovary but is now believed to be the fallopiantube. Subsequently, the cancer cells float in the peritoneal

cavity and attach to secondary sites of implantation, mostnotably the omentum, which is the most common site ofovarian cancer metastasis (2). All of the organs within theperitoneal cavity are lined with a single layer of mesothelialcells which cover an underlying stroma composed of extracel-lular matrices (ECM) and stromal cells (3). Once ovariancancer cells attach to this layer, they invade through the barrierof mesothelial cells into the peritoneum, omentum, or bowelserosa. Therefore, inhibiting the initial interaction of cancerwith mesothelial cells could be a means of preventing ovariancancer progression and metastasis.

Mesothelial cells were first thought to function only as amechanical barrier. This view was supported by histologicstudies of ovarian cancer nodules attached to peritoneal organsthat showed an absence of mesothelial cells under the tumor,suggesting that the mesothelial cells had failed as a protectivebarrier and contributed to successful tumor cell implantation(4–7). Recent reports suggest that mesothelial cells activelyparticipate in establishing an ovarian cancer metastatic nicheand that they are reprogrammed by ovarian cancer cells tofacilitate tumor growth becoming cancer-associated mesothe-lial cells (2, 8, 9).

Exosomes are 30- to 100-nm membrane vesicles of endocyticorigin, mediating diverse biologic functions, including tumor cellinvasion, cell–cell communication, and antigen presentation

1Department of Obstetrics and Gynecology, Osaka University Graduate Schoolof Medicine, Suita, Osaka, Japan. 2Department of Obstetrics and Gynecology,Gifu University Graduate School of Medicine, Gifu, Japan. 3Osaka Medical Centerand Research Institute for Maternal and Child Health, Izumi, Osaka, Japan.4Department of Obstetrics and Gynecology/Section of Gynecologic Oncology,University of Chicago, Chicago, Illinois.

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

E. Lengyel and T. Kimura shared senior authorship.

Corresponding Author: Kenjiro Sawada, Department of Obstetrics and Gyne-cology, Osaka University Graduate School of Medicine, 2–2, Yamadaoka Suita,Osaka 5650871, Japan. Phone: 81-6-6879-3351; Fax: 81-6-6879-3359; E-mail:[email protected].

doi: 10.1158/1541-7786.MCR-16-0191

�2016 American Association for Cancer Research.

MolecularCancerResearch

Mol Cancer Res; 15(1) January 201778

on June 22, 2020. © 2017 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst October 6, 2016; DOI: 10.1158/1541-7786.MCR-16-0191

http://mcr.aacrjournals.org/

through transfer of proteins, mRNAs, and microRNAs (10). Thevery recently characterized mechanism of cell–cell communica-tion through exosomes raises the possibility that the transfer ofgenetic information or proteins via exosomes modulates cellularactivities in recipient cells (10).

Given that the role of mesothelial cells during ovarian cancermetastasis is unclear and that exosomes are able to transferinformation between different cell types, we set out to determinewhether exomes play a role in communication between meso-thelial cells and ovarian cancer cells.

Materials and MethodsMaterials

GW4869 (#92613), 5-(N-ethyl-N-isopropyl)-amiloride (EIPA;#A3058), PKH67 (MIDI67), and Ponceau-S (P3045-10G) werepurchased from Sigma-Aldrich. E-Cadherin antibody (#610181)and the integrin antibodies Sampler Kit (#611435; integrin-a2,-a5, -aV, -b1, and -b3) were purchased from BD Biosciences.Antibodies against integrin-a1 (sc-10728), cytokeratin-18 (sc-6259), and Oct-3/4 (sc-9018) were obtained from Santa CruzBiotechnology. Antibodies against b-actin (#4967) and CD44(156-3C11) were obtained from Cell Signaling. Alexa Fluor633 Goat Anti-Mouse IgG (A21050), Lipofectamine 2000(#11668019), TRIzol (#15596-018), and CellTracker GreenCMFDA (C2925) were obtained from Life Technologies. DonkeyAnti-Mouse IgG 10 nm gold (ab39593), Alexa Fluor 488 GoatAnti-Mouse IgG H&L (ab150113), Alexa Fluor 555 Donkey Anti-Rabbit IgGH&L (ab150074), antibody againstMMP9 (ab53296),and antibody against CD63 (ab134045) were from Abcam. Anti-body against CD63 (EXOAB-CD63A-1) was purchased from SBI.Antibody against CD63 (#11-343-C025) and CD81 (#11-558-C025) was purchased from EXBIO. Antibody against vimentin(M0725) was purchased from DAKO. Antibody against calretinin(RB-9002-P0) was purchased from Thermo Fisher Scientific Inc.Antibody against MMP9 (GTX-100458) was purchased fromGenetex. Recombinant human TGFb1 (#100-21) was purchasedfrom PeproTech. LY2157299 (#120-05991) was purchased fromWako Pure Chemical Industries.

Cell cultureThe HeyA8 cell line was generously provided by Dr. Anil Sood

(MD Anderson Cancer Center, Houston, TX). The IOSE cell linewas generously provided by Dr. Masaki Mandai (Kinki Univer-sity, Osaka, Japan). Briefly, ovarian surface epithelial cells werepurified from normal ovary and transfected with SV40 large Tantigen and the human telomerase reverse transcriptase (hTERT;ref. 11). TYK-nu cell lines were purchased from Health ScienceResearch Resources Bank (Osaka, Japan). RMG-1, RMUG-S,OVSAHO, and OVKATE cell lines were purchased from HealthScience Research Resources Bank. OVCAR-3 cell line was pro-vided by the Cell Resource Center for Biomedical ResearchInstitute of Development, Aging and Cancer of Tohoku Uni-versity (Sendai, Japan). ES-2 and Caov-3 cell lines were pur-chased from ATCC. Cells were authenticated by short tandemrepeat DNA profiling at Takara-Bio Inc. and were used for thisstudy within 6 months of resuscitation.

Isolation of exosomesExosomes were isolated from cell culture medium as pre-

viously described (12) with minor modifications. Conditioned

medium (CM) containing exosome-depleted FBS (preparedby overnight ultracentrifugation at 100,000 � g at 4oC) wasprepared by incubating cells grown at subconfluence for48 hours. CM was centrifuged at 2,000 � g for 20 minutesat room temperature. Supernatant fractions were filteredusing 200-nm pore size filters. The resulting cell-free mediumwas ultracentrifuged at 100,000 � g for 70 minutes to generatean exosome pellet that was washed once with PBS usingthe same ultracentrifugation conditions, followed by resus-pension in PBS. The amount of exosomal protein was assessedby the Lowry method (Bio-Rad).

Electron microscopyElectron microscopy was performed as described (12) using a

transmission electron microscope (H-7650; Hitachi, Ltd.).

Western blottingWestern blotting was performed as described (13). The

primary antibodies were used in the following dilutions: CD44(1:1,000), CD63 (1:250), CD81 (1:250), E-cadherin (1:1,000),b-actin (1:2,000), integrin-a1 (1:200), integrin-a2 (1:1,000),integrin-a5 (1:2,500), integrin-aV (1:2,500), integrin-b1(1:2,500), integrin-b3 (1:1,000), Oct-3/4 (1:200), and calreti-nin (1:500).

ImmunohistochemistryImmunohistochemistry of ovarian cancer paraffin-embedded

tissuewas performed as previously described (14). The slideswereincubated with the primary antibody against CD44 (#156-3C11,1:200), calretinin (1:200), or MMP9 (1:50). For human perito-nealmesothelial cell (HPMC) staining, cells were plated on 8-wellchamber slides and were fixed with 4% paraformaldehyde. Slideswere stained with an antibody against cytokeratin-18 (1:50),vimentin (1:50), or calretinin (1:100).

Immunofluorescent analysisHPMCs were fixed with 4% paraformaldehyde and stained

with anti-CD44 antibody (#156-3C11, 1:200) at room temper-ature for 1 hour (15). After washing, samples were incubated withAlexa Fluor 633–labeled goat anti-mouse IgG (1:100) and stainedwith DAPI. The samples were imaged using an FV1000-D LaserScanning Confocal microscope (Olympus).

