|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
Blood, 15 November 2004, Vol. 104, No. 10, pp. 3257-3266. Prepublished online as a Blood First Edition Paper on July 29, 2004; DOI 10.1182/blood-2004-03-0824.
IMMUNOBIOLOGY Endocytosis, intracellular sorting, and processing of exosomes by dendritic cellsFrom the Thomas E. Starzl Transplantation Institute; the Department of Surgery and the Department of Dermatology, University of Pittsburgh Cancer Institute; the Department of Immunology and the Department of Cell Biology; and the Physiology and Center for Biologic Imaging, University of Pittsburgh Medical Center, Pittsburgh, PA.
Exosomes are nanovesicles released by leukocytes and epithelial cells. Although their function remains enigmatic, exosomes are a source of antigen and transfer functional major histocompatibility complex (MHC)–I/peptide complexes to dendritic cells (DCs) for CD8+ T-cell activation. Here we demonstrate that exosomes also are internalized and processed by immature DCs for presentation to CD4+ T cells. Endocytosed exosomes are sorted into the endocytic compartment of DCs for processing, followed by loading of exosome-derived peptides in MHC-II molecules for presentation to CD4+ T cells. Targeting of exosomes to DCs is mediated via milk fat globule (MFG)–E8/lactadherin, CD11a, CD54, phosphatidylserine, and the tetraspanins CD9 and CD81 on the exosome and  v/ 3 integrin, and CD11a and CD54 on the DCs. Circulating exosomes are internalized by DCs and specialized phagocytes of the spleen and by hepatic Kupffer cells. Internalization of blood-borne allogeneic exosomes by splenic DCs does not affect DC maturation and is followed by loading of the exosome-derived allopeptide IE52-68 in IAb by host CD8+ DCs for presentation to CD4+ T cells. These data imply that exosomes present in circulation or extracellular fluids constitute an alternative source of self- or allopeptides for DCs during maintenance of peripheral tolerance or initiation of the indirect pathway of allorecognition in transplantation.
Dendritic cells (DCs) are antigen (Ag)–presenting cells (APCs) that function as biosensors of the cellular microenvironment by detecting the presence of signals that determine T-cell tolerance or immunity.1,2 To accomplish this task, DCs acquire extracellular Ags by receptor-mediated endocytosis, macropinocytosis, or phagocytosis3-5; by incorporation of microvesicles shed from the surface of neighboring cells,6,7 and by their recently described interaction with nanovesicles (  100 nm) termed "exosomes."8-12 Exosomes are formed by reverse budding of the membrane of late endosomes13-15 or multivesicular bodies (MVBs) and are released to the extracellular space by fusion of MVB with the plasma membrane.13-15 Originally described in neoplastic cell lines,16 exosomes also are produced by leukocytes and epithelial cells.17-22 Although the function of exosomes still is poorly understood, exosomes are a source of Ag for APCs and participate in Ag presentation to T lymphocytes.11,12 High concentrations of exosomes expressing major histocompatibility complex (MHC) and costimulatory molecules activate T-cell clones and T-cell lines weakly10,13 and fail to stimulate naive T cells.9,11 This impaired naive T-cell stimulatory ability of exosomes has been attributed to their low T-cell receptor–cross-linking capacity (inadequate for naive T-cell activation) and their small size and membrane composition.10 However, in the presence of DCs, exosomes increase their ability to stimulate T cells.10,11,23,24 The mechanism of interaction of extracellular exosomes with DCs is unknown. Although there is evidence that exosomes may transfer functional MHC-I/peptide complexes to DCs,24 it is unclear whether exosomes cluster or fuse with DCs or if they are internalized and processed, as occurs with vesicles derived from apoptotic cells.2-5
