of May 10, 2015.This information is current as

Invasion via Fas Signaling PathwayActivated T Cell Exosomes Promote Tumor

Jianli WangandQingqing Wang, Yunshan Yang, Lie Wang, Xuetao Cao

Zhijian Cai, Fei Yang, Lei Yu, Zhou Yu, Lingling Jiang,

http://www.jimmunol.org/content/188/12/5954doi: 10.4049/jimmunol.11034662012;

2012; 188:5954-5961; Prepublished online 9 MayJ Immunol 

Referenceshttp://www.jimmunol.org/content/188/12/5954.full#ref-list-1

, 30 of which you can access for free at: cites 46 articlesThis article

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2012 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Activated T Cell Exosomes Promote Tumor Invasion via FasSignaling Pathway

Zhijian Cai,*,1 Fei Yang,*,1 Lei Yu,* Zhou Yu,* Lingling Jiang,* Qingqing Wang,*

Yunshan Yang,† Lie Wang,* Xuetao Cao,* and Jianli Wang*

Activated T cells release bioactive Fas ligand (FasL) in exosomes, which subsequently induce self-apoptosis of T cells. However,

their potential effects on cell apoptosis in tumors are still unknown. In this study, we purified exosomes expressing FasL from

activated CD8+ T cell from OT-I mice and found that activated T cell exosomes had little effect on apoptosis and proliferation of

tumor cells but promoted the invasion of B16 and 3LL cancer cells in vitro via the Fas/FasL pathway. Activated T cell exosomes

increased the amount of cellular FLICE inhibitory proteins and subsequently activated the ERK and NF-kB pathways, which

subsequently increased MMP9 expression in the B16 murine melanoma cells. In a tumor-invasive model in vivo, we observed that

the activated T cell exosomes promoted the migration of B16 tumor cells to lung. Interestingly, pretreatment with FasL mAb

significantly reduced the migration of B16 tumor cells to lung. Furthermore, CD8 and FasL double-positive exosomes from tumor

mice, but not normal mice, also increased the expression of MMP9 and promoted the invasive ability of B16 murine melanoma

and 3LL lung cancer cells. In conclusion, our results indicate that activated T cell exosomes promote melanoma and lung cancer

cell metastasis by increasing the expression of MMP9 via Fas signaling, revealing a new mechanism of tumor immune

escape. The Journal of Immunology, 2012, 188: 5954–5961.

Fas (also known as CD95/Apo-1) is a transmembrane proteinbelonging to the TNF/nerve growth factor receptor super-family, which transmits an apoptotic signaling in suscep-

tible cells once triggered by its natural ligand FasL (1). Fas-mediatedapoptosis plays important roles in various biological processes,including activation-induced cell death in T lymphocytes orcell-mediated cytotoxicity against virus infection (2, 3). Almostall kinds of tumor cells express Fas, and the Fas/FasL system playsan important role in controlling survival or growth of tumor cells(4). Although Fas signaling was reported to induce tumor apo-ptosis and inhibit tumor growth in vivo when triggered by FasL(5–7), accumulated evidence also demonstrated that, in the pres-ence of certain levels of FasL, Fas signaling promotes, notinhibits, tumor growth in Fas-resistant tumor cells (8, 9). In ad-dition to inducing tumor cell proliferation, Fas ligation was provedto advance cell cycle and increase tumor motility and invasiveness(10–14). All of these studies suggest that Fas signaling plays animportant role in tumor progress in Fas-resistant tumors.

Activation of T cells is a pivotal step in the process of host an-titumor immunity. During this course, T cells release large numbersof membrane vesicular bodies termed exosomes (15), which wereinitially described by Johnstone et al. (16). Exosomes are nano-scale membrane vesicles (50–100 nm) released by various livecells, which have a lipid bilayer structure (17). They are formed bymembrane budding into the lumen of an endocytic compartment,leading to the formation of multivesicular bodies. Fusion of mul-tivesicular bodies to the plasma membrane leads to the extracel-lular release of exosomes (18). The exosomes secreted by activatedhuman T cells express bioactive FasL and can induce the apoptosisof Jurkat T cells (15). Exosomes purified from CD8+ T cells,which are generated by cultivation of OVA-pulsed dendritic cells(DCs) with naive CD8+ T cells derived from OT-I mice, also ex-press FasL and can inhibit DCs to stimulate CD8+ CTL responses(19). However, the potential effects of FasL in activated T cellexosomes on Fas-resistant tumor cells are not well understood.Matrix metalloproteinase (MMP)9 plays a critical role in the

breakdown of extracellular matrix in normal physiological pro-cesses, such as embryonic development, reproduction, and tissueremodeling, as well as in disease processes, such as tumor me-tastasis (20). MMP9 was reported to be an important factor infacilitating lung cancer invasion and metastases in the B16 mel-anoma model (21). It was also demonstrated that the activation oftumor Fas signaling can induce the expression of MMP9 (22).Therefore, we hypothesize that FasL in activated T cell exosomesprobably upregulates MMP9 expression and promotes the invasiveability of Fas-resistant tumor cells, which may provide a novelview for understanding tumor escape from the immune system.

Materials and MethodsMice and cell lines

C57BL/6-Tg (Tcra Tcrb) 1100Mjb/J (OT-I) mice were obtained from theInstitute of Immunology, National Key Laboratory ofMedical Immunology,Second Military Medical University (Shanghai, China). Female C57BL/6J(6–8-wk-old) mice were purchased from Joint Ventures Sipper BKExperimental Animal (Shanghai, China). Tnfsf6gld mice on a C57BL/6

*Institute of Immunology, School of Medicine, Zhejiang University, Hangzhou310058, Zhejiang, China; and †Zhejiang Cancer Hospital, Hangzhou 310022, Zhe-jiang, China

1Z.C. and F.Y. contributed equally to this work.

Received for publication December 2, 2011. Accepted for publication April 5, 2012.

This work was supported by the National Key Basic Research Program of China(2007CB512400), the National Natural Science Foundation of China (30671909 and30972725), and the Natural Science Foundation of Zhejiang Province (Z2090042).

Address correspondence and reprint requests to Dr. Jianli Wang, Institute of Immu-nology, School of Medicine, Zhejiang University, Hangzhou 310058, China. E-mailaddress: jlwang@zju.edu.cn

Abbreviations used in this article: c-FLIP, cellular FLICE inhibitory protein;COX2, cyclooxygenase-2; DC, dendritic cell; EXO, exosome derived from acti-vated CD8+ OT-I T cell; EXOcont, exosome derived from naive CD8+ OT-I T cell;EXOgld, exosome derived from Con A-activated CD8+ T cell of gld mice; EXOwt,exosome derived from Con A-activated CD8+ T cell of C57BL/6 mice; FasL, Fasligand; MMP, matrix metalloproteinase; OT-I, C57BL/6-Tg (Tcra Tcrb) 1100Mjb/J;PDTC, ammonium pyrrolidinedithiocarbamate; siRNA, small interfering RNA;WT, wild-type.