Plasmid and siRNA transfectionOverexpression of CD44s in HPMCs was performed by tran-

sient transfection using the CD44s pBabe-puro retroviral vector(Plasmid#19127, Addgene) at a concentration of 1 mg/mL. Twen-ty-four hours after transfection, HPMCs were used for experi-ments. The retroviral vector pBABEpuro gateway (Plas-mid#51570, Addgene) served as a control. CD44 expression inHeyA8 cells was transiently knocked down using a predesignedsiRNA duplex directed against CD44 at a concentration of 200nmol/L. The sequences of the siRNA duplexes (Japan Bioservice)were as follows: CD44 siRNA #1, 50-AAAUGGUCGCUACAG-CAUCTT-30 (sense) and 50-GAUGCUGUAGCGACCAUUUTT-30

(antisense); CD44 siRNA #2, 50-GUAUGACACAUAUUG-CUUCTT-30 (sense) and 50-GAAGCAAUAUGUGUCAUACTT-30

(antisense). MISSION siRNA Universal Negative Control #1(SCI001, Sigma) was used as a control.

Exosomes Promote Ovarian Cancer Cell Invasion

www.aacrjournals.org Mol Cancer Res; 15(1) January 2017 79




C

CD63

β-Actin

CD81

TYK

Cell Exo ExoExoCell Cell-nuHeyA8 IOSE

25

50

40

35

50

(kDa)

E

contrastCytokeratin

Vimentin

control

HPMC

Calretinin

Phase

Negative

18

F

HP

MC

HeyA8-exo

IOSE-exo

Exo

som

e tr

eatm

ent

DAPIPKH67

Merged contrast(Exosome)Phase

PKH67 Dye only

TYK-nu-exo

B

0

15

30

45

60

Exo

som

al p

rote

in (

µg)

Dcellular RNA

Small RNARibosomal RNA

18s 28s

Exosomal RNA

Small RNARibosomal RNA

18s 28s

A

CD63

Negative control

HeyA8 TYK-nu ES-2 Caov-3 RMUG-S IOSE

Exosome isolated from cell culture media of cell lines

OVSAHO

Epithelial ovarian cancer cell linesNonmalignantcontrol cell line

25 200 1,000 4,000 (nt) 25 200 1,000 4,000 (nt)

(FU)

40

20

0

(FU)40

20

0

Nakamura et al.

Mol Cancer Res; 15(1) January 2017 Molecular Cancer Research80




Profiling of cellular and exosomal RNATotal RNA was extracted using TRIzol. RNA isolated from cells

and exosomes were analyzed using the Agilent 2100 Bioanalyzer(Agilent Technologies, Inc.).

Exosome labeling and treatmentFor exosome tracking, exosomes were fluorescently labeled

using the PKH67 dye according to manufacturer's protocol.

HPMC-coated in vitro invasion assayHPMCs (2 � 104) were plated onto 25-mg Matrigel-coated

24-well FluoroBlok cell culture inserts with an 8-mm poremembrane (351152 Corning Incorporated). Because Fluoro-Blok membranes almost completely block the passage of lightfrom 490 to 700 nm, only fluorescently labeled cells that haveinvaded are detected by bottom-reading of fluorescence (16).After HPMCs grew to confluence, 5 � 104 fluorescently(CMFDA)-labeled epithelial ovarian cancer (EOC) cells (HeyA8and TYK-nu) cultured in 0.1% BSA/DMEM were incubated onthe monolayer of HPMCs and allowed to invade for 48 hours.Cell culture medium consisting of 10%FBS/DMEM was used asa chemoattractant and was kept in the lower chamber. Fluo-rescence of invaded cells was read at 494/517 nm (Ex/Em) on abottom-reading fluorescent plate reader (SH-9000Lab, CoronaElectric Co., Ltd.), and the number of invaded cells was quan-titated by imaging an average of 25 random areas per well. Inexperiments using exosomes, culture media were changed to0.1% BSA/DMEM, once HPMCs grew to confluence, and iso-lated exosomes were added 24 hours before the treatment withcancer cells.

Matrigel in vitro invasion assayCMFDA-labeled EOC cells were directly plated onto Matrigel-

coated inserts. When exosomes were used, isolated exosomes in0.1% BSA/DMEM were added to the Matrigel-coated inserts 24hours before treatment with cancer cells.

Cell proliferation assayEOC cells were plated onto 96-well plates (5 � 103 cells/well)

in 10%FBS/DMEM for 24 hours and incubated in 0.1% BSA/DMEMwith orwithout exosomes (100mg/mL) orGW4869 for 48hours. Primary HPMCs were plated onto 96-well plates (1 � 104

cells/well) in 20%FBS/RPMI for 24 hours and incubated in 0.1%BSA/RPMI with or without exosomes (1, 10, 100 mg/mL) for 24hours. Cell proliferationwas evaluatedby amodifiedMTSassay aspreviously described (15).

Mesothelial clearance assayPrimary human HPMCs (5 � 104) were plated on a 24-well

plate and incubated until confluent (7). Culture media wereswitched to 0.1%BSA/DMEM, and 100 mg/mL of isolated exo-somes was added. Twenty-four hours following the incubation,the cells were fixed and stained with Acti-Stain 555–conjugatedphalloidin (#PHDH1, Cytoskeleton). Mesothelial clearance areawas defined as cell-free area, and mesothelial clearance wascalculated as the ratio of cell-free area to total area using theImageJ Software (NIH, Bethesda, MD).

ZymographySerum-free conditioned medium was collected from 1 �

105 HPMCs with or without 100 mg/mL of EOC-derivedexosomes, and a gelatin zymogram was performed as previ-ously described (17).

Quantitative real-time PCRTotal RNA was extracted using TRIzol, and cDNA was synthe-

sized using ReverTra Ace qPCR RT Master Mix (FSQ-201;TOYOBO). qRT-PCR was performed using the StepOnePlusReal-Time PCR System. The following probes were used: humanmatrix metalloproteinase-9 (MMP9: Hs00234579_m1) andhuman GAPDH (4326317E). Relative expression of CD44 wasassayed using the SYBR Green assay (Applied Biosystems:#4309155). The designed primers in this experiment were asfollows: CD44: forward primer: 50-AGTCACAGACCTGCC-CAATGCCTTT-3, reverse primer: 50-TTTGCTCCACCTTCTT-GACTCCCATG-30; GAPDH: forward primer: 50-CATGAGAAG-TATGACAACAGCCT-30, reverse primer: 50-AGTCCTTCCACGA-TACCAAAGT-30 (Applied Biosystems). PCR was performed intriplicate, and the relative levels of CD44 and MMP9 expressionwere calculated using the 2�DDCt method.

ELISAThe concentration of TGFb in exosomes was analyzed using

Human/Mouse TGFb 1 ELISA Ready-SET-Go! (#88-8350;eBioscience).

GW4869 treatmentGW4869 (18) was stored at �20�C as a 3 mmol/L stock

suspension in DMSO. Right before use, the suspension wassolubilized by the addition of 5% methane sulfonic acid (MSA;Wako).

Measurement of particle size and concentration distributionNanoparticles in exosome suspensions were analyzed using a

NanoSight LM10V-HS (Malvern Instruments Ltd.).

Figure 1.EOC-derived exosomes are transferred to human primary mesothelial cells. A, Electron microscopy: exosomes were immunogold labeled with anti-CD63antibody and transmission electron micrographs of purified exosomes secreted from several EOC cell lines taken. IOSEs served as control. Scale bar,100 nm B, Protein content: 2 � 108 EOC cells were incubated for 48 hours, exosomes collected, and protein determined using the Lowry proteinassay. C, Western blot analysis on isolated exosomes and parental cellular lysates (10 mg/lane). Exosomes were isolated from 2 EOC cell lines (HeyA8 andTYK-nu) and IOSE. D, RNA was extracted from EOC-derived exosomes and analyzed by a Bioanalyzer to detect ribosomal subunits 28S and 18S (left).A profile of cellular RNA is shown as a control (right). Arrows indicate the position of small RNA, 18S and 28S. E, Primary HPMCs isolated fromnormal human omentum. HPMCs stained for anti–cytokeratin-18, vimentin, and calretinin (negative control omitted primary antibody). Scale bar,100 mm. F, Confocal microscopy. HPMCs were cultured for 24 hours in the presence of PKH67, a fluorescent dye labeling exosomes. HPMCs are identifiedby the DAPI-stained, blue nuclei. Scale bar, 20 mm. Data in A, C, D, E, F Are representative of 3 experiments with similar results. Data in B are shown asmean � SEM of 3 independent experiments.