Herein we demonstrate that exosomes are internalized efficiently by DCs. Targeting of exosomes to DCs depends on ligands on the exosome and DC surface and is independent of complement factors. Once internalized by DCs, exosomes are sorted into recycling endosomes and then through late endosomes/lysosomes. By this mechanism, DCs process and present peptides derived from the internalized exosomes to T cells. In vivo, blood-borne exosomes are captured by DCs and specialized phagocytes of the spleen and by hepatic Kupffer cells. In the steady state, uptake of circulating exosomes by splenic DCs does not induce DC maturation and does not prevent CD40-induced DC activation in vivo. Our results demonstrate that blood-borne allogeneic exosomes are efficiently targeted, internalized, and processed by splenic DCs in vivo, a phenomenon followed by presentation of exosome-derived allopeptides by CD8
Mice and reagents
C57BL/10 (B10) and BALB/c mice were from The Jackson Laboratory (Bar Harbor, ME). 1H3.1 TCR- Generation of DCs
Bone marrow (BM) DCs were generated as described.25 BM cells from femurs of B10 mice were depleted of erythrocytes by hypotonic lysis. Erythroid cells, T and B lymphocytes, natural killer (NK) cells, and granulocytes were removed by incubation with mAbs (TER-119, CD3 Generation of BMDC-derived exosomes
BALB/c bone marrow–dendritic cells (BMDCs) were generated as described in "Generation of DCs." On day 4, medium was replaced by fresh medium with cytokines and 10% volume/volume exosome-free FCS obtained by overnight ultracentrifugation (100 000g). DC supernatants were collected on days 6 and 8 and centrifuged at 4°C at 300g (10 minutes), 1 200g (20 minutes), 10 000g (30 minutes), and 100 000g (60 minutes).13 Exosomes were washed in phosphate-buffered saline (PBS) and pelleted by overnight ultracentrifugation (100 000g). The amount of protein in the exosome preparation was assessed by Bradford assay (BioRad, Hercules, CA). For flow cytometric analysis, 500 µg exosomes were incubated with a fixed number of 4.5-µm beads (Dynabeads, Dynal, Lake Success, NY) coated with CD11b or IAd mAb. Beads coated with exosomes were labeled with the following phycoerythrin (PE) mAbs (BD PharMingen): H2Dd,IAd, IE (BioDesign International, Saco, ME), CD8 Electron microscopy Suspensions of exosomes were fixed with 4% paraformaldehyde (PF) and placed on grids for examination. BALB/c BMDC exosomes were labeled with 5-nm gold IAd mAb and incubated with B10 BMDCs. Then, BMDCs were fixed in PF, incubated in 3% gelatin, resuspended in 2.3 M sucrose, and frozen in liquid nitrogen. Ultrathin cryosections were labeled with rat anti–lysosome-associated membrane protein-1 (LAMP-1) mAb (1D4B [PDB] , BD PharMingen) followed by 12-nm gold anti–rat IgGs (Jackson ImmunoResearch Laboratories, West Grove, PA). All transmission electron micrographs were obtained using JEM1210 electron microscopy (JEOL, Peabody, MA) at 80 kv. Images were printed directly from negatives (Kodak TEM Film; Kodak, Rochester, NY) and printed onto Polycontrast paper (Kodak). Images were digitalized using a ScanJet 5300C flatbed scanner (Hewlett Packard, Palo Alto, CA) at 1000 dots per inch. Internalization of exosomes by BMDCs BALB/c BMDC exosomes were labeled with PKH67, mixed with 5 x 105 B10 BMDCs (1 hour at 37°C). Thereafter, the cells were washed with cold PBS, labeled with PE CD11c mAb, and fixed in PF. The percentage of CD11c+ DCs with PKH67+ exosomes was analyzed by flow cytometry. For blocking experiments, BMDCs were preincubated (30 minutes, 4°C) with the following mAbs (10-25 µg/mL, BD PharMingen): CD9 (KMC8), CD11a (M17/4), CD11b (M1/70), CD11c (HL3), CD18 (GAME-46), CD51 (H9.2B8), CD543E2, CD612G9.G2, CD812F7, or MFG-E8 (2422).26 Some assays were performed with 10 mM O-phospho-L-serine, 10 mM O-phospho-D-serine (Sigma), 1 mg/mL H-Gly-Arg-Gly-Asp-Thr-Pro-OH (GRGDTP) or 1 mg/mL H-Gly-Arg-Ala-Asp-Ser-Pro-OH (GRADSP) (Calbiochem, La Jolla, CA). Confocal microscopy For identification of early and late endosomes, BMDCs were incubated with 25 µg/mL of Texas red–transferrin or 10 µg/mL Dil-low-density lipoprotein (Dil-LDL; Molecular Probes, Eugene, OR) for 30 minutes at 37°C in medium without FCS and washed in PBS. BMDCs then were mixed with PKH67-labeled exosomes. The uptake of exosomes by BMDCs was stopped by washing in cold 0.1% sodium azide PBS, followed by fixation in PF. BMDCs were attached to poly-L-lysine–coated slides and imaged with a Leica TCS-NT confocal microscope (Leica Microsystems, Deerfield, IL). Assay of Ag presentation
