Melanoma-Derived Extracellular Vesicles Bear the Potential for the Induction of Antigen-Specific Tolerance

Background: Cancer-induced immunosuppression is antigen-specific rather than systemic and the mechanisms for the antigen specificity are incompletely understood. Here we explore the option that tumor-associated antigens (TAAs) may be transferred to antigen-presenting cells (APCs), together with immunosuppressive molecules, through cancer-derived small extracellular vesicles (sEVs), such as exosomes. Stimulation of a suppressive phenotype in the very same APCs that take up TAAs may yield antigen-specific tolerance. Methods: sEVs isolated from patient-derived or well-established melanoma cell lines were used to demonstrate the transfer of major histocompatibility complex (MHC) molecules to the surface of APCs. The immunosuppressive influence of sEVs was assessed by flow cytometry analysis of activation markers, cytokine expression, and mixed lymphocyte reactions. Results: MHC class I molecules were transferred from melanoma cells to the cell surface of APCs by sEVs. Concomitantly, CD86 and CD40 co-stimulatory molecules were down-regulated and IL-6 production was strongly induced. TGF-β transported by sEVs contributed to the promotion of a suppressive phenotype of APCs. Conclusion: The presented results indicate the existence of a hitherto undescribed mechanism that offers an explanation for antigen-specific tolerance induction mediated by cancer-derived sEVs.


Introduction
Melanoma belongs to the group of tumors that are highly sensitive to immunotherapies. The therapeutic antibodies targeting the immune checkpoints, ipilimumab for cytotoxic T-lymphocyte antigen-4 (CTLA-4) and nivolumab and pembrolizumab for programmed death 1 (PD-1), greatly increased the average life expectancy for patients with metastatic melanoma. However, about half of all patients demonstrate primary or acquired resistance [1,2]. A better understanding how melanoma cells modulate the immune system is clearly needed. Here, we provide evidence that melanoma-derived sEVs contribute to antigen-specific immunosuppression.
Cancer development can be suppressed by the host immune system [3]. Cytotoxic T cells constitute a part of the adaptive immune response that very efficiently kills cancerous cells. The activation of cytotoxic T cells depends on their direct interactions with antigen presenting cells (APCs) which take up tumor-associated antigens (TAA), process them into peptides, and present them on MHC (major

Antigen-Presenting Cells (APCs)
The human monocytic leukemia cell line THP-1 (kindly provided by Prof. J. Dziadek, Polish Academy of Sciences, Lodz, Poland) was cultured at 5 × 10 5 -1 × 10 6 cells/mL in RPMI 1640 medium, supplemented with 10% FBS, 2 mM l-glutamine, 0.168 mM penicillin, 0.172mM streptomycin, 0.05 mM β-mercaptoethanol, and 1mM sodium pyruvate in a humidified incubator at 37 • C and 5% CO 2 (Binder GmbH, Tuttlingen, Germany). Primary human monocytes were isolated from the blood of healthy donors by a magnetic cell separation procedure using anti-CD14 antibodies (MACS system, Miltenyi Biotech, Bergisch Gladbach, Germany). The MACS system was also used to isolate T cells with the pan T cell isolation kit (MACS system, Miltenyi Biotech). The differentiation of primary monocytes or THP-1 cells into dendritic cells (DCs) was achieved by culturing the cells in the presence of recombinant human IL-4 (100 ng/mL) and GM-CSF (100 ng/mL) for three to five days. For the stimulation of cytokine production, cells were incubated at a density of 5 × 10 5 cells per mL in serum-free medium (AIM V, Gibco-Thermo Fisher Scientific, Life Technologies Polska, Warsaw, Poland) for 24 h with LPS (Lipopolysaccharide (E. coli 0111:B4), 1 µg/mL; Sigma-Aldrich, St Louis, MO, USA) and CD40 ligand (100 ng/mL, soluble, human, recombinant Mega CD40L, Enzo Life Sciences, Lausen, Switzerland).

