Redox-signaling transmitted in trans to neighboring cells by melanoma-derived TNF-containing exosomes

Free Radical Biology & Medicine 43 (2007) 90 – 99 www.elsevier.com/locate/freeradbiomed

Original Contribution

Redox-signaling transmitted in trans to neighboring cells by melanoma-derived TNF-containing exosomes Anita Söderberg a,1 , Ana María Barral a,1 , Mats Söderström a , Birgitta Sander b , Anders Rosén a,⁎ a

b

Department of Biomedicine and Surgery, Division of Cell Biology, Linköpings Universitet, SE-58185 Linköping, Sweden Department of Laboratory Medicine, Division of Pathology, Karolinska University Hospital Huddinge, Karolinska Institutet, Stockholm, Sweden Received 31 January 2007; accepted 29 March 2007 Available online 31 March 2007

Abstract Hydrogen peroxide is known to be involved in redox signaling pathways that regulate normal processes and disease progression, including cytokine signaling, oxidative stress, and cancer. In studies on immune surveillance against cancer, hydrogen peroxide was found to disrupt cytotoxic T-cell function, thus contributing to tumor escape. In this study, secretion of TNF-containing vesicles of rab9+ endosomal origin, termed exosomes, was investigated using GFP-TNF constructs. We observed a polarized intracellular trafficking and apical secretion of TNF-positive nanovesicles. Cell-to-cell transfer of TNF was observed in exosomes in real-time microscopy, occurring separate from the melanin/melanosome compartment. Exosomes were prepared by ultracentrifugation or immunoisolation on anti-β2-microglobulin magnetic beads. TNF as well as TNF receptors 1 and 2 were present in the exosomes as determined by Western blot, flow cytometry, and deconvolution microscopy. The functional significance of melanoma-derived exosomes was established by their signaling competence with ability to generate significantly higher ROS levels in T cells compared with sham exosomes (P = 0.0006). In conclusion, we report here, for the first time, that TNF is found in tumor cellderived exosomes and that these exosomes transmit redox signaling in trans to neighboring cells. The results are of importance for a better understanding of tumor escape mechanisms. © 2007 Elsevier Inc. All rights reserved. Keywords: TNF; Exosomes; Melanoma; Hydrogen peroxide; Redox signaling; Tumor escape mechanisms

Introduction Hydrogen peroxide (H2O2) is generated in phagocytes as a host defense to inflict damage to engulfed microorganisms, but H2O2 also supports proliferation, differentiation, and migration and affects numerous intracellular signaling pathways including growth factors and cytokines such as plateletAbbreviations: β2m, β2-microglobulin; carboxy-H2DCFDA, (5-(and-6)carboxy-2′,7′-dichlorodihydrofluorescein diacetate; FCS, fetal calf serum; GFP, green fluorescent protein; H2O2, hydrogen peroxide; Ig, immunoglobulin; IP, immunoprecipitation; IL, interleukin; MFI, mean fluorescence intensity; mAb, monoclonal antibody; PBMC, peripheral blood mononuclear cells; PFA, paraformaldehyde; PMA, phorbol 12-myristate 13-acetate; ROS, reactive oxygen species; TNF, tumor necrosis factor; TACE, TNF-α-converting enzyme/ADAM17; TNFR, tumor necrosis factor receptor; Trx, thioredoxin; TrxR, thioredoxin reductase; WB, Western blot. ⁎ Corresponding author. Fax: +46 13 22 4314. E-mail address: [email protected] (A. Rosén). 1 These authors contributed equally to this study. 0891-5849/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.freeradbiomed.2007.03.026

derived growth factor, epidermal-derived growth factor, insulin, TNF, and interleukin-1 (IL-1) [1]. The role of cytokines, particularly TNF in malignant melanoma, has been studied extensively. While most studies address the production of TNF by immune cells that infiltrate the tumor tissue, melanoma cells have been shown to express TNF both in vitro and in vivo [2]. Intracellular expression of TNF in melanoma cells confers resistance against the TNF-mediated killing launched by invading immune cells, partly due to the down-regulation of TNF receptors (TNFRs) on the melanoma cells [3]. In vivo, the development of tumor-related anergy could be linked to the appearance of hydrogen peroxide induced oxidative stress that blocks effector T-cell function [4]. In addition, tumor-derived exosomes that inhibit T-cell functions have been isolated from patients' sera [5]. Exosomes are nanovesicles of 40 to 90 nm in size that are produced by various cells in the body under normal physiological conditions such as immune suppression during pregnancy [6–8] as well as during disease conditions, e.g., inflammation

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

and cancer [9]. Exosomes are also explored in dendritic cellbased tumor vaccines [10]. The expression of TNF in primary human melanomas in situ has been correlated to the decreased infiltration of CD3-positive cells [11]. Trying to explore the biological role of TNF in melanomas, we previously studied the expression of various cytokines and redox-active proteins such as thioredoxin 1 (Trx1) and Trx reductase 1 (TrxR1) in a panel of cancer cell lines [12–14], and found that despite a high intracellular expression, soluble TNF was not released into the supernatants [15] in contrast to macrophages that secreted high levels of TNF as well as redox-active proteins. In the present study, we wanted to explore (i) if TNF was correctly processed in malignant melanoma cells; (ii) if and how it was exported from melanoma cells; and (iii) a possible mechanism of immune regulation. We found that in the melanoma cell lines studied, the TNF-α-converting enzyme (TACE)/a disintegrin and metalloprotease domain 17 protein (ADAM17) was indeed present and TNF was correctly cleaved. However, melanoma-produced TNF was not released directly into the supernatant, but was concealed in exosome vesicles. The presence of both TNFR1 and TNFR2 was also revealed in these exosomes, confirming a previous study that found TNFR1 in exosomes released from human vascular endothelial cells [16]. Furthermore, we found that melanomareleased exosomes induced high levels of reactive oxygen species (ROS) in lymphocytes, suggesting a possible role of “immune counterattack” for tumor-derived exosomes. Materials and methods Cell cultures FM3 and FM55M2 malignant melanoma cell lines [17], THP1 monocytic cell line (ATCC: TIB-202), and Jurkat T-cell line (clone E6-1, ATCC: TIB-152) were maintained in an RPMI 1640 medium supplemented with 10% fetal calf serum (FCS). For stimulation, cells were treated with 50 nM phorbol 12myristate 13-acetate (PMA) for the indicated time. For microscopic studies, if not otherwise indicated, cells were seeded at a cell density of 0.5 × 105 cells/ml, onto round coverslips of 19 mm diameter, and placed in 12-well culture plates. Molecular biology reagents and materials Restriction endonucleases, oligonucleotides, enzymes, and cell culture reagents were purchased from Invitrogen (Paisley, UK), unless otherwise indicated. Deep Vent DNA polymerase was purchased from New England Biolabs (Ipswich, MA), and the vectors, pEGFP-N3 from Clontech (Mountain View, CA). The following mouse monoclonal antibodies (mAbs) were used in this study: anti-human TNF clone CY-014, purchased from Innogenetics (presently distributed by Triolabs AB, Gothenburg, Sweden); biotinylated anti-human TNF (clone BAF210, R&D Systems, Minneapolis, MN); antiTACE, clone M222, was donated by Dr. R. Black (Immunex

