The α_vβ₆ Integrin Is Transferred Intercellularly via Exosomes

Background: The α_vβ₆ integrin is up-regulated in cancer progression and metastasis.

Results: This integrin is transferred via exosomes from prostate cancer cells expressing α_vβ₆ into α_vβ₆-negative cancer cells.

Conclusion: Exosomes containing the α_vβ₆ integrin impact neighboring cells by enhancing their ability to migrate.

Significance: Exosome transfer of α_vβ₆ integrin among cancer cells may contribute to a horizontal propagation of more aggressive integrin-associated phenotypes.

Abstract

Exosomes, cell-derived vesicles of endosomal origin, are continuously released in the extracellular environment and play a key role in intercellular crosstalk. In this study, we have investigated whether transfer of integrins through exosomes between prostate cancer (PrCa) cells occurs and whether transferred integrins promote cell adhesion and migration. Among others, we have focused on the α_vβ₆ integrin, which is not detectable in normal human prostate but is highly expressed in human primary PrCa as well as murine PrCa in Pten^pc−/− mice. After confirming the fidelity of the exosome preparations by electron microscopy, density gradient, and immunoblotting, we determined that the α_vβ₆ integrin is actively packaged into exosomes isolated from PC3 and RWPE PrCa cell lines. We also demonstrate that α_vβ₆ is efficiently transferred via exosomes from a donor cell to an α_vβ₆-negative recipient cell and localizes to the cell surface. De novo α_vβ₆ expression in an α_vβ₆-negative recipient cell is not a result of a change in mRNA levels but is a consequence of exosome-mediated transfer of this integrin between different PrCa cells. Recipient cells incubated with exosomes containing α_vβ₆ migrate on an α_vβ₆ specific substrate, latency-associated peptide-TGFβ, to a greater extent than cells treated with exosomes in which α_vβ₆ is stably or transiently down-regulated by shRNA or siRNA, respectively. Overall, this study shows that exosomes from PrCa cells may contribute to a horizontal propagation of integrin-associated phenotypes, which would promote cell migration, and consequently, metastasis in a paracrine fashion.

Introduction

Prostate cancer (PrCa)³ is the most common cancer among males in the United States with 233,000 new cases and 29,480 deaths estimated for 2014 (1). For PrCa patients, the aggressive metastatic disease reflects the most adverse clinical outcome and is difficult to study given the heterogeneous nature of this malignant tumor (2, 3). At metastatic sites, predominantly bone, altered interaction with the prostate microenvironment promotes the ability of tumor cells to migrate (4, 5). Understanding the molecular events involved in the development of metastatic PrCa has the potential to identify new biological targets that can aid in prognosis and development of more effective therapies.

Exosomes are small vesicles derived from the intraluminal membrane budding of multivesicular bodies and are constitutively released into the extracellular environment through fusion with the plasma membrane (6, 7). Exosomes are defined predominantly by size (40–100 nm) and enrichment of specific lipid-raft proteins, such as tetraspanins (CD9, CD63, CD81) and flotillin-1 (FLOT1), and by their cup-shaped morphology as evaluated by transmission electron microscopy (8). Most cells secrete exosomes, which mediate local and systemic cell communication through the horizontal transfer of DNA, microRNA, mRNA, lipids, and proteins (9–12). In particular, Peinado et al. (9) showed that melanoma-associated exosomes promote metastasis by carrying proteins that affect bone marrow progenitor cells. Two general mechanisms have been hypothesized to explain the transfer of exosomal content between cells; both mechanisms propose that exosomes incorporate transmembrane proteins into the plasma membrane of the recipient cell and release their lumen content into the cytoplasm (13, 14).

Integrins are transmembrane receptors that are composed of an α-subunit and a β-subunit involved in regulating a variety of cellular processes, including adhesion, migration, proliferation, and differentiation. Integrins are also known to be deregulated as PrCa progresses to advanced stages (15, 16). Overexpression of α_vβ₆, an epithelium-specific integrin, has been reported to correlate with malignant progression and poor clinical prognosis in a variety of carcinomas, and to promote metastasis (17, 18). α_vβ₆ expression is not detectable in normal human prostate but is highly expressed in human primary PrCa (19),⁴ as well as murine PrCa in Pten^pc−/− mice (20) and thus, is a promising therapeutic target in PrCa. Furthermore, this integrin has multiple regulatory functions in tumor cell biology. The extracellular domain of α_vβ₆ has been shown to bind latency-associated peptide (LAP)-TGFβ1 and subsequently facilitates the release of active TGFβ1, which is a prometastatic cytokine (21). The cytoplasmic domain of the β₆ subunit has been implicated in mediating cell proliferation and migration (22–25). Finally, our group has recently shown that the α_vβ₆ integrin promotes an autonomous osteolytic program in PrCa cells by up-regulation and activation of the matrix metalloproteinase-2 (MMP2) (26).