Melanoma Cell Culture for sEV Production
The human melanoma cell line A375 (a gift from Prof. Piotr Laidler, Jagiellonian University, Krakow, Poland) was maintained in RPMI 1640 medium supplemented with 10% FBS and antibiotics as before. For the elimination of bovine serum derived sEVs, growth media were subjected to ultracentrifugation at 100,000× g for 2.5 h at 10 • C (full acceleration and maximal breaking). For sEV production, cells were grown in this sEV-depleted medium. Melanoma cell lines from drug-naïve patients were also used as source of sEVs. For that, an approval by Ethical Commission of Medical University of Lodz was obtained and patients consented to tissue acquisition. Patient-derived melanoma cell lines, DMBC12 and DMBC21, were grown in non-adherent flasks in stem cell medium as described [27].

Isolation and Characterization of Extracellular Vesicles
Tumor cell-derived sEVs were isolated by differential centrifugation [28]. After the cell cultures reached high density (about 90% confluency for adherent cells, about 10 6 cells/mL for suspension cells), cells were pelleted by centrifugation at 300× g for 4 min. The remaining supernatants were centrifuged for 30 min at 10,000× g (10 • C) to precipitate cell debris and organelles, then sEVs were pelleted by ultracentrifugation at 100,000× g for 2.5 h (10 • C, full acceleration and breaking power). The sEV pellet was re-suspended in 11 mL phosphate-buffered saline (PBS) and the last centrifugation step was repeated. The pellet was again re-suspended in PBS and the protein concentration was determined by the Bradford assay (Bio-Rad Polska, Warsaw, Poland). The identity of sEVs was confirmed through the detection of characteristic surface markers [28]. EVs were adsorbed to latex beads, stained with anti-CD63 or anti-CD9 antibodies, and analyzed by flow cytometry. Staining with not-specific isotypic antibodies served as a negative control. Melanoma-derived sEVs (A375) in a concentration of 0.1 µg/µL were further analyzed by atomic force microscopy using an NTEGRA SPECTRA instrument (NT-MDT, Spectrum Instruments Ltd., Limerick, Ireland) [29]. For the visualization of EVs by Transmission Electron Microscopy (TEM, Tesla BS 512 with YAG camera, Brno, Czech Republic) the vesicles were fixed with 2% paraformaldehyde. The sEV suspension was loaded onto formvar carbon coated electron microscopy grids (FCF200-Cu-50, 200 mesh, Electron Microscopy Sciences, Hatfield, PA, USA), fixed in 1% (v/v) glutaraldehyde, and contrasted with 2% (w/v) aqueous uranyl acetate (Polysciences) [30].

Flow Cytometry
Flow cytometry measurements were performed on a FACSCalibur (Becton-Dickinson, East Rutherford, NJ, USA). For each sample, a minimum of 10 4 cells or beads was measured. Results were analyzed using the CellQuest software (Becton-Dickinson). On myeloid cells, the surface markers CD40, CD80, CD86, HLA-DR, and HLA-ABC were assessed. Antibodies labeled with fluorescein (HLA-DR), phycoerythrin (CD80, HLA-ABC), or apocyanin (CD86, CD40) were purchased from Biolegend (San Diego, CA, USA) and Becton-Dickinson. For the flow cytometric analysis of sEVs, the vesicles were adsorbed to aldehyde-sulfate latex beads (3.8 µm size, Life Technologies Polska, Warsaw) for 20 min at room temperature. BSA was used to saturate the binding sites of the beads, followed by incubation with glycine (100 mM) in PBS for 20 min at room temperature. The beads were washed twice with FACS buffer (PBS, 0.5% BSA, 0.1% Na-azide) incubated with FITC-conjugated or PE-conjugated monoclonal antibodies for 1h in the dark, washed again with cold FACS buffer, and analyzed. Cytokines (IL-12, IL-6, VEGF, TGF-β) from cell culture supernatants were quantified using cytometric bead arrays (Becton-Dickinson), according to the accompanying instructions.