91

Corp., Seattle, WA); anti-FLAG (clone M2, Sigma-Aldrich Co, St. Louis, MO); anti-TNFR1, clone htr-9 and anti-TNFR2, clone utr 1 (BMA Biomedicals, Augst, Switzerland); antimelanosome, clone HMB-45 (DAKO, Glostrup, Denmark); anti-rab9, late endosomal marker (Abcam Ltd, Cambridge, UK); anti-golgin-97, Golgi apparatus marker, clone CDF4, A21270 (Molecular Probes/Invitrogen). As negative controls, mouse IgG1 and IgG2a Abs, as well as normal rabbit Ig (DAKO), were used. As secondary reagents, anti-mouse and anti-rabbit Ig conjugated with Alexa594 (Molecular Probes), FITC-conjugated anti-mouse Ig, biotinylated goat anti-mouse IgG, and FITC-conjugated streptavidine (DAKO) or streptavidine-Alexa488 (Molecular Probes) were used. For exosome isolation, rabbit anti-β2-microglobulin (anti-β2m) (DAKO, A0072) was used. Plasmid constructs and transfection Fig. 1B shows a schematic drawing of the plasmid constructs. (1) pEGFP-TNF fusion protein. A cDNA fragment containing the entire coding sequence of human TNF was generated by PCR using the plasmid pATHTNF (Accession Number LMBP2529, BCCM/LMBP, http://www.belspo.be/ bccm/lmbp.htm) as a template with the following primers: sense 5′-TGGAATTCACACCATGAGCACTGAAAGC-3′ (EcoRI site underlined) and antisense 5′-GTCGGGTACCCAGGGCAATGATCCCAAAG-3′ (KpnI site underlined). The resulting 722-bp product was digested with EcoRI and KpnI and subsequently subcloned in-frame into the pEGFP-N3 vector, digested in the same manner. (2) pEGFP-TNF-FLAG was generated by introducing a (FLAG)2 epitope at the Nterminal of the TNF sequence. The two following complementary oligonucleotides, encoding the sequence of the (FLAG)2 epitope, and containing restriction site overhangs for subcloning were used: sense sequence, 5′TCGACCATGGGAGACTACAAGGACGACGACGACAAGGAC TACAAGGACGACGACGACAAG-3′ (XhoI overhang underlined); antisense sequence, 5′-AATTCTTGTCGTCGTCGTCCTTGTAGTCCTTGTCGTCGTCGTCCTTGTAG TCTCCCATGG-3′ (EcoRI overhang underlined). After annealing, the double-stranded oligonucleotide was inserted in-frame into the pEGFP-TNF plasmid digested with EcoRI and XhoI. For transient transfection of melanoma cell lines, Lipofectamine (Invitrogen) was used according to the manufacturer's instructions. Cells were transfected with 1 μg DNA for 5 h and assayed 48 h posttransfection. Alternatively, stable transfectants were selected using an optimal concentration (600 μg/ml for FM55M2) of geneticin (Sigma-Aldrich Co.). Immunofluorescence and flow cytometry The paraformaldehyde (PFA)-fixation/saponin permeabilization method was used for intracellular staining, as previously described [18]. Cells were stained either in suspension for analysis by flow cytometry or on coverslips for microscopy. Primary Ab was used at a 5–10 μg/ml optimal concentration and an isotype control Ab was employed in parallel, at the same

92

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

concentration. For microscopy, the Alexa-conjugated secondary Ab was used. For flow cytometry, FITC-conjugated anti-mouse Ig was also employed. Coverslips were mounted face down with ProLong anti-fading mounting medium (Molecular Probes) on slides and sealed with nail varnish. For surface staining by flow cytometry, melanoma cells were detached with 10 mM EDTA. Cells were washed in PBS with 2% FCS and 0.1% NaN3. All washes and incubations were carried out on ice. Cells were sequentially incubated for 30 min with primary Ab

or negative control, biotinylated goat anti-mouse IgG, followed by a streptavidine conjugate. Cells or exosomes (isolated on beads) were processed for immunofluorescence as described above. Ten thousand cells (or 4 × 104 beads) were analyzed in a FACS Calibur flow cytometer (BD). Data were processed using the CellQuest software. Cells or exosome beads were observed under a Zeiss Axiovert 200 M inverted microscope, equipped with a 63×, 1.4 NA Plan Apochromat oil objective, or a 63×, 1.2 NA Plan Apochromat water objective. Image stacks of 20

