Exosomes are secreted vesicles formed in late endocytic compartments. Mature dendritic cells (DCs) secrete exosomes bearing functional MHC-peptide complexes and high levels of ICAM-1. Such exosomes can activate Ag-specific naive T cells but only after recapture by recipient APCs. In this study, we addressed the molecular mechanisms of interaction between exosomes and recipient DCs. We show that exosomes can be presented by mouse DCs without the need for internalization and processing. Exosomes interact with DCs through a specific saturable receptor. Although the two major ligands of ICAM-1, LFA-1 and Mac-1, are expressed by lymphoid organ DCs, only LFA-1 is required for exosome capture by these cells. Accordingly, we show that CD8+ DCs express higher levels of LFA-1 than CD8− DCs, and that they are the main recipients of exosomes in vivo. We propose a new role for LFA-1 on DCs, as a receptor for exosomes to favor Ag transfer between DCs in vivo.
Exosomes are small membrane vesicles (diameter, 50–100 nm) that form within late multivesicular endosomal compartments (multivesicular bodies) (1, 2). Exosomes are released in the extracellular environment by fusion of multivesicular bodies with the plasma membrane. Exosome secretion has been reported for numerous cell types (listed in Ref. 2): reticulocytes, mast cells, dendritic cells (DCs),3 platelets, B lymphocytes, T lymphocytes, tumor cells, intestinal epithelial cells, and, more recently, the immune cells of the nervous system, i.e., microglia (3) but also neurons (4).
DCs are key actors in the initiation of immune responses (5, 6), and their exosomes have been extensively studied for their potential immune functions. DC-derived exosomes bear functional MHC class I- and class II-peptide complexes and can induce activation of specific T cells in vivo (7, 8). In vitro, it has been shown that DC-derived exosomes cannot directly interact with T cells but have to be recaptured by other DCs that use Ag or preformed MHC-peptide complexes transferred by exosomes to activate specific T cells (7, 9, 10). Indeed, exosomes also bear adhesion molecules, such as CD11a and CD11b integrins, ICAM-1/CD54 or MFG-E8, which could be involved in their targeting and uptake by recipient cells (10, 11, 12, 13). However, the exact mechanisms of interaction between exosomes and recipient cells remain unclear. Exosomes could either be captured via a specific receptor or be adsorbed nonspecifically to the plasma membrane. Captured exosomes could then either be internalized within recipient cells, or they could fuse with the plasma membrane or remain attached to the cell surface. Exosome endocytosis has been reported in a context where exosomes transfer Ag, which is processed into lysosomal compartments and presented on endogenous MHC molecules of recipient DCs (10). Tumor Ags transferred by tumor-derived exosomes to DCs are also processed and presented on endogenous MHC molecules (14, 15). But DC-derived exosomes also bear preformed MHC-peptide complexes that can be used by recipient DCs (7). Another mechanism of interaction between such exosomes and DCs could thus take place to allow T cell activation. In vivo, for instance, exosomes have been observed attached to the surface of follicular DCs in tonsil (16).
We have previously demonstrated that ICAM-1 on mature DC-derived exosomes is essential for their functional activity (13). ICAM-1 is widely expressed on endothelium and leukocytes. ICAM-1 is a costimulatory ligand of particular importance in DC-T cell interactions. ICAM-1 is also an adhesion molecule and is involved in each step of the leukocyte-endothelial cell adhesion cascade (rolling, firm adhesion, transmigration, and subendothelial migration) (17, 18). As a ligand for LFA-1 (CD11a/CD18) (19) and Mac-1 (CD11b/CD18) (20) integrins, ICAM-1 could either help the capture of exosomes by recipient APCs, or favor T cell binding to the recipient APC bearing exosomes on its surface.