Tissue samplesSpecimens of human omentum were obtained following writ-

ten informed consent from patients undergoing gynecologicalsurgeries at Osaka University Hospital (Osaka, Japan). Sampleswere obtained in accordance with the Declaration of Helsinki(Institutional Review Board of Osaka University Graduate Schoolof Medicine approval no.: 10064). HPMCs were isolated aspreviously described (19). Briefly, benign omental specimenswere minced and incubated with 1:1 solution of PBS and0.25% trypsin/25 mmol/L EDTA at 37�C for 30 minutes. Afterthe solution containing cells was centrifuged, the pellet wassuspended and cultured in 20% FBS/RPMI-1640 with 100 U/mLpenicillin/streptomycin. Cells at passage 1 were used for thesubsequent experiments. EOC samples were collected from 15patients with FIGO stage III–IV epithelial serous ovarian cancerconsecutively treated at Osaka University Hospital between 2011and2015.All caseswere histologically confirmedby a gynecologicpathologist as being high-grade serous carcinomas.

Animal experimentsFemale athymic BALB/c nude mice were purchased from CLEA

Japan, Inc. All animal experiments were approved by the Institu-tional Animal Care and Use Committee of Osaka University.HeyA8 cells indicated in the figures were injected intraperitone-ally. Seven days after the inoculation, themice were sacrificed andomental tissues were carefully dissected. No statistical methodswere used to predetermine sample size. The experiments werenot randomized, and the investigators were not blinded to allo-cation during experiments and outcome assessment.

Statistical analysisJMP software version 10.0.2 (SAS Institute Japan Ltd.) was

used for statistical analysis. All data had normal distribu-tion and equal variances were tested by F-test. Data areexpressed as mean � SEM. Differences were analyzed usingone-way ANOVA and Bonferroni correction for multiple com-parisons. Differences were considered statistically significantat P < 0.05.

ResultsOvarian cancer–derived exosomes can be transferred tomesothelial cells

To elucidate the effects of exosomes derived from EOC cells onperitoneal dissemination, exosomes were characterized by elec-tron microscopy, Western blotting, and capillary electrophoresisof RNA. Electronmicroscopy revealed that exosomes had a round-shapedmorphology andwere approximately 50 to 100 nm in sizeand stained positive for CD63, an exosome marker (ref. 20;Fig. 1A). Exosome secretion was measured in several EOC cells,and HeyA8 and TYK-nu were found to produce the highestquantity of exosomes (Fig. 1B) and were used in subsequentexperiments. As a nonmalignant control, exosomes from immor-talized ovarian surface epithelial cells (IOSE) were used. Immu-noblotting showed that CD63 (21)was seen in exosomes from allcell lines used in these experiments. CD81 (21), another exosomemarker, was also enriched in exosomes from HeyA8 cells andIOSEs, whereas whole-cell lysates expressed minimal CD63 andCD81 (Fig. 1C). In contrast, b-actin was expressed in whole-celllysates, whereas exosomes showed low expression (Fig. 1C). TotalRNAs extracted from whole cells using capillary electrophoresis

expressed the 18S and 28S ribosomal subunits, whereas they werenot detected in exosomes which expressed small RNAs, such asmiRNAs (ref. 22; Fig. 1D).

We then studied the internalization of EOC-derived exosomesby HPMCs. Primary HPMCs displayed a cobblestone-like mor-phology and expressed cytokeratin-18, vimentin (19), and calre-tinin, a discriminatory marker for mesothelial cells (Fig., 1E;ref. 23). Exosomes fromEOCcell lines and IOSEwere internalizedby HPMC and detected in the cytoplasm of the HPMCs usingconfocal laser microscopy (Fig. 1F), indicating that EOC-derivedexosomes can be internalized by HPMCs.

EOC-derived exosomes promote ovarian cancer invasionthrough increased mesothelial clearance and MMP induction

The role of EOC-derived exosomes in ovarian cancer inva-sion was studied using an in vitro invasion assay. To mimic themesothelial layer covering the omentum and other peritonealorgans, a monolayer of HPMCs was plated onto growth factor–reduced Matrigel (Fig. 2A). Pretreatment with HeyA8- or TYK-nu–derived exosomes on HPMCs significantly stimulated theinvasion of ovarian cancer cells, whereas pretreatment withIOSE-derived exosomes did not affect EOC invasion. EOCexosomes did not affect proliferation of HPMCs (Fig. 2B andC). In the absence of HPMCs (Matrigel-only coating), exosomesderived from either HeyA8 or TYK-nu cells did not promoteinvasion (Fig. 2D), nor did treatment with exosomes alonepromote ovarian cancer cell proliferation (Fig. 2E), indicatingthat exogenous treatment of EOC-derived exosomes affectedHPMCs but not cancer cells.

Next, we aimed to identify the changes in HPMC surfacereceptors induced by exosomes (Fig. 3A). The expression pat-tern of key proteins related to tumor cell adhesion or invasion(24) was examined by Western blotting (Fig. 3B). Uptake ofEOC-derived exosomes into HPMCs induced expression of thestandard form of CD44 (CD44s) and downregulation of E-cadherin, whereas the expression patterns of integrin-a1, -a2,-a5, -av, -b1, -b3 were unchanged (Fig. 3B). CD44 has severalisoforms generated by alternative mRNA splicing (25). Onlythe expression of CD44s (80 kDa) was increased in HPMCs(but not in IOSE), whereas the expression of other CD44variants (CD44v, 120–200 kDa) was very low (Fig. 3B). West-ern blotting as well as electron microscopy showed that CD44swas expressed by ovarian cancer–derived exosomes but notIOSE-derived exosomes (Fig. 3C and D). Three minutes afterthe uptake of PKH67-labeled exosomes, HPMCs were immu-nostained with Alexa Fluor 555–labeled CD44 antibody. CD44was colocalized with exosomes, indicating that CD44 derivedfrom ovarian cancer cells was actually internalized in HPMCsvia exosomes (Fig. 3E). There was no significant change inCD44 mRNA levels in HPMCs following treatment with EOC-derived exosomes, indicating that exosomes do not regulateCD44 expression on the transcriptional level (Fig. 3F). Giventhat E-cadherin expression was decreased in HPMCs, we con-sidered the possibility that mesothelial cells undergo epithe-lial–mesenchymal transition (EMT; ref. 2). Confocal micros-copy revealed that HPMCs treated with EOC-derived exosomesbecome elongated and spindle-shaped, whereas HPMCs trea-ted with IOSE-derived exosomes retained their cobblestone-like morphology (Fig. 1F). Because EOC-derived exosomesinduced the morphologic transition to a mesenchymal phe-notype, we determined HPMC cell clearance (6). Twenty-four

Nakamura et al.





hours after treatment with EOC-derived exosomes, cell–cellcontact between HPMCs was reduced and the cell-free areas (i.e., mesothelial clearance) were significantly greater than those

found with mesothelial cells treated with IOSE-derived exo-somes (Fig. 3G). Because CD44 is associated with a proteolyticform of the MMP9 on the cancer cell membrane and CD44-

A

C

Exosomes

CMFDA-labeled ovarian cancer cells

HPMCMonolayer

Matrigel

Fluorescence-blockingPET membrane Measurement of invading ovarian

cancer cells using fluorescence

24 h 48 hDMEM with10% FBS

DMEM with0.1% BSA

D E

HPMC-coated in vitro invasion assay

Rel

ativ

e in

vasi

ve

ovar

ian

canc

er c

ells

HeyA8

*

**

*

0.5

1

1.5

TYK-nu

*****

(Exosome: µg/mL)

*

*

0

0.5

1

1.5

0

Hey-exo10 50 100 10 50 100 10 50 100 10 50 100 10 50 10010 50 100

TYK-exo IOSE-exo Hey-exo TYK-exo IOSE-exo

PBS PBS

**

**

**

**

Exosomes added to HPMC monolayer Exosomes added to HPMC monolayer

Rel

ativ

e in

vasi

veov

aria

n ca

ncer

cel

ls

HPMC Coat (–) in vitro invasion assay

HeyA8 TYK-nu

0

0.5

1

0

0.5

1

Exosomes added onto Matrigel (100 µg/mL)

Cel

l via

bilit

y

HeyA8 TYK-nu

0

0.5

1

0

0.5

1

Exosomes added to ovarian cancer cells (100 µg/mL)

0

0.5

1

HeyA8-exo

0 1 10 1000

0.5

1

IOSE-exo

0 1 10 100Rel

ativ

e H

PM

C v

iabi

lity

(µg/mL)

HPMC + ExosomeB

Figure 2.