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

planes with a Z-distance of a maximum 300 nm were collected with an Axiocam MRm CCD camera, using Zeiss AxioVision software. For certain analyses, images were collected in a Leica TCS NT confocal microscope equipped with a 63× water objective. The images were processed for deconvolution analysis in the Huygens software version 2.3.1a-64 (Scientific Volume Imaging), using an Octane Workstation (Silicon Graphic Institute). The maximum likelihood estimation (MLE) algorithm was used. Images were viewed and reconstructed in 3D, using an Imaris 3. 1. 3, from Bitplane AG. Final presentation and mounting of the images were performed using Adobe Photoshop version 7.0. Movies were edited using Adobe ImageReady. Time-lapse microscopy on live cells FM3 cells were grown on coverslips and transfected with pEGFP-TNF, as described. Forty-eight hours posttransfection, cells were gently washed with Krebs-Ringers buffer, and the coverslips were placed face down on a microculture slide with a well (Kebo, Stockholm, Sweden) containing 100 μl KRB with 50 nM PMA. The coverslips were sealed with nail varnish and observed under a microscope. The objective was equipped with an object heater to maintain the temperature at 37°C. Fluorescence images consisting of 20 Z-planes were acquired every 10 min. Immunoprecipitation (IP) and Western blot (WB) Proteins from cells or exosome preparations were solubilized in a lysis buffer containing protease inhibitors, IP was performed using Sepharose-protein A as previously described [13]. Shortly, immunoprecipitated proteins were separated on 4–20% SDS-PAGE gels and transferred to 0.2 μm PVDF membranes, blocked with StartingBlock buffer (Pierce Biotechnology) with 0.05% Tween 20 for 20 min. IP proteins were detected by WB using primary Abs: biotinylated anti-TNF, antiTNFR1, or anti-TNFR2 at a concentration of 0.2 μg/ml in StartingBlock for 1 h. The filters were washed three times with

93

PBS with 0.05% Tween 20 (wash solution) and incubated with streptavidine-HRP (20 ng/ml) in StartingBlock for 1 h followed by three washes. SuperSignal WestDura chemiluminescent substrate (Pierce) was used according to the manufacturer's protocol. Exosome preparations Exosomes were purified by filtration/ultracentrifugation as previously described [19]. Briefly, FM55M2 and FM3 human melanoma cells were grown to confluence in RPMI 1640 with 10% exosome-free FCS (prepared by ultracentrifugation at 100,000g overnight and 0.22 μm sterile-filtered). Supernatants from confluent cultures of the adherent melanoma cells were harvested (200 ml of 24 h medium). The supernatants were centrifuged for 5 min at 300g and filtered through a 0.22-μm sterile filter to remove large cellular debris. The supernatants were then centrifuged for 1 h at 100,000g and then removed. The pellets were washed by resuspension in 1 ml PBS and collected by centrifugation for 1 h at 100,000g. The pellets were resuspended in a small volume (50–100 μl) of PBS and the total protein concentration was determined. The exosomes were kept in aliquots at −70°C for Western blot analysis. As a control, we used a medium with FCS that had not been in contact with the cells, but was processed identically. Secondly, exosomes were purified by anti-β2m-conjugated magnetic beads according to an established method in which exosome purity was verified with transmission electron microscopy [20,21]. In short, 15 μg of exosomes was isolated from eighteen 75-cm2 flasks (with 25-ml medium) of confluent FM55M2 cells. Briefly, we used protein-A-coupled Dynabeads (Dynatech, Oslo, Norway) incubated with anti-β2m Ab in a phosphate buffer, pH 8.1 (106 beads to 3 μg Ab). Control beads with normal rabbit IgG were also prepared. For 24 h, the medium from 6 × 106 confluent FM55M2 cells was sequentially centrifuged at 100g, for 8 min and 2000g for 15 min. Supernatants were then filtered through a 0.22-μm sterile filter. Beads (3 × 105 anti-β2m) were incubated end over end for 2 h at room temperature with a medium from 6 × 106 cells (optimized

Fig. 1. Expression of TNF and pro-TNF in melanoma cell lines. (A) pro-TNF is correctly cleaved. FM3 cells were stimulated with 50 nM PMA for 15 and 60 min. Immunoprecipitations were performed with anti-TNF mAbs. These were then analyzed in SDS-PAGE and developed by Western blot for the presence of pro-TNF cleavage. Unstimulated cells (0 min) served as control. Pro-TNF (26 kDa) and mature cleaved TNF (17 kDa) are indicated by arrows. (B) Schematic diagram of TNFGFP constructs used in these studies. The gray box at the carboxy terminus of TNF indicates GFP fusion epitope that was used for localization experiments. An amino terminal prosequence is indicated (arrow). The solid box at the amino terminus of TNF indicates the presence of an epitope (FLAG)2 tag that was used to facilitate immunolocalization experiments. (C) TNF trafficking in endosomal compartments to apical dendrites analyzed before and after PMA stimulation. Upper left panel, control, FM3 cells prior to stimulation. The cells were grown on coverslips and transfected with pEGFP-TNF (green). After 48 h, cells were stained with the antiTACE mAb and anti-mouse Ig-Alexa594 (red). TACE is present on the membrane of the cells and in elongated microfilament structures. Scale bar: 10 μm. Upper right panel, TNF trafficking in FM3 cells after PMA stimulation. TNF redistribution to dendrites at cell–cell interphase positions. Triple-color immunofluorescence: TNF was detected in the endosomal compartment and at paracrine/juxtacrine positions as green distinct dots localized to the tips of the dendrites. TACE was stained by antiTACE plus Alexa594 conjugate (red). Nuclei were counterstained with DAPI (blue). Scale bar: 15 μm. Mid-panel left, TNF trafficking to apical dendrite positions after PMA stimulation (same conditions as described in upper right panel. TNF colocalizes with TNFR1 to apical dendrite positions. FM3 cells were stimulated for 1 h. Mid-panel right, TNFR1 in red (anti-TNFR1 and anti-mouse IgG-Alexa594), the green color is TNF-GFP, and the yellow color denotes areas of colocalization. Arrows indicate nanovesicles at apical positions. Lower left panel, TNF and melanosome compartments are distinct and separately localized. FM3 cells were PMA-stimulated for 1 h followed by fixation and two-color immunofluorescence development. Melanosomes were stained with the anti-melanosome mAb (HMB 45) and anti-mouse IgG-Alexa594 (red) and TNF-GFP was visualized in green. Distinct and separate localization patterns, no overlapping or colocalization between melanosomes and TNF could be observed (arrows). Lower right panel, TNF localizes to late endosome compartment. FM55M2 were stained with anti-Rab9, a late endosome marker, anti-mouse IgG-Alexa594 (red). TNF were stained with anti-TNF biotin and streptavidine-Alexa488 (green) and analyzed by deconvolution microscopy and 3D reconstitution by Imaris software.