In this study, we have analyzed the molecular mechanisms of interaction between DCs and exosomes bearing preformed MHC-peptide complexes. We show that exosomes can be captured and presented by DCs without the need for internalization and reprocessing. Exosomes interact with recipient cells through a specific saturable receptor. ICAM-1 being essential for DC-derived exosome functional activity, we investigated the role of ICAM-1 ligands on DCs in the interaction with exosomes. We demonstrate that LFA-1, but not Mac-1, on DCs is important for presentation of exosome-borne MHC-peptide complexes, both in vitro and in vivo. We show that CD8+ DCs, which do not express Mac-1, express high levels of LFA-1, and capture exosomes in vivo in lymphoid organs more efficiently than CD8− DCs, through interactions with this integrin. In this study, we describe a new function for LFA-1 on DCs, and a specific role of CD8+ DCs in acquiring exosome-derived Ag from other DCs, thus potentially enhancing the efficiency of the immune response.
Materials and Methods
Mice were obtained from Charles River Laboratories (C57BL/6), the Curie Institute animal facility (Marilyn (21), Icam-1−/− (22)), and Cancer Research United Kingdom London Research Institute (Mac-1−/− (23), Lfa-1−/− (24), and control wild type (WT)). Mice were housed in specific pathogen-free conditions. Experiments were done in accordance with the guidelines of the French Veterinary Department.
The DC line D1 (25) was cultured as previously described (12) in complete IMDM, supplemented with 15% GM-CSF-containing, J558-conditioned medium (11). Bone marrow-derived DCs were generated by 10–14 days of culture in complete IMDM with 30% J558-conditioned medium. Maturation was induced by 16-h treatment with 5 μg/ml LPS (Sigma-Aldrich). Splenic and lymph node DCs were purified from spleens or lymph nodes with anti-CD11c-coupled magnetic beads (N418 clone) according to the manufacturer’s instructions (Miltenyi Biotec) after 45-min treatment in DNase (10 μg/ml)- and collagenase (1 mg/ml)-containing medium at 37°C. Purity of DC preparations was >80%, as assessed by CD11c expression by FACS analysis. For experiments with CD8+ and CD8− DCs, cells were further purified by cell sorting on a FACSAria (BD Pharmingen) after staining with anti-CD8, anti-CD11c, and anti-B220 Abs. Purity of sorted DCs was 99%.
Exosomes were purified from the supernatant of D1 cells, WT, or Icam1−/− bone marrow-derived DCs, cultured in medium depleted of FCS-derived exosomes by overnight centrifugation at 100,000 × g (depleted medium) (12). To prepare exosomes bearing I-Ab/HY complexes, LPS-pretreated DCs were pulsed for 3 h with 250 nM HY peptide, NAGFNSNRANSSRSS; the culture medium was replaced with fresh depleted medium without peptide; and the supernatant was collected 24 h later. Exosomes were purified by filtration on 0.22-μm pore filters, followed by ultracentrifugation at 100,000 × g as described (12). In each exosome preparation, the concentration of total proteins was quantified by Bradford assay (Bio-Rad). A total of 3 μg of exosomal proteins was loaded on SDS-PAGE for Western blotting to check exosome protein composition (26).
Abs and reagents
Fluorophore-coupled Abs used for FACS analysis were as follows (all from BD Pharmingen): anti-CD11c (HL3 clone), CD11b (M1/70), LFA-1 (2D7), CD69 (H1.2F3), CD45.1 (A20), CD4 (RM4-5), CD44 (IM7), CD8 (53-6.7), B220 (RA3-6B2), CD40 (3/23), and the corresponding isotype controls. CFSE was from Molecular Probes. HY (NAGFNSNRANSSRSS) peptide was from Neosystem. DNase and collagenase were from Roche Diagnostics. Propidium iodide and paraformaldehyde were from Sigma-Aldrich.