Intercellular Transfer of Peptide-MHC-I Complexes
To demonstrate the transfer of MHC class I molecules to the cell surface of APCs, fluorescently labeled antibodies recognizing HLA-ABC were bound to melanoma sEVs in PBS (30 min at 0 • C). The mixture was diluted 10-fold and insoluble antibody aggregates were removed by centrifugation at 12,000× g for 10 min. The supernatant was further diluted (10-fold) and sEVs were pelleted by ultracentrifugation at 100,000× g for 2.5 h. The sEV pellet was re-suspended in cell culture medium and added to APCs at a concentration of 10 µg/mL. The transported fluorescence was measured 16 h later by flow cytometry. APC-labeled goat anti-mouse secondary antibodies (Biolegend) were used to demonstrate the appearance of the MHC-antibody complexes at the cell surface.

RNA Isolation and Real-Time Quantitative Reverse Transcription PCR (qRT-PCR)
Total RNA was isolated according to standard protocols using the TriPure Isolation Reagent, (Roche Diagnostics, Mannheim, Germany). The mRNA levels of CD40, CD80, CD86, HLA-A, ALA-B, HLA-C, and HLA-DRalpha were evaluated by real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) using the LC RNA amplification kit SYBR Green I and a LightCycler Instrument 1.0 (Roche Diagnostics). The target gene expression levels were related to the house-keeping reference gene GADPH and to 5S ribosomal RNA. All primers used are listed in the supplementary Table S1. Products of amplification were identified by the thermal dissociation method.

Allogeneic Mixed Lymphocyte Reaction
PBMCs were isolated from the blood of healthy donors by Ficoll cushion centrifugation. PBMCs from one donor were re-suspended in cell culture medium (RPMI 1640 supplemented with 10% FBS, 2 mM l-glutamine, 100 IU/mL penicillin, 100 µg/mL streptomycin). Cell density was adjusted to 1 × 10 6 cells/mL. PBMCs of a second donor were re-suspended in PBS and labeled with 5 µM CFSE (carboxyfluorescein diacetate succinimidyl-ester, Molecular Probes, InVitrogen) for 10 min at 37 • C. Labeled cells were washed four times with cell culture medium and suspended at 1 × 10 6 cells/mL. A total number of 2 × 10 5 labeled cells were mixed with the same amount of unlabeled cells from the first donor. The test was set up in triplicates (three independent experiments). The sEVs isolated from A375 melanoma cells were added at a concentration of 20 µg/mL. Maximal stimulation of T cell proliferation was achieved using PMA (Phorbol-12-myristate-13-acetate, 10 ng/mL) and ionomycin (500 ng/mL) in cultures containing only cells from the second donor. After 72 h, cells were harvested, stained with APC-labelled anti-CD3 antibodies and analyzed by flow cytometry.

Statistical Analysis
Results from at least three independent experiments are presented as mean ± SD. Data were analyzed using Student's paired t-test after confirming normal distribution by the Shapiro-Wilk test and a p-value below 0.05 was considered significant.

Isolation and Characterization of Melanoma Cell-Derived sEVs
Melanoma cell lines were grown in serum-free (DMBC cell lines) or exosome-depleted medium (A375) and sEVs were isolated by differential centrifugation. About 3 × 10 7 particles per µg of protein were obtained. The efficiency of removal of sEVs from the culture medium was confirmed by demonstrating decreased RNA concentration in the depleted medium to a level lower than in commercially available exosome-depleted media (not shown). The sEVs were characterized by detection of CD9 and CD63 proteins at their surface by flow cytometry after adsorption of the vesicles to latex beads ( Figure 1a). Atomic force microscopy (AFM) analysis ( Figure 1b) and transmission electron microscopy ( Figure 1c) showed vesicular structures in the size range typical for exosomes (50-130 nm). Immunoblot analysis was used to show the presence of Alix, a soluble protein characteristic for small EVs, as well as the absence of Calnexin, a marker protein found in endoplasmic reticulum membranes in our vesicle preparations.

