of April 10, 2019.This information is current as

Block EBV Infection In VitroSelectively Target B Cells through CD21 andReleased by EBV-Transformed B Cells Exosomes Containing Glycoprotein 350

GabrielssonGeorge, Eva Klein, Annika Scheynius and SusanneNagy, Mandira Paul, Qin Li, Sherree Friend, Thaddeus C. Helen Vallhov, Cindy Gutzeit, Sara M. Johansson, Noémi

ol.1001145http://www.jimmunol.org/content/early/2010/11/23/jimmun

published online 24 November 2010J Immunol 

MaterialSupplementary

5.DC1http://www.jimmunol.org/content/suppl/2010/11/24/jimmunol.100114

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

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

Print ISSN: 0022-1767 Online ISSN: 1550-6606. All rights reserved.1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Exosomes Containing Glycoprotein 350 Released byEBV-Transformed B Cells Selectively Target B Cellsthrough CD21 and Block EBV Infection In Vitro

Helen Vallhov,*,1 Cindy Gutzeit,*,1 Sara M. Johansson,* Noemi Nagy,† Mandira Paul,*

Qin Li,† Sherree Friend,‡ Thaddeus C. George,‡ Eva Klein,† Annika Scheynius,*

and Susanne Gabrielsson*

Exosomes are nano-sized membrane vesicles released from a wide variety of cells, formed in endosomes by inward budding of the

endosomal limiting membrane. They have immune stimulatory-, inhibitory-, or tolerance-inducing effects, depending on

their cellular origin, which is why they are investigated for use in vaccine and immune therapeutic strategies. In this study, we

explored whether exosomes of different origins and functions can selectively target different immune cells in human peripheral

blood. Flow cytometry, confocal laser scanning microscopy, and multispectral imaging flow cytometry (ImageStream) revealed that

exosomes derived from human monocyte-derived dendritic cells and breast milk preferably associated with monocytes. In contrast,

exosomes from an EBV-transformed B cell line (LCL1) preferentially targeted B cells. This was not observed for an EBV2B cell line

(BJAB). Electron microscopy, size-distribution analysis (NanoSight), and a cord blood transformation assay excluded the presence

of virions in our LCL1 exosome preparations. The interaction between LCL1-derived exosomes and peripheral blood B cells could

be blocked efficiently by anti-CD21 or anti-gp350, indicating an interaction between CD21 on B cells and the EBV glycoprotein

gp350 on exosomes. The targeting of LCL1-derived exosomes through gp350–CD21 interaction strongly inhibited EBV infection

in B cells isolated from umbilical cord blood, suggesting a protective role for exosomes in regulating EBV infection. Our finding

also suggests that exosome-based vaccines can be engineered for specific B cell targeting by inducing gp350 expression. The

Journal of Immunology, 2011, 186: 000–000.

Exosomes (30–100 nm in size) were first described aswaste products of maturing reticulocytes (1) and havesince been found to be released from a wide variety of

normal and malignant cells (2, 3). They have been isolated frombiological fluids, including plasma and human breast milk (4, 5).Exosomes are formed by inward budding of the endosomal-limiting membrane (2, 6) and access the extracellular spacewhen the endosome is emptied upon its fusion with the outer cellmembrane. The composition of exosomes varies depending ontheir cell type of origin and their activation state. Exosomes fromdendritic cells (DCs) express MHC class I and II molecules and

are enriched in tetraspanins (e.g., CD63 and CD81), costimulatorymolecules (2), and integrins and sphingolipids (7). They areconsidered key players in cell–cell communication because theycan transfer Ag/MHC complexes (8–10) between cells. It was alsoshown that exosomes are involved in cell-to-cell transfer ofmRNA and microRNA (11, 12), as well as infectious agents, in-cluding prions (13) and HIV (14). We previously reported thatmilk exosomes might be involved in tolerance induction (5),which was also observed for exosomes derived from intestinalepithelial cells (15). In contrast, B cell- (16, 17) and DC-derivedexosomes (18) have mainly been implicated in T cell stimulation.However, a suppressive effect was recently demonstrated forexosomes derived from EBV-transformed B cell lines (LCLs)carrying functional viral microRNAs that repress EBV-targetedgenes in noninfected bystander cells (12). The therapeutic po-tential of exosomes is being evaluated in clinical studies in whichtumor peptide-loaded DC exosomes are used in cancer immuno-therapy (19), and different types of exosomes are being in-vestigated in vaccine strategies against various pathogens (13).However, fundamental mechanistic questions remain to be eluci-dated to understand the role of exosomes in vivo and the fullpotential of exosomes in clinical treatment.In this study, we explored the nature of exosome–cell com-

munication by investigating whether exosomes of various cellularorigins with different functions target immune cells differently.Therefore, we compared the fate of tolerizing (breast milk), stim-ulatory (DCs), and repressive (LCL) exosomes with respect totheir binding properties to immune cells in peripheral blood. Ourresults revealed that exosomes derived from human DCs and hu-man breast milk preferably associated with monocytes, whereasexosomes from an LCL selectively targeted B cells. The molecule

*Clinical Allergy Research Unit, Department of Medicine Solna and †Department ofMicrobiology, Tumour and Cell Biology, Karolinska Institutet, Stockholm, Sweden;and ‡Amnis Corporation, Seattle, WA 98121

1H.V. and C.G. contributed equally to this paper.

Received for publication April 12, 2010. Accepted for publication October 25, 2010.

This work was supported by the Swedish Research Council Medicine, the SwedishCouncil for Working Life and Social Research, the Swedish Cancer and AllergyFund, the Swedish Cancer Society, the Center for Allergy Research, KarolinskaInstitutet, the Swedish Heart and Lung Foundation, the Swedish Asthma and AllergyAssociation’s Research Foundation, and the Hesselman, Ake Wiberg, and MagnusBergvall Foundations. N.N. is a recipient of a Cancer Research Fellowship of theCancer Research Institute (New York)/Concern Foundation (Los Angeles).

Address correspondence and reprint requests to Dr. Cindy Gutzeit, Clinical AllergyResearch Unit, Department of Medicine Solna, Karolinska University HospitalL2:04, 171 76 Stockholm, Sweden. E-mail address: cindy.gutzeit@ki.se

The online version of this article contains supplemental material.

Abbreviations used in this paper: CBMC, cord blood mononuclear cell; CLSM,confocal laser-scanning microscopy; DC, dendritic cell; EBNA2, EBV-encoded nu-clear Ag 2; LMP1, latent membrane protein 1; RT, room temperature; TEM, trans-mission electron microscopy.

Copyright� 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1001145

Published November 24, 2010, doi:10.4049/jimmunol.1001145 by guest on A

pril 10, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

responsible for the exosome binding to B cells was found to be theEBV-derived glycoprotein gp350. The gp350-bearing exosomesefficiently blocked EBV infection in vitro. These findings suggesta role for exosomes in controlling EBV infection, as well as thepossibility of developing exosome-based therapies in which B celltargeting is desired.