In vitro T cell stimulation assay
Different quantities of exosomes or HY peptide were incubated in duplicate wells in 96-well round-bottom plates with 2 × 104 recipient cells (i.e., purified spleen DCs) in 50 μl of complete IMDM for 3 h at 37°C. For fixation experiments, DCs were washed twice in PBS and resuspended in 3% paraformaldehyde and incubated 3 min at 4°C. Fixation was stopped by addition of PBS/0.2 M glycine, and cells were washed in PBS/0.2 M glycine and then in complete IMDM. Fixed cells were counted and incubated with HY peptide or exosomes. Cells were washed twice with 100 μl of complete IMDM, and then 5 × 104 Marilyn lymph node cells were added in 100 μl of complete IMDM. Cells were harvested 18 h after addition of Marilyn T cells and stained with anti-CD4 and anti-CD69 Abs before FACS analysis on a FACSCalibur (BD Pharmingen).
In vivo T cell stimulation assay
One microgram of exosomes or 0.025 nmol of HY peptide were injected s.c. in the hindfoot pad of female WT or Lfa-1−/− mice (2–4 mice/group) that had been injected i.v. 24 h before with 106 lymph node cells from CD45.1+ Marilyn mice. Marilyn lymph node cells were labeled with CFSE (5 μM in PBS/0.1% BSA, 8 min, 37°C) before injection in host mice. Cells from the draining lymph nodes (i.e., pooled ipsilateral popliteal and inguinal lymph nodes) were analyzed 4 days later by FACS, after staining with anti-CD45.1, anti-CD44, and anti-CD4 Abs.
In vivo injection in mice
Five to 10 female WT or Lfa-1−/− mice were injected in the four footpads with exosomes (5 μg) or HY peptide (0.025 nmol). Noninjected mice were also included in the experiment. Three hours after injection, draining lymph nodes (popliteal and axillary lymph nodes) were harvested, pooled, and treated with DNase (10 μg/ml) and collagenase (1 mg/ml) (45 min, 37°C). Cells were directly purified by cell sorting on a FACSAria (BD Pharmingen) after staining with anti-CD8, anti-CD11c, and anti-B220 Abs. Cells were maintained at 4°C during the purification process. A total of 5 × 103 or 104 CD11chighB220−CD8high and CD11chighB220−CD8− cells was incubated in 96-well plates with 2 × 104 Marilyn lymph node cells. Eighteen hours later, cells were stained with anti-CD4 and anti-CD69 Abs for FACS analysis.
Results are expressed as means ± SEM. Statistical analysis was performed by using Student’s test. FACS data were analyzed with CellQuest software (BD Pharmingen).
Presentation of exosomes by DCs does not require internalization and reprocessing
Like tumor-derived exosomes (14), DC-derived exosomes contain Ag from the producing cell that can be transferred to DCs for processing and presentation (10). However, a specificity of DC-derived exosomes is that they also bear functional preformed MHC-peptide complexes (7). In this context, we addressed whether internalization and reprocessing of exosomes by recipient cells is required to induce T cell activation.
Throughout this work, MHC class II-peptide-bearing exosomes were purified from the supernatant of LPS-treated DCs incubated with HY peptide as previously described (12, 13). These exosomes do not contain free HY peptide (13). To address the need for internalization, we analyzed exosome presentation by fixed DCs, which cannot perform any endocytosis or phagocytosis. Splenic DCs, fixed with paraformaldehyde or not, were incubated with HY peptide or HY-bearing exosomes during 3 h and washed before addition of I-Ab-HY-specific naive CD4 T cells, obtained from Marilyn mice (21). T cell activation was assessed by measuring up-regulation of the early T cell activation marker CD69. When incubated with HY peptide, fixed DCs could not induce any Marilyn T cell activation as opposed to fresh DCs (Fig. 1⇓A). This observation may be explained by the fact that, in the strong fixation conditions used here, endogenous peptides become covalently associated to surface MHC class II molecules, preventing any peptide exchange, and/or that preprocessed antigenic peptides are loaded on MHC class II molecules inside endocytic compartments, rather than directly at the cell surface, as proposed by others (27, 28). By contrast, in the presence of HY-bearing exosomes, fixed DCs could trigger T cell activation, well above background activation (i.e., in the absence of DCs) (Fig. 1⇓B). These results show that exosomes bearing preformed MHC-peptide complexes can efficiently activate T cells when exposed at the cell surface without the need for reprocessing, and suggest that exosomes can be captured by DCs through interactions with cell surface molecules.