sEVs Transport MHC Class I Molecules from Melanoma Cells to APCs
To demonstrate the potential of sEVs to transfer melanoma-derived MHC molecules to APCs, sEVs were incubated with fluorescently labeled monoclonal mouse antibodies recognizing HLA-ABC (MHC class I) or CD63. Contaminating antibody complexes were eliminated by centrifugation, and unbound soluble antibodies were removed by pelleting the sEVs after diluting the antibody binding reaction about 100-fold with PBS. Primary human monocytes (Figure 2a,b) or THP-1 cells (Figure 2c,d) from a human monocytic cell line were incubated for 16 hours with the antibody-coated sEVs in cell culture. Flow cytometry measurements showed that the labeled antibodies bound to the MHC molecules were transferred to the target cells ( Figure 2, shift along the x-axis). To demonstrate the specificity of the transfer, two kinds of negative controls were included, a sample containing sEVs mixed with isotype control antibodies ( Figure 2a) and another sample containing HLA-ABC antibodies, but no sEVs. The isotype control confirmed that antibodies reached the target cells only when bound to specific antigens on the surface of sEVs. The negative control lacking sEVs ('-sEVs') showed that no detectable amount of free antibody was transferred by the applied procedure. To demonstrate that the sEV-transported MHC/antibody complexes appeared at the cell surface, the recipient cells were incubated with anti-mouse secondary antibodies ( Figure 2, shift along the y-axis). These antibodies could bind only to the primary antibodies which appeared on the cell surface. The results demonstrated that the transported MHC class I complexes became exposed at the cell membrane. When an antibody to CD63 was transported by sEVs, the fluorescent label of the primary antibody was transferred to the cells, but the antibody did not appear at the cell surface as visualized by the low percentage (4.1%) of cells stained by the secondary anti-mouse antibody (Figure 2d).

sEVs Transport MHC Class I Molecules from Melanoma Cells to APCs
To demonstrate the potential of sEVs to transfer melanoma-derived MHC molecules to APCs, sEVs were incubated with fluorescently labeled monoclonal mouse antibodies recognizing HLA-ABC (MHC class I) or CD63. Contaminating antibody complexes were eliminated by centrifugation, and unbound soluble antibodies were removed by pelleting the sEVs after diluting the antibody binding reaction about 100-fold with PBS. Primary human monocytes (Figure 2a,b) or THP-1 cells membrane. When an antibody to CD63 was transported by sEVs, the fluorescent label of the primary antibody was transferred to the cells, but the antibody did not appear at the cell surface as visualized by the low percentage (4.1%) of cells stained by the secondary anti-mouse antibody (Figure 2d).

sEV Mediated Modulation of Immune Receptor Surface Expression on APCs
To demonstrate the immunosuppressive capacity of melanoma-derived sEVs, human monocytes, and DCs derived thereof by in vitro culture were incubated with these sEVs. In preliminary experiments (not shown) an optimal sEV concentration of 20 µg/mL for the treatment of target cells (corresponding to approximately 1000-1250 vesicles/cell) was found to induce marked and consistent effects. Incubation of primary human monocytes (HuMo) with sEVs derived from A375 melanoma cells decreased the expression of co-receptor protein CD40 and HLA-DR, while the mean fluorescence intensity of HLA-ABC molecules and CD86 remained largely unchanged ( Figure  3a, upper panel). In similar experiments using THP-1 cells, down-regulation was observed only for HLA-DR, while CD40 and CD86 cell surface expression was increased (Figure 3b). When CD14+ primary human monocytes were first differentiated into DCs (Figure 3a, lower panel), exposure to A375 cell-derived sEVs down-regulated the surface expression levels of MHC proteins class I and II, CD86 and CD40. DCs derived from THP-1 cells treated with A375-derived sEVs showed downregulation of HLA molecules, while CD40 was up-regulated and CD86 stayed unchanged ( Figure   Figure 2. Melanoma cell-derived sEVs transfer MHC class I and class II molecules to the surface of APCs. The sEVs from A375 cells were incubated with PE-labelled isotype (a) or anti-HLA-ABC antibodies (b,c), purified by centrifugation, and incubated overnight with healthy donor monocytes (HuMo; a,b) or THP-1 cells (c,d). The next day, the cells were stained with APC-labelled secondary anti-mouse antibodies to demonstrate the appearance of the transferred MHC/antibody complexes on the surface of the recipient cells. A negative control following the same procedure but lacking sEVs is shown in the left part of (b,c). (d) Transfer of PE-labelled anti-CD63 antibodies through sEVs, followed by staining with anti-mouse antibodies. The percentages of events in the four quadrants are indicated. The results of one representative experiment out of three are shown.