Materials and MethodsExosome sources

Buffy coats from healthy blood donors at the blood bank of KarolinskaUniversity Hospital were used for Ficoll-Paque (Amersham PharmaciaBiotech, Uppsala, Sweden) separation of PBMCs. Monocyte-derived DCswere generated as previously described (20), and their supernatants werecollected between days 6 and 8. The EBV-transformed LCL1 cells, derivedfrom a patient with birch pollen allergy, was a kind gift from Dr. BarbaraBohle (Medical University of Vienna, Vienna, Austria) (16). The EBV2

BJAB cell line (21) was a kind gift from Dr. Mikael Karlsson (KarolinskaInstitutet). All cell lines were cultured in complete medium consisting ofRPMI 1640 (Life Technologies, Invitrogen, Paisley, U.K.) supplementedwith 25 mg/ml gentamicin (Life Technologies), 10% heat-inactivated andexosome-depleted FCS (Hyclone, Logan, UT), 2 mmol/l L-glutamine, 100IU/ml penicillin (Life Technologies), and 100 mg/ml streptomycin (LifeTechnologies). LCL1 cells were only used for up to 10 passages after theywere thawed from a frozen stock. All cells were checked for mycoplasmainfection using the MycoAlert mycoplasma detection kit (Lonza, Rock-land, ME). Mature human breast milk was prepared as previously de-scribed (5) and stored at 280˚C within 24 h of sampling. The study wasapproved by the local ethics committee; for milk collection, informedwritten consent was obtained from the donors.

Isolation of exosomes

Exosomes were isolated from cell-culture supernatants (DC or B cell lines)or from human cell-free milk by differential centrifugation, as previouslydescribed (5, 22). All three kinds of pelleted exosomes were resuspendedimmediately after the last wash in a small volume of PBS. If aggregatesremained, solutions were filtered through a small-volume 0.2-mm syringefilter, before protein concentrations were determined using the Bio-Rad Dc

assay (Bio-Rad, Hercules, CA), according to the manufacturer’s instruc-tions. The same amount of protein was used from each of the four exosomesources.

Staining of exosomes

Exosomes were stained with the green fluorescent membrane dye PKH67(Sigma-Aldrich, St. Louis, MO). This was done by transferring exosomesfrom PBS to diluent C solution (Sigma-Aldrich) by centrifugation at100,000 3 g for 70 min. PKH67, diluted to 4 mM, and the exosomes (300mg/ml) were filtered through small 0.2-mm syringe filters before mixing at1:1 for 5 min, followed by the addition of 5% BSA and washing. Pre-liminary microscopic analysis showed that exosomes formed aggregatesafter staining (data not shown). To avoid these aggregates, the exosomeswere resuspended in cell-culture media and filtered through a small-volume 0.2-mm syringe filter immediately before addition to cells. Asa control, the same concentration of PKH67 was centrifuged in parallel tocreate a background control for possible pelleted unbound dye.

Coculture of PKH67-stained exosomes with PBMCs

Prefiltered PKH67-stained exosomes were added to PBMCs for 1 and 4 h at37˚C, as well as 4˚C during temperature studies. Ten micrograms of ex-osomes per 53 105 PBMCs was added. The PKH67 dye pellet centrifugedin parallel was used as a background control.

Transmission and immune electron microscopy

Exosomes, which were precaptured on anti-HLA class II magnetic beads(23), and B cells cocultured with and without exosomes for 4 h were an-alyzed by transmission electron microscopy (TEM) as previously de-scribed (5). Exosomes without beads, the 10,000 3 g pellet from LCL1supernatants, and centrifuged (16,000 3 g) supernatants from the EBV-producing B95-8 B cell line were processed by negative staining; an ali-quot of 3 ml was added to a cupper grid with a carbon-coated supportingfilm for 10 min. The excess solution was soaked off by a filter paper, andthe grid was stained with 2% uranyl acetate in water for 10 s, dipped indistilled water, and air-dried. Samples were examined using a Tecnai 10transmission electron microscope (Fei, Acht, The Netherlands) at 80 kV.

Digital images were captured by a Mega View III digital camera (SoftImaging System, Munster, Germany).

Flow-cytometry analysis

PBMCs were incubated with PKH67-stained exosomes and subsequentlystained using two mouse mAb panels: 1) “allophycocyanin” anti–HLA-DR(MHC class II) PECy5, anti–CD14-PE, and anti-CD19 Pacific Blue (orPE-Texas Red) and 2) “T cells” anti-CD8 allophycocyanin (or PECy5),anti-CD4 PE, and CD3 Pacific Blue (BD Biosciences, San Jose, CA). Thesamples were run on a FACS Aria flow cytometer and analyzed by FACSDiva (both from BD Biosciences). Gating was done first on lymphocyte/monocytes in forward light scatter/side scatter. For the allophycocyaninpanel, the HLA-DR+CD14+ and HLA-DR+CD142 populations were gatedfirst. Thereafter, the CD19+ cells were gated from the HLA-DR+CD142

cells. For the T cell panel, the populations were selected as CD3+CD4+

versus CD3+CD8+ directly out of the lymphocyte/monocyte gate.

Confocal laser-scanning microscopy

PBMCs were incubated with PKH67-stained exosomes, centrifuged ontoobject slides (ShandonCytospin3,BlockScientific, Bohemia,NY), andfixedin 4% formaldehyde for 15 min. Staining was carried out using mAbs toCD3, CD14, or CD19 (BD Biosciences), according to the manufacturer’sinstructions. A secondary goat anti-mouse mAb labeled with Alexa Fluor546 (Molecular Probes, Eugene, OR) was used for detection. After staining,slides were mounted with 90% glycerol. Fluorescent images were acquiredon a confocal laser-scanning microscope (TCS SP2; Leica Microsystems,Mannheim, Germany).

ImageStream analysis

Exosomes were stained and coincubated with PBMCs, as described for flowcytometry. Cells were run on the ImageStream multispectral imaging flowcytometer, and images were analyzed using IDEAS image-analysis software(Amnis). Ten thousand events were collected in each sample, and single-stained compensation controls were used to compensate fluorescence be-tween channel images on a pixel-by-pixel basis. Gating was done accordingto the principle for FACS. The cellular location of the PKH67 fluorescencewas measured using the Internalization feature. The Internalization featureis defined as a ratio of the intensity inside the cell/intensity of the entire cell.The higher the score, the greater the concentration of intensity insidethe cell. The inside of the cell is defined by an erosion of a mask that fits themembrane of the cell. The feature is invariant to cell size and can ac-commodate concentrated bright regions and small dim spots. The ratiois mapped to a log scale to increase the dynamic range to values between{2inf, inf}. Cells with primarily internal fluorescence have positive scores,whereas cells showing little internalization have negative scores.

Blocking assays

PBMCs (106/ml) were coincubated with mAbs to CD18 (clone MEM48),CD21 (clone B-E5), or isotype-matched Abs (20 mg/ml; Nordic BioSite,Taby, Sweden) for 30 min in cell-culture media at room temperature (RT);they were washed with PBS before exosomes were added at 10 mg/53 105

cells at 37˚C for 1 or 4 h. FACS analysis was performed as previouslydescribed. LCL1 exosomes were treated with anti-CD23 (clone 9P25, 30mg/ml; Beckman Coulter, Bromma, Sweden) or with the supernatant (30%of total volume) from a mouse hybridoma culture producing the gp350/250-neutralizing mAb 72A1 (DSMZ, Braunschweig, Germany). An irrel-evant isotype control or hybridoma supernatant was used as control. After30 min in cell-culture media at RT, pretreated LCL1 exosomes were addedto B cells (10 mg/2.5 3 105 cells) for 4 h, before performing FACSanalysis. B cells had been isolated from PBMCs using a B cell isolation kitII (Miltenyi Biotec/Fisher Scientific, Goteborg, Sweden).

Intracellular flow-cytometry staining

LCL1 and BJAB cells (5 3 104) were fixed with 4% formaldehyde for 5min at RT. After three washes with PBS, cells were incubated for 10 min in1% saponin (Sigma-Aldrich, Stockholm, Sweden) solution at RT. Cellswere stained with the primary mAb 72A1 against gp350/250 by adding thesupernatant from the mouse hybridoma culture at RT for 1 h. Afterwashing twice with 0.1% saponin solution, cells were incubated with thesecondary Alexa Fluor 488 (Invitrogen, Taastrup, Denmark) Ab for 45 minat RT, washed, and analyzed by flow cytometry.