94

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

ratio). Beads were washed three times and fixed in PBS with 4% PFA prior to immunofluorescence or FACS experiments. Exosome labeling and binding to lymphocytes Exosomes purified by anti-β2m-conjugated magnetic beads were stained with PKH-26 (PKH26 Red Fluorescent Cell Linker Kit) (Sigma-Aldrich) according to the manufacturer's protocol. The labeled exosomes were eluted from the magnetic beads by lowering the pH to 4.0. The exosomes were then dialyzed against PBS on a floating filter for 1 h (Millipore 0.025 μm). The exosomes were added to 50,000 lymphocytes (Jurkat, peripheral blood mononuclear cells (PBMCs), or purified T cells from healthy blood donors) and cultured for 1 to 5 h. The cells were washed three times with PBS. ROS detection in exosome-exposed T cells ROS were detected in T cells by Image-iT LIVE green ROS detection kit 5-(and-6)-carboxy-2′,7′-dichlorodihydrofluorescein diacetate (carboxy-H2DCFDA) from Molecular Probes. T cells were purified by negative selection using the Dynal T-cell negative isolation kit (Dynal) according to the manufacturer's instruction and incubated with the ROS marker. Exosomes, purified from 18 × 106 FM55M2 cells, were labeled with PKH26, as described above, and then added to 1 × 105 T cells for 1 h. The cells were then physically sorted for high ROS level (top 10%) in a FACSAria flow cytometer (BD). The cells expressing high levels of ROS were visually analyzed in microscopy (multicolor immunofluorescence) for exosome uptake (red color) and burst of ROS (green color). In immunofluorescence overlay image dual-labeling (red and green) was represented by yellow color. Alternatively, ROS was detected by chemiluminescence (CL) as previously described [22]. Briefly, cells were incubated at 37°C in a six-channel Biolumat LB9505 (Berthold Co.) instrument. PBMC (95% lymphocytes) were isolated by FicollHypaque (GE Healthcare/Pharmacia Biotech, Uppsala, Sweden). The reaction mixture contained 1 ml of cell suspension with 1 × 106 PBMC in PBS, 5 × 10-5 M luminol, and 4 units/ml HRP. The cells were stimulated with TNF (1 or 100 ng/ml) or exosomes from 24 h supernatants of FM55M2 melanoma cells prepared in the same way as described above, with rabbit antiβ2m-coated beads. As control, we used rabbit normal IgGcoated beads as well as sham exosomes that were derived (under identical conditions to exosome isolation) from cell-free FCScontaining medium. In addition, T cells (not exposed to exosomes) were included as controls. The chemiluminescence response is expressed as cpm and calculated from the integral surface area under each curve. Statistics Statistical evaluation of the data was carried out with the JMP statistics package (SAS Institute, Cary, NC) operating on Macintosh and Windows computers. Each experiment was repeated 3 times.

Results In a majority of cells that secrete TNF, the cytokine is reported to be produced as a 26-kDa precursor and cleaved by TACE to its mature 17-kDa form [23]. To investigate if this also occurs in malignant melanoma, we analyzed the cleavage pattern of TNF using Western blots of PMA-stimulated FM3 cells (Fig. 1A), where both the unprocessed and the processed form of TNF can be observed. The ratio between the mature 17kDa TNF and the 26-kDa proTNF form increased on PMA stimulation in a time-dependent manner (0, 15, 60 min). However, very little TNF was secreted into the culture medium, in line with our earlier findings [15]. TACE could be detected both intracellularly in permeabilized cells and on the surface of nonpermeabilized FM3 melanoma cells using flow cytometry. The mean fluorescence intensity of TACE in FM3 cells was 154 and 96, for permeabilized and nonpermeabilized cells, respectively. The THP-1 monocytic cell line was used as a positive control for TACE (supplemental data Fig. 1). In order to understand the apparent secretory block, we analyzed the TNF and TNFR intracellular trafficking and distribution. Fig. 1B shows the plasmid constructs used to study the intracellular topology. The GFP-TNF pattern was identical to endogenous TNF as verified by double staining with an anti-TNF mAb (data not shown). No TNF was secreted into the supernatant as measured by ELISA (< 2 pg/107 confluent cells). The TNF topological position in unstimulated cells (Fig. 1C, upper left panel) was imaged in two-color immunofluorescence of FM3 cells: TACE is shown in red (Alexa594) and TNF in green (GFP-TNF). On stimulation with 50 nM PMA, TNF was promptly redistributed and after 10 min, a polarized apical trafficking to the dendrites was observed (Fig. 1C, upper right, middle left image). Seventy to 80% of dendrites were positive for GFPTNF after stimulation, whereas unstimulated cells showed a very low frequency of TNF-positive endosomal vesicles (< 5%). TNFR1 was analyzed in melanoma cells and found to be localized to the dendrites following 10 min of PMA stimulation (Fig. 1C, middle right image). We then asked whether TNF trafficking was separate from the melanosome compartment. Melanin pigment synthesis is a unique function of melanocytes. Melanin is stored and transported in special melanosome organelles that move toward the dendrites and are transferred to neighboring keratinocytes, a process promoted by the keratinocyte growth factor [24]. We explored whether TNF followed a similar pathway, and stained GFP-TNF-transfected FM3 cells with the HMB-45 melanosome marker. Colocalization of TNF and melanosomes was not observed either in unstimulated or in stimulated FM3 cells (Fig. 1C, lower left image). Moreover, transmission electron microscopy using immunogold staining for TNF in FM55M2 cells corroborated that TNF was localized outside the melanosomes (results not shown). Endogenous TNF (in nontransfected FM55M2 cells) was localized to late endosome compartment as judged by its colocalization with rab9 (a marker for late endosomes). Fig. 1C, lower right panel, shows a deconvolved 3D-image reconstituted