Exosomes interact with DCs through a specific receptor
To determine whether exosomes adhere nonspecifically to the plasma membrane of DCs or are captured by a specific saturable receptor, we performed a competition experiment between exosomes bearing MHC-peptide complexes or not using our in vitro T cell early activation assay as a readout for exosome capture by DCs. Although indirect, this readout gave in our hands more reliable results than binding assays using fluorescently labeled exosomes, and excluded the risk of altering the structure of exosome membrane by labeling with lipid markers. DCs were incubated with HY peptide or HY-bearing exosomes along with different doses of exosomes derived from DCs that were not exposed to HY peptide (control exosomes) (Fig. 2⇓). Before addition of Marilyn T cells, DCs were washed to remove noncaptured exosomes. In the presence of excess control exosomes, HY-exosome-induced Marilyn T cell activation was inhibited. By contrast, excess control exosomes did not impair the presentation of HY peptide by recipient DCs. These results indicate that control exosomes were competing with HY-bearing exosomes for a saturable specific receptor.
Role of ICAM-1 ligands in exosome-DC interactions
We have shown previously that exosomes secreted by mature DCs bear increased amounts of the adhesion molecule ICAM-1, compared with immature DC-derived exosomes, and that ICAM-1 on exosomes is essential to induce efficient Marilyn T cell proliferation in vitro in the presence of recipient DCs (13). ICAM-1 expression on exosomes is also required for early activation of naive T cells, because recipient DCs incubated with exosomes from mature Icam-1−/− DCs induce little CD69 up-regulation by Marilyn T cells, compared with WT exosomes (Fig. 3⇓A). We thus examined the role of ICAM-1 ligands, Mac-1 and LFA-1, in exosome-DC interactions. We performed our in vitro T cell activation assay using DCs purified from the spleens of Mac-1−/− or Lfa-1−/− mice as recipient cells. Mac-1−/−, Lfa-1−/−, and WT DCs were similar in terms of surface expression of MHC class II molecules and purity in splenic cell preparations, as assessed by the percentage of CD11c-positive cells (Fig. 3⇓B). Furthermore, Mac-1−/− (Lfa-1−/−, respectively) DCs expressed normal levels of LFA-1 (Mac-1, respectively) (Fig. 3⇓B). DCs were incubated with HY peptide or HY-bearing exosomes and washed before Marilyn T cell addition. Up-regulation of CD69 on T cells was analyzed after 18 h. In the presence of HY peptide, Mac-1−/− and Lfa-1−/− DCs induced T cell activation as efficiently as WT DCs (Fig. 3⇓C). This result shows that Mac-1−/− and Lfa-1−/− DCs have no defect in their ability to present preprocessed Ag to T cells and is consistent with similar observations by others (29). By contrast, in the presence of HY-bearing exosomes, Lfa-1−/− DCs were less efficient at inducing Marilyn T cell activation, compared with Mac-1−/− and WT DCs (Fig. 3⇓D). These results indicate that LFA-1 expression by DCs is important, but Mac-1 is dispensable, for exosome-induced T cell activation.