Sucrose gradient

Fractions of exosome preparations were collected by sucrose gradient aspreviously described (5). Fractions were directly loaded onto anti-MHC

2 LCL EXOSOMES TARGET B CELLS VIA CD21

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

class II Dynabeads (Dynal, Oslo, Norway) for flow-cytometry analysis orcentrifuged at 200,000 3 g for 35 min at 4˚C for immunoblot analysis.

Immunoblot analysis

Each pelleted exosome fraction was separated by SDS-PAGE (12%) andtransferred to polyvinylidene difluoride membranes (Millipore, Solna,Sweden). Membranes were stained with mAbs to latent membrane protein 1(LMP1) (clone CS. 1.4; DakoCytomation, Stockholm, Sweden), gp350(clone 2L10; Millipore), CD81 (clone H-121; Santa Cruz Biotechnology,Heidelberg, Germany), or HLA-DR (clone TAL.1B5; DakoCytomation),according to the manufacturer’s instructions. Membranes were visualizedwith ECL Advance Western blotting Detection kit and exposed on Hyper-films (GE Healthcare, Uppsala, Sweden).

Cord blood transformation assay

Heparinized cord blood samples, obtained from the Karolinska UniversityHospital and approved by the local ethics committee, were subjected toFicoll-Paque density centrifugation. One million cord blood mono-nuclear cells (CBMCs) were incubated with the 10,000 3 g pellet fromBJAB or LCL1 supernatants or with 24 mg BJAB or LCL1 exosome prep-arations for 1.5 h in a humidified 37˚C, 5% CO2 incubator. CBMCs werewashed and resuspended in complete RPMI 1640 at 106 cells/ml and seededin quintuple, at 23 105 cells per well/200ml, in 96-well plates. As a positivecontrol for EBV-induced cell transformation, CBMCs were exposed toEBV (B95-8 virus)-containing supernatant. Culture medium served asnegative control. CBMCs were fed weekly with fresh medium. On day 33,the transformation was registered visually by the appearance of typical cellaggregates and by thymidine-incorporation assay. One microCurie of[3H]thymidine (GE Healthcare, Uppsala, Sweden) was added to the cul-tures and incubated for 16 h. CBMCs were harvested onto glass fiber filters,and radioactivity was measured in a scintillation counter (1205 Beta-plate, PerkinElmer, Upplands Vasby, Sweden).

NanoSight

Size distribution within exosome preparations was analyzed by measuringthe rate of Brownian motion using a NanoSight LM10 system, which isequipped with a fast video capture and particle-tracking software (Nano-Sight, Amesbury, U.K.) (24). EBV (strain B95-8) was used as a control.Prior to analysis with NanoSight, EBV was heat inactivated for 20 min at56˚C.

EBV-infection assay/EBV-encoded nuclear Ag 2-inductionassay

B cells were isolated from human umbilical cord blood (approved by thelocal ethics committee) using a B cell isolation kit II (Miltenyi Biotec). Atotal of 5 3 105 B cells were left untreated or were preincubated for 2 h at37˚C with 100 mg BJAB or LCL1 exosomes. Thereafter, cells were infectedwith EBV by incubating them for 1 h at 37˚C with B95-8 virus-containingsupernatant. Cells were washed once with medium and centrifuged at3003 g for 10 min. B cells were seeded onto a 48-well plate and cultured in1 ml complete medium. Three days postinfection, cells were harvested andwashed once in PBS, and the cell pellet was stored at 220˚C. Cell pelletswere lysed in loading buffer (16% glycerol, 100 mM Tris-HCl [pH 6.8], 3%SDS, 0.1% bromophenol blue, 4% 2-ME) and incubated for 10 min at100˚C, and proteins were separated through SDS-PAGE and analyzed byimmunoblot for the expression of EBV-encoded nuclear Ag 2 (EBNA2;clone PE2, 1/5,000; Novocastra Laboratories, Newcastle, U.K.) and b-actin(clone AC-15, 1/400,000; Sigma-Aldrich). Polyclonal goat anti-mouseHRP (1/2000; DakoCytomation) was used as secondary Ab. Membraneswere visualized with ECL Plus solution (GE Healthcare) and exposed onHyperfilms (GE Healthcare).

Statistical analysis

The Wilcoxon matched-pairs test was applied using GraphPad Prismsoftware, version 4.03 (Graphpad Software, San Diego, CA; http://www.graphpad.com); p values ,0.05 were considered significant.

ResultsExosomes derived from human DC and breast milk targetmonocytes, whereas LCL1 exosomes prefer B cells

To elucidate the binding of exosomes to immune cells, we comparedexosomes isolated from human monocyte-derived DCs, from anEBV-transformed lymphoblastoid B cell line (LCL1), and from

human breast milk with regard to their adherence to PBMCs.Exosomes were stained with a green fluorescent membrane dye(PKH67) to be detectable by flow cytometry. To determine whetherexosomes were properly labeled and whether their structures wereretained after staining with PKH67, the exosomes were bound tomagnetic anti-MHC class II beads (23). Flow-cytometry analysisshowed that green fluorescent MHC class II-containing vesicleshad been captured by the beads (Supplemental Fig. 1A), and TEMdisplayed nanovesicles with intact lipid bilayers, indicating that thePKH67 labeling did not interfere with exosome morphology(Supplemental Fig. 1B). The different exosomes were coincubatedwith PBMCs for 1 or 4 h and analyzed bymulticolor flow cytometryto evaluate the association pattern of exosomes with cells. After 1 h,DC-derived exosomes mainly interacted with monocytes (HLA-DR+CD14+; average 46%), whereas only 17% bound to B cells(HLA-DR+CD142CD19+; Fig. 1A). In contrast, LCL1 exosomesshowed the reverse pattern of association, with a strong preferencefor B cells. Sixty-three percent of the B cells were positive for LCL1exosomes, whereas, on average, only 17% of the monocytes hadassociated with these exosomes at 1 h (Fig. 1B). After 4 h of coin-cubation there was a general increase in the percentages of exo-some+ cells within each cell population, but the distinct associationpatterns with the different cell populations remained (Fig. 1D–F).Milk exosome interactions were generally low after 1 h of coincu-bation (Fig. 1C). However, at 4 h, 55% of themonocytes and 18% ofthe B cells had associatedwithmilk exosomes (Fig. 1C), resemblingthe pattern for DC exosomes at 1 h. Previously, we found that milkexosome preparations have a lower content of exosomal vesicles inrelation to the total protein amount in the exosome pellet comparedwith pellets from other exosome types (5). Accordingly, when in-creasing the amount of milk exosomes 5-fold (50 mg/5 3 105

PBMCs), the level of milk exosome interaction at 1 h reached anaverage of 64% with monocytes and 26% with B cells (n = 3; datanot shown). These levels are comparable to those seen at 1 h whenincubating with DC exosomes at 10 mg/5 3 105 PBMCs, sugges-ting higher amounts of nonexosomal proteins in the milk exosomepreparations compared with the other exosome preparations.Although T cells constitute the majority of PBMCs (∼70% in

this study),,8% of CD4+ or CD8+ T cells showed association withany of the exosomes after 4 h (Fig. 1). No consistent differences inthe preferences for CD4+ compared with CD8+ T cells with thedifferent exosome types were observed.