Copyright� 2012 by TheAmericanAssociation of Immunologists, Inc. 0022-1767/12/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1103466

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

background (henceforth termed gld) were obtained from The JacksonLaboratory, bred under specific pathogen-free conditions, and used at 6–8wk of age. B16 melanoma cell line and 3LL Lewis lung cancer cell lineoriginating from C57BL/6 mice were purchased from American TypeCulture Collection (Manassas, VA).

Peptides, reagents, and Abs

OVA257–264 peptide was synthesized by Chinese Peptide (Hangzhou,China). Matrigel matrix basement membrane was from BD Biosciences(San Diego, CA). DAPI was from BIOMOL Research Laboratories (NJ,CA). INTERFERin siRNA transfection reagent was from Polyplus (NY,CA). NF-kB inhibitor ammonium pyrrolidinedithiocarbamate (PDTC)and ERK/MAPK inhibitor U0126 were from Sigma-Aldrich (St. Louis,MO). Abs against FasL allophycocyanin (MFL3), CD8 PE (53-6.7),CD9 PE (eBioKMC8), CD11c PE (N418), and TCR FITC (B20.1) werefrom eBioscience (San Diego, CA). Abs against Tsg 101 (C-2), GRP 94(H-212), P-ERK (E-4), ERK1/2 (MK1), MMP9 (M-17), FLIPs/L (G-11),HSP70 (3A3), CD8 (YTS169.4), FasL (C-178), and Fas (5F9) were pur-chased from Santa Cruz Biotechnology (Santa Cruz, CA). Abs againstphospho–NF-kB (Ser536) (93H1) and NF-kB (C22B4), as well as HRP-linked secondary Abs, were from Cell Signaling Technology (MI, CA).

Preparation and purification of CD8+ T cell-released exosomes

A total of 2 3 106 cells/ml naive lymphocytes derived from OT-I mice wascultured in RPMI 1640 complete medium (containing 50 mM 2-ME, 1 mMsodium pyruvate, 10 mM HEPES, and 10 U/ml IL-2) and stimulated or notwith OVA257–264 peptide (2 mg/ml) for 48 h. Active CD8+ T cells werepurified using CD8+ Dynabeads (Invitrogen, Carlsbad, CA). Purified OVA-specific active CD8+ T cells were then cultured in exosome-free RPMI1640 complete medium for 24 h. Exosomes in T cell culture supernatantswere isolated, as previously described (23). Briefly, collected culture super-natants were differentially centrifuged at 300 3 g for 10 min, 1,200 3 gfor 20 min, and 10,000 3 g for 30 min at 4˚C. The supernatant from thefinal centrifugation was ultracentrifuged at 100,000 3 g for 1 h at 4˚C.After removing the supernatant, the exosome pellets were washed in largevolume of ice-cold PBS and centrifuged at 100,000 3 g for another 1 h at4˚C. To isolate the FasL+ exosomes released by CD8+ T cells of C57BL/6and FasL-deficient gld mice, CD8+ T cells were stimulated with 50 mg/mlCon A for 15 min and then washed and cultured in exosome-free RPMI1640 complete medium for 1 h. Supernatant was collected, and exosomeswere purified, as previously described (23). Exosomes in the pellet wereresuspended in PBS. Exosomes from CD8+ T cells, stimulated or not withOVA257–264 peptide, were termed EXO or EXOcont; exosomes from ConA-activated CD8+ T cells of C57BL/6 or gld mice were termed EXOwt orEXOgld, respectively.

Electron microscopy

For electron microscopy observations, exosome pellets were fixed in 4%paraformaldehyde at 4˚C for 1 h. Then, the pellets were loaded ontoelectron microscopy grids coated with formvar carbon, contrasted, andembedded in a mixture of uranyl acetate and methylcellulose. Sectionswere observed with a Philips Tecnai-10 transmission electron micro-scope operating at 80 kV (Phillips Electronic Instruments, Mahwah,NJ).

FACS analysis of exosomes

For FACS analysis, exosomes were coated onto 4-mm-diameter aldehyde/sulfate latex beads, using a previously described method (24). Briefly, 20mg exosomes was incubated with 5 ml 4-mm-diameter aldehyde/sulfatelatex beads for 15 min at room temperature in PBS, with 20 ml finalvolume. The mixture was then transferred to 1 ml PBS with gentleshaking for 1 h. After centrifugation, the pellet was blocked by incuba-tion with 20 ml FCS for 30 min. Exosome-coated beads were washedthrice in PBS and resuspended in 50 ml PBS. Afterwards, beads wereincubated with corresponding fluorescent Abs for 1 h at room tempera-ture in the dark. Beads were analyzed by flow cytometry using a FACS-Calibur flow cytometer (Becton Dickinson, Mountain View, CA) andFlowJo software.

Western blot

A total of 50 mg exosomes or crude proteins extracted from cell lysateswas separated by 12% SDS-PAGE and transferred onto polyvinylidenedifluoride membrane (Millipore, Billerica, MA). Membrane was blockedwith 5% BSA in TBST and then incubated with corresponding primaryAbs overnight at 4˚C. After incubating with HRP-coupled secondary Ab

for 1 h, the membranes were scanned using Tanon 4500 (Shanghai, China),according to the manufacturer’s instructions.

Tumor invasion assay in vitro

After rehydration using a 6-fold volume of serum-free media, 50 mlMatrigel was added on a 8-mm polycarbonate membrane in 24-wellTranswell plates, and the Matrigel was solidified at 37˚C. A total of 1 3106 cells/ml was stimulated with exosomes for 2 h in cryovials underconstant rotation at 37˚C. Then 23 105 B16 or 23 104 3LL tumor cells in100 ml serum-free media were seeded into the top chamber. The bottomchamber was filled with 800 ml media containing 20% serum. Forty-eighthours after culture at 37˚C, cells were fixed with methanol for 20 min andwashed with PBS thrice for 20 min each. The fixed cells were stained with10 mg/ml DAPI for 30 min and washed with PBS. The stained cells wereexamined using a fluorescence microscope. To confirm the role of Fassignal, MMP9, NF-kB, and ERK signal in the tumor invasion, 20 mg/mlanti-mouse FasL-neutralized Ab, 10 nM MMP9 inhibitor, 10 mM PDTC,or 10 mM U0126 was added to the culture medium.

Detection of cell apoptosis

For detection of apoptosis induced by exosomes, thymocytes (1 3 106/ml)were stimulated with 20 mg/ml EXOwt or EXOgld for 8 h. For detection ofB16 or 3LL cell apoptosis induced by exosomes, viable tumor cells (2 3105/ml) were stimulated with 5, 10, or 20 mg/ml EXOcont or EXO for 24h. The apoptotic cells were stained with FITC Annexin V (BD Pharmin-gen, San Diego, CA) and propidium iodide (Sigma-Aldrich) for 5 min at4˚C in the dark. The stained cells were analyzed by FACS, as describedpreviously (25).

Cell proliferation assay

Cells were stimulated with 5, 10, or 20 mg/ml exosomes for 24, 48, or 72 hin a 96-well plate, and 20 ml alamarBlue (Invitrogen) was added per wellfor 6 h. The fluorescent intensity was detected using a DTX 880 multimodedetector (Beckman Coulter, Palo Alto, CA).