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

micrograph of TNF (green) colocalizing with rab9 (red). Similar distribution was found in GFP-TNF-transfected cells. (3D viewing is available in the online supplemental video 1). Golgin-97, a Golgi apparatus marker, did not colocalize with TNF (data not shown). Viable FM55M2 cells were transfected with GFP-TNF and stimulated with PMA. The live cells were observed in real-time under an inverted microscope at 630× magnification. Transfer of GFP-TNF nanovesicles to neighboring cells could be observed after 50 min (online supplemental video 2). Tumor cells, particularly melanomas, are known to secrete nanovesicles of 40–90 nm size, called exosomes. We investigated whether FM3 and FM55M2 melanoma cell linesecreted exosomes contained TNF. Surprisingly, mature TNF and both TNF receptors could be detected in exosome preparations of the cells (Figs. 2A, B, and C) using a previously published purification protocol that yields high-quality exosomes as certified by electron microscopy and proteomic analysis [20]. Only the full-length forms of both receptors

95

could be detected in the Western blots, suggesting an alternative mechanism of soluble receptor generation. To further corroborate this finding, exosomes were immunoisolated using magnetic beads conjugated with rabbit anti-β2m according to a previously established method [21] (or control beads conjugated with normal rabbit IgG), followed by staining for TNF, TNFR1, and TNFR2. Fig. 2D shows microscope images of magnetic bead-isolated melanoma-derived exosomes. Control beads containing exosomes, but stained with an isotype control mouse IgG, as a primary Ab, were completely negative (left panel). Also, control beads conjugated with normal rabbit IgG were negative, showing background bead autofluorescence only. TNF-stained exosomes showed several distinct fluorescentlabeled dots of 40–90 nm size. Fluorescent microspheres (Molecular Probes) of 200 nm were used as size markers. FACS analysis of the beads, in parallel, confirmed that TNFR2 was most positive with mean fluorescence intensity = 13.6 compared to TNFR1 (MFI = 6.1) and TNF (MFI = 5.2) (supplemental data Fig. 2).

Fig. 2. Exosomes contain TNF, TNFR1, and TNFR2 as analyzed by immunoprecipitation. (A) Exosomes purified from 24 h supernatant of confluent cultures of FM3 melanoma cells. Exosomes were immunoprecipitated and analyzed in SDS-PAGE and WB using anti-TNF. Lane 1 (rTNF), recombinant human TNF. Lane 2 (Exo), exosome preparation. Lane 3 (Ctr), medium control with FCS (sham exosomes prepared in an identical manner to melanoma exosomes). (B) Exosomes purified as above and analyzed by IP by anti-TNFR1. Lane 1 (Ctr), medium control with FCS (sham exosomes). Lane 2 (Exo), FM3 exosome preparation. (C) Exosomes purified as above and analyzed by IP in SDS-PAGE by anti-TNFR2. Lane 1 (Ctr), sham exosomes. Lane 2 (Exo), FM3 exosome preparation. (D) Micrograph showing FM55M2derived exosomes that express TNF, TNFR1, and TNFR2. Exosomes were isolated from a conditioned medium of 6 × 106 confluent FM55M2 cells by binding to Dynatech magnetic beads conjugated with anti-β2m (106 beads per 3 μg Ab). The exosomes were stained with anti-TNF, anti-TNFR1, and anti-TNFR2 mAbs, respectively, and developed with Alexa conjugate Ab. Arrowheads indicate TNF-, TNFR1-, and TNFR2-positive exosomes. Control (IgG) show beads with exosomes stained with an isotype IgG primary Ab (negative control). Fluorescent microspheres (Molecular Probes) of 200 nm were used as size markers (not shown in micrograph). Magnetic beads are visible as dim green autoflourescent spheres. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

96

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

Fig. 3. Exosome uptake and ROS generation in T-lymphocytes. (A) Exosomes captured by T-lymphocytes. Exosomes, immunoisolated from FM55M2 melanoma cells, on magnetic beads were labeled with red PKH26. The photograph shows the capture of exosomes by T-lymphocytes (Jurkat) after 5 h incubation at 37°C, observed in two-color immunofluorescence with differential interference contrast (DIC) microscopy. (B) Chemiluminescence response generated by exosomes. Exosomes (isolated from cell-free supernatant of FM55M2 melanoma cells) were added to Ficoll-Hypaque-isolated peripheral blood lymphocytes at 37°C in a Biolumat instrument and the chemiluminescence (CL) response was monitored for 40 min. The control preparation was derived (under identical conditions to exosome isolation) from cell-free FCS-containing medium (sham exosomes). (C) Flow cytometry determination of ROS in T cells after exposure to melanoma exosomes. Analysis of peripheral blood T-lymphocytes was performed in a BD FACSAria instrument using the carboxy-H2DCFDA reagent. Cells were exposed to: (1) FM55M2 melanoma exosomes isolated by anti-β2m magnetic beads. (2) Control eluate from beads coated with isotype control IgG. (3) Control buffer, no exosomes. (4) Sham exosomes isolated from FCS (10% in RPMI 1640 medium). Statistical diagram diamonds were generated in JMP program indicating mean values and 95% upper and lower confidence interval. Analysis of variance was used for P value calculations. Three separate experiments were performed. (D) Micrograph showing multicolor fluorescence for detection of ROS in T-lymphocytes after exosomal uptake. Peripheral T-lymphocytes from healthy blood donors were exposed to exosomes isolated from FM55M2 melanoma cells and labeled with red-emitting PKH26. The burst of ROS in T-lymphocytes were measured by Image-iT LIVE Green and the cells were sorted in a BD FACSAria. The top 10% cells were sorted for observation in a Zeiss Axiovert 200M microscope equipped with differential interference contrast (DIC) device using a 40× objective. Merged image: yellow color indicates coexpression of ROS-increase and exosome-uptake. Scale bars, 10 μm.