1H3.1 TCRtg CD4+ T lymphocytes and the T-T cell hybrids BE Immunofluorescence Cryostat sections (8 µm) were fixed in 4% PF, blocked with goat serum, and incubated with the following biotin-mAbs: CD11c, H2Dd (BD PharMingen), MOMA-1 (Bachem, King of Prussia, PA), F4/80 (Bachem), or ER-TR9 (Bachem). Then, slides were incubated with 1:3000 cyanine 3 (Cy3)–streptavidin (Jackson ImmunoResearch Laboratories). For triple labeling, sections were incubated with biotin Y-Ae, hamster CD11c, and rat B220 mAbs. As a second step, slides were incubated with Cy3-streptavidin, Cy2 anti–hamster IgGs, and Cy5 anti–rat IgGs. Cytospins (230g) were fixed in 4% PF, blocked with goat serum, and incubated overnight (4°C) with biotin H2Dd, biotin IAb, or rat LAMP-1 mAbs followed by biotin anti–rat Igs and Cy3-streptavidin. Nuclei were stained with 4'6-diamidino-2-phenylindole 2HCl (DAPI) (Molecular Probes). RNAse protection assay
The analysis of cytokine mRNAs was performed by RNAse protection assay (RPA) as described.5,25 Briefly, RNA was isolated using a total RNA Isolation Kit (BD PharMingen) from CD11c+ BMDCs isolated by magnetic sorting. cDNAs encoding mouse IL-1 Statistical analysis Results are expressed as means ± SD. Comparisons between means were performed by analysis of variance (ANOVA), followed by the Student Newman Keuls test. Comparison between 2 means was performed by Student t test. A P value less than .05 was considered significant.
DCs capture extracellular exosomes: role of surface molecules We analyzed whether murine BMDCs (B10) internalize exosomes. BALB/c BMDC exosomes consisted of 65-100 nm membrane vesicles expressing MHC-I/II, CD71 (transferrin receptor), CD80, CD86,8,21 and ligands probably involved in docking or internalization of exosomes by DCs21 [CD11a-c, CD54 (intercellular cell adhesion molecule-1; ICAM-1), milk fat globule (MFG)–E8/lactadherin, CD9, CD81, and externalized phosphatidylserine (PS) (Figure 1A-C).
In vitro, 30% ± 7% of DCs internalized PKH67-labeled (green) exosomes within 2 hours at 37°C (Figure 2A). The uptake of exosomes decreased with cytochalasin D, EDTA, or at 4°C, suggesting that exosomes were internalized actively rather than attached to the DC surface (Figure 2A). We investigated the role of molecules expressed by the surface of exosomes and DCs in the endocytosis of exosomes. A decrease in exosome uptake by BMDCs was caused by simultaneous inhibition of
Exosomes are internalized efficiently by immature DCs
To test if the ability to capture exosomes differed between immature (CD11c+ CD86–) and mature (CD11c+ CD86+) BM-DCs,25 BMDCs were labeled with CyChrome-CD11c and PE-CD86 mAbs after phagocytosis of PKH67-labeled exosomes. Exosomes were internalized mostly by immature (CD86–) BMDCs (Figure 2C). Next, we analyzed the ability of different splenic DC subsets to internalize exosomes. In the steady state, splenic DCs include 2 main DC populations: (1) CD11c+ CD8 Intracellular sorting of internalized exosomes by DCs We studied the intracellular traffic of internalized exosomes in BMDCs. Immature (CD86–) BMDCs (B10) were preincubated with Texas red–transferrin (red; to label early endosomes) or with Dil–low-density lipoprotein (Dil-LDL; red; to stain late endosomes/lysosomes) and cocultured with PKH67-labeled (green) exosomes from DCs (BALB/c) and analyzed by confocal microscopy. As early as 5 minutes, PKH67 was detected in early endosomes (Figure 3A-B) and after 2 hours, PKH67 was found in Dil-LDL+ late endosomes/lysosomes (Figure 3C-D), a result confirmed on cytospins of BMDCs labeled with lysosome-associated membrane protein-1 (LAMP-1) mAb (Figure 3E). The traffic of internalized exosomes within BMDCs was further analyzed by immunoelectron microscopy. BMDC exosomes (BALB/c) were surface labeled with 5-nm gold IAd mAb (Figure 3F) and incubated with BMDCs (B10). After 20 minutes, 5-nm gold-labeled exosomes were detected inside late endosomes expressing LAMP-1 on their limiting membrane (Figure 3G-I). At later time points (1-2 hours), 5-nm gold particles were detected in electron-dense lysosomal vesicles expressing LAMP-1 (Figure 3J). We did not find 5-nm gold-labeled exosomes or 5-nm gold particles on the surface of BMDCs (Figure 3G, asterisk).