Ovarian cancer–derived exosomes promote ovarian cancer invasion through mesothelial cells. A, Schematic of the human peritoneal mesothelial cells in vitroinvasion assay. HPMCs were plated onto 24-well cell culture inserts coated with Matrigel. Twenty-four hours after exosomes were added to culturemedium, fluorescently (CMFDA) labeled ovarian cancer cells were incubated on HPMCs and allowed to invade for 48 hours. Cells that migrated to the trans-side of the membrane were quantitated using immunofluorescent microscopy. B, Cell proliferation assay. Primary HPMCs were plated onto 96-wellplates for 24 hours with or without exosomes (1, 10, 100 mg/mL). C, HPMC-coated in vitro invasion assay. Exosomes were added to culture medium at theindicated concentrations and invasion of CMFDA-labeled HeyA8 or TYK-nu cells was evaluated. D, Invasion assay. Exosomes (100 mg/mL) were addedto a Matrigel-coated insert. Twenty-four hours later, 5 � 104 fluorescently (CMFDA) labeled ovarian cancer cells (HeyA8 and TYK-nu) were added. E, Cellproliferation assay. HeyA8 or TYK-nu cells (5 � 103 cells) plated onto 96-well plates were treated with 100 mg/mL exosomes. Forty-eight hours later,proliferation was assessed by an MTS assay. B and C, cancer cell invasion was measured relative to control (PBS). Data in B and C are shown as mean� SEM of3 independent experiments. � , P < 0.05; �� , P < 0.01.






A

HGF

(100 µg/mL)

24 h

Exosome

HPMC Monolayer

Applied for each experiment

CR

elat

ive

CD

44 m

RN

A

PB

S

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

+ Exosome

0

0.5

1

HPMC

CD44s

CD44v

E-Cadherin

Integrin α1fusion protein

Integrin α2

Integrin αV

Integrin α5

Integrin β1

Integrin β3

β-Actin

PB

S

Hey

-exo

TY

K-e

xo

IOS

E- e

xo

140

RM

G-1

HPMC + Exosome

240

140

10070

140100

50

140

140

140

100

100

100

50

B

(kDa)

CD44s

CD63

CD44v

240

140

50

40

35

100

70

(kDa)

Rel

ativ

e M

MP

-9 m

RN

A **

PB

S

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

+ Exosome

******

HPMC

0

2

4

6

8

Pro MMP-9

Pro MMP-2

PB

S

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

+ Exosome

100

70

HPMC(kDa)

HPMC + IOSE-exo

HPMC + Hey-exo

Cel

l-fre

e ar

ea (

%)

PB

S

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

*****

+ Exosome

0

10

20

30

HPMC

HP

MC

HeyA8-exo

IOSE-exo

Exo

som

e tr

eatm

ent

PKH67 Dye only

DAPI CD44PKH67

(Exosome) Merged

CD44

HeyA8 TYK-nu IOSE

Exosome isolated from cell culture mediaD

E

I

Nakamura et al.





associated cell surface MMP9 is known to promote cell-medi-ated degradation of ECM (26), MMP9 expression in HPMCswas assessed after exosome uptake. Quantitative real-time PCRconfirmed that MMP9 mRNA expression was significantlyincreased in HPMCs treated with EOC-derived exosomes ascompared with those treated with PBS only or IOSE-derivedexosomes (HeyA-8; 5-fold, TYK-nu; 6.3-fold; Fig. 3H). Gelatinzymography was performed to analyze whether EOC-derivedexosomes induce HPMCs to secrete MMP. Conditionedmedium from HPMCs treated with EOC-derived exosomesincreased the gelatinolytic activity of pro-MMP9, whereasthey did not affect gelatinolytic activity of MMP-2 (Fig. 3I).The involvement of MMP9 activity in EOC invasion wasanalyzed using an anti-MMP9 antibody after its inhibitoryeffect against MMP9 activity was confirmed by gelatinzymography (Supplementary Fig. S1A). EOC invasion wassignificantly decreased by MMP9 inhibition, especiallywhen HPMCs were pretreated with EOC-derived exosomes(Supplementary Fig. S1B). Collectively, these data show thatEOC-derived exosomes promote ovarian cancer invasion byaffecting HPMCs through the increase of mesothelial clearanceand the induction of CD44-mediated MMP9 secretion byHPMCs.

Inhibition of exosome secretion attenuated CD44 acquisitionby HPMCs and blocked ovarian cancer invasion

GW4869, a neutral sphingomyelinase inhibitor, inhibits exo-some secretion (18). Addition of GW4869 to HeyA8 (Fig. 4A andB) and TYK-nu (Supplementary Fig. S2A) ovarian cancer cellsblocked exosome secretion in a dose-dependentmanner, withoutaffecting cell viability (Fig. 4C, Supplementary Fig. S2B). GW4869treatment significantly inhibited (49%) the invasion of both celllines in the presence of HMPC coating, whereas in the absence ofHPMCs, invasion was unchanged (Fig. 4D, Supplementary Fig.S2C). Consistent with a regulation of CD44 by exosomes,GW4869 attenuated CD44 uptake in HPMCs cocultured withEOC cells in a dose-dependent manner. E-cadherin expressionwas inversely correlated with CD44, suggesting that the inhibitorblocked EMT in the HPMCs (Fig. 4E, Supplementary Fig. S2D)which was also confirmed by phase-contrast microscopy (Fig. 4F;Supplementary Fig. S2E). When EIPA, a known exosome uptakeinhibitor that blocks macropinocytosis (27), was treated withHPMCs cocultured with EOC cells (Supplementary Fig. S3),CD44 expression in HPMCs was similarly inhibited.

In addition, GW4869 restored the epithelial appearance ofHPMC in a dose-dependent manner (Fig. 4F). These resultsindicate that the inhibition of exosomes released from EOC cellsinhibit ovarian cancer invasion by attenuating CD44 expressionand reversing EMT in HPMCs.

EOC-derived exosomes induce morphologic change in HPMCsindependent of TGFb signaling

Because TGFb signaling is one of the key regulators of EMT (28)and is also correlated with enhanced CD44 expression (29), westudied the involvement of TGFb in the acquisition of CD44 byHPMCs. At 10 ng/mL, TGFb induced CD44 expression anddownregulated E-cadherin expression, effects which were abro-gated by use of the potent TGFb receptor I kinase inhibitorLY2157299 (Fig. 5A). In addition, TGFb induced themorphologicchange of HPMCs to a spindle shape, an effect which was alsoinhibited by LY2157299 (Fig. 5B). The TGFb receptor I kinaseinhibitor did not interfere with the ability of EOC-derived exo-somes to upregulate CD44s and downregulate E-cadherin inHPMCs (Fig. 5C). ELISA assays revealed that each exosomecontains a low level of TGFb (6–162 pg/100 mg exosomes)and that this concentration was similar between EOC cellsand IOSEs (Fig. 5D). Coculture with EOC cells not only inducedCD44 expression in HPMCs but also resulted in acquisition of aspindle cell, EMT-type morphology, despite LY2157299 treat-ment (Fig. 5E and F). These experiments suggest that CD44acquisition by HPMCs via exosomes is independent of TGFbsignaling.

Downregulation of CD44 expression in EOC-derived exosomesattenuated ovarian cancer invasion

To further understand the role of exosomal CD44 in ovariancancer progression, CD44 expression in HeyA8 cells was knockeddown using siRNAs (Fig. 6A and B), which resulted in a reductionof CD44 expression in the HeyA8-derived exosomes (Fig. 6C).Treatment of HPMCs with exosomes derived from HeyA8 CD44knockdown (KD) cells abrogated the upregulation of CD44s andthe downregulation of E-cadherin in HPMCs (Fig. 6D). Thepretreatment of HPMCs with exosomes from HeyA8 CD44 KDcells significantly blocked invasion (Fig. 6E) and mesothelial cellclearance (Fig. 6F) and reduced the gelatinolytic activity ofMMP9(Fig. 6G).

In the opposite experiment, HPMCs were transfected with aCD44 expression vector and their effect on EOC invasion wasdetermined. Immunoblotting showed that CD44 was overex-pressed in HPMC, leading to the downregulation of E-cadherinexpression (Fig. 6H). CD44-overexpressing HPMCs displayed aspindle cell morphology and lost cell–cell contact, whereas con-trol vector–transfected HPMCs retained a cobblestone-likeappearance (Fig. 6I). CD44-overexpressing HPMCs promotedinvasion of EOC cells when compared with control vector–trans-fected HPMCs (Fig. 6J). The mesothelial clearance area wassignificantly increased in CD44-overexpressing HPMCs (Fig.6K), and the conditioned medium of CD44-overexpressingHPMCs increased the gelatinolytic activity of MMP9 (Fig. 6L).