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

The functional significance of exosomes was investigated by analysis of H2O2 redox signaling. First, we investigated exosomal uptake into lymphocytes. T-lymphocytes (Jurkat and healthy blood donor T cells) were found to actively bind melanoma-derived exosomes. Fig. 3A shows that prelabeled exosomes (red fluorescent PKH26) bound to Jurkat T cells. The kinetics of the uptake revealed that after 1 h, 16% of the cells had captured exosomes, increasing to 26% of the cells after 24 h. Blood donor T cells showed uptake in the same range (data not shown). Secondly, we investigated the functional consequence of exosome binding to target cells by analyzing H2O2 generation. Freshly isolated lymphocytes were exposed to FM55M2 exosomes. We found that high ROS levels were generated during a 40-min time interval with FM55M2 exosomes compared with sham exosomes isolated from FCS (Fig. 3B). Control experimental samples run in parallel showed that 1 ng and 100 ng/ml of recombinant human TNF induced a ROS burst with peak values of 2.9 × 106 and 5.7 × 106 cpm, respectively. The amount of 0.5 μg of exosomes (total protein) showed a 5.9 × 106 cpm peak response value, which is comparable to the TNFinduced levels. Thirdly, isolated exosomes were labeled with PKH-26 (red emission) and added to freshly isolated carboxy-H2DCFDA labeled T-lymphocytes for 1 h. The T-lymphocytes were analyzed and physically sorted in a FACSAria flow cytometer for ROS content (green emission). Significantly higher ROS levels were detected in T cells exposed to exosomes as compared with controls that were exposed to sham exosomes or buffer controls (P = 0.0006) (Fig. 3C). Microscopic inspection of these isolated T cells with high ROS levels confirmed that the majority of these cells had captured exosomes, i.e., duallabeled cells (Fig. 3D). In comparison, in controls T cells (sham exosome treated), only a minor fraction were double-labeled. Discussion The main and novel finding in this study is that TNF, TNFR1, and TNFR2 were exported from melanoma cells in exosomes and actively transferred to neighboring cells, and that these exosomes were functional in generating a significant ROS burst. Our previous finding that cultured normal human melanocytes and a panel of human melanoma cell lines expressed high levels of intracellular TNF suggested a conventional secretory pathway for this cytokine. However, we were not able to detect TNF in the supernatant of PMA-stimulated cells by ELISA, in contrast to control monocyte/macrophage cell lines U937 and THP-1 that secreted considerable quantities [15]. Therefore, we investigated the mechanism behind what we considered a secretory block. We have clarified in this study that the topological compartment localization of TNF is mainly late endosomes, as detected by rab9 colocalization and to a much lesser extent Golgi apparatus (golgin-97 colocalization) in unstimulated cells. Correct processing of TNF by the TACE enzyme was also found in that the 26-kDa pro-form of TNF cleaved into a 17-kDa

97

mature form. Preliminary studies using double-tagged (FLAGGFP) TNF constructs suggest that the TNF reaching the dendrites is already cleaved (results not shown). However, where this cleavage exactly takes place intracellularly remains unconfirmed. Interestingly, recent parallel studies show that synovial fibroblast-derived exosomes contain a membrane form of TNF [25]. Real-time imaging of PMA-stimulated GFP-TNF-transfected melanoma cells showed exosome transfer of TNF to the neighboring cells (video available online). This transfer follows a separate intracellular route, distinct from the melanin/ melanosome pathway, as shown by the nonoverlapping localization patterns of TNF and the HMB-45 melanosome marker (Fig. 1C, lower left panel). The size of the GFP-TNF dots corresponded to the size of the exosome nanovesicles. Exosomes are nanoparticles with a diameter of 40–90 nm that originate in the multivesicular bodies of the endosomal system [8]. Among other functions, they are considered a novel way of delivering cytokines and soluble cytokine receptors [16]. Melanomas, in particular, are known to release exosomes rich in tumor-associated antigens [26]. In the present study, TNF was found in exosome preparations from supernatants of melanoma cell lines, suggesting a novel mechanism of cell–cell transfer of this cytokine. Uptake of exosomes by dendritic cells seems to be mediated by diverse molecules such as milk fat globule-E8/lactadherin, CD11a, CD54, phosphatidylserine, and the tetraspanins, CD9 and CD81, on the exosome; as well as alpha(v)/beta(3) integrin and CD11a and CD54 on dendritic cells [27]. Further studies are required to determine the types of molecules that are involved in this process in melanomas. To our knowledge, this is the first time the presence of TNF and full-length TNFR2 in melanoma-derived exosomes is described. We also confirmed the uptake of exosomes into human lymphocytes. TNFR1 has been found recently in exosomes derived from endothelial cells and human serum [16]. Tumor-derived exosomes were previously shown to carry molecules of the TNFR superfamily such as FasL that inhibit T-cell function [5]. In this study the founding superfamily members TNFR1 and TNFR2 were found. In particular, TNFR2, which has been shown to play an important role in T-cell activation [28], seems to form part of the tumor “counterattack” against the immune system. Exosomes contained processed TNF, whereas TNF receptors were present in full-length form. This might suggest different processing mechanisms or a particular set of molecular exosome composition. The functional significance of the exosomes was revealed by the generation of H2O2 in lymphocytes that had captured the exosomes. Several lymphocyte receptors such as TNF-superfamily receptors [29] including FasL [30], dopamine receptors [31], Toll-like receptors [1,32], and B-cell receptors [33] are known to generate H2O2 as a signaling mediator [34]. Several reports describe the functional down-regulation of the T-lymphocyte receptor (TCR) signaling caused by a zeta-chain inactivation by ROS [4,35]. Conceptually, melanoma-derived TNF could play two roles, in line with the remarkable functional duality of TNF that has been reported [36]. First, the transfer and delivery of TNF as a

98

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99

cytokine and growth factor to neighboring melanoma cells in a paracrine/juxtacrine manner may be important for the induction of survival signals via the NF-κB pathway, similar to that described for IL15 and IL15R [37]. On the other hand, the release of TNF might affect the tumorinfiltrating T-lymphocytes, as strongly indicated by the uptake of exosomes into T-lymphocytes from healthy blood donors or Jurkat T-cell line (Fig. 3). In conclusion, exosomes contain TNF, as shown in this study, and FasL, as shown by others [5]. The relative contribution of TNF and FasL to the ROS generation renders further studies. The exosomes are instrumental in the induction of oxidative stress, a condition recently reported to uncouple the TCR signaling pathway, which in turn leads to down-regulated T-cell numbers. Our previous finding that TNF-positive melanomas in vivo showed a decreased infiltration of CD3+ T lymphocytes [11] supports this notion. The biological role of TNF in the tumor microenvironment of malignant melanomas may thus partly be explained by the release of ROS-generating exosomes, which may promote a tumor escape mechanism.