DCs process alloAgs derived from internalized exosomes for presentation to CD4+ T cells
We investigated the ability of BMDCs to process and load exosomal allopeptides into MHC class II. After 24 hours of incubation of BMDCs (B10; IAb; IE–) with allogeneic exosomes (BALB/c, IAd; IE
Next, we investigated if DC maturation enhanced the ability of BMDCs to present exosomal allopeptides to CD4+ T cells. BMDCs (B10) were treated with allogeneic exosomes (BALB/c) followed by stimulation with agonistic anti–CD40 IgM mAb (10 µg/mL, 16 hours) and used as stimulators of 1H3.1 TCRtg CD4+ T cells. 1H3.1 CD4+ T cells are specific for the IAb-IE Traffic of exosomes in vivo
Two hours after intravenous injection of PKH67-labeled allogeneic exosomes in B10 mice, PKH67 was detected in MOMA-1+ metallophillic macrophages, ER-TR9+ macrophages, and CD11c+ DCs of the splenic marginal zone (MZ, Figure 5A-C). PKH67 colocalized with H2Dd (PKH67 and H2Dd from the BALB/c exosomes) and accumulated in LAMP-1+ vesicles of splenic B10 DCs (Figure 5D-E). After 24 to 48 hours, CD11c+ DCs with PKH67+ inclusions were found in the center of the splenic follicle (Figure 5F). The kinetics of internalization of circulating exosomes by subsets of splenic DCs was assessed by flow cytometry. Two hours after intravenous exosome administration, splenic CD11c+ DCs (most of them CD8
In vivo circulating PKH67+ exosomes also were captured by hepatic (F4/80+) Kupffer cells (Figure 5H-J). Intravenous administration of exosomes does not induce maturation of splenic DCs
We investigated if capture of blood-borne exosomes affected maturation of splenic DCs in vivo. Twenty-four hours after intravenous injection of allogeneic exosomes (BALB/c) into B10 mice (200 µg/mouse), splenic DCs had not up-regulated expression of the DC-maturation markers IAb, CD86, or CD54 (Figure 6A). Administration of exosomes did not interfere with splenic DC maturation induced by agonistic CD40 mAb (intraperitoneal) in vivo (FGK4.5, Figure 6A). The effect of exosomes on cytokine mRNA expression by DCs was analyzed by RPA. Due to the high number of DCs required, we used immature BMDCs (B10) selected by magnetic sorting (CD86–DCs
Presentation of exosomal Ags to CD4+ T cells by splenic DCs in vivo
The ability of splenic DCs to process alloAgs derived from internalized exosomes and to load exosomal allopeptides into MHC II in vivo for CD4+ T-cell presentation was analyzed. Twenty-four to 36 hours after intravenous injection of exosomes (BALB/c) into B10 mice, splenic DC expressing IAb-IE
The immunologic role of exosomes is still unclear. T-cell exosomes expressing molecules of the TNF superfamily participate in cytotoxicity and activation-induced T-cell death.22,33 In other experimental models, exosomes have been used to promote T-cell immunity. Thus, in mice, subcutaneous administration of tumor peptide-pulsed DC exosomes induces cytotoxic T cells (CTLs) and tumor rejection,8 and subcutaneous injection of male DC exosomes into females activates CD4+ T cells specific for an H-Y peptide.11 The mechanism by which exosomes interact with DCs (or other APCs) is unknown. However, there is evidence that exosomes require the presence of DCs to activate naive T cells9,11 and that exosomes transfer functional MHC-I/peptide to acceptor DCs for presentation to CD8+ T cells.24 Thus, exosomes may potentially bind to the DC surface, fuse with the plasma membrane of DCs, or be internalized by DCs. The only known physiologic targets for exosomes are follicular DCs. B-cell–derived MHC II+ exosomes bind to the surface of follicular DCs with a possible role in development of high-affinity effector/memory B cells.32 Here, we demonstrate that exosomes are internalized by DCs through a mechanism that requires participation of the DC cytoskeleton and is calcium and temperature dependent. Our observation that the ability of BMDCs to capture exosomes decreases with cell maturation agrees with the fact that immature DCs exhibit higher endocytic capacity than mature DCs.1