sEV Mediated Modulation of Immune Receptor Surface Expression on APCs
To demonstrate the immunosuppressive capacity of melanoma-derived sEVs, human monocytes, and DCs derived thereof by in vitro culture were incubated with these sEVs. In preliminary experiments (not shown) an optimal sEV concentration of 20 µg/mL for the treatment of target cells (corresponding to approximately 1000-1250 vesicles/cell) was found to induce marked and consistent effects. Incubation of primary human monocytes (HuMo) with sEVs derived from A375 melanoma cells decreased the expression of co-receptor protein CD40 and HLA-DR, while the mean fluorescence intensity of HLA-ABC molecules and CD86 remained largely unchanged (Figure 3a, upper panel). In similar experiments using THP-1 cells, down-regulation was observed only for HLA-DR, while CD40 and CD86 cell surface expression was increased (Figure 3b). When CD14+ primary human monocytes were first differentiated into DCs (Figure 3a, lower panel), exposure to A375 cell-derived sEVs down-regulated the surface expression levels of MHC proteins class I and II, CD86 and CD40. DCs derived from THP-1 cells treated with A375-derived sEVs showed down-regulation of HLA molecules, while CD40 was up-regulated and CD86 stayed unchanged (Figure 3b). Using DCs derived from the peripheral blood of healthy donors, the expression of the investigated receptors was also assessed at the mRNA level (Figure 3c). Exposure to A375 melanoma sEVs reduced the HLA-DR mRNA level to about 80% of the untreated control, while class I MHC molecules and CD40 were slightly up-regulated (<1.5-fold). CD86 mRNA levels remained unchanged. 3b). Using DCs derived from the peripheral blood of healthy donors, the expression of the investigated receptors was also assessed at the mRNA level (Figure 3c). Exposure to A375 melanoma sEVs reduced the HLA-DR mRNA level to about 80% of the untreated control, while class I MHC molecules and CD40 were slightly up-regulated (<1.5-fold). CD86 mRNA levels remained unchanged.  Overlaid are the histograms after exposure to A375 melanoma-derived sEVs for 48 hours (filled violet peaks; representative of n = 6). Lower panel: HuMo were differentiated into DCs (HuDC) and exposed to sEVs for 48 hours and the cell surface expression was analyzed by flow cytometry, as above (representative of n = 3). (b) The mean fluorescence intensities (MFI) of the surface expression of MHC class I, II, CD40, and CD86 after exposure to melanoma-derived sEVs was related to the MFI of unexposed cells. Both monocytes and DCs derived from either the THP-1 cell line (left diagram) or primary human CD14+ cells (right diagram) were used in the experiments. At least three independent experiments were conducted for each cell type and significant differences were marked by asterisks (Student's t-test). (c) The mRNA expression levels in DC cells derived from the peripheral blood of healthy donors treated with A375-derived sEVs relative to untreated control cells. Mean values and standard deviations are shown from three independent experiments. The differences between samples after sEV exposure and untreated controls did not reach statistical significance (Student's t-test).

sEVs Affect the Cytokine Expression of APCs
THP-1 cells were differentiated into immature DCs and stimulated with CD40L and LPS [28]. The production of cytokines was assessed by cytometric bead assays using the cell culture supernatants. Concomitant incubation with sEVs that had been isolated from the A375 cell line, or patient-derived melanoma cells DMBC12 and DMBC21, increased the secretion of IL-12, as well as IL-6, while the levels of VEGF were slightly reduced (Figure 4). diagram) were used in the experiments. At least three independent experiments were conducted for each cell type and significant differences were marked by asterisks (Student's t-test). (c) The mRNA expression levels in DC cells derived from the peripheral blood of healthy donors treated with A375derived sEVs relative to untreated control cells. Mean values and standard deviations are shown from three independent experiments. The differences between samples after sEV exposure and untreated controls did not reach statistical significance (Student's t-test).

sEVs Affect the Cytokine Expression of APCs
THP-1 cells were differentiated into immature DCs and stimulated with CD40L and LPS [28]. The production of cytokines was assessed by cytometric bead assays using the cell culture supernatants. Concomitant incubation with sEVs that had been isolated from the A375 cell line, or patient-derived melanoma cells DMBC12 and DMBC21, increased the secretion of IL-12, as well as IL-6, while the levels of VEGF were slightly reduced (Figure 4).