[8] [9]

[10]

[11]

[12]

[13]

[14]

[15]

Acknowledgments The authors thank Ms. Birgitta Axelsson, Ms. Petra Cassel, Dr. Maria Kvarnström for their excellent technical assistance. Dr. Birger Christensson for the stimulating discussions, and Dr. Florence Sjögren for excellent assistance with the FACSAria. This work was supported by funds from the Swedish Research Council, County Council of Östergötland, and Östergötland County Cancer Research Fund (A.M.B.), and the Swedish Cancer Association, Grants 3930-200-04xBB (M.S.) and 3171-B00-12xAC (A.R.). GSD.

[16]

[17]

[18]

[19]

Appendix A. Supplementary data [20]

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.freeradbiomed.2007.03.026. References [1] Rhee, S. H2O2, a necessary evil for cell signaling. Science 312: 1882–1883; 2006. [2] Moretti, S.; Pinzi, C.; Spallanzani, A.; Berti, E.; Chiarugi, A.; Mazzoli, S.; Fabiani, M.; Vallecchi, C.; Herlyn, M. Immunohistochemical evidence of cytokine networks during progression of human melanocytic lesions. Int. J. Cancer 84:160–168; 1999. [3] Decoster, E.; Cornelis, S.; Vanhaesebroeck, B.; Fiers, W. Autocrine tumor necrosis factor (TNF) and lymphotoxin (LT) alpha differentially modulate cellular sensitivity to TNF/LT-alpha cytotoxicity in L929 cells. J. Cell Biol. 143:2057–2065; 1998. [4] Takahashi, A.; Hanson, M. G.; Norell, H. R.; Havelka, A. M.; Kono, K.; Malmberg, K. J.; Kiessling, R. V. Preferential cell death of CD8+ effector memory (CCR7-CD45RA-) T cells by hydrogen peroxide-induced oxidative stress. J. Immunol. 174:6080–6087; 2005. [5] Taylor, D. D.; Gercel-Taylor, C. Tumour-derived exosomes and their role in cancer-associated T-cell signalling defects. Br. J. Cancer 92:305–311; 2005. [6] Taylor, D. D.; Akyol, S.; Gercel-Taylor, C. Pregnancy-associated exosomes and their modulation of T cell signaling. J. Immunol. 176: 1534–1542; 2006. [7] Akyol, S.; Gercel-Taylor, C.; Reynolds, L. C.; Taylor, D. D. HSP-10 in

[21]

[22]

[23]

[24]

[25]

[26]

ovarian cancer: expression and suppression of T-cell signaling. Gynecol. Oncol. 101:481–486; 2006. Barral, A.; von Herrath, M. Exosomes: specific intercellular nano-shuttles? Curr. Immunol. Rev. 1:1–6; 2005. Whiteside, T. L. Tumour-derived exosomes or microvesicles: another mechanism of tumour escape from the host immune system? Br. J. Cancer 92:209–211; 2005. Thery, C.; Regnault, A.; Garin, J.; Wolfers, J.; Zitvogel, L.; RicciardiCastagnoli, P.; Raposo, G.; Amigorena, S. Molecular characterization of dendritic cell-derived exosomes. Selective accumulation of the heat shock protein hsc73. J. Cell Biol. 147:599–610; 1999. Sander, B.; Boeryd, B. Tumor necrosis factor-alpha expression in human primary malignant melanoma and it relationship to tumor infiltration by CD3+ cells. Int. J. Cancer 66:42–47; 1996. Sahaf, B.; Söderberg, A.; Ekerfelt, C.; Paulie, S.; Rosén, A. Enzyme-linked immunospot assay for detection of thioredoxin and thioredoxin reductase secretion from cells. Methods Enzymol. 353:22–35; 2002. Söderberg, A.; Sahaf, B.; Rosén, A. Thioredoxin reductase, a redox-active selenoprotein, is secreted by normal and neoplastic cells: presence in human plasma. Cancer Res. 60:2281–2289; 2000. Sahaf, B.; Söderberg, A.; Spyrou, G.; Barral, A. M.; Pekkari, K.; Holmgren, A.; Rosén, A. Thioredoxin expression and localization in human cell lines: detection of full-length and truncated species. Exp. Cell Res. 236:181–192; 1997. Barral, A.; Källström, R.; Sander, B.; Rosén, A. Thioredoxin, thioredoxin reductase and TNFα expressed in melanoma cells: correlation to resistance against cytotoxic attack. Melanoma Res. 10:331–343; 2000. Hawari, F. I.; Rouhani, F. N.; Cui, X.; Yu, Z. X.; Buckley, C.; Kaler, M.; Levine, S. J. Release of full-length 55-kDa TNF receptor 1 in exosome-like vesicles: a mechanism for generation of soluble cytokine receptors. Proc. Natl. Acad. Sci. USA 101:1297–1302; 2004. Kirkin, A. F.; Petersen, T. R.; Olsen, A. C.; Li, L.; thor Straten, P.; Zeuthen, J. Generation of human-melanoma-specific T lymphocyte clones defining novel cytolytic targets with panels of newly established melanoma cell lines. Cancer Immunol. Immunother. 41:71–81; 1995. Sander, B.; Andersson, J.; Andersson, U. Assessment of cytokines by immunofluorescence and the paraformaldehyde-saponin procedure. Immunol. Rev. 119:65–93; 1991. Raposo, G.; Nijman, H. W.; Stoorvogel, W.; Liejendekker, R.; Harding, C. V.; Melief, C. J.; Geuze, H. J. B lymphocytes secrete antigenpresenting vesicles. J. Exp. Med. 183:1161–1172; 1996. Clayton, A.; Court, J.; Navabi, H.; Adams, M.; Mason, M. D.; Hobot, J. A.; Newman, G. R.; Jasani, B. Analysis of antigen presenting cell derived exosomes, based on immuno-magnetic isolation and flow cytometry. J. Immunol. Methods 247:163–174; 2001. Clayton, A.; Turkes, A.; Dewitt, S.; Steadman, R.; Mason, M. D.; Hallett, M. B. Adhesion and signaling by B cell-derived exosomes: the role of integrins. FASEB J. 18:977–979; 2004. Johansson, A.; Dahlgren, C. Characterization of the luminol-amplified light-generating reaction induced in human monocytes. J. Leukoc. Biol. 45:444–451; 1989. Black, R. A.; Rauch, C. T.; Kozlosky, C. J.; Peschon, J. J.; Slack, J. L.; Wolfson, M. F.; Castner, B. J.; Stocking, K. L.; Reddy, P.; Srinivasan, S.; Nelson, N.; Boiani, N.; Schooley, K. A.; Gerhart, M.; Davis, R.; Fitzner, J. N.; Johnson, R. S.; Paxton, R. J.; March, C. J.; Cerretti, D. P. A metalloproteinase disintegrin that releases tumour-necrosis factoralpha from cells. Nature 385:729–733; 1997. Cardinali, G.; Ceccarelli, S.; Kovacs, D.; Aspite, N.; Lotti, L. V.; Torrisi, M. R.; Picardo, M. Keratinocyte growth factor promotes melanosome transfer to keratinocytes. J. Invest. Dermatol. 125:1190–1199; 2005. Zhang, H. G.; Liu, C.; Su, K.; Yu, S.; Zhang, L.; Zhang, S.; Wang, J.; Cao, X.; Grizzle, W.; Kimberly, R. P. A membrane form of TNF-alpha presented by exosomes delays T cell activation-induced cell death. J. Immunol. 176: 7385–7393; 2006. Mears, R.; Craven, R. A.; Hanrahan, S.; Totty, N.; Upton, C.; Young, S. L.; Patel, P.; Selby, P. J.; Banks, R. E. Proteomic analysis of melanomaderived exosomes by two-dimensional polyacrylamide gel electrophoresis and mass spectrometry. Proteomics 4:4019–4031; 2004.