Figure 3.EOC-derived exosomes induce HPMCs to secrete MMP9 and increase HPMC clearance. A, Schematic. B and C, Western blotting. B, Expression of majoradhesion and invasion markers. RMG-1 was used as a positive control for CD44 variant forms. C, CD44 expression in exosomes. CD63 was used as aninternal control. D, Electron microscopy. Exosomes were immunogold labeled with anti-CD44 antibody as Fig. 1A. Bars represent 100 nm. E, Confocalmicroscopy. HPMCs were cultured for 3 minutes in the presence of PKH67-labeled exosomes from IOSE or HeyA8. HPMCs were stained with DAPI (blue nuclei)and anti-CD44 antibody followed by Alexa Fluor 633–conjugated secondary antibody (red). Bars represent 20 mm. F, Real-time qRT-PCR for CD44using SYBR Green. G,Mesothelial clearance assay. Twenty-four hours after exosome treatment, HPMCs were fixed and stained with Acti-Stain 555–conjugatedphalloidin (red). The percentage of cell-free area induced by exosome treatment (indicated by white dashed line) was measured using ImageJ software(left). Representative images of fluorescence microscopy (right). Bars indicate 50 mm. H, Real-time qRT-PCR for MMP9 using SYBR Green. I, Gelatinzymography. The supernatants of HPMC treated with exosomes were added for the assay. Images in B–E, G, I are representative of 3 independent experimentswith similar results. Data in F–H are mean � SEM of 3 independent experiments. � , P < 0.05; �� , P < 0.01.






Taken together, the data in Fig. 6 show that exogenous expressionof CD44 in HPMCs promotes ovarian cancer invasion by increas-ing mesothelial clearance and inducing MMP9 secretion.

CD44 is overexpressed in mesothelial cells in omentalmetastasis

Next, we investigated whether the results found using celllines have correlates in human tissue from patients with EOC;specifically whether CD44 expression is increased in HPMCsfrom omental metastasis. Omental tissue harboring tumor wasobtained from 15 patients and immunostained with anti-CD44antibody (Supplementary Table S1). Normal HPMCs lining theomentum surface were identified using positive calretinin stain-ing (Fig. 7A). In HPMCs from normal omentum, CD44 expres-sion was mostly negative (Fig. 7A). In contrast, in 9 of 15 (60%)tumors, strong CD44 expression was seen in the primary ovarianhigh-grade serous performed (Supplementary Fig. S4) andshowed CD44 to be colocalized with CD63 carcinomas(Fig. 7B). To confirm exosomes contain CD44 in these primary

ovarian high-grade serous carcinomas, double immunostainingwith CD44 and CD63, a representative exosome marker, wasconducted. In omental tissue harboring disseminated ovariancancer cells, CD44 was strongly expressed in metastatic implantsincluding HPMCs. However, HPMCs did not express CD44 inareas further away from the tumor cells (Fig. 7C). HPMCs inter-twined with metastatic cancer cells displayed spindle-like mor-phology, and the mesothelial cell barrier was disrupted (Fig. 7C).In serial sections, CD44-positive HPMCs harboring tumorsexpressed MMP9, whereas normal HPMCs were mostly MMP9-negative (Supplementary Fig. S5).

CD44 expression in mesothelial cells was also studied usingan ovarian cancer xenograft model (30). To mimic early ovariancancer metastasis to omental tissue, HeyA8 cells transfectedwith control siRNA or CD44 siRNA were injected intraperito-neally into female BALB/c nu/nu mice (31). One week afterinoculation, mice showed microscopic tumor dissemination onthe omental surface. Human CD44 expression was seen at theleading edge of invading HeyA8 cells transfected with controlsiRNA and mouse mesothelial cells harboring cancer cells

D

GW4869 (µmol/L)0E

xoso

mal

pro

tein

(µg

) **

0

5

10

15

20

1 10

A

CD44s

E-Cadherin

β-Actin

70

140

GW4869 (µmol/L)0

Coculture

HPMC

100

100

50

0 1 10 10

with HeyA8

FE

Coculture with HeyA8

GW4869 (µmol/L)

0

HPMC

0 1 10 10

(kDa)

Rel

ativ

e in

vasi

veH

eyA

8 ce

lls

GW4869 (µmol/L)

HPMC coat (+)

**HPMC coat (–)

1 100 1 100

1

0.5

0

1

0.5

0

C

Hey

A8

Cel

l via

bilit

y

GW4869 (µmol/L)

0

1

10

0.5

10

1.0

0.5

00 100 200 300

Con

cent

ratio

n (1

×101

3pa

rtic

les/

mL)

Particle size (nm)

DMSOGW4869 10 µmol/L

B

Figure 4.

Inhibiting exosome secretion attenuates CD44 uptake by HPMCs and suppresses ovarian cancer invasion. A, HeyA8 cells (5 � 107) were treated with thesphingomyelinase inhibitor GW4869 (48 hours), exosomes collected, and exosomal proteins concentration quantified (Lowry). B, Nanoparticle analysis.Concentration and size distribution of nano-sized particles in exosome suspension were measured using NanoSight system. Concentration was indicated bythe number of nano-sized particles per 1-mL culture medium. C, Proliferation assay. HeyA8 cells (5 � 103) were treated with GW4869. Cell viability was assessed48 hours after incubation. D, Left, invasion assay. After an HPMC monolayer was formed on Matrigel-coated inserts, CMFDA-labeled HeyA8 cells were platedonto Matrigel and were allowed to invade for 48 hours in the presence of GW4869. Right, Matrigel in vitro invasion assay without HPMC. E, Western blotting.HPMC growing as monolayer on the bottom was cocultured with HeyA8 cells in an insert in the presence or absence of GW4869 (24 hours). Thereafter, HPMC celllysates were collected, lysed and immunoblotting was performed. F, Phase contrast microscopy. HPMCs were treated with GW4869 and morphologicchanges evaluated. Scale bar, 100 mm. Data in A (left), B, C are shown as mean � SEM of 3 independent experiments. Data in A (right) are shown as mean of5 measurements. Images in D, E are representative of 3 experiments with similar results. �� , P < 0.01.

Nakamura et al.





displayed positive human CD44 expression (Fig. 7D, top).However, HeyA8 cells transfected with CD44 siRNA showedfaint CD44 expression and CD44 was negative in mouseomental tissues harboring cancer cells (Fig. 7D, bottom). Tofurther quantify the induction of CD44 expression, we injectedeither 1 � 107 or 1 � 106 HeyA8 cells intraperitoneally intofemale athymic BALB/c nude mice and compared human CD44expression in omental tissues from the 2 groups of mice(Supplementary Fig. S6A–S6C). In the tissue of mice injectedwith 1 � 107 HeyA8 cells, strong CD44 expression was seen(Supplementary Fig. S6D) as compared with tissue from mice

injected with 1 � 106 HeyA8 cells. These data support our invitro findings indicating that CD44 acquisition by mesothelialcells is induced by direct contact with EOC cells.

DiscussionCancer-derived exosomes can act as mediators of metastasis,

and, as a result, a minority population of cancer cells can repro-gram stromal cells through the intercellular transfer of proteins,mRNAs, and miRNAs (32). Here, we show that EOC-derivedexosomes promoted ovarian cancer invasion by transferring

D

A

E

B

TGFβ (ng/mL)LY2157299 (µmol/L)

00

100

101

01

HPMC

C

Con

cent

ratio

n of

Hey

-exo

TY

K-e

xo

IOS

E-e

xo

ExosomeT

GF

β (p

g/m

L)

0

30

60

90

120

150

(100 µg/mL)

FHPMC

LY21

5729

9(µ

mol

/L)

CocultureIOSEHeyA8 TYK-nu

1

0

with

TGFβ (ng/mL)LY2157299 (µmol/L)

00

10

100

00.1

01

100.1

101

70

100CD44s

E-Cadherin

β-Actin

HPMC

140

10050

(kDa)

Exosome (100 µg/mL)TGFβ (10 ng/mL)

LY2157299 (1 µmol/L)––

+–

–+

++

–+

–+

–+

CD44s

E-Cadherin

β-Actin

HPMC

70

140

100

10050

(kDa)

E-Cadherin

Coculture IOSEHeyA8 TYK-nu

LY2157299(µmol/L) 0 0.1 1

CD44s

β-Actin

70

140

100

100

50

HPMC

0 0.1 1 0 0.1 1 0 0.1 1

with

(kDa)

Figure 5.