sEVs Reduce T Cell Proliferation in Allogeneic Mixed Lymphocyte Reactions
To investigate whether the reduced expression of immune receptors would also have functional consequences, mixed lymphocyte reactions (MLRs) were performed. Co-culturing the PBMCs

sEVs Reduce T Cell Proliferation in Allogeneic Mixed Lymphocyte Reactions
To investigate whether the reduced expression of immune receptors would also have functional consequences, mixed lymphocyte reactions (MLRs) were performed. Co-culturing the PBMCs isolated from healthy donors resulted in the stimulation of T cell proliferation based on allogeneic reactivity. Labeling the cells of one donor with CFSE allowed for determining the proportion of cells undergoing cell division, as each cell division halves the mean fluorescence intensity. By gating on CD3-positive cells, the analysis was restricted to T cells. As shown in Figure 5, the presence of A375 sEVs reduced the percentage of proliferating T cells in MLRs slightly, but significantly (p = 0.045).
isolated from healthy donors resulted in the stimulation of T cell proliferation based on allogeneic reactivity. Labeling the cells of one donor with CFSE allowed for determining the proportion of cells undergoing cell division, as each cell division halves the mean fluorescence intensity. By gating on CD3-positive cells, the analysis was restricted to T cells. As shown in Figure 5, the presence of A375 sEVs reduced the percentage of proliferating T cells in MLRs slightly, but significantly (p = 0.045).

sEVs Transfer TGF-β with Immunosuppressive Activity
sEVs can transfer immunosuppressive cytokines, such as TGF-β. Control experiments using recombinant TGF-β showed that this cytokine was able to down-regulate the expression of MHC class I and class II and CD40 molecules (Figure 6a, compare the first two columns for each antigen). Therefore, we measured TGF-β directly in sEV preparations (Figure 6b). TGF-β could be identified in quantities sufficient to induce the observed effects on APCs. Recombinant TGF-β was used in a concentration of 20 pg/mL. In experiments assessing the inhibitory effect on immune receptor expression, the sEVs were used at a concentration of 20 µg protein per mL of medium. Consequently, about 1 pg TGF-β per µg protein should be sufficient to obtain similar effects as with the recombinant cytokine. The amount of TGF-β measured in the sEV preparations was in the range of 10-15 pg/µg protein (Figure 6b). TGF-β neutralizing antibodies, along with sEVs, reduced the sEV-mediated down-regulation of HLA-ABC, HLA-DR, CD40, and CD86 (Figure 6a, third and fourth columns). These results suggest that TGF-β participated in the down-regulation of immune receptors during the incubation with sEVs.  3.6. sEVs Transfer TGF-β with Immunosuppressive Activity sEVs can transfer immunosuppressive cytokines, such as TGF-β. Control experiments using recombinant TGF-β showed that this cytokine was able to down-regulate the expression of MHC class I and class II and CD40 molecules (Figure 6a, compare the first two columns for each antigen). Therefore, we measured TGF-β directly in sEV preparations (Figure 6b). TGF-β could be identified in quantities sufficient to induce the observed effects on APCs. Recombinant TGF-β was used in a concentration of 20 pg/mL. In experiments assessing the inhibitory effect on immune receptor expression, the sEVs were used at a concentration of 20 µg protein per mL of medium. Consequently, about 1 pg TGF-β per µg protein should be sufficient to obtain similar effects as with the recombinant cytokine. The amount of TGF-β measured in the sEV preparations was in the range of 10-15 pg/µg protein (Figure 6b). TGF-β neutralizing antibodies, along with sEVs, reduced the sEV-mediated down-regulation of HLA-ABC, HLA-DR, CD40, and CD86 (Figure 6a, third and fourth columns). These results suggest that TGF-β participated in the down-regulation of immune receptors during the incubation with sEVs.
undergoing cell division, as each cell division halves the mean fluorescence intensity. By gating on CD3-positive cells, the analysis was restricted to T cells. As shown in Figure 5, the presence of A375 sEVs reduced the percentage of proliferating T cells in MLRs slightly, but significantly (p = 0.045).