RNA interference assay

For transient silencing of cellular FLICE inhibitory protein (c-FLIP), 21-ntsequences of small interfering RNA (siRNA) duplexes were synthesized(GenePharma, Shanghai, China): 59-CCUCCUGGAUAGCUUAAGUTT-39(sense) and 59-ACUUAAGCUAUCCAGGAGGTT-39 (antisense). 59-UU-CUCCGAACGUGUCACGUTT-39 (sense) and 59-ACGUGACACGUUC-GGAGAATT-39 (antisense) were synthesized as siRNA scramble control.A total of 40 nM siRNA duplexes was transfected into cells (2 3 105/well)using 3 ml INTERFERin siRNA transfection reagent on a 24-well plate. Theefficiency of c-FLIPL transient silencing was confirmed by Western blot.

Real-time PCR

Total RNAwas extracted with TRIzol reagent (Invitrogen), according to themanufacturer’s instructions. The specific primers for b-actin for real-timePCR were 59-CGTTGACATCCGTAAAGACC-39 (sense) and 59-AACA-GTCCGCCTAGAAGCAC-39 (antisense). The specific primers for MMP9for real-time PCR were 59-CGGCACGCCTTGGTGTAGCA-39 (sense)and 59-GGCGCACCAGCGGTAACCAT-39 (antisense). The followingPCR conditions were used: 1 cycle at 95˚C for 30 s and then 40 cyclesof 5 s at 95˚C and 34 s at 60˚C. Real-time PCR was performed on anApplied Biosystems 7500 real-time PCR system (Foster City, CA).

Lung invasion assay of B16 tumor cells in vivo

A total of 5 3 105 B16 tumor cells with 30 mg EXOcont, EXO, 20 mg/mlanti-FasL–preincubated EXO, or PBS was injected into 8-wk-old C57BL/6mice via the tail vein. Two weeks after injection, mice were sacrificed, andthe lungs were detached. After the number of lung nodules was counted,the excised lungs were photographed and fixed in 10% formaldehyde forfurther H&E staining.

Purification of an in vivo counterpart of EXO fromtumor-activated CD8+ T cells

A total of 1 3 106 B16 tumor cells was injected s.c. into 8-wk-old C57BL/6 mice. The mice were sacrificed 10 d after tumor inoculation. Lymphocytesuspensions from normal and tumor mice were prepared by crushing thelymph nodes and spleens between two pieces of ground glass. For prep-aration of tumor tissue suspensions, tumor tissues were detached anddissociated enzymatically for 2 h with 1 mg/ml type I collagenase (Sigma-Aldrich, St. Louis, MO) in the presence of 50 U/ml RNase and DNase

The Journal of Immunology 5955

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

(Sigma-Aldrich). Exosomes from all kinds of suspensions were filtratedusing a 0.22-mm-filter device and then purified by differential centrifu-gation, as mentioned above. For purification of exosomes from CD8+

T cells, the exosomes were sorted through CD8+ Dynabeads (Dynal;Invitrogen). Briefly, 200 mg exosomes was mixed with 100 ml Dynabeadsat 4˚C by gentle shaking overnight, and the exosome-coated beads weredetached with 200 ml DETACHaBEAD at room temperature for 45 min.Bead-free supernatant was colleted after precipitating beads via a magnet.CD8+ exosomes were washed in a large volume of ice-cold PBS andcentrifuged at 100,000 3 g for 1 h.

Statistical analysis

Experimental data were analyzed using one-way ANOVA and the t testusing SPSS 11.5. Differences were considered significant when the pvalues were , 0.05.

ResultsFasL+ EXO showed no effect on B16 tumor cell apoptosis andproliferation

To evaluate the isolated EXO integrity from activated CD8+ T cells,we inspected EXO morphology by electron microscopy. The ob-servation showed that isolated EXO displayed a typical roundmorphology, with a diameter of 50–100 nm (Fig. 1A). Like theirparent CD8+ T cells, EXO also expressed CD8+ T cell–specific

molecules, including CD8 and TCR. In addition, they expressedspecific EXO markers, such as CD9, HSP70, and TSG101,whereas endoplasmic reticulum-residing protein GRP94 was ab-sent in EXO. Moreover, EXO, but not EXOcont, showed positiveexpression of FasL (Fig. 1B, 1C). These results demonstrate theintegrity of isolated positive EXO from activated CD8+ T cells inthis study. With the successful isolation of Fas+ EXO, we firstexamined whether EXO could induce apoptosis of thymocytes.Apoptosis assay by FACS showed that EXO, but not EXOcont,could significantly induce the apoptosis of thymocytes, and theability of EXO to induce the apoptosis of thymocytes was greatlyinhibited when preneutralized by FasL mAbs (Fig. 1D), whichsuggest the biological activity of FasL in EXO. To further confirmthe relationship between apoptosis induction and FasL in EXO,we isolated exosomes from activated CD8+ T cells of C57BL/6and gld mice and confirmed their expression of FasL (data notshown). After treatment with EXOwt or EXOgld, increasing ap-optosis was found in EXOwt-treated, but not EXOgld-treated,thymocytes (Fig. 1E). These results indicate that the apoptosis-inducing function of EXO is FasL dependent. Then we detectedwhether EXO could also induce the apoptosis of Fas-expressingB16 or 3LL tumor cells. The results showed that FasL+ EXO has

FIGURE 1. FasL+ EXO showed no effect on B16

tumor cell apoptosis and proliferation. (A) Electron

micrograph of activated CD8+ T cell EXO. Scale bar,

200 nm. (B) For analysis of expression of surface

molecules, 20 mg of EXO was incubated with 5 ml of

4-mm-diameter aldehyde/sulfate latex beads and

stained with a panel of Abs (solid lines) or negative

control Abs (dotted lines). (C) Western blot analysis of

EXO/EXOcont and the whole-cell lysate of activated

CD8+ OT-I T cells/naive CD8+ OT-I T cells using the

Abs indicated. (D) A total of 1 3 106/ml thymocytes

was stimulated with 20 mg/ml exosomes, 20 mg/ml of

FasL-neutralized mAbs, or exosomes preincubated

with isotype mAbs (ISO) for 8 h. The apoptosis of

thymocytes was analyzed by FACS (n = 3). **p ,0.01 versus Cont, EXOcont, or EXO+anti-FasL. (E) A

total of 13 106/ml thymocytes was stimulated with 50

mg/ml Con A-activated CD8+ T cell exosomes from

WT or gld mice for 8 h. The apoptosis of thymocytes

was analyzed by FACS (n = 3). **p , 0.01 versus

Cont or EXOgld. (F) B16 tumor cells were incubated

with a serial concentration of EXO or EXOcont for

24 h. The apoptosis of B16 tumor cells was analyzed

by FACS (n = 3). (G) B16 tumor cells were incubated

with a serial concentration of EXO or EXOcont

for 24, 48, or 72 h. The proliferation of B16 tumor cells

was measured by alamarBlue assay. Data are repre-

sentative of three independent experiments.