There is indirect evidence that ligands present on the exosome surface are required to dock exosomes to the surface of target cells or to extracellular matrix proteins.15,34 Blocking of CD91 impairs presentation of exosomes, probably by interfering with uptake of exosomal Ags by APCs,12 and reticulocyte exosomes bind fibronectin via the Once internalized, exosomes were sorted into the endosomal pathway by DCs. Late endosomes/MVB store MHC II and are known as MHC II–enriched compartments (MIICs),42,43 sites where Ag processing occurs. In this endosomal compartment, exosome-derived allopeptides were loaded into MHC II and transported to the DC surface for presentation to CD4+ T cells. The presence of internalized exosomes within MIICs is not surprising since the MIICs are positioned strategically in the endocytic route of DCs.43,44 MIICs are proteolytically active organelles, exhibit low pH and are the site where endocytosed proteins are processed into peptides for MHC II loading for Ag presentation.45 The fact that in this study, a pH increase reversibly impaired the capacity of DCs to present an allopeptide-derived form internalized exosomes, implies that the endocytosed vesicles must be at least partially processed within MVB. Proteins like epidermal growth factor receptor are first sorted into exosomes within late endosomes/lysosomes for proteolytic degradation.46 In a similar way, we found that internalized exosomes were processed proteolytically for Ag presentation by DCs. There is evidence that exosomes transfer peptide-MHC complexes to the surface of APCs.10,11 However, our observations and a previous report indicate that extracellular exosomes do not cluster or fuse with the plasma membrane of myeloid DCs, as occurs with follicular DCs.32 Therefore, if extracellular exosomes are targeted to late endosomes/MVB by DCs, how can peptide-MHC complexes embedded in the membrane of internalized exosomes reach the DC surface? Different groups have demonstrated recently that during the translocation of newly synthesized MHC II from late endosomes/MVB to the cell surface, endogenous exosomes bearing MHC II back fuse with the limiting membrane of MVB to mediate rapid transfer of MHC II to the cell surface via tubular vesicles.47,48 A similar mechanism may apply to internalized exosomes once they reach MVB.
Importantly, we also provide in vivo evidence that circulating exosomes are captured efficiently by DCs and specialized macrophages of the spleen and by hepatic Kupffer cells. Exosomes were internalized initially by splenic CD8
Exosomes (likely from different cellular sources) have been found in human, rodent, and fetal calf sera.34,52 The physiologic relevance of targeting circulating exosomes to splenic DCs and hepatic Kupffer cells is unknown, as is the role played by presentation of exosomal Ags by splenic CD8
We have found that circulating exosome-derived alloAgs are presented to T cells by CD8
The 1H3.1 TCRtg mice, the 1.3H1 T-T cell hybrid, and the mAb Y-Ae were kindly provided by the late Dr C. Janeway and Dr C. Viret (Yale University, New Haven, CT). We thank Dr Philippa Marrack (Howard Hughes Medical Institute, National Jewish Medical and Research Center, Denver, CO) for the BE 20.6 T-T cell hybrid. The anti–mouse MFG-E8 mAb 2422 was kindly provided by Dr S. Nagata (Osaka University, Medical School, Osaka, Japan).
Submitted March 4, 2004; accepted May 28, 2004.
Prepublished online as Blood First Edition Paper, July 29, 2004; DOI 10.1182/blood-2004-03-0824.
Supported by grants R01 075512, R21 HL69725, and R21 AI55027 (A.E.M); R01 CA100893 and R21 AI57958 (A.T.L.); R01 AI43916 and P01 CA73743 (L.D.F.); and R01 DK49745 and R01 AI41011 (A.W.T.) from the National Institutes of Health, Bethesda, MD.
The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked "advertisement" in accordance with 18 U.S.C. section 1734.
Reprints: Adrian E. Morelli, E1546 Biomedical Science Tower, 200 Lothrop St, Pittsburgh, PA 15213-2582; e-mail: morelli{at}imap.pitt.edu.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
, B220, NK-1.1, Gr1, and IAb; BD PharMingen, San Diego, CA) followed by rabbit complement (Cedarlane, Hornby, Ontario, Canada). BM cells were cultured with RPMI-1640 (Life Technologies, Grand Island, NY), 10% volume/volume fetal calf serum (FCS), glutamine, nonessential amino acids, sodium pyruvate, HEPES [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid], 2-ME, and antibiotics plus mGM-CSF and mIL-4 (1000 U/mL). Splenic DCs were isolated as described.
92%). 
, tumor necrosis factor-