Exosomes are mainly associated with the cell membrane ofB cells but are internalized by monocytes

Next, we askedwhere the exosomes localizewithin the different celltypes. We coincubated PKH67-labeled exosomes with PBMCs andanalyzed exosome association by confocal laser-scanning micros-copy (CLSM). At 1 h, in general, no or only weak exosome signalscould be detected in association with cells (data not shown),probably as a result of the lower detection level with CLSM com-pared with flow cytometry. After 4 h, DC exosomes were mainlyinternalized by monocytes (CD14+) and, to a lesser degree, byB cells (CD19+), which often had cell membrane-associated exo-somes (Fig. 2A). In contrast, LCL1 exosomes interacted to a higherdegreewith B cells and were mainly localized to the cell membrane(Fig. 2B). Monocytes showed weaker signals for LCL1 B cellexosomes (Fig. 2B). In general, milk exosomes showed weak sig-nals, as seen in flow cytometry, and were detected as internalized inmonocytes or associated with the cell membrane of B cells (Fig.2C). In the few cases in which an interaction between exosomes andCD3+ T cells occurred (∼1 of 50 cells), the exosomes were mainlylocalized near or in contact with the cell membrane (data notshown). Thus, these CLSM data are consistent with our flow-

The Journal of Immunology 3

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

cytometry data, showing that exosomes from EBV-transformedB cells preferentially target B cells, whereas DC-derived exo-somes associate more with monocytes. However, these results alsosuggest that exosomes interact differently with different cell types;they are mostly associated with the cell membrane of B cells andT cells but are internalized by monocytes.To verify the association and localization of different exosomes

with immune cells by another technical approach, we decided toexplore the ImageStream system. This system is a combined flowcytometer and fluorescence microscope that automatically capturesmultispectral images of each cell that passes through the flow cell atvery high rates, enabling image-based analysis of large numbersof cells per sample. Using the internalization feature, we distin-guished monocytes (HLA-DR+CD14+) that had surface-associatedexosomes, internalized exosomes, or both (intermediate; Fig. 2D).For technical reasons, no B cell data were obtained, but themonocyte data received were consistent with our conventionalflow-cytometry and CLSM data. The ImageStream showed thatDC and milk exosomes preferentially associated with monocytes,

whereas LCL1 exosomes did not (Fig. 2E). Furthermore, themajority of the monocytes had internalized the various exosomes,reinforcing our previous findings (Fig. 2E). In this study, we alsoshowed that this method is feasible for quantifying exosome lo-calization in large numbers of cells.

The binding between LCL1-derived exosomes and B cells istemperature independent

To investigate whether the association of different exosomes withPBMCs is receptor mediated or dependent on active internaliza-tion, we cocultured PKH67+ exosomes with PBMCs at 4 and 37˚C.For all three exosome types, the association with monocytes de-creased when incubation was performed at 4˚C for 1 and 4 h (Fig.3), indicative of an active, probably phagocytic, uptake by this celltype. DC (Fig. 3A, 3D) andmilk exosome (Fig. 3C, 3F) associationswith B cells were similarly diminished during cold conditions. Incontrast, only a slight decrease in interaction between LCL1 exo-somes and B cells was seen at 4˚C at both time points (Fig. 3B, 3E).Thus, the interaction between LCL1 exosomes and B cells waslargely independent of temperature; therefore, it is more likelymediated through adhesionmolecules or surface receptors, which iswhy we set out to dissect this interaction further.

The binding between LCL1-derived exosomes and B cells isdependent on CD21 expressed on the B cells

B cells express cell-surface receptors, such as the human com-plement receptor 2 (CD21), which, together with a distinct patternof adhesion molecules [e.g., ICAM-1 or integrins, such as LFA-1(CD11a/CD18), Mac-1 (CD11b/CD18), or p150,95 (CD11c/CD18)(17)], may mediate the observed strong B cell preference of B cellexosomes. To reveal the specificity of the exosome binding, mAbsagainst CD21 or CD18 were added to PBMCs before incubationwith exosomes. The interaction between LCL1 exosomes andperipheral blood B cells was blocked efficiently by anti-CD21,whereas no blocking effect was observed with anti-CD18 Abs(Fig. 4A).To elucidatewhether the B cell targeting observedwas dependent

on the EBV transformation of the B cells, we compared it withexosomes from the EBV2 B cell line BJAB (22). Results showedthat BJAB exosomes bound to B cells 10 times less than LCL1exosomes, suggesting the involvement of EBV-derived or -inducedproteins in the binding (Fig. 4B). Taken together, these data indicatethat the binding between LCL1 exosomes and B cells is dependenton an interaction with CD21 and not on LFA-1, Mac-1, or p150,95,and suggest a selective B cell exosome targeting, which was spe-cific for the exosomes derived from the EBV-transformed B cells.

gp350 on LCL1 exosomes mediates the binding to B cells

Next, we aimed to elucidate the exosomal ligand involved in theCD21 binding on B cells. Known ligands to CD21 include the low-affinity receptor for IgE (CD23), the EBV envelope glycoproteingp350, and the complement factor C3d (25). It was shown thatexosomes released from B cells and macrophages contain C3fragments (26), but because heat-inactivated FCS was used in allof our cell cultures, C3d is unlikely to make a difference. CD23,which is highly expressed on EBV-transformed B cell lines (27),was detected on our LCL1 exosomes but not on BJAB exosomes(Supplemental Fig. 2). The EBV glycoprotein gp350 is a structuralprotein that is critical for viral attachment to B cells (28), but it hasnot been shown to be present on exosomes. Hence, gp350 andCD23 were the first candidates to be investigated for their possibleinvolvement in the exosome–B cell interaction. These ligandswere blocked by anti-gp350/220 (29) or anti-CD23 Abs (25). The

FIGURE 1. Exosomes target specific cell populations in PBMCs. PKH67-

stained DC (A, D), LCL1 (B, E), and breast milk (C, F) exosomes were

incubated with PBMCs from healthy blood donors for 1 h (A–C, n = 9) or

4 h (D–F, n = 8). Ten micrograms of exosomes per 5 3 105 PBMCs was

added. As a background control, the PKH67 dye pellet centrifuged in

parallel was used, which showed low to undetectable fluorescence (data

not shown). Association of exosomes was measured by collection of 104

events per sample in four-color flow cytometry. Data are expressed as the

percentage of PKH67+ cells in each subpopulation. Horizontal lines rep-

resent mean values. Different blood donors are indicated by individual

symbols.

4 LCL EXOSOMES TARGET B CELLS VIA CD21

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

binding of LCL1 exosomes to B cells was substantially reducedwhen gp350 was blocked; interestingly, however, no reduction inexosome binding was seen when CD23 was blocked (Fig. 4C).This observation suggests the presence of gp350 on the surface ofLCL1 exosomes and that this EBV glycoprotein, and not CD23,mediates the exosome binding. As a control, we also addeda nonneutralizing anti-gp350 mAb (2L10) (29), which did notblock exosome binding (data not shown), reinforcing the notion ofa specific blocking of gp350 via the neutralizing mAb (72A1).Herpesvirus-infected cells express viral glycoproteins on their

cell surface during the lytic phase of the viral life cycle. There-fore, we investigated the presence of gp350 on LCL1 cells by flowcytometry (Fig. 4D). Interestingly, we observed that a high pro-portion of LCL1 cells expressed gp350 on the cell surface, as wellas intracellularly. To verify the presence of gp350 on LCL1 exo-somes, exosomes were purified through sucrose gradient, andfractions were analyzed by immunoblotting (Fig. 4E). Raposoet al. (22) demonstrated that LCL-derived exosomes float in a su-crose gradient at a density of 1.08–1.22 g/ml. Immunoblot anal-ysis for exosome-enriched HLA-DR and CD81 within a density of1.10–1.19 g/ml for LCL1 and BJAB exosome preparations verifiedthe presence of exosomes. LMP1 was found on exosomes secretedby LCLs (30), and we also demonstrated its presence on the LCL1exosomes. Intriguingly, gp350 was also detected on LCL1 exo-somes. However, gp350 expression on LCL1 exosomes did notfully overlap with LMP1 expression, suggesting a heterogenousexosome population. As expected, neither gp350 nor LMP1 wasdetected on BJAB exosomes (Fig. 4E).