5956 ACTIVATED T CELL EXOSOMES PROMOTE TUMOR INVASION

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

little effect on B16 apoptosis (Fig. 1F) or 3LL apoptosis (data notshown). Because it was demonstrated that Fas signaling can pro-mote the proliferation of thyroid carcinomas (11), we next in-vestigated whether EXO could promote the proliferation of B16 or3LL tumor cells. The results showed that, with prolonged treat-ment time or increased concentration, EXO did not affect theproliferation of B16 (Fig. 1G) or 3LL (data not shown) comparedwith control.

EXO promoted tumor cell invasion via Fas signaling in vitro

Despite the observation that FasL+ EXO showed little effecton apoptosis and proliferation of B16 and 3LL tumor cells, wefurther investigated whether EXO can cause other functionalchanges in tumor cells. Because it was reported that activationof Fas signaling can induce motility and invasion of apoptosis-resistant tumor cells (13), we reasoned that EXO may increase theinvasive ability of tumor cells in vitro. To validate this, inva-siveness assays in vitro were performed. In this assay, B16 or 3LLtumor cells were placed in the top chamber in serum-free mediawith stimulus (EXO, EXOcont, or PBS) for 2 h. The bottomchamber contained 20% FCS. Forty-eight hours after culture, weobserved an increasing number of B16 or 3LL tumor cells inthe bottom chamber in the EXO group but not in the EXOcont orPBS group (Fig. 2A). Interestingly, once preneutralized by FasLmAbs, the ability of EXO to promote tumor cell invasion wascompletely abrogated (Fig. 2B). These results suggest that EXOmay increase the invasive ability of tumor cells via Fas signalingin vitro.

EXO-mediated activation of Fas signaling inducedupregulation of MMP9 expression in tumor cells

It was reported that CD95L mediates the migration of myeloidcells via MMP9 activation (26). We assume that the effect of EXOon tumor cell invasion is also related to MMP9 activation. Asshown in Fig. 3A, after incubation with EXO, but not EXOcont,the mRNA and protein levels of MMP9 in B16 were significantlyincreased; the optimal concentration to stimulate was 10 mg/ml(Fig. 3A, 3B). After blockage of Fas signaling by anti-FasL, theexpression of MMP9 induced by EXO decreased greatly (Fig.3C), which provides solid evidence that upregulation of MMP9induced by EXO is Fas/FasL dependent. To further confirm therelationship between EXO-mediated promotion of tumor cell in-vasion and MMP9 upregulation, MMP9 inhibitor was added toB16 tumor cell-invasiveness assays in vitro. A total of 10 nM ofMMP9 inhibitor completely abolished the ability of EXO topromote tumor cell invasion (Fig. 3D). These results indicate thatthe tumor cell invasion induced by FasL+ EXO was associatedwith MMP9 expression.

EXO promoted MMP9 expression and tumor cell invasionthrough ERK and NF-kB pathways

The stimulation of CD95L on tumor cells can activate the NF-kBand ERK pathways (13), and the activation of NF-kB and ERKincreases MMP9 expression (8, 27). Similarly, EXO stimulationactivated the NF-kB and ERK pathways in B16 cancer cells,which was demonstrated by the increased phosphorylation of NF-kB and ERK in response to EXO stimulation but not EXOcontstimulation (Fig. 4A). Nevertheless, it seems that the expressionof ERK and NF-kB is in parallel pathways, because when ERKexpression was disrupted by its specific inhibitor U0216, littlechange in the expression of NF-kB was observed (data notshown). To further reveal the roles of activated NF-kB and ERK inMMP9 expression and tumor cell invasion, specific inhibitors forthe NF-kB–and ERK-signaling pathways were used to treat B16cancer cells before EXO stimulation; MMP9 expression wassubsequently examined. We found that both NF-kB and ERKinhibitors (PDTC and U0126) markedly suppressed EXO-inducedMMP9 expression and tumor cell invasion (Fig. 4B, 4C). Col-lectively, these results showed that FasL+ EXO-activated NF-kB– and ERK-signaling pathways were responsible for the in-creased MMP9 expression and tumor cell invasion.

c-FLIP was critical to the increase in MMP9 expression andtumor invasion induced by EXO

c-FLIP is an endogenous inhibitor of death receptor-induced ap-optosis that acts through the caspase-8 pathway and is widelyexpressed in tumors (28, 29). c-FLIP exists as a long (c-FLIPL)or a short (c-FLIPS) variant (30, 31). c-FLIPL shows overallstructural homology to caspase-8, containing two death effectordomains that interact with Fas-associated death domain protein, aswell as an inactive caspase-like domain. c-FLIPS contains only thetwo death effector domains and has lower antiapoptotic capacity(32). In B16 tumor cells, we found that Fas engagement with EXOled to the increased accumulation of c-FLIPL but not c-FLIPS (Fig.5A). The increase in c-FLIPL was detected at 5 min after EXOstimulation, and the high level was maintained for 24 h (Fig. 5A).It was reported that c-FLIP can promote activation of the ERK-and NF-kB–signaling pathways (32). Therefore, we investigatedwhether their activation in B16 tumor cells is mediated by c-FLIPLupon EXO stimulation. We inhibited c-FLIPL expression bysiRNA and confirmed the inhibitory efficiency by Western blot.c-FLIPL expression in c-FLIPL siRNA-transfected B16 cells de-creased ∼50% compared with c-FLIPL expression in control cells(Fig. 5B). Then we wanted to know whether the knock down ofc-FLIPL leads to the apoptosis of B16 cells stimulated by EXO. Wefound that the apoptosis of B16 cells with c-FLIPL knock downshowed few differences in comparison with control siRNA knock

FIGURE 2. EXO promoted tumor cell invasion via Fas signaling in vitro. (A) B16 or 3LL tumor cells were incubated with 10 mg of EXO, EXOcont, or

PBS for 2 h, and the cells were plated in the bottom chamber precoated with 50 ml of Matrigel. Forty-eight hours later, the number of cells on the bottom of

the Transwell filter was imaged and quantified. (B) Same experimental procedure as in (A), except that the cells were preincubated with 20 mg/ml of FasL-

neutralized mAbs or ISO for 2 h (n = 5). Data are representative of three independent experiments. *p , 0.05 versus Cont and EXOcont (A) or EXO+anti-

FasL (B).

The Journal of Immunology 5957

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

down (Fig. 5C), which excludes the possibility that the changesin signal pathways may be caused by increasing cell apoptosis.Subsequent Western blot analysis showed that the phosphorylationof ERK and NF-kB decreased significantly when the expressionof c-FLIPL was inhibited by siRNA (Fig. 5D). Moreover, cellinvasion and the expression of MMP9 were also inhibited whenc-FLIPL expression was knocked down (Fig. 5E, 5F). It is knownthat NF-kB signals can induce the expression of c-FLIP (33),which is related to the expression of MMP9 (8). We propose thata similar pattern might also appear in Fas signaling-activated tu-mor cells induced by EXO. As shown in Fig. 5G, inhibition of theNF-kB–signaling pathway decreased the expression of c-FLIPL.Interestingly, the same result was observed when B16 tumor cellswere treated by specific inhibitors of ERK (Fig. 5G). These resultssuggest that c-FLIPL plays an important role in the increase inMMP9 and tumor invasion induced by EXO.