LFA-1 is involved in exosome-DC interactions in vivo
To confirm the relevance of these results in vivo, we next analyzed the efficiency of exosome-induced T cell activation after direct injection in Lfa-1−/− mice. A total of 1 μg of control exosomes or HY-bearing exosomes or 0.025 nmol of HY peptide was injected in the footpads of female WT or Lfa-1−/− mice adoptively transferred 1 day before with CFSE-labeled Marilyn lymph node cells. Because transferred Marilyn T cells express LFA-1 (data not shown), this assay allows analysis of the role of LFA-1 molecules on the host’s leukocytes, especially APCs. Draining lymph nodes were harvested 4 days later, and Marilyn T cell proliferation was evidenced by a 2-fold decrease of CFSE staining intensity in daughter cells at each division cycle (Fig. 4⇓A). In mice injected with HY peptide, we did not observe any significant difference in T cell proliferation in WT or Lfa-1−/− animals, showing that there is no major defect in those mice for presentation of preprocessed peptide, LFA-1+ T cell migration and activation (Fig. 4⇓B). The profiles of T cell proliferation obtained in mice injected with HY-bearing exosomes were different in WT and Lfa-1−/− mice (Fig. 4⇓B): the proportion of undivided (hence unactivated) cells in Lfa-1−/− mice was twice that observed in WT mice, and the proportion of cells resulting from five division cycles or more was reduced by 40% (Fig. 4⇓B). Exosome-induced T cell activation is therefore less efficient in Lfa-1−/− mice, showing that the presence of LFA-1 on recipient APCs is also important for exosome-induced T cell activation in vivo.
LFA-1 is differentially expressed on subsets of lymphoid organ DC
These results suggest that LFA-1 could be a receptor for exosomes on DCs. DCs are a heterogenous population of phenotypically and functionally different subtypes that can be distinguished by differential expression of surface molecules, in particular Mac-1 (30, 31, 32). We examined the expression of LFA-1 on different conventional DC subsets from spleen (Fig. 5⇓A) and cutaneous lymph nodes (Fig. 5⇓B). In the spleen, where DCs are phenotypically immature and express low levels of CD40, CD8+ DCs express higher levels of LFA-1 than CD8− DCs (Fig. 5⇓A). In lymph nodes, high levels of LFA-1 are only found on CD8+ DCs, whereas LFA-1 is expressed at low levels by both resident (CD40−) and skin-derived (CD40+) CD8− DCs (Fig. 5⇓B). Consistent with published data, we observed that Mac-1 was expressed at very low levels by CD8+ DCs from both organs (Fig. 5⇓). High expression of LFA-1 is therefore a feature unique to CD8+ DCs.
CD8+ DCs preferentially capture exosomes in vivo
Because CD8+ and CD8− subtypes of lymphoid organ DCs express different levels of LFA-1, we next examined whether their ability to capture and present exosomes in vivo was different. HY-bearing exosomes or HY peptide was injected in the footpads of female WT mice. Draining lymph nodes were harvested 3 h later, and cells were sorted based on their expression of B220, CD11c, and CD8. At this early time point, only Ag transported through the afferent lymph can be detected in the lymph node, whereas DCs migrating from the injection site have not yet reached this location (33, 34, 35). Sorted B220−CD11chighCD8high (CD8+ DCs) and B220−CD11chighCD8− cells (CD8− DCs) were cultured in vitro with Marilyn T cells, and T cell activation was assessed by measuring CD69 up-regulation. When mice were injected with HY peptide, we observed that both CD8+ and CD8− DCs could present the Ag (Fig. 6⇓A). This result shows that both CD8+ and CD8− DCs captured efficiently free peptide entering the lymph nodes. By contrast, when mice were injected with exosomes, only CD8+ DCs displayed I-Ab-HY complexes allowing T cell activation, 3 h after injection (Fig. 6⇓B). Thus, lymph node CD8+ DCs capture exosomes in vivo more efficiently than CD8− DCs. This unique ability of CD8+ DCs to capture exosomes in the lymph nodes depends on their expression of LFA-1, because when exosomes were injected in Lfa-1−/− mice, CD8+ DCs showed drastically reduced ability to activate T cells (Fig. 6⇓B). Capture and presentation of HY peptide, by contrast, was identical in Lfa-1−/− and WT mice (Fig. 6⇓A). These results show that CD8+ DCs use LFA-1 to capture and present exosomes in the lymph nodes.
In this study, we show that exosomes secreted by mature DCs, and bearing functional MHC class II-peptide complexes and high amounts of ICAM-1 molecules, are captured and presented by CD8+ DCs in vivo through specific interactions with LFA-1.