Uptake of exosomes by HPMC is independent of TGFb signaling. A, Western blot analysis of HPMCs treated with TGFb and/or the TGFb receptor I kinaseinhibitor (LY2157299). B, Representative morphologic changes of HPMCs described in A. C, Western blotting of HPMCs treated with the indicatedcombinations of TGFb (10 ng/mL), LY2157299 (1 mmol/L), and EOC cell exosomes (100 mg/mL) for 48 hours. D, ELISA for TGFb in exosomes. E, Westernblot analysis of HPMCs cocultured with EOC cells or IOSEs in the presence or absence of LY2157299. HPMC growing as monolayer on the bottom werecocultured with 5 � 105 EOC cells or IOSEs in the presence or absence of LY2157299 (48 hours). F, Phase contrast images of HPMC from E. Scale bar, 100 mm.Images in A–C, E, F are representative of 3 experiments with similar results. Data in E are mean � SEM of 3 experiments.






D

E

HeyA8 siRNA

control siRNACD44 siRNA #1CD44 siRNA #2

48 h

Isolation of

ctrl-exo#1-exo#2-exo

Exosome

24 h

HPMC Monolayer

Applied for

B

ACells Transfection exosomes (100 µg/mL) each experiment

C

CD44s

β-Actin

240

50

14010070

CD44v

HeyA8 Transfectedwith CD44 siRNA

(kDa)

CD63

40

Exosome

CD44s

50

240

140

10070

CD44v

35

(kDa)

E-Cadherin

HPMC + Exosome

CD44s

β-Actin

240

50

14010070

CD44v

140

(kDa)

F G

H

CD44s

E-

β-Actin

140

CD44v

50

140

100

240

70

HPMC

Cadherin

(kDa)

I

DAPIAlexa Fluor

merged

ctrl

CD44+

HPMC

633 (CD44)

J

K L

Pro MMP-9

Pro MMP-2

100

70

HPMC

(kDa)

Pro MMP-9

Pro MMP-2

100

70

HPMC+ Exosome

(kDa)

Rel

ativ

e in

vasi

ve

HPMC Coat (+)

***

Hey

A8

cells

0

0.5

1

0

0.5

1

HPMC Coat (–)

Rel

ativ

e in

vasi

ve

ovar

ian

canc

er c

ells

HeyA8 TYK-nu

* **

HPMC Coated on Matrigel

0

2

1

0

1

2

Cel

l-fre

e ar

ea (

%)

ctrl-exo #1-exo**

0

10

20

30

HPMC

ctrl CD44+*

Cel

l-fre

e ar

ea (

%)

0

10

20

30

HPMC

Nakamura et al.





exosomes containing CD 44 to HPMCwhich reprograms them toa more EMT phenotype, thereby promoting ovarian cancer cellinvasion andmetastasis. EOC cells, by releasing exosomes, repro-gram HPMCs, which then facilitate EOC invasion supporting amodel of bidirectional communication (Fig. 7E).

Several studies have shown that cancer cell–derived exosomesreprogram or educate other cells to aid tumor survival andpromotemetastasis (33); however, to date, only a limited numberof reports have investigated the cancer-promoting effect of aparticular protein or miRNA in cancer-derived exosomes. Forinstance, melanoma-derived exosomes reprogram bone marrowprogenitor cells to support tumor growth and metastasis throughMET (34) andbreast cancer–derived exosomalmiR-105promotesmetastasis by destroying endothelial cell barriers (35). Recently,Singh and colleagues revealed that integrin-avb3 is transferredfrom tumorigenic to nontumorigenic cancer cells via exosomesand its de novo expression in recipient cells promotes cell migra-tion on its ligand (36).

It is well-known that the tumor microenvironment, includingstromal cells, plays an important role in cancer progression (37).Under normal conditions, mesothelial cells act as a mechanicalbarrier to protect intra-abdominal organs (3). However, reportshave shown that mesothelial cells can also serve as a metastaticniche, promoting cancer cell adhesion and invasion, which areessential steps of the peritoneal dissemination of ovarian cancer(2, 38). Since several previous reports have focused on TGFbsignaling as a means of communication between EOC cells andHPMCs, we inhibited TGFb signaling using a potent inhibitor,LY2157299. While this TGFb inhibitor completely abrogated themorphologic changes to a mesenchymal phenotype and thedownregulation of E-cadherin induced by TGFb, treatmentwith EOC-derived exosomes induced these phenomena in thepresence of LY2157299, indicating that the mechanism is TGFb-independent.

Several adhesion molecules such as integrins or ECM com-ponents have been identified as important mediators of cancermetastasis to the mesothelium (39). Among these, hyaluronan(HA) and its receptor, CD44, have been shown to play criticalroles in ovarian cancer metastasis (40). It is reported thatovarian cancer cell adhesion to mesothelial cell monolayers ismediated, at least in part, by the interaction between HA andCD44, which further mediates ovarian cancer progression (41).

A recent immunohistochemical study showed that CD44expression was seen in 42 of 72 (58.3%) ovarian tumors andthat CD44 (þ)/E-cadherin (�) ovarian tumors have a moreaggressive phenotype and are associated with poor patientsurvival (42). We found a similar strong expression of CD44in tumors analyzed. CD44 is also known to be a key regulator ofEMT in several cancer types (43–45). Consistent with thesereports, our findings show that HPMCs treated with EOC-derived exosomes displayed a mesenchymal phenotype andreduced E-cadherin expression. CD44 has also been shown torecruit MMP9 to the cell surface and mediate tumor cellinvasion (25), which is consistent with our zymographic anal-yses (Fig. 6). Given the important role of CD44 in cancerinvasion, the molecule holds promise as a therapeutic targetin ovarian cancer treatment. However, clinical trials of systemictreatment with CD44 neutralizing antibodies have been termi-nated because of unacceptable levels of toxicity (46). Nontoxicalternative therapies need to be developed for future trials.Capturing CD44-expressing exosomes in body fluid appears tobe less toxic and has the potential to be a possible anticancertherapy.

Interestingly, inhibition of exosomes released from EOC cellsusing GW4869 attenuated CD44 acquisition in HPMCs andaccordingly suppressed ovarian cancer invasion. Kosaka andcolleagues suggested that cancer cell–derived exosomes circu-lating in the blood not only promote cancer malignancythrough the formation of premetastatic niches but also interferewith treatment (47). Indeed, Ciravolo and colleagues reportedthat the exosomes released by the HER2-overexpressing breastcancer cell lines express a full-length HER2 molecule and thatthese exosomes bound to trastuzumab and inhibited its anti-cancer cell proliferative activity (48). On the basis of thispreclinical data, Aethlon Medical Inc. has developed HER2o-some, as a therapeutic strategy to combat HER2-positive breastcancer through the capture of circulating HER2-positive exo-somes (49).

There are some limitations to this study. First, as a largeamount of exosomes were needed to stably reproduce theexperimental data, HeyA8 and TYK-nu cells, which expressexosomes most abundantly, were used. However, a recent geno-mic profiling of ovarian cancer cell lines (50) suggests thatHeyA8 seems not to be typical of high-grade serous ovarian

Figure 6.CD44 expression in HPMC promotes ovarian cancer invasion. A–G, Downregulation of exosomal CD44 suppresses ovarian cancer invasion. A, Schematic ofsiRNA transfection and exosome treatment of HPMCs. Exosomes were isolated from HeyA8 cells transfected with control siRNA or 2 different CD44siRNAs. B–D, Western blot analyses. B, HeyA8 cells transfected with control siRNA (ctrl) or 2 different CD44 siRNAs (#1 and #2). C, Exosomes describedin A. D, HPMCs treated with exosomes described in A. E, Left, HPMC-coated in vitro invasion assay. After the HPMC monolayer was formed, Matrigel-coatedcell culture inserts and exosomes were added followed by CMFDA-labeled HeyA8 cells 24 hours later. Right, Matrigel in vitro invasion assay. Culture mediumwith 100 mg/mL of exosomes was added to Matrigel-coated inserts. Twenty-four hours later, 5 � 104 fluorescence (CMFDA)-labeled HeyA8 cells wereincubated in a 24-well FluoroBlok cell culture insert and allowed to invade for 48 hours. F, Mesothelial clearance assay. HPMCs were treated with exosomesand the percentage of cell-free area induced by exosome treatment (indicated by white dashed line) was measured using ImageJ software (left).Representative images of HPMCs (right). G, Gelatin zymography. HPMC supernatant treated with exosomes was added to a gelatin-containing gel.Proteolysis was detected as a white band. H–L, Overexpression of CD44 in HPMCs promotes ovarian cancer invasion. H, Western blotting. HPMCs weretransfected with wild-type CD44 plasmid (CD44þ) or control vector (ctrl). I, confocal microscopy. HPMCs were stained with DAPI (blue nuclei) and anti-CD44antibody followed by Alexa Fluor 633–conjugated secondary antibody (red). Bars represent 20 mm. J, Invasion assay. HPMCs transfected with control orCD44 plasmid were plated onto cell culture inserts. Thereafter, an in vitro invasion assay was performed as described in E. K, Mesothelial clearanceassay. HPMCs were transfected with control or CD44 plasmid and incubated for 24 hours. Cells were added to the assay as described above. The percentage ofcell-free area is shown (left). Representative images of HPMCs (right). Bars indicate 50 mm. L, Gelatin zymography. HPMCs supernatants from HPMCtransfected with control or CD44 plasmid were used for gelatin zymography. Images in B–D, F–I, K, L are representative of 3 experiments with similar results.Data in F, J, K are mean � SEM of 3 experiments. � , P < 0.05; �� , P < 0.01.






cancer (HGSC). While TYK-nu conforms more closely to HGSC,this cell line was not transplantable to mice and, therefore, wecould use it in our in vivo experiments. Second, it was notconclusively demonstrated that EOC-derived exosomes containenriched CD44. Ultracentrifugation-based exosome isolation iscommonly used; however, it is not possible to completely

eliminate contamination pulled down from other sources inthe culture medium (51).