EXO increased the tumor cell lung invasion via Fas signalingin vivo

To further clarify the effect of EXO on tumor invasion, we per-formed tumor invasive analysis with the B16 tumor cell-inducedmurine lung cancer model. We found that EXO greatly in-creased the tumor cell lung invasion, as indicated by gross mor-phology of the lungs (Fig. 6A, left panels) and the number ofinvasive nodules (Fig. 6A, right panel). Histological assessment oflungs showed few tumor colonies in EXOcont and PBS groups butmore numerous tumor colonies in EXO groups (Fig. 6B). Afterpreincubation with FasL mAbs, the ability of EXO to promote

tumor lung invasion was markedly decreased (Fig. 6). Theseresults strongly support that EXO increase tumor cell lung inva-sion via Fas signaling in vivo.

Identification of a counterpart of EXO in vivo fromtumor-activated CD8+ T cells

To further confirm the existence of CD8 and FasL+ exosomes andtheir effects on tumor metastasis in tumor mice, we isolated CD8+

exosomes from spleens and draining lymph nodes of control andtumor mice using CD8+ magnetic beads. The CD8+ exosomesfrom tumor tissues were also isolated. As shown in Fig. 7A,exosomes from 3LL and B16 tumor mice, but not control mice,expressed FasL; its expression level in exosomes from tumortissues was greater than that in tumor lymphocytes. All exosomesexpressed CD8 molecules and exosomal mark proteins, such asHSP70 and TSG101 (Fig. 7A). After stimulation with exosomesisolated from tumor mice, but not control mice, 3LL and B16tumor cells showed increase expression of MMP9 and enhancedinvasive ability. Consistent with their expression of FasL, exo-somes from tumor tissue showed stronger stimulatory effect thandid those from tumor lymphocytes (Fig. 7B, 7C). Based on ourresults, it can be predicted that the number of lung tumor nodulesin the B16 tumor model of wild-type (WT) mice will be greaterthan that in gld mice. We examined the difference in the numberof lung tumor nodules between WT and gld mice; as expected, thenumber was greater in WT mice (Fig. 7D). These results indicatethat CD8+ and FasL+ exosomes are naturally generated with tumorprogression and exert an exacerbated effect in tumor invasion.

FIGURE 3. EXO-mediated activation of Fas signaling induced upregulation of MMP9 expression in tumor cells. (A) B16 tumor cells were treated with

the indicated doses of EXO for 24 h or 10 mg/ml EXO for the indicated times. mRNA expression levels of MMP9 were assayed by quantitative PCR. (B)

Western blot analysis of MMP9 expression in B16 tumor cells stimulated with 10 mg/ml of EXO for the indicated times. (C) Western blot analysis of MMP9

expression in B16 tumor cells stimulated with 10 mg/ml of EXO, EXO preincubated with 20 mg/ml of FasL-neutralized mAbs, or isotype mAbs (ISO) for

2 h. (D) Invasive assay of B16 tumor cells after stimulation with 10 mg/ml of EXO or EXO plus MMP9 inhibitor (MMP9i) (n = 5). Data are representative

of three independent experiments. *p , 0.05, **p , 0.01.

FIGURE 4. EXO promoted MMP9 expression and tumor cell invasion through ERK and NF-kB pathways. (A) B16 tumor cells were stimulated with

10 mg/ml of EXO or EXOcont for the indicated times and harvested; the extracted crude proteins were subjected to Western blot with Abs against NF-kB,

phospho–NF-kB, ERK, or phospho-ERK. (B) B16 tumor cells were pretreated with NF-kB–specific inhibitor PDTC or ERK1/2-specific inhibitor U0126 for

30 min and then stimulated with 10 mg/ml EXO for 24 h. The expression of MMP9 in B16 tumor cells was detected by Western blot. (C) Using the same

experimental conditions as in (B), the number of cells that migrated through Matrigel and adhered to the bottom of Transwell filter was quantified (n = 5).

Data are representative of three independent experiments. **p , 0.01.

5958 ACTIVATED T CELL EXOSOMES PROMOTE TUMOR INVASION

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

DiscussionStimulation of FasR normally induces an apoptotic death signal.However, the interaction between Fas and FasL in tumor cells doesnot always induce a death signal, and conversely, promotes tu-morigenesis (11). Activated CD8+ T cells secrete FasL+ exosomes,which can induce DC apoptosis and inhibit their ability to stim-ulate CD8+ CTL responses (19). Until now, the effect of FasL+

exosomes from activated CD8+ CTLs on Fas-resistant tumors wasunknown. In this study, we found that FasL+ EXO did not affectthe apoptosis and proliferation of tumor cells but accelerated theirinvasion in vivo via upregulation of MMP9 expression. Moreover,we identified a counterpart of EXO in tumor-bearing mice that

also possessed the ability to increase MMP9 expression and

cancer invasion.Increasing the expression of FasL during activation is one of the

properties of T lymphocytes. Tumor Ag-specific CTLs and CD8+

T cells activated by tumor Ag are generally the most importanteffector cells to attack tumor cells (34). Activated CD8+ T cells

are present within the tumor site, but after being educated by tu-

mor cells, they are in an unresponsive state (35). Our results showthat CD8 and FasL+ EXO could be isolated from tumor tissues,

suggesting that, in the tumor environment, CTLs continuouslysecrete FasL+ EXO to promote tumor growth rather than eliminate

tumor tissues (i.e., CTLs in the tumor site are converted from

FIGURE 5. c-FLIP was critical to the increase in MMP9 expression and tumor invasion induced by EXO. (A) Western blot analysis of c-FLIPL and

c-FLIPS expression in B16 tumor cells stimulated with 10 mg/ml of EXO or EXOcont for the indicated times. (B) B16 cells were transfected with c-FLIPL(Si) or control (Cont Si) siRNA duplexes for 24 h, and c-FLIPL expression in the cells was detected by Western blot. (C) The apoptosis of Si- or Cont Si-

transfected B16 tumor cells treated by EXO were analyzed by FACS (n = 3). (D) After inhibition of c-FLIPL expression by siRNA, the phosphorylation of

NF-kB and ERK in EXO- or EXOcont-stimulated B16 tumor cells was detected by Western blot. (E) After inhibition of c-FLIPL expression by siRNA,

MMP9 expression was detected by Western blot in B16 cells stimulated with 10 mg/ml of EXO or EXOcont for 24 h. (F) After inhibition of c-FLIPLexpression by siRNA, the number of B16 tumor cells, stimulated with 10 mg/ml of EXO or EXOcont for 2 h, which migrated through Matrigel to the bottom

of the Transwell filter was quantified (n = 5). (G) Western blot analysis of c-FLIPL expression in EXO-stimulated B16 tumor cells pretreated with NF-kB–

specific inhibitor PDTC or ERK1/2-specific inhibitor U0126. Data are representative of three independent experiments. **p, 0.01 versus Cont or Si+EXO.