sEVs Transfer TGF-β with Immunosuppressive Activity
sEVs can transfer immunosuppressive cytokines, such as TGF-β. Control experiments using recombinant TGF-β showed that this cytokine was able to down-regulate the expression of MHC class I and class II and CD40 molecules (Figure 6a, compare the first two columns for each antigen). Therefore, we measured TGF-β directly in sEV preparations (Figure 6b). TGF-β could be identified in quantities sufficient to induce the observed effects on APCs. Recombinant TGF-β was used in a concentration of 20 pg/mL. In experiments assessing the inhibitory effect on immune receptor expression, the sEVs were used at a concentration of 20 µg protein per mL of medium. Consequently, about 1 pg TGF-β per µg protein should be sufficient to obtain similar effects as with the recombinant cytokine. The amount of TGF-β measured in the sEV preparations was in the range of 10-15 pg/µg protein (Figure 6b). TGF-β neutralizing antibodies, along with sEVs, reduced the sEV-mediated down-regulation of HLA-ABC, HLA-DR, CD40, and CD86 (Figure 6a, third and fourth columns). These results suggest that TGF-β participated in the down-regulation of immune receptors during the incubation with sEVs.

Discussion
The suppression of anti-cancer immune responses by tumors is central to their survival. Cancer-induced immunosuppression is antigen-specific rather than systemic and we provide evidence that melanoma-derived sEVs are able to induce TAA-specific tolerance. sEVs have to meet two requirements to function as mediators of TAA-specific immune suppression, as follows: They have to transfer MHC molecules from cancer cells to APCs and, simultaneously, they have to promote an immunosuppressive phenotype. We demonstrate that MHC class I receptor proteins are transported to APCs by sEVs where they appear at the cell surface ( Figure 2). In contrast, CD63 carried by sEVs did not reach the target cell surface (Figure 2d), suggesting that the intracellular distribution of vesicular proteins is regulated and only specific proteins become exposed at the cell surface.
The second requirement, the immunosuppressive activity of melanoma-derived extracellular vesicles, was demonstrated by down-regulation of MHC molecules and co-stimulatory receptors on monocytes and DCs (Figure 3), induced production of IL-6 ( Figure 4), and reduced T cell proliferation in MLRs ( Figure 5). The immunosuppressive activity of tumor-derived EVs has been well documented earlier. Exosomes triggered apoptosis of CD8 T cells, impaired cytotoxic T cell functions, and induced conversion of naïve T cells into Treg cells [31]. The latter can occur either directly through transported cytokines and metabolic enzymes [32,33] or indirectly involving MDSCs [32]. TGF-β and PgE2 are able to induce MDSCs in mice [34] and humans [35] and both factors were found to be transported by exosomes [34]. In humans, the additional contribution of Hsp70 and miRNAs was demonstrated [36][37][38]. Exosomes were implicated in antigen-specific tolerance, as exosomes isolated from the blood of antigen-fed mice or tumor-bearing mice were able to suppress delayed-type hypersensitivity (DTH) in an antigen-specific manner [39][40][41]. When isolated from the blood of tumor-bearing mice, exosomes suppressed the inflammation caused by a defined tumor antigen. The presence of MHC class II and FasL on these exosomes was of crucial importance for their suppressive activity [40].
In our experimental setting, the treatment with cancer-derived sEVs changed the receptor expression on APCs in different ways. First, the sEVs are allogeneic to the APCs, thus they exert a non-specific stimulation [42]. APC activation usually includes up-regulation of MHC molecules and co-stimulatory receptors, similar to incubation with LPS. Indeed, an elevated expression of CD40 and CD86 was observed in THP-1 cells exposed to melanoma-derived sEVs (Figure 3). In line with these results, Bretz et al. demonstrated that the stimulation of THP-1 cells by exosomes of cancerous and non-cancerous origin occurred through a TLR (toll-like receptor) dependent mechanism [42]. Furthermore, tumor-derived sEVs have been shown to activate primary monocytes [43]. We used sEVs isolated from a well-established melanoma cell line (A375), as well as from patient-derived primary melanoma cell lines. A second mechanism changing the expression level of MHC receptors is the direct transfer