In conclusion, we propose a novel role for EOC-derivedexosomes in targeting adjacent peritoneal mesothelial cells(Fig. 7E). Our data identified exosome-mediated transfer ofCD44 as a key regulator of mesothelial cell reprogrammingand the metastatic progression of ovarian cancer. In the light

A B High-grade serous ovarian cancerNormal omentum

C

Calretinin CD44 CD44

D

HPMC

EOvarian cancer cell Mesothelial cell

CD44

MMP-9

CD44-containing

Degradation of extracellular matrix

Mesothelial clearance

E-Cadherin

Cancer invasion

exosomes

Production/secretion

Human omentum in early peritoneal dissemination Calretinin

HPMC

CD44

CD44+ HPMCCD44– HPMCCell-free area

Ovarian cancer

Omentum

Ovarian cancer

Omentum

Human CD44 in mouse xenograft model

HeyA8 cells

Control siRNA

HeyA8 cells

CD44 siRNA

Mouse omentum

Mouse omentum

Human CD44+ mouse peritoneal mesothelial cell

Figure 7.

CD44 is strongly expressed inperitoneal mesothelial cellsinteracting with ovarian cancer cells.A–C, Immunohistochemistry for CD44and calretinin, a marker of mesothelialcells, in representative tissuesobtained from patients with EOC.A, Normal omental tissues. Blackarrowheads indicate HPMCs. HPMCsare calretinin-positive (left) and CD44negative (right). B, Primary tumorsfrom 2 different human patients withHGSC. Both tumors overexpressedCD44. C, Serial sections of an omentaltissue with serous EOC metastasisobtained from a 51-year-old womanwith advanced HGSC. Calretininstaining identified HPMCs (top).Cancer cells were strongly CD44-positive whereas they were calretinin-negative. Red arrowheads indicateCD44-positive HPMCs adjacent tocancer cells. Blue arrowheads indicateCD44-negative HPMCs remote fromcancer cells. Black arrows indicatemesothelial clearance. D,Immunohistochemical analysisshowing that 1 � 105 human HeyA8cells transfected with control siRNA orCD44 siRNA were injectedintraperitoneally into female BALB/cnu/nu mice. One week after theinjection, they were sacrificed andomental tissues were collected. Redarrowheads indicate human CD44-positive mouse mesothelial cells intumor implants to the omentum. Barsindicate 100 mm. E, Model.Bidirectional communication betweenEOC andmesothelial cells. See text fordetails.

Nakamura et al.





of several publications highlighting the importance of exo-somes in cancer biology and the results described here, wehypothesize that targeting exosomes might have a therapeuticapplication in ovarian cancer treatment and recommend that itshould be explored further for future clinical use.

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

Authors' ContributionsConception and design: K. Sawada, T. KimuraDevelopment of methodology: K. Nakamura, K. Sawada, K. HashimotoAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): Y. Kinose, A. Yoshimura, A. Toda, E. Nakatsuka,K.-i. MorishigeAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K. Nakamura, E. Nakatsuka, E. Lengyel, T. KimuraWriting, review, and/or revision of the manuscript: K. Nakamura, K. Sawada,H. Kurachi, E. LengyelAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): E. Nakatsuka, K.-i. Morishige, H. Kurachi

Study supervision: K. Sawada, S. Mabuchi, T. Kimura

AcknowledgmentsWe thank Mami Morikawa for her secretarial assistance and Ayako

Okamura for her technical assistance. We also thank Gail Isenberg (Univer-sity of Chicago, Chicago, IL) for critical editing of this article.

Grant SupportThis work was supported by a Grant-in-Aid for Scientific Research from

the Ministry of Education, Science, Sports and Culture of Japan(24592515, 26670725, and 26293360 to K. Sawada and 24249080 toT. Kimura). This work was also supported by Takeda Science Foundation(to K. Sawada) and Osaka Medical Research Foundationfor Intractablediseases (to K. Nakamura).

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 indicatethis fact.

Received June 2, 2016; revised September 18, 2016; accepted September 24,2016; published OnlineFirst October 6, 2016.

References1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin

2016;66:7–30.2. Kenny HA, Chiang CY, White EA, Schryver EM, Habis M, Romero IL, et al.

Mesothelial cells promote early ovarian cancer metastasis through fibro-nectin secretion. J Clin Invest 2014;124:4614–28.

3. Mutsaers SE.The mesothelial cell. Int J Biochem Cell Biol 2004;36:9–16.4. Kenny HA, Nieman KM, Mitra AK, Lengyel E. The first line of intra-

abdominal metastatic attack: breaching the mesothelial cell layer. CancerDiscov 2011;1:100–2.

5. Davidowitz RA, Selfors LM, Iwanicki MP, Elias KM, Karst A, Piao H,et al. Mesenchymal gene program-expressing ovarian cancer spher-oids exhibit enhanced mesothelial clearance. J Clin Invest 2014;124:2611–25.

6. Iwanicki MP, Davidowitz RA, NgMR, Besser A, Muranen T, Merritt M, et al.Ovarian cancer spheroids use myosin-generated force to clear the meso-thelium. Cancer Discov 2011;1:144–57.

7. Lengyel E.Ovarian cancer development and metastasis. Am J Pathol2010;177:1053–64.

8. Rieppi M, Vergani V, Gatto C, Zanetta G, Allavena P, Taraboletti G, et al.Mesothelial cells induce the motility of human ovarian carcinoma cells.Int J Cancer 1999;80:303–7.

9. Satoyoshi R, AibaN, Yanagihara K, YashiroM, TanakaM. Tks5 activation inmesothelial cells creates invasion front of peritoneal carcinomatosis.Oncogene 2015;34:3176–87.

10. Lee JK, Park SR, Jung BK, Jeon YK, Lee YS, KimMK, et al. Exosomes derivedfrom mesenchymal stem cells suppress angiogenesis by down-regulatingVEGF expression in breast cancer cells. PLoS One 2013;8:e84256.

11. Miyanishi M, Mandai M, Matsumura N, Yamaguchi K, Hamanishi J,Higuchi T, et al. Immortalized ovarian surface epithelial cells acquiretumorigenicity by Acrogranin gene overexpression. Oncol Rep 2007;17:329–33.

12. L€asser C, Eldh M, L€otvall J. Isolation and Characterization of RNA-Containing Exosomes. J Vis Exp 2012;59:e3037.

13. Kinose Y, Sawada K, Makino H, Ogura T, Mizuno T, Suzuki N, et al. IKKbregulates VEGF expression and is a potential therapeutic target forovarian cancer as an antiangiogenic treatment. Mol Cancer Ther 2015;14:909–19.

14. Kinose Y, Sawada K, Nakamura K, Sawada I, Toda A, Nakatsuka E, et al. Thehypoxia-related microRNA miR-199a-3p displays tumor suppressor func-tions in ovarian carcinoma. Oncotarget 2015;6:11342–56.

15. Sawada K, Morishige K, Tahara M, Kawagishi R, Ikebuchi Y, Tasaka K, et al.Alendronate inhibits lysophosphatidic acid-induced migration of humanovarian cancer cells by attenuating the activation of rho. Cancer Res2002;62:6015–20.

16. Partridge J, Flaherty P. An in vitro FluoroBlok tumor invasion assay. J VisExp 2009;29:1475.

17. Lengyel E, Schmalfeldt B, Konik E, Sp€athe K, H€arting K, Fenn A,et al. Expression of latent matrix metalloproteinase 9 (MMP-9)predicts survival in advanced ovarian cancer. Gynecol Oncol 2001;82:291–8.