Exosome signals are derived from exosomes and not virions

The expression of gp350 on LCL1 cells and exosomes raises thequestion of whether some of our LCL1 cells are in a lytic stage of

the EBV life cycle, which would suggest that EBV may also reside

in the EBV-transformed B cell cultures as free virus particles.

Therefore, there is a possibility that virions might contaminate our

exosome preparations and, thereby, give a false-positive signal for

exosomes in our experiments. In preliminary experiments, we

detected viral DNA by PCR; however, the presence of DNA does

not always correspond to the presence of complete virions. Hence,

the more sensitive cord blood-transformation assay was performed

to investigate the presence of infectious EBV, as well as TEM

analysis to detect EBV particles in our exosome preparations. The

10,0003 g pellets (obtained during an intermediate centrifugation

step during the exosome-isolation procedure), as well as exosome

preparations from BJAB and LCL1 supernatants, were added to

CBMCs to monitor the outgrowth of EBV-transformed primary

B cells (Fig. 5A). Infection with the B95-8 virus served as a pos-

itive control. The addition of pooled 10,000 3 g pellets from

BJAB supernatants did not induce the outgrowth of LCLs, as

quantified by [3H]thymidine incorporation. In addition, visual ex-

amination of the cultures did not reveal any typical aggregates of

transformed B cells. In contrast, the addition of pooled 10,0003 g

pellets from LCL1 supernatants to the CBMCs induced the out-

growth of LCLs, indicating the production of infectious virus

particles by LCL1 cells. This finding is in line with the ob-

served cell-surface expression of gp350 on LCL1 cells (Fig. 4D),

FIGURE 2. Exosomes are mainly internalized by monocytes, whereas they associate with the cell membrane of B cells. PBMCs were incubated for 4 h

with PKH67-stained DC (A), LCL1 (B), and breast milk (C) exosomes (10 mg of exosomes per 5 3 105 PBMCs) and then stained for anti-CD14 or anti-

CD19, followed by Alexa Fluor 546 labeling (red). The cells were analyzed by CLSM. As a background control, the PKH67 dye pellet centrifuged in

parallel was used, which showed undetectable fluorescence (data not shown). Scale bars, 10 mm. Images of PBMCs are shown as one representative

experiment of two. D, ImageStream analysis shows that HLA-DR+CD14+ cells mainly internalize exosomes. As an example, a histogram for PKH67+

breast milk exosome (10 mg of exosomes per 53 105 PBMCs) interaction with HLA-DR+CD14+ cells after 4 h is plotted (upper panel). Composite images

of cells with surface (,0) or internal (.0) exosomes are shown (lower panels): HLA-DR (pink), CD14 (orange), and PKH67+ exosomes (green). Scale bar,

10 mm. Results are shown as one representative experiment of two using different blood donors. At least 104 total events were collected for each sample. E,

Percentages of exosomes associated with the HLA-DR+CD14+ cells, as well as the percentages of those cells with internalized exosomes, which were

calculated by the Internalization feature, are shown. Two thousand events were analyzed with the Internalization feature.

The Journal of Immunology 5

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

indicating that LCL1 cells are in a lytic, virus-producing stage.However, no outgrowth of LCLs was observed after the additionof exosome preparations (100,000 3 g) from BJAB and LCL1supernatants to CBMCs, indicating the absence of infectious EBVwithin our LCL1 exosome preparations. Furthermore, we per-formed TEM analysis of the 10,000 3 g pellet (Fig. 5B) and the100,000 3 g exosome pellet (Fig. 5C) from LCL1 supernatantsand compared the morphology of the observed particles andvesicles with TEM photographs of virions from supernatants ofB95-8 cells (Fig. 5D). TEM analysis of the 10,000 3 g pelletconfirmed the presence of viral particles similar to those observedin B95-8 supernatants and as demonstrated for another virus-producing cell line (31). In contrast, no viral particles weredetected in the 100,000 3 g exosome pellet from LCL1 super-natants (Fig. 5C). In addition, we compared the size distribution ofexosome preparations from BJAB and LCL1 cells with EBV-containing B95-8 supernatants by nanoparticle tracking analysis(NanoSight). This method quantitatively confirmed the lower sizerange within the LCL1 and BJAB exosome preparations (average∼100 nm) compared with EBV, which had a size range of 150–200 nm (Fig. 5E), further supporting the absence of virions in ourexosome preparations. Finally, we confirmed the presence of

FIGURE 3. The binding between LCL1-derived exosomes and B cells is

independent of temperature. PBMCs were incubated with PKH67-labeled

exosomes from DCs (A, D), EBV-transformed B cells (LCL1) (B, E), or

breast milk (C, F) for 1 or 4 h at 4 or 37˚C and analyzed as described in

Fig. 1. Data are percentages of exosome+ cells within each subpopulation

of HLA-DR+ cells measured as PKH67+ by flow cytometry. Mean 6 SD

for PBMCs from three donors are shown.

FIGURE 4. The interaction between B cells and exosomes derived from

EBV-transformed B cells is mediated by interactions between CD21 and

gp350. PBMC cultures were treated with anti-CD21, anti-CD18, or iso-

type-matched controls (20 mg/ml), before LCL1 exosomes (A) or BJAB

exosomes (B) were added for 4 h. C, LCL1 exosomes were preincubated

with anti-gp350 supernatant from the 72A1 mouse hybridoma (30% of

total volume), anti-CD23 (30 mg/ml), or isotype-matched controls and then

added to purified B cells for 4 h. Representative flow-cytometry dot plots

(right panels), including percentage numbers, show LCL1 exosome asso-

ciation with B cells treated with isotype control (upper panel) or anti-

gp350 mAbs (lower panel). Analysis was performed as described in Fig. 1.

Mean 6 SD for PBMCs from three donors are shown (A–C). D, Flow-

cytometry graphs show extracellular (left panel) and intracellular (right

panel) expression of gp350, comparing BJAB cells with LCL1 cells. Ten

thousand cells were analyzed. E, Pellets of sucrose gradient fractions from

LCL1 (left panel) and BJAB (right panel) exosome preparations were

analyzed by immunoblot using Abs against LMP1, gp350, CD81, and

HLA-DR. The density of each fraction was determined by refraction index

measurements. One representative experiment of three is shown.

6 LCL EXOSOMES TARGET B CELLS VIA CD21

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

exosomes by immunogold labeling for CD63 and HLA-DR (Fig.5F). Exosomes attached to primary B cells (Fig. 5G) or exosomepreparations alone (Fig. 5C, 5F) did not reveal any EBV particlesin the size of infectious EBV (200 nm) or naked high-density viruscapsids.

gp350-containing LCL1 exosomes reduce EBV infection ofprimary B cells

To determine the functional consequences of gp350 exosomebinding to B cells, we investigated whether LCL1 exosome bindinghas an impact on EBV infection. Therefore, B cells isolated fromumbilical cord blood were preincubated with BJAB or LCL1exosomes or left untreated prior to EBVexposure. Successful EBV

infection was monitored after 3 d by immunoblotting for expressionof latent protein EBNA2. Intriguingly, preincubation of B cells withLCL1 exosomes strongly reduced subsequent EBV infection, be-cause reduced EBNA2 protein levels were observed (Fig. 6). Thisstrong inhibitory effect was not observed when B cells werepreincubated with control BJAB exosomes; only a slight reductionin EBNA2 expression was seen.