The way exosomes interact with cells of the immune system, and induce T cell-dependent immune responses in vivo, is still not entirely clear. In this study, we demonstrate that neither capture nor presentation by DCs in vitro of mature DC-derived exosomes bearing preformed MHC-peptide complexes requires internalization and reprocessing, although both may still happen in vivo. In our in vitro assay, exosomes can act as Ag-presenting microdomains at the surface of fixed DCs to activate T cells, but the fate of exosomes captured by live DCs remains unclear. Exosomes could remain attached at the cell surface, as occurs with follicular DCs (16), and present their MHC-peptide complexes directly to T cells. It cannot be excluded that exosomes are endocytosed; they could then fuse with the limiting membrane of endocytic compartments and their MHC-peptide complexes recycled and exposed at the cell surface for T cell stimulation. Indeed, MHC class II molecule transport in late endocytic compartments is very dynamic (36, 37), and it has been shown that MHC class II molecules from lysosomal compartments can be exposed at the cell surface (38). Finally, exosomes could also fuse with the plasma membrane.
Exosomes are a source of preformed functional MHC-peptide complexes (7, 9) but can also transfer material that can be degraded and reprocessed for presentation on endogenous MHC molecules. Morelli et al. (10) have shown that immature DC-derived exosomes are internalized by DCs and degraded into lysosomal compartments, and that blocking various surface molecules on DCs partially reduces exosome uptake. Tumor exosomes are also uptaken by DCs and their content degraded for T cell presentation (14). The exosomes used in those two studies bear little if any ICAM-1 molecules, as opposed to exosomes secreted by mature DCs (13). These results suggest different mechanisms of interaction between DCs and exosomes depending on their surface protein composition: capture through ICAM-1/LFA-1 interactions and presentation of preformed MHC-peptide complexes or internalization and degradation of exosomes for presentation of their Ags on endogenous MHC molecules.
An important result from this study is the description of a new role for expression of the integrin LFA-1 by DCs. The role of LFA-1 in cell migration or cell-cell interactions has been extensively studied on T cells (24, 39, 40), neutrophils (41, 42, 43), or B cells (44, 45, 46). By contrast, the role of LFA-1 in the specific DC functions had hardly been analyzed so far. Only one study reports a role for LFA-1 expressed by DCs in the formation of the immunological synapse (47). Two recent studies have analyzed the functional state of LFA-1 on human monocyte-derived DCs (48) or murine bone marrow-derived DCs (29), and reported that this integrin was not active. These results are in apparent contradiction with our work, because we show here that LFA-1 is involved in exosome capture by lymphoid organs DCs, therefore suggesting that these DCs bear activated or activable LFA-1 molecules. Indeed, we observed specific binding in vitro of ICAM-1-coated fluorescent beads to freshly purified mouse lymphoid organ DCs, but not to immature murine bone marrow-derived DCs (data not shown): in vitro-generated DCs may thus not always reflect exactly the physiological situation of DCs present in lymphoid organs. Our results suggest that ICAM-1/LFA-1 interactions between exosomes and recipients cells could allow specific targeting of DC-derived exosomes toward cells that express an activated form of LFA-1.
ICAM-1/LFA-1 interactions have been observed previously in another cell-vesicle interaction mechanism. HIV-1 virus can acquire host ICAM-1 molecules and then become more infectious for target T cells (49). Moreover, target cells expressing activated LFA-1 molecules are more susceptible to ICAM-1-bearing virus infection (50). Interaction of ICAM-1-bearing virus with LFA-1-expressing target cells triggers viral material release in the cells, probably after fusion of the limiting membrane of HIV-1 virus with the plasma membrane (51). In these studies, ICAM-1-bearing viruses interact with CD4 T cells. T cells express LFA-1, but we have demonstrated that DC-derived exosomes cannot directly interact with or be captured by naive CD4 T cells (7). A likely explanation of this observation is that LFA-1 molecules expressed by circulating leukocytes have low affinity for ICAM-1 (19), and need to be activated by engagement of CD2 or TCR (52), or various mechanisms (53, 54), before they could bind ICAM-1-bearing exosomes.