18. KosakaN, IguchiH, Yoshioka Y, Takeshita F,Matsuki Y,Ochiya T. Secretorymechanisms and intercellular transfer of microRNAs in living cells. J BiolChem 2010;285:17442–52.

19. Kenny HA, Krausz T, Yamada SD, Lengyel E. Use of a novel 3D culturemodel to elucidate the role of mesothelial cells, fibroblasts and extra-cellular matrices on adhesion and invasion of ovarian cancer cells to theomentum. Int J Cancer 2007;121:1463–72.

20. Valadi H, Ekstr€om K, Bossios A, Sj€ostrand M, Lee JJ, L€otvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism ofgenetic exchange between cells. Nat Cell Biol 2007;9:654–9.

21. Zhang HG, Grizzle WE. Exosomes: a novel pathway of local and distantintercellular communication that facilitates the growth and metastasis ofneoplastic lesions. Am J Pathol 2014;184:28–41.

22. Grange C, TapparoM, Collino F, Vitillo L, Damasco C,DeregibusMC, et al.Microvesicles released from human renal cancer stem cells stimulateangiogenesis and formation of lung premetastatic niche. Cancer Res2011;71:5346–56.

23. Barberis MC, Faleri M, Veronese S, Casadio C, Viale G. Calretinin. Aselective marker of normal and neoplastic mesothelial cells in serouseffusions. Acta Cytol 1997;41:1757–61.

24. Zigler M, Dobroff AS, Bar-Eli M. Cell adhesion: implication in tumorprogression. Minerva Med 2010;101:149–62.

25. Ponta H, Sherman L, Herrlich PA. CD44: from adhesion molecules tosignaling regulators. Nat Rev Mol Cell Biol 2003;4:33–45.

26. YuQ, Stamenkovic I. Localization of matrix metalloproteinase 9 to the cellsurface provides a mechanism for CD44-mediated tumor invasion. GenesDev 1999;13:35–48.

27. Tian T, Zhu YL, Zhou YY, Liang GF,Wang YY, Hu FH, et al. Exosome uptakethrough clathrin-mediated endocytosis and macropinocytosis and medi-ating miR-21 delivery. J Biol Chem 2014;289:22258–67.

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

29. Z€oller M. CD44: Can a cancer-initiating cell profit from an abundantlyexpressed molecule? Nat Rev Cancer 2011;11:254–67.

30. SawadaK, Radjabi AR, ShinomiyaN, Kistner E, KennyH, Becker AR, et al. c-Met overexpression is a prognostic factor in ovarian cancer and an effectivetarget for inhibition of peritoneal dissemination and invasion. Cancer Res2007;67:1670–9.






31. Lengyel E, Burdette JE, Kenny HA, Matei D, Pilrose J, Haluska P, et al.Epithelial ovarian cancer experimental models. Oncogene 2014;33:3619–33.

32. Balaj L, Lessard R,Dai L, Cho YJ, Pomeroy SL, Breakefield XO, et al. Tumourmicrovesicles contain retrotransposon elements and amplified oncogenesequences. Nat Commun 2011;2:180.

33. Vaksman O, Trop�e C, Davidson B, Reich R. Exosome-derived miRNAs andovarian carcinoma progression. Carcinogenesis 2014;35:2113–20.

34. PeinadoH,Ale�ckovi�cM,LavotshkinS,Matei I, Costa-SilvaB,Moreno-BuenoG, et al.Melanomaexosomeseducate bonemarrowprogenitor cells towardapro-metastatic phenotype through MET. Nat Med 2012;18:883–91.

35. Zhou W, Fong MY, Min Y, Somlo G, Liu L, Palomares MR, et al. Cancer-secreted miR-105 destroys vascular endothelial barriers to promote metas-tasis. Cancer Cell 2014;25:501–15.

36. Singh A, Fedele C, Lu H, Nevalainen MT, Keen JH, Languino LR. Exosome-mediated transfer of alphavbeta3 integrin from tumorigenic to non-tumorigenic cells promotes a migratory phenotype. Mol Cancer Res.2016 Jul 20. [Epub ahead of print].

37. Hanahan D, Coussens LM. Accessories to the crime: functions ofcells recruited to the tumor microenvironment. Cancer Cell 2012;21:309–22.

38. Sugiyama K, Kajiyama H, Shibata K, Yuan H, Kikkawa F, Senga T.Expression of the miR200 family of microRNAs in mesothelial cellssuppresses the dissemination of ovarian cancer cells. Mol Cancer Ther2014;13:2081–91.

39. Sawada K, Ohyagi-Hara C, Kimura T, Morishige K. Integrin inhibitors as atherapeutic agent for ovarian cancer. J Oncol 2012;2012:915140.

40. Ween MP, Oehler MK, Ricciardelli C. Role of versican, hyaluronan andCD44 in ovarian cancer metastasis. Int J Mol Sci 2011;12:1009–29.

41. Gardner MJ, Catterall JB, Jones LM, Turner GA. Human ovarian tumourcells can bind hyaluronic acid via membrane CD44: a possible step inperitoneal metastasis. Clin Exp Metastasis 1996;14:325–34.

42. RyabtsevaOD, LukianovaNY, Shmurakov YA, Polishchuk LZ,Antipova SV.Significance of adhesion molecules expression for estimation of serousovarian cancer prognosis. Exp Oncol 2013;35:211–8.

43. Brown RL, Reinke LM, Damerow MS, Perez D, Chodosh LA, Yang J, et al.CD44 splice isoform switching in human and mouse epithelium isessential for epithelial-mesenchymal transition and breast cancer progres-sion. J Clin Invest 2011;121:1064–74.

44. Mima K, Okabe H, Ishimoto T, Hayashi H, Nakagawa S, Kuroki H, et al.CD44s regulates the TGF-b-mediated mesenchymal phenotype and isassociated with poor prognosis in patients with hepatocellular carcinoma.Cancer Res 2012;72:3414–23.

45. Cho SH, Park YS, KimHJ, Kim CH, Lim SW, Huh JW, et al. CD44 enhancesthe epithelial-mesenchymal transition in association with colon cancerinvasion. Int J Oncol 2012;41:211–8.

46. Orian-Rousseau V.CD44, a therapeutic target for metastasising tumours.Eur J Cancer 2010;46:1271–7.

47. KosakaN, Yoshioka Y, TominagaN,Hagiwara K, Katsuda T,Ochiya T.Darkside of the exosome: the role of the exosome in cancer metastasis andtargeting the exosome as a strategy for cancer therapy. Future Oncol2014;10:671–81.

48. Ciravolo V, Huber V, Ghedini GC, Venturelli E, Bianchi F, Campiglio M,et al. Potential role of HER2-overexpressing exosomes in counteringtrastuzumab-based therapy. J Cell Physiol 2012;227:658–67.

49. Aethlon Medical Inc. A medical device strategy to inhibit HER2þ breastcancer progression [updated 2011 Sep27; cited 2015Oc5]. Available from:http://www.aethlonmedical.com/assets/001/5080.pdf.

50. Domcke S, Sinha R, LevineDA, Sander C, Schultz N. Evaluating cell lines astumour models by comparison of genomic profiles. Nat Commun2013;4:2126.

51. Lobb RJ, Becker M, Wen SW, Wong CS, Wiegmans AP, Leimgruber A, et al.Optimized exosome isolation protocol for cell culture supernatant andhuman plasma. J Extracell Vesicles 2015;4:27031.


Nakamura et al.




2017;15:78-92. Published OnlineFirst October 6, 2016.Mol Cancer Res Koji Nakamura, Kenjiro Sawada, Yasuto Kinose, et al. of CD44 to Peritoneal Mesothelial CellsExosomes Promote Ovarian Cancer Cell Invasion through Transfer

Updated version

10.1158/1541-7786.MCR-16-0191doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2017/06/16/1541-7786.MCR-16-0191.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/15/1/78.full#ref-list-1

This article cites 49 articles, 11 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/15/1/78.full#related-urls

This article has been cited by 9 HighWire-hosted articles. Access the articles 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/15/1/78To 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-16-0191

http://mcr.aacrjournals.org/content/suppl/2017/06/16/1541-7786.MCR-16-0191.DC1

http://mcr.aacrjournals.org/content/15/1/78.full#ref-list-1

http://mcr.aacrjournals.org/content/15/1/78.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/15/1/78


Documents

Exosomes Promote Ovarian Cancer Cell Invasion through ... · Exosomes Promote Ovarian Cancer Cell Invasion through Transfer of CD44 to Peritoneal Mesothelial Cells Koji Nakamura1,