FIGURE 6. EXO increased tumor cell

lung invasion via Fas signaling in vivo.

(A) A total of 5 3 105 B16 tumor cells

with 30 mg of EXOcont, EXO, 20 mg/ml

of ISO-preincubated EXO, 20 mg/ml

of FasL-neutralized mAbs preincubated

EXO or PBS was injected into C57BL/6

mice through the tail vein. The invasion

of B16 tumor cells into lung (left panels)

and the number of lung nodules were

assessed (right panel) (n = 5). (B) The

invasion of B16 tumor cells into lung was

visualized in H&E-stained pathological

sections. Original magnification 3100.

Data are representative of three inde-

pendent experiments. **p , 0.01 versus

PBS, EXOcont, and EXO+anti-FasL.

The Journal of Immunology 5959

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

a guard into an accessory of tumor progression under such cir-cumstances). In the FACS results in Fig. 1A, EXO is adsorbedonto latex with an intact membrane structure, and if FasL can bedetected, it is probably membrane-bound. Moreover, in the resultsfrom Western blotting, we detected a band at ∼40 kDa, which issimilar to the molecular mass of membrane-bound FasL, whereaswe detected no signal at 26 kDa, which is similar to the molecularmass of soluble FasL (data not shown). CTLs themselves alsoexpress FasL and may trigger Fas signal in tumor cells, whichprobably promotes MMP9 expression and the invasive capabilityof tumor cells. However, there are very limited numbers of CTLsin tumor sites (35). Furthermore, unlike exosomes, whose diam-eters are,100 nm, allowing them to freely pass through basementmembrane (36), immigration of CTLs from spleen and lymphnodes into the tumor site is more complicated. It can be presumedthat more FasL+ EXO will accumulate in the tumor location thanCTLs, which will probably exert a stronger effect of invasivepromotion on tumors than their parent CTLs.Many molecules were reported to be associated with tumor in-

vasion. Among them, MMP2, MMP9, cyclooxygenase-2 (COX2),b-catenin, and urokinase plasminogen activator, are consideredthe primary tumor-invasive effectors (13, 20, 37–39). In our study,we found that, after stimulation by EXO, expression of the mRNAlevels of MMP2, MMP9, and COX2 was significantly elevated in3LL and B16 tumor cells. However, we could not detect anychange in the protein expression of MMP2 in either cell type (datanot shown). Although the expression of COX2 mRNA increased inboth cells, almost no PGE2 secretion could be detected with orwithout EXO stimulation in B16 tumor cells (data not shown).Interestingly, urokinase plasminogen activator, which is reportedto be related to tumor invasion mediated by Fas signaling (13),showed no increase in either cell type after EXO stimulation (datanot shown). Among those tumor-invasive effectors, we detectedonly elevated mRNA and protein levels of MMP9 in both cancercell lines. It is likely that MMP9 is a universal effector moleculethat promotes tumor invasion induced by FasL in EXO.Although increasing tumors have been found to possess Fas

resistance, the signaling mechanisms of tumor Fas resistance re-main unknown. c-FLIP is an endogenous inhibitor of deathreceptor-induced apoptosis, and it plays a critical role in tumor-

apoptosis resistance via the caspase-8 pathway (28, 40). In thisstudy, we found that, only 5 min after FasL+ EXO stimulation, the

amount of c-FLIPL increased markedly, and this was sustained for

$24 h. c-FLIP is extremely unstable and can be quickly degraded

by the ubiquitination pathway (41, 42). The early increase in c-

FLIPL probably results from the inhibition of its degradation by

ubiquitin. Although it was reported that protein kinase C-mediated

phosphorylation regulates c-FLIP ubiquitylation and stability

(43), the mechanism underlying the ubiquitylation inhibition of

c-FLIPL induced by EXO remains to be elucidated. As a result

of positive feedback, NF-kB and ERK activation triggered by

c-FLIPL probably promotes the synthesis of c-FLIPL, which may

explain why it remains at high levels for a long time after EXO

stimulation. Although the Fas-mediated apoptosis after c-FLIP

siRNA inhibition in B16 cells showed no difference compared

with control siRNA, we cannot exclude that c-FLIP plays a role in

the Fas resistance of B16 cells, because the effect of c-FLIP knock

down is mild, and the expression of c-FLIP decreases only ∼50%.

It was reported that activation of the ERK pathway and subsequent

induction of antiapoptotic proteins, such as Bcl-2, play important

roles in the survival of CD8+ T and NK cells (44, 45) In B16 tumor

cells stimulated by FasL+ EXO, we found that Bcl-2 was upreg-

ulated after EXO stimulation, and specific inhibitors of ERK and

NF-kB activation could decrease the expression of Bcl-2. But

there was no increasing apoptosis in the B16 cells treated with

Bcl-2 inhibitor compared with the group treated with DMSO or

control, which suggests that Bcl-2 may not be involved in the Fas

resistance of B16 tumor cells induced by FasL-expressed EXO

(data not shown). Tumor cells were also demonstrated to syn-

thesize a protective protein to resistance apoptosis induced by Fas

signaling (46). A high level of a protein tyrosine phosphatase, Fas-

associated phosphatase-1, which can interfere with transduction of

the apoptotic signal by Fas (47) upon interaction of its third PDZ

domain with the C-terminal three amino acids of Fas (48), has

been implicated as a possible agent causing preferential survival

of human pancreatic adenocarcinomas stimulated by FasL (46). It

is possible that Fas-associated phosphatase-1 contributes to the

Fas resistance in B16 tumor cells stimulated by EXO, but this

needs to be explored further.

FIGURE 7. Identification of a counterpart of EXO in vivo from tumor-activated CD8+ T cells. CD8+ exosomes from spleens and draining lymph nodes of

control (EXOcln) and tumor mice (EXOtln), as well as tumor tissue of tumor mice (EXOts), were isolated by CD8+ magnetic beads. (A) Western blot

analysis of FasL, CD8, HSP70, and TSG101 expression in EXOcln, EXOtln, and EXOts. (B) Western blot analysis of MMP9 expression in tumor cells after

stimulation with 10 mg/ml of EXOcln, EXOtln, and EXOts for 24 h. (C) The number of invasive tumor cells stimulated with EXOcln, EXOtln, and EXOts

through Matrigel to the bottom of the Transwell filter was quantified (n = 5). *p , 0.05, **p , 0.01 versus Cont and EXOcln. (D) A total of 5 3 105 B16

tumor cells was injected into C57BL/6 (WT) or gld (GLD) mice via the tail vein. The number of lung nodules was measured (n = 5). Data are representative

of three independent experiments. **p , 0.01.

5960 ACTIVATED T CELL EXOSOMES PROMOTE TUMOR INVASION

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

In conclusion, we report that CTLs are diverted from tumorinhibition to tumor promotion through secreting FasL+ EXO. Ourdata showed that EXO could activate c-FLIPL and ERK- and NF-kB–signaling pathways to increase MMP9 expression in tumorcells, which is Fas/FasL dependent. The upregulation of MMP9leads to exacerbated tumor invasion, and these tumor-promotingexosomes are proved to exist in tumor-bearing mice. Our resultsreveal a novel mechanism of tumor immune escape and suggestthat strategies to inhibit, rather than to promote, Fas activityshould be considered during cancer therapy.