DiscussionThis study showed that exosomes from a LCL line that expressesthe viral structural protein gp350 on the surface are capable ofspecifically targeting B cells. We also demonstrated that theseexosomes can block EBV infection in vitro. This suggests a rolefor exosomes in the regulation of EBV infection.

FIGURE 5. Exosome signals are derived from exosomes and not virions. A, To investigate the presence of infectious EBV in exosome preparations, the

10,0003 g pellets, as well as exosome preparations (100,0003 g) from BJAB and LCL1 supernatants, were added to CBMCs, and the outgrowth of LCLs

was monitored by [3H]thymidine incorporation after 33 d. Supernatant from EBV-producing B95-8 cells served as a positive control, and cell culture

medium alone served as a negative control. The 10,000 3 g pellets from three BJAB and LCL1 supernatant preparations were tested in one CBMC donor.

Five BJAB and LCL1 exosome preparations were tested in five CBMC donors, which are shown as mean values (right panel). TEM images from the

10,000 3 g pellet (B), the 100,000 3 g exosome pellet of LCL1 preparations (C), and EBV (B95-8 virus) (D), which were processed by negative staining.

Arrows indicate exosomes. E, NanoSight measurement of particle-size distribution in preparations from LCL1 exosomes, BJAB exosomes, and EBV. Data

are mean values (n = 3) and were normalized to 1 for size comparison. F, In immune electron microscopy, mAbs against CD63 and HLA-DR were added to

LCL1 exosome preparations and detected by gold-conjugated secondary Abs, 10 and 15 nm, respectively. Arrow 1 indicates CD63+ exosomes, and arrow 2

indicates HLA-DR positive exosomes. G, A representative image of exosomes associated with a B cell surface. B–D, Original magnification 387,000; F,

original magnification 375,000; G, original magnification 360,000.

The Journal of Immunology 7

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

It is known that different cell types produce exosomes withphenotypes that mainly reflect their cells of origin (32, 33). In thisstudy, we looked at the other side of the exosome communicationpathway and demonstrated that the exosomes tested, frommonocyte-derived DCs and breast milk, seem to have the samemain target in human peripheral blood (i.e., monocytes). Reachingthe monocytes, the exosomes seem to be actively engulfed,probably by phagocytosis, which was also recently demonstratedto be the mechanism of exosome uptake by other phagocytes, suchas macrophages (34). However, this paper also showed that theselectivity of exosomes to target monocytes might change if theexosome-producing cell carries pathogen-specific molecules,which we demonstrated in this study by comparing exosomesfrom an EBV2 (BJAB) cell line with an EBV+ B cell line (LCL1).The LCL1-derived exosomes mainly targeted B cells; however,this was not seen for BJAB exosomes. The interaction betweenLCL1 exosomes and B cells was efficiently blocked by Abs toCD21 on B cells or Abs to gp350 on exosomes but not by anti-CD18 or anti-CD23, demonstrating the involvement of EBV. Thereason why anti-CD23 had no effect on the cell–exosome in-teraction, although CD23 is more abundant on LCL1 comparedwith BJAB (Supplemental Fig. 2), might be due to a lower affinityinteraction between CD23 and CD21 compared with the gp350–CD21 interaction. This indicates that a very high molecular af-finity (35) is needed for rapid (1 h) binding of exosomes in vitroand might be even more important in vivo.The interaction of the various exosomes with T cells was rather

low (,8% of T cells positive for exosomes), although DC andB cell exosomes are known to display the ICAM-1 that binds toLFA-1 expressed on, for example, T cells (17, 36). However, it wasrecently reported that activated T cells with an upregulated ex-pression of LFA-1 attract DC exosomes to a higher degree com-pared with inactive T cells (37). Therefore, our low degree ofexosome–T cell interaction (Fig. 1) might be explained by the factthat we used ex vivo PBMCs, likely to contain mainly non-stimulated T cells. The activation state of the exosome-reci-pient cell, as well as the maturation status of the cell producing theexosomes, is important. Our preliminary data showed that exo-somes from LPS-activated DCs attached to monocytes to an evenhigher extent compared with the exosomes from immature DCs(data not shown).The observation that gp350 was present on the LCL1 exosomes

raised the question whether we had infectious and/or pleiomorphicEBV particles in our exosome preparations, which could be re-sponsible for the PKH67 signal seen on B cells. The binding of EBV

to CD21 is well established (27). By using the sensitive cord blood-transformation assay, as well as TEM analysis (Fig. 5A–D), we didnot find any evidence for virions in our exosome preparations.Thus, exosome preparations were cleared of virions by centrifu-gation; therefore, we consider it unlikely that virions were re-sponsible for the fluorescence signal seen on B cells.It should be mentioned that the LCL1 used in this study does not

represent a typical LCL, because their cells have a regular roundshape, and it produces virions. However, this is an advantage whenthe aim is to study exosomes produced during the lytic phase ofEBV infection. Further investigations are warranted to virologi-cally characterize and compare the LCL1 cell line with other LCLswith respect to the expression of viral and cellular gene products,which was already shown to vary in different LCLs (38).The finding that binding of gp350 containing LCL1 exosomes to

B cells inhibited subsequent EBV infection (Fig. 6) suggests thatexosomes play a role during the early phases of EBV infection.Thus, lytically infected B cells secrete gp350-containing exosomesthat bind to CD21 on bystander B cells, thereby protecting themfrom subsequent EBV infection (39). This innate-like immunemechanism might control EBV infection, in addition to the pow-erful adaptive immune responses. We also observed a slight in-hibition of EBV infection when B cells were preincubated withBJAB exosomes. A possible explanation is that these exosomesunspecifically bind to added EBV particles and, thereby, stericallyhinder them from infecting B cells.In the current study, we also observed that the EBV-transformed

B cells seemed to produce more exosomes, measured as proteinconcentration, compared with the EBV2 B cells (data not shown).This may also mirror the situation in vivo, in which a high numberof exosomes may contribute to control the spread of EBV in-fection. Alternatively, the induction of exosome production andthe exosomal expression of gp350 may contribute to the immu-nomodulatory potential of EBV. This was recently suggested forHSV-1, for which glycoprotein B is able to manipulate the classII-processing pathway, whereby the protein hijacks HLA-DR fromits normal transport route and directs it into the exosomal pathway(40). The recent finding of viral microRNA in LCL exosomes mayalso indicate that exosomes are used to deliver viral messagesbetween cells (12). Further investigations are needed to under-stand the extent to which exosomes are involved in the regulationof EBV infection or spreading.Our findings also suggest how exosomes can be engineered

(e.g., by inducing the expression of gp350) to redirect their cellulartargeting to B cells, which may potentiate their therapeutic use-fulness. A role for B cells in producing a complete T cell responsewas suggested in the 1980s (41). Later, Ding et al. (42) showedthat targeting of Ags to B cells can potentiate specific T cellresponses and break immune tolerance. Furthermore, B cells areparticularly important in achieving long-term T cell immunity(43) and, recently, we showed that exosomes require the supportof activated B cells for generating Ag-specific T cell responsesin vivo (44). Hence, by targeting B cells in cancer vaccines,tolerance could be broken, and a more long-lasting T cell im-munity might be achieved. Furthermore, CD21 is expressed byB cells, as well as by follicular DCs (45). This implies thatexosomes with surface-associated gp350 may also target follicularDCs in vivo, thereby enhancing a possible immune activation andmemory in vivo.The direct use of exosomes from EBV-transformed B cells is

a possibility; however this approach could have limitations ina clinical setting because of the presence of LMP1. LMP1 is knownto inhibit T cell proliferation (30) and was reported to have onco-genic properties (46). Still, several studies showed the potential of

FIGURE 6. gp350-containing LCL1 exosomes strongly reduce EBV

infection of primary B cells. Umbilical cord blood B cells were pre-

incubated for 2 h with BJAB or LCL1 exosomes or were left untreated (2)

prior to EBVexposure. Three days postinfection, B cells were analyzed by

immunoblot for EBNA2 and b-actin expression. Phase-contrast micros-

copy images of B cells preincubated with BJAB exosomes, LCL1 exo-

somes, or left untreated prior to subsequent EBV infection, and

noninfected B cells (from left to the right). One representative experiment

of three is shown. Original magnification 35. Scale bar, 100 mm.