of HLA-ABC by sEVs to the surface of APCs. Third, vesicle-transported mediators, including cytokines and miRNAs, can change the expression of immune receptors. We demonstrated that TGF-β, one of the most powerful immunosuppressive cytokines, was present in sEV preparations in sufficient amounts to cause the observed reduction of MHC and CD40 expression ( Figure 6). Exosomes carry a latent form of TGF-β with full biological activity in their membrane [44]. We confirmed the potential of TGF-β to mediate the down-regulation of immune receptors and show that the presence of anti-TGF-β neutralizing antibodies reversed the sEV-mediated down-regulation of MHC molecules CD40 and CD86 ( Figure 6). The expression levels of HLA-ABC, CD40, and CD86 in the presence of anti-TGF-β neutralizing antibodies even exceeded the one of control cells, indicating that sEVs by themselves exert immune stimulatory effects that are outperformed by TGF-β mediated suppression. TGF-β transported by exosomes was recently shown to induce Treg cells in gastric cancer [45]. Furthermore, depletion of TGF-β from leukemia cell-derived exosomes increased the immune stimulatory potential of these vesicles, inducing potent anti-leukemic immunity [46]. We suggest that inhibition of DC maturation visualized by immune receptor down-regulation contributed to the observed immunosuppressive effects. Corroborating our findings, impairment of DC differentiation upon exposure to lung carcinoma-derived exosomes was recently demonstrated in mice [47].
While the protein expression was markedly modulated by sEV exposure, corresponding changes in the mRNA levels for HLA class I and II, CD40, and CD86 were not detected. The modulation of immune receptors seems to occur at a post-transcriptional level and might be caused by miRNAs contained in sEVs. MiRNAs inhibit translation of specific target mRNAs by binding to complementary sequences. MiRNAs from exosomes released by melanoma cells have been shown to play multiple roles in promoting tumor development [48]. In our microarray analysis, published recently, 386 common miRNAs were detected in sEVs derived from four primary melanoma cell lines [49].
Interestingly, contact with sEVs also led to the induction of cytokine secretion (Figure 4). We observed a strong up-regulation of pro-inflammatory IL-6 secretion by sEV-exposed APCs. This interleukin seems to play a major role in melanoma pathogenesis. Through an autocrine loop, including the activation of STAT3, IL-6 signaling contributes to survival and proliferation of melanoma cells, increases the metastatic potential, and facilitates immune evasion [50]. More specifically, IL-6 mediated STAT3 activation in DCs was shown to down-regulate MHC class II molecules, resulting in the inhibition of the functional maturation of DCs [51]. IL-6 promotes MDSC generation [52] and has been shown to impair T helper 1 cell differentiation [53]. In contrast to Il-6, which was virtually absent in control cell cultures, VEGF was secreted in relatively high amounts, also in the absence of sEVs. The incubation with sEVs left VEGF levels mostly unchanged. VEGF was shown to correlate with MDSC infiltration into tumor tissue [54].
In line with a predominant immunosuppressive function of melanoma-derived EVs, reduced T cell proliferation in MLRs was also measured ( Figure 5). Some of the observed effects were relatively mild, raising the question about their relevance in vivo. However, in our experiments, sEVs were applied only once while, in tumor-bearing organisms, a constant flow of sEVs may cause accumulative effects. Furthermore, in the patients, the sEVs produced by cancer cells are autologous and exert no allogeneic stimulation. We tried to avoid extreme sEV concentrations, which might raise non-specific effects. However, exploration of dose effects could further improve our understanding and, for the future, experiments are planned to demonstrate that the immunosuppressive effects observed in our study are highly relevant to the disease.

Conclusions
In summary, exposure to melanoma-derived sEVs shifts APCs towards an immunosuppressive phenotype, with down-regulated expression of immune receptors that are required for the optimal stimulation of a cytotoxic T cell response. Immunosuppression as cancer-associated deregulation involves many mechanisms and the antigen-specific immunosuppression through sEVs represents one plausible strategy.