DisclosuresThe authors have no financial conflicts of interest.

References1. Peter, M. E., and P. H. Krammer. 2003. The CD95(APO-1/Fas) DISC and be-

yond. Cell Death Differ. 10: 26–35.2. Green, D. R., N. Droin, and M. Pinkoski. 2003. Activation-induced cell death in

T cells. Immunol. Rev. 193: 70–81.3. Schulze-Osthoff, K., D. Ferrari, M. Los, S. Wesselborg, and M. E. Peter. 1998.

Apoptosis signaling by death receptors. Eur. J. Biochem. 254: 439–459.4. Nagata, S. 1997. Apoptosis by death factor. Cell 88: 355–365.5. Li, Y., A. J. Raffo, L. Drew, Y. Mao, A. Tran, D. P. Petrylak, and R. L. Fine.

2003. Fas-mediated apoptosis is dependent on wild-type p53 status in humancancer cells expressing a temperature-sensitive p53 mutant alanine-143. CancerRes. 63: 1527–1533.

6. Peng, S. L., M. E. Robert, A. C. Hayday, and J. Craft. 1996. A tumor-suppressorfunction for Fas (CD95) revealed in T cell-deficient mice. J. Exp. Med. 184:1149–1154.

7. Rokhlin, O. W., R. A. Glover, and M. B. Cohen. 1998. Fas-mediated apoptosis inhuman prostatic carcinoma cell lines occurs via activation of caspase-8 andcaspase-7. Cancer Res. 58: 5870–5875.

8. Chen, L., S. M. Park, A. V. Tumanov, A. Hau, K. Sawada, C. Feig, J. R. Turner,Y. X. Fu, I. L. Romero, E. Lengyel, and M. E. Peter. 2010. CD95 promotestumour growth. Nature 465: 492–496.

9. Zhang, Y., Q. Liu, M. Zhang, Y. Yu, X. Liu, and X. Cao. 2009. Fas signalpromotes lung cancer growth by recruiting myeloid-derived suppressor cells viacancer cell-derived PGE2. J. Immunol. 182: 3801–3808.

10. Li, H., X. Fan, C. Stoicov, J. H. Liu, S. Zubair, E. Tsai, R. Ste Marie, T. C. Wang,S. Lyle, E. Kurt-Jones, and J. Houghton. 2009. Human and mouse colon cancerutilizes CD95 signaling for local growth and metastatic spread to liver. Gas-troenterology 137: 934–944, 944.e1–4.

11. Mitsiades, C. S., V. Poulaki, G. Fanourakis, E. Sozopoulos, D. McMillin,Z. Wen, G. Voutsinas, S. Tseleni-Balafouta, and N. Mitsiades. 2006. Fas sig-naling in thyroid carcinomas is diverted from apoptosis to proliferation. Clin.Cancer Res. 12: 3705–3712.

12. Shinohara, H., H. Yagita, Y. Ikawa, and N. Oyaizu. 2000. Fas drives cell cycleprogression in glioma cells via extracellular signal-regulated kinase activation.Cancer Res. 60: 1766–1772.

13. Barnhart, B. C., P. Legembre, E. Pietras, C. Bubici, G. Franzoso, and M. E. Peter.2004. CD95 ligand induces motility and invasiveness of apoptosis-resistant tu-mor cells. EMBO J. 23: 3175–3185.

14. Nijkamp, M. W., F. J. Hoogwater, E. J. Steller, B. F. Westendorp, T. A. van derMeulen, M. W. Leenders, I. H. Borel Rinkes, and O. Kranenburg. 2010. CD95 isa key mediator of invasion and accelerated outgrowth of mouse colorectal livermetastases following radiofrequency ablation. J. Hepatol. 53: 1069–1077.

15. Martınez-Lorenzo, M. J., A. Anel, S. Gamen, I. Monleon, P. Lasierra, L. Larrad,A. Pineiro, M. A. Alava, and J. Naval. 1999. Activated human T cells release bio-active Fas ligand and APO2 ligand in microvesicles. J. Immunol. 163: 1274–1281.

16. Johnstone, R. M., M. Adam, J. R. Hammond, L. Orr, and C. Turbide. 1987.Vesicle formation during reticulocyte maturation. Association of plasma mem-brane activities with released vesicles (exosomes). J. Biol. Chem. 262: 9412–9420.

17. Chaput, N., and C. Thery. 2011. Exosomes: immune properties and potentialclinical implementations. Semin. Immunopathol. 33: 419–440.

18. Mincheva-Nilsson, L., and V. Baranov. 2010. The role of placental exosomes inreproduction. Am. J. Reprod. Immunol. 63: 520–533.

19. Xie, Y., H. Zhang, W. Li, Y. Deng, M. A. Munegowda, R. Chibbar, M. Qureshi,and J. Xiang. 2010. Dendritic cells recruit T cell exosomes via exosomal LFA-1leading to inhibition of CD8+ CTL responses through downregulation ofpeptide/MHC class I and Fas ligand-mediated cytotoxicity. J. Immunol. 185:5268–5278.

20. Fridman, R., M. Toth, D. Pena, and S. Mobashery. 1995. Activation of proge-latinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res. 55: 2548–2555.

21. van Deventer, H. W., Q. P. Wu, D. T. Bergstralh, B. K. Davis, B. P. O’Connor,J. P. Ting, and J. S. Serody. 2008. C-C chemokine receptor 5 on pulmonaryfibrocytes facilitates migration and promotes metastasis via matrix metal-loproteinase 9. Am. J. Pathol. 173: 253–264.

22. Kleber, S., I. Sancho-Martinez, B. Wiestler, A. Beisel, C. Gieffers, O. Hill,M. Thiemann, W. Mueller, J. Sykora, A. Kuhn, et al. 2008. Yes and PI3K bindCD95 to signal invasion of glioblastoma. Cancer Cell 13: 235–248.

23. Chen, W., J. Wang, C. Shao, S. Liu, Y. Yu, Q. Wang, and X. Cao. 2006. Efficientinduction of antitumor T cell immunity by exosomes derived from heat-shockedlymphoma cells. Eur. J. Immunol. 36: 1598–1607.

24. Dai, S., T. Wan, B. Wang, X. Zhou, F. Xiu, T. Chen, Y. Wu, and X. Cao. 2005.More efficient induction of HLA-A*0201-restricted and carcinoembryonic an-tigen (CEA)-specific CTL response by immunization with exosomes preparedfrom heat-stressed CEA-positive tumor cells. Clin. Cancer Res. 11: 7554–7563.

25. Zhang, X., H. Huang, J. Yuan, D. Sun, W. S. Hou, J. Gordon, and J. Xiang. 2005.CD4-8- dendritic cells prime CD4+ T regulatory 1 cells to suppress antitumorimmunity. J. Immunol. 175: 2931–2937.