8 LCL EXOSOMES TARGET B CELLS VIA CD21

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

exosomes, derived from EBV-transformed B cells, to activateT cells (16, 22). In addition, the relative expression of LMP1 andgp350 on exosomes may differ during the latent and lytic phase;therefore, the exosome cultures may be fine-tuned for optimalclinical applications. An alternative strategy would be to modifyDC exosomes, which have been implicated in clinical studies (19),to express gp350.In conclusion, we showed that exosomes, found in breast milk

and produced by humanmonocyte-derivedDCs and anEBV2Bcellline, do not preferably associate with B cells. Instead, they mainlytarget monocytes, which actively engulf exosomes, as demon-strated for milk and DC exosomes. However, if B cells harborEBV in their lytic stage, the produced exosomes change theirpreference from monocytes toward B cells, whereby exosome-associated gp350 binds to the EBV entry receptor CD21 onB cells. The specific binding of these exosomes to B cells stronglyreduced EBV infection and suggests a protective role for exosomesduring EBV infection. Furthermore, exosomes targeting B cellsmight be efficient in inducing long-term cellular and humoral im-mune responses; hence, they should be considered potential toolsin the treatment of cancer and inflammatory diseases.

AcknowledgmentsWe thank Prof. Ingemar Ernberg, Department of Microbiology, Tumour and

Cell Biology, and Dr. Mikael Karlsson, Department ofMedicine Solna, Kar-

olinska Institutet, for helpful discussions. We also thank Dr. Kjell Hultenby

and staff at the Electron Microscopy Unit, Karolinska University Hospital,

Huddinge, Sweden, and Sven Petersen, School of Cancer Sciences, Univer-

sity of Birmingham, Birmingham, U.K., for excellent support with TEM.

Prof. Peter Biberfeld, Department of Pathology and Tumor Biology, Kar-

olinska Institutet, shared EBV TEM expertise. We also thank Dr. Andrew

Malloy for analyzing the exosome preparations with NanoSight.

DisclosuresS.F. and T.C.G. are employees of Amnis Corporation, which manufactures

ImageStream. S.G. has a patent pending on exosome-based treatment of

cancer.

References1. Johnstone, R.M., M. Adam, J. R. Hammond, L. Orr, and C. Turbide. 1987. Vesicle

formation during reticulocyte maturation. Association of plasma membrane ac-tivities with released vesicles (exosomes). J. Biol. Chem. 262: 9412–9420.

2. Thery, C., L. Zitvogel, and S. Amigorena. 2002. Exosomes: composition, bio-genesis and function. Nat. Rev. Immunol. 2: 569–579.

3. McLellan, A. D. 2009. Exosome release by primary B cells. Crit. Rev. Immunol.29: 203–217.

4. Caby, M. P., D. Lankar, C. Vincendeau-Scherrer, G. Raposo, and C. Bonnerot.2005. Exosomal-like vesicles are present in human blood plasma. Int. Immunol.17: 879–887.

5. Admyre, C., S. M. Johansson, K. R. Qazi, J. J. Filen, R. Lahesmaa, M. Norman,E. P. Neve, A. Scheynius, and S. Gabrielsson. 2007. Exosomes with immunemodulatory features are present in human breast milk. J. Immunol. 179: 1969–1978.

6. Stoorvogel, W., M. J. Kleijmeer, H. J. Geuze, and G. Raposo. 2002. The bio-genesis and functions of exosomes. Traffic 3: 321–330.

7. Laulagnier, K., C. Motta, S. Hamdi, S. Roy, F. Fauvelle, J. F. Pageaux,T. Kobayashi, J. P. Salles, B. Perret, C. Bonnerot, and M. Record. 2004.Mast cell- and dendritic cell-derived exosomes display a specific lipid compo-sition and an unusual membrane organization. Biochem. J. 380: 161–171.

8. Bedford, P., K. Garner, and S. C. Knight. 1999. MHC class II molecules trans-ferred between allogeneic dendritic cells stimulate primary mixed leukocytereactions. Int. Immunol. 11: 1739–1744.

9. Denzer, K., M. J. Kleijmeer, H. F. Heijnen, W. Stoorvogel, and H. J. Geuze.2000. Exosome: from internal vesicle of the multivesicular body to intercellularsignaling device. J. Cell Sci. 113: 3365–3374.

10. Montecalvo, A., W. J. Shufesky, D. B. Stolz, M. G. Sullivan, Z. Wang,S. J. Divito, G. D. Papworth, S. C. Watkins, P. D. Robbins, A. T. Larregina, andA. E. Morelli. 2008. Exosomes as a short-range mechanism to spread alloantigenbetween dendritic cells during T cell allorecognition. J. Immunol. 180: 3081–3090.

11. Valadi, H., K. Ekstrom, A. Bossios, M. Sjostrand, J. J. Lee, and J. O. Lotvall.2007. Exosome-mediated transfer of mRNAs and microRNAs is a novelmechanism of genetic exchange between cells. Nat. Cell Biol. 9: 654–659.

12. Pegtel, D. M., K. Cosmopoulos, D. A. Thorley-Lawson, M. A. van Eijndhoven,E. S. Hopmans, J. L. Lindenberg, T. D. de Gruijl, T. Wurdinger, andJ. M. Middeldorp. 2010. Functional delivery of viral miRNAs via exosomes.Proc. Natl. Acad. Sci. USA 107: 6328–6333.

13. Schorey, J. S., and S. Bhatnagar. 2008. Exosome function: from tumor immu-nology to pathogen biology. Traffic 9: 871–881.

14. Nguyen, D. G., A. Booth, S. J. Gould, and J. E. K. Hildreth. 2003. Evidence thatHIV budding in primary macrophages occurs trough the exosome releasepathway. J. Biol. Chem. 278: 52347–52354.

15. Karlsson, M., S. Lundin, U. Dahlgren, H. Kahu, I. Pettersson, and E. Telemo. 2001. “To-lerosomes” are produced by intestinal epithelial cells. Eur. J. Immunol. 31: 2892–2900.

16. Admyre, C., B. Bohle, S. M. Johansson, M. Focke-Tejkl, R. Valenta, A. Scheynius,and S. Gabrielsson. 2007. B cell-derived exosomes can present allergen peptidesand activate allergen-specific T cells to proliferate and produce TH2-like cytokines.J. Allergy Clin. Immunol. 120: 1418–1424.

17. Clayton, A., A. Turkes, S. Dewitt, R. Steadman, M. D. Mason, and M. B. Hallett.2004. Adhesion and signaling by B cell-derived exosomes: the role of integrins.FASEB J. 18: 977–979.

18. Thery, C., L. Duban, E. Segura, P. Veron, O. Lantz, and S. Amigorena. 2002.Indirect activation of naıve CD4+ T cells by dendritic cell-derived exosomes.Nat. Immunol. 3: 1156–1162.