Finally, we show preferential capture of exosomes by lymph node CD8+ DCs in vivo. Three hours after injection of exosomes, only CD8+ DCs could present exosome-borne Ag. In this setting, lack of LFA-1 expression on lymph node-resident CD8+ DCs severely impaired their ability to capture and present exosomes. At a later time point (24 h), preliminary observations show that a population of CD8− DCs recovered from the lymph nodes, which may be either lymph node-resident DCs, or dermal DCs that have migrated from the injection site, could also present exosome-derived MHC-peptide complexes, although less efficiently than CD8+ DCs (data not shown). We therefore show in this study that, in lymph nodes, resident CD8+ DCs are the main recipient of exosomes through specific interactions with LFA-1. In vitro, both CD8+ and CD8− DCs can efficiently present immature (7) (i.e., bearing little ICAM-1) or mature exosomes (data not shown), but these observations do not reflect physiological conditions with limiting access to exosomes. In our in vitro (Fig. 3⇑) and in vivo (Fig. 4⇑) Ag presentation assays, lack of LFA-1 on DCs impaired their ability to present exosome-borne Ags, but only partially, which suggest the existence of another receptor for exosomes. Candidate molecules could be surface receptors that are more strongly expressed by CD8+ DCs, such as, for instance, CD205, CD36, or CD24 (55, 56).
These findings have interesting implications in view of the recent literature. Indeed, transfer of preformed MHC-peptide complexes from injected DCs to resident lymph node DCs has been suggested to participate to CD4 T cell activation in the lymph nodes in several studies (57, 58). More strikingly, several recent reports have shown that transfer of antigenic information occurs in vivo between CD8− DCs from the periphery and CD8+ DCs residing in lymphoid organs, leading to presentation of MHC class II (59, 60)- or MHC class I-peptide complexes (61) by both DC populations. The mechanism of such Ag transfer remains unknown. More stringently, in a skin infection model, Allan et al. (35, 62) have shown that Langerhans cells and dermal DCs migrate from the epidermis where they encounter the virus but cannot present the Ag in the skin draining lymph nodes. Only resident CD8+ DCs can stimulate specific CD8+ T cells. In this work, the authors do not address how viral Ag is presented to CD4+ T cells in the lymphoid organs. One possibility is that migrating DCs only transfer antigenic material to CD8+ DCs for processing and presentation to CD8+ T cells and interact themselves with CD4+ T cells. Our results offer an alternative model in which exosomes could transfer to CD8+ DCs Ag that will be cross-presented on endogenous MHC class I molecules, along with preformed MHC class II-peptide complexes. This mechanism would also allow concomitant Ag presentation to CD4+ and CD8+ T cells on the same APC, which is often required for efficient T cell priming (63, 64, 65).
Our results suggest that exosomes could represent a mean for Ag transfer between DCs in vivo. Although this possibility is attractive, the demonstration of a physiological role for exosomes awaits further evidence. By allowing a better understanding of the molecular mechanisms of DC-exosome interactions, our results will contribute to address this question, as well as improving therapeutic strategies involving exosomes.
We thank the team of the Curie Institute’s animal facility, Dr. O. Lantz for providing Marilyn mice, and Dr. C. Hivroz for critical reading of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by Ligue Nationale contre le Cancer, Institut National de la Santé et de la Recherche Médicale, Institut Curie, European Grant “DC for New Immunotherapies” 512074, and Cancer Research United Kingdom. E.S. was a fellow of Ministère de l’Education et de la Recherche.
↵2 Address correspondence and reprint requests to Dr. Clotilde Théry, Institut Curie, Pavillon Pasteur, Institut National de la Santé et de la Recherche Médicale, Unité 653, 26 rue d’Ulm, 75005 Paris, France. E-mail address:
↵3 Abbreviations used in this paper: DC, dendritic cell; WT, wild type.
- Received February 26, 2007.
- Accepted May 22, 2007.
- Copyright © 2007 by The American Association of Immunologists