26. Letellier, E., S. Kumar, I. Sancho-Martinez, S. Krauth, A. Funke-Kaiser,S. Laudenklos, K. Konecki, S. Klussmann, N. S. Corsini, S. Kleber, et al. 2010.CD95-ligand on peripheral myeloid cells activates Syk kinase to trigger theirrecruitment to the inflammatory site. Immunity 32: 240–252.

27. Liu, L. P., H. F. Liang, X. P. Chen, W. G. Zhang, S. L. Yang, T. Xu, and L. Ren.2010. The role of NF-kappaB in Hepatitis b virus X protein-mediated upregu-lation of VEGF and MMPs. Cancer Invest. 28: 443–451.

28. Sharp, D. A., D. A. Lawrence, and A. Ashkenazi. 2005. Selective knockdown ofthe long variant of cellular FLICE inhibitory protein augments death receptor-mediated caspase-8 activation and apoptosis. J. Biol. Chem. 280: 19401–19409.

29. Rippo, M. R., S. Moretti, S. Vescovi, M. Tomasetti, S. Orecchia, G. Amici,A. Catalano, and A. Procopio. 2004. FLIP overexpression inhibits death receptor-induced apoptosis in malignant mesothelial cells. Oncogene 23: 7753–7760.

30. Chang, D. W., Z. Xing, Y. Pan, A. Algeciras-Schimnich, B. C. Barnhart,S. Yaish-Ohad, M. E. Peter, and X. Yang. 2002. c-FLIP(L) is a dual functionregulator for caspase-8 activation and CD95-mediated apoptosis. EMBO J. 21:3704–3714.

31. Golks, A., D. Brenner, C. Fritsch, P. H. Krammer, and I. N. Lavrik. 2005.c-FLIPR, a new regulator of death receptor-induced apoptosis. J. Biol. Chem.280: 14507–14513.

32. Kataoka, T., R. C. Budd, N. Holler, M. Thome, F. Martinon, M. Irmler, K. Burns,M. Hahne, N. Kennedy, M. Kovacsovics, and J. Tschopp. 2000. The caspase-8inhibitor FLIP promotes activation of NF-kappaB and Erk signaling pathways.Curr. Biol. 10: 640–648.

33. Micheau, O., S. Lens, O. Gaide, K. Alevizopoulos, and J. Tschopp. 2001. NF-kappaB signals induce the expression of c-FLIP. Mol. Cell. Biol. 21: 5299–5305.

34. Armstrong, A. C., D. Eaton, and J. C. Ewing. 2001. Science, medicine, and thefuture: Cellular immunotherapy for cancer. BMJ 323: 1289–1293.

35. Bronte, V., T. Kasic, G. Gri, K. Gallana, G. Borsellino, I. Marigo, L. Battistini,M. Iafrate, T. Prayer-Galetti, F. Pagano, and A. Viola. 2005. Boosting antitumorresponses of T lymphocytes infiltrating human prostate cancers. J. Exp. Med.201: 1257–1268.

36. van Niel, G., and M. Heyman. 2002. The epithelial cell cytoskeleton and intracel-lular trafficking. II. Intestinal epithelial cell exosomes: perspectives on their struc-ture and function. Am. J. Physiol. Gastrointest. Liver Physiol. 283: G251–G255.

37. Kim, J., J. A. Pyun, S. Won Cho, K. Lee, and K. Kwack. 2011. Lymph nodemetastasis of gastric cancer is associated with the interaction between poly(ADP-ribose) polymerase 1 and matrix metallopeptidase 2. DNA Cell Biol. 30:1011–1017.

38. Kim, H., J. Y. Song, J. Y. Cho, Y. S. Yoon, H. S. Han, H. S. Lee, H. S. Ryu, andG. Choe. 2010. Strong cytoplasmic expression of COX2 at the invasive fronts ofgallbladder cancer is associated with a poor prognosis. J. Clin. Pathol. 63: 1048–1053.

39. Fu, L., C. Zhang, L. Y. Zhang, S. S. Dong, L. H. Lu, J. Chen, Y. Dai, Y. Li,K. L. Kong, D. L. Kwong, and X. Y. Guan. 2011. Wnt2 secreted by tumourfibroblasts promotes tumour progression in oesophageal cancer by activation ofthe Wnt/b-catenin signalling pathway. Gut 60: 1635–1643.

40. Mathas, S., A. Lietz, I. Anagnostopoulos, F. Hummel, B. Wiesner, M. Janz,F. Jundt, B. Hirsch, K. Johrens-Leder, H. P. Vornlocher, et al. 2004. c-FLIPmediates resistance of Hodgkin/Reed-Sternberg cells to death receptor-inducedapoptosis. J. Exp. Med. 199: 1041–1052.

41. Kim, Y., N. Suh, M. Sporn, and J. C. Reed. 2002. An inducible pathway fordegradation of FLIP protein sensitizes tumor cells to TRAIL-induced apoptosis.J. Biol. Chem. 277: 22320–22329.

42. Toivonen, H. T., A. Meinander, T. Asaoka, M. Westerlund, F. Pettersson,A. Mikhailov, J. E. Eriksson, and H. Saxen. 2011. Modeling reveals that dynamicregulation of c-FLIP levels determines cell-to-cell distribution of CD95-mediated apoptosis. J. Biol. Chem. 286: 18375–18382.

43. Kaunisto, A., V. Kochin, T. Asaoka, A. Mikhailov, M. Poukkula, A. Meinander,and J. E. Eriksson. 2009. PKC-mediated phosphorylation regulates c-FLIPubiquitylation and stability. Cell Death Differ. 16: 1215–1226.

44. D’Souza, W. N., C. F. Chang, A. M. Fischer, M. Li, and S. M. Hedrick. 2008.The Erk2 MAPK regulates CD8 T cell proliferation and survival. J. Immunol.181: 7617–7629.

45. Subramanian, M., and C. Shaha. 2007. Up-regulation of Bcl-2 through ERKphosphorylation is associated with human macrophage survival in an estrogenmicroenvironment. J. Immunol. 179: 2330–2338.

46. Ungefroren, H., M. Voss, M. Jansen, C. Roeder, D. Henne-Bruns, B. Kremer, andH. Kalthoff. 1998. Human pancreatic adenocarcinomas express Fas and Fas li-gand yet are resistant to Fas-mediated apoptosis. Cancer Res. 58: 1741–1749.

47. Sato, T., S. Irie, S. Kitada, and J. C. Reed. 1995. FAP-1: a protein tyrosinephosphatase that associates with Fas. Science 268: 411–415.

48. Yanagisawa, J., M. Takahashi, H. Kanki, H. Yano-Yanagisawa, T. Tazunoki,E. Sawa, T. Nishitoba, M. Kamishohara, E. Kobayashi, S. Kataoka, and T. Sato.1997. The molecular interaction of Fas and FAP-1. A tripeptide blocker of hu-man Fas interaction with FAP-1 promotes Fas-induced apoptosis. J. Biol. Chem.272: 8539–8545.

The Journal of Immunology 5961

by guest on May 10, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from