A. Söderberg et al. / Free Radical Biology & Medicine 43 (2007) 90–99 [27] Morelli, A. E.; Larregina, A. T.; Shufesky, W. J.; Sullivan, M. L.; Stolz, D. B.; Papworth, G. D.; Zahorchak, A. F.; Logar, A. J.; Wang, Z.; Watkins, S. C.; Falo, L. D., Jr.; Thomson, A. W. Endocytosis, intracellular sorting, and processing of exosomes by dendritic cells. Blood 104:3257–3266; 2004. [28] Kim, E. Y.; Teh, H. S. Critical role of TNF receptor type-2 (p75) as a costimulator for IL-2 induction and T cell survival: a functional link to CD28. J. Immunol. 173:4500–4509; 2004. [29] Xue, X.; Piao, J. H.; Nakajima, A.; Sakon-Komazawa, S.; Kojima, Y.; Mori, K.; Yagita, H.; Okumura, K.; Harding, H.; Nakano, H. Tumor necrosis factor alpha (TNFalpha) induces the unfolded protein response (UPR) in a reactive oxygen species (ROS)-dependent fashion, and the UPR counteracts ROS accumulation by TNFalpha. J. Biol. Chem. 280: 33917–33925; 2005. [30] Lu, B.; Wang, L.; Stehlik, C.; Medan, D.; Huang, C.; Hu, S.; Chen, F.; Shi, X.; Rojanasakul, Y. Phosphatidylinositol 3-kinase/Akt positively regulates Fas (CD95)-mediated apoptosis in epidermal Cl41 cells. J. Immunol. 176: 6785–6793; 2006. [31] Meredith, E. J.; Holder, M. J.; Rosén, A.; Lee, A. D.; Dyer, M. J.; Barnes, N. M.; Gordon, J. Dopamine targets cycling B cells independent of

[32]

[33] [34] [35]

[36] [37]

99

receptors/transporter for oxidative attack: implications for non-Hodgkin's lymphoma. Proc. Natl. Acad. Sci. USA 103:13485–13490; 2006. Matsuzawa, A.; Saegusa, K.; Noguchi, T.; Sadamitsu, C.; Nishitoh, H.; Nagai, S.; Koyasu, S.; Matsumoto, K.; Takeda, K.; Ichijo, H. ROS-dependent activation of the TRAF6-ASK1-p38 pathway is selectively required for TLR4-mediated innate immunity. Nat. Immunol. 6:587–592; 2005. Reth, M. Hydrogen peroxide as second messenger in lymphocyte activation. Nat. Immunol. 3:1129–1134; 2002. Tonks, N. K. Redox redux: revisiting PTPs and the control of cell signaling. Cell 121:667–670; 2005. Kono, K.; Salazar-Onfray, F.; Petersson, M.; Hansson, J.; Masucci, G.; Wasserman, K.; Nakazawa, T.; Anderson, P.; Kiessling, R. Hydrogen peroxide secreted by tumor-derived macrophages down-modulates signaltransducing zeta molecules and inhibits tumor-specific T cell-and natural killer cell-mediated cytotoxicity. Eur. J. Immunol. 26:1308–1313; 1996. Aggarwal, B. B. Signalling pathways of the TNF superfamily: a doubleedged sword. Nat. Rev. Immunol. 3:745–756; 2003. Dubois, S.; Mariner, J.; Waldmann, T. A.; Tagaya, Y. IL-15Ralpha recycles and presents IL-15 In trans to neighboring cells. Immunity 17:537–547; 2002.