19. Viaud, S., M. Terme, C. Flament, J. Taieb, F. Andre, S. Novault, B. Escudier,C. Robert, S. Caillat-Zucman, T. Tursz, et al. 2009. Dendritic cell-derivedexosomes promote natural killer cell activation and proliferation: a role forNKG2D ligands and IL-15Ralpha. PLoS ONE 4: e4942.

20. Sallusto, F., and A. Lanzavecchia. 1994. Efficient presentation of soluble antigenby cultured human dendritic cells is maintained by granulocyte/macrophagecolony-stimulating factor plus interleukin 4 and downregulated by tumor ne-crosis factor alpha. J. Exp. Med. 179: 1109–1118.

21. Clements, G. B., G. Klein, and S. Povey. 1975. Production by EBV infection ofan EBNA-positive subline from an EBNA-negative human lymphoma cell linewithout detectable EBV DNA. Int. J. Cancer 16: 125–133.

22. Raposo, G., H. W. Nijman, W. Stoorvogel, R. Liejendekker, C. V. Harding,C. J. Melief, and H. J. Geuze. 1996. B lymphocytes secrete antigen-presentingvesicles. J. Exp. Med. 183: 1161–1172.

23. Clayton, A., J. Court, H. Navabi, M. Adams, M. D. Mason, J. A. Hobot,G. R. Newman, and B. Jasani. 2001. Analysis of antigen presenting cell derivedexosomes, based on immuno-magnetic isolation and flow cytometry. J. Immunol.Methods 247: 163–174.

24. Sheldon, H., E. Heikamp, H. Turley, R. Dragovic, P. Thomas, C. E. Oon,R. Leek, M. Edelmann, B. Kessler, R. C. Sainson, et al. 2010. New mechanismfor Notch signaling to endothelium at a distance by Delta-like 4 incorporationinto exosomes. Blood 116: 2385–2394 .

25. Pochon, S., P. Graber, M. Yeager, K. Jansen, A. R. Bernard, J. P. Aubry, andJ. Y. Bonnefoy. 1992. Demonstration of a second ligand for the low affinityreceptor for immunoglobulin E (CD23) using recombinant CD23 reconstitutedinto fluorescent liposomes. J. Exp. Med. 176: 389–397.

26. Papp, K., P. Vegh, J. Prechl, K. Kerekes, J. Kovacs, G. Csikos, Z. Bajtay, andA. Erdei. 2008. B lymphocytes and macrophages release cell membrane de-posited C3-fragments on exosomes with T cell response-enhancing capacity.Mol. Immunol. 45: 2343–2351.

27. Garnier, J. L., N. R. Cooper, and M. J. Cannon. 1993. Low expression of CD20and CD23 in Epstein-Barr virus-induced B cell tumors in SCID/hu mice. Am. J.Pathol. 142: 353–358.

28. Young, K. A., A. P. Herbert, P. N. Barlow, V. M. Holers, and J. P. Hannan. 2008.Molecular basis of the interaction between complement receptor type 2 (CR2/CD21) and Epstein-Barr virus glycoprotein gp350. J. Virol. 82: 11217–11227.

29. Khyatti, M., A. Ahmad, M. Blagdon, R. Frade, and J. Menezes. 1998. Binding ofthe endogenously expressed Epstein-Barr virus (EBV) envelope glycoproteingp350 with the viral receptor masks the major EBV-neutralizing epitope andaffects gp350-specific ADCC. J. Leukoc. Biol. 64: 192–197.

30. Flanagan, J., J. Middeldorp, and T. Sculley. 2003. Localization of the Epstein-Barr virus protein LMP 1 to exosomes. J. Gen. Virol. 84: 1871–1879.

31. Molnar, Z., P. Biberfeld, and G. Klein. 1974. Immuno-fluorescence and immuno-electron microscopy on EBV-associated membrane antigen of a nasopharyngealcarcinoma derived lymphoblastoid cell line (LY-28). Eur. J. Cancer 10: 633–637.

32. Johansson, S. M., C. Admyre, A. Scheynius, and S. Gabrielsson. 2008. Differenttypes of in vitro generated human monocyte-derived dendritic cells releaseexosomes with distinct phenotypes. Immunology 123: 491–499.

33. Segura, E., S. Amigorena, and C. Thery. 2005. Mature dendritic cells secreteexosomes with strong ability to induce antigen-specific effector immuneresponses. Blood Cells Mol. Dis. 35: 89–93.

34. Feng, D., W. L. Zhao, Y. Y. Ye, X. C. Bai, R. Q. Liu, L. F. Chang, Q. Zhou, andS. F. Sui. 2010. Cellular internalization of exosomes occurs through phagocy-tosis. Traffic 11: 675–687.

35. Moore, M. D., R. G. DiScipio, N. R. Cooper, and G. R. Nemerow. 1989. Hy-drodynamic, electron microscopic, and ligand-binding analysis of the Epstein-Barr virus/C3dg receptor (CR2). J. Biol. Chem. 264: 20576–20582.

36. Segura, E., C. Nicco, B. Lombard, P. Veron, G. Raposo, F. Batteux, S. Amigorena,and C. Thery. 2005. ICAM-1 on exosomes from mature dendritic cells is cri-tical for efficient naive T-cell priming. Blood 106: 216–223.

37. Nolte-’t Hoen, E. N., S. I. Buschow, S. M. Anderton, W. Stoorvogel, andM. H. Wauben. 2009. Activated T cells recruit exosomes secreted by dendriticcells via LFA-1. Blood 113: 1977–1981.

38. Taylor, K. A., S. Wetzel, D. S. Lyles, and B. A. Pollok. 1994. Dual EBNA1promoter usage by Epstein-Barr virus in human B-cell lines expressingunique intermediate cellular phenotypes. J. Virol. 68: 6421–6431.

The Journal of Immunology 9

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

39. Fingeroth, J. D., J. J. Weis, T. F. Tedder, J. L. Strominger, P. A. Biro, andD. T. Fearon. 1984. Epstein-Barr virus receptor of human B lymphocytes is theC3d receptor CR2. Proc. Natl. Acad. Sci. USA 81: 4510–4514.

40. Temme, S., A. M. Eis-Hubinger, A. D. McLellan, and N. Koch. 2010. The herpessimplex virus-1 encoded glycoprotein B diverts HLA-DR into the exosomepathway. J. Immunol. 184: 236–243.

41. Ron, Y., and J. Sprent. 1987. T cell priming in vivo: a major role for B cellsin presenting antigen to T cells in lymph nodes. J. Immunol. 138: 2848–2856.

42. Ding, C., L. Wang, J. Marroquin, and J. Yan. 2008. Targeting of antigens toB cells augments antigen-specific T-cell responses and breaks immune toleranceto tumor-associated antigen MUC1. Blood 112: 2817–2825.

43. Whitmire, J. K., M. S. Asano, S. M. Kaech, S. Sarkar, L. G. Hannum,M. J. Shlomchik, and R. Ahmed. 2009. Requirement of B cells for generatingCD4+ T cell memory. J. Immunol. 182: 1868–1876.

44. Qazi, K. R., U. Gehrmann, E. Domange Jordo, M. C. Karlsson, andS. Gabrielsson. 2009. Antigen-loaded exosomes alone induce Th1-type memorythrough a B-cell-dependent mechanism. Blood 113: 2673–2683.

45. Roozendaal, R., and M. C. Carroll. 2007. Complement receptors CD21 andCD35 in humoral immunity. Immunol. Rev. 219: 157–166.

46. Middeldorp, J. M., and D. M. Pegtel. 2008. Multiple roles of LMP1 inEpstein-Barr virus induced immune escape. Semin. Cancer Biol. 18: 388–396.

10 LCL EXOSOMES TARGET B CELLS VIA CD21

by guest on April 10, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from