Although some controversy exists regarding the exact nature of exosomal protein content (27), different groups have demonstrated that integrins are found in exosomes from different cell types including PrCa cells (28–31). Proteomic analysis has shown that α₂, α₃, α₆, β₁, and β₄ integrin subunits are found in microvesicles from PrCa cells (28). In addition, a recent study has shown that the presence of the α₃ integrin subunit in PrCa cell derived-exosomes interferes with non-cancerous prostate cell functions (29). Moreover, Clayton et al. (30) have shown that B cell-derived exosomes express functional β₁ and β₂ integrins that are capable of mediating anchorage to the extracellular matrix (ECM). Furthermore, α_vβ₆ has been shown to be expressed in exosomes, and when co-expressed with ovalbumin in gut epithelial cell-derived exosomes, it causes activation of different immune system cell types (31). As a result, LAP-TGFβ is converted to the active form, TGFβ1, within immune system cells, thus conferring tolerogenic properties. However, this mechanism is not strictly exosome-dependent because it is also mediated by α_vβ₆ and ovalbumin in a soluble form. Another study shows the presence of the integrin β₄ subunit in exosomes from pancreatic ductal adenocarcinoma; this integrin was shown to be necessary for plectin inclusion in the exosomes (32). However, the authors proposed only a structural role for this integrin in the exosomes. All these studies failed to investigate whether or not exosomes were internalized and recycled by the recipient cells and whether there was a real transfer of integrins between the different cell lines.

In the present work, we provide the first evidence that exosomes are able to transfer a specific integrin and its related functions between different subsets of PrCa cells. We observe internalization and surface expression of the α_vβ₆ integrin mediated by PC3 cell derived-exosomes. Surface expression of α_vβ₆ integrin confers a gain of function in the α_vβ₆-negative recipient DU145 cells, which show increased cell adhesion and migration on LAP-TGFβ, a specific α_vβ₆ substrate. Overall, this study shows that exosomes from a subset of cancer cells may contribute to the horizontal propagation of integrin-associated phenotypes to a different subset of cancer cells in a paracrine fashion.

EXPERIMENTAL PROCEDURES

Cell Lines

PC3, DU145, C4-2B, and RWPE-2 (designated here RWPE) cell lines, culture conditions, and generation of cell transfectants have been previously described (26, 33).

Exosome Isolation and Characterization

Cells were washed with PBS and grown in serum-free medium for 48 h. Exosomes secreted into the medium were purified by differential ultracentrifugation (8). Briefly, culture supernatants were centrifuged at 2000 × g for 20 min at 4 °C to clear cells and large debris. This supernatant was then centrifuged at 10,000 × g for 30 min at 4 °C to remove residual membranous debris. The remaining supernatant was then subjected to ultracentrifugation at 100,000 × g for 70 min at 4 °C to pellet the exosomes. The exosomes were resuspended in PBS and re-pelleted at 100,000 × g for 70–120 min at 4 °C to remove contaminating proteins, and the final pellet was re-suspended in PBS for further analysis.

Transmission Electron microscopy analysis was performed as described (8). Briefly, 5–10 μl of exosomes suspended in PBS were placed on a Formvar carbon-coated grid and negatively stained with 2% uranyl acetate solution. Images were taken using a JEM-100CXII electron microscope operated at 80.0 kV. Biochemical exosome characterization was carried out by immunoblotting (IB) analysis. Exosomes and cells were lysed with radioimmunoprecipitation assay buffer (10 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1 mm EDTA, 0.1% SDS, 1% Triton X-100, and 1% sodium deoxycholate) supplemented with protease inhibitors as described previously (33). The amount of protein present in exosomes was determined using a BCA protein assay kit (Pierce). Equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes (Millipore). Membranes were probed with the following primary mouse mAbs to CD63 and CD81 (Abcam), β₁ (BD Transduction Laboratories), and β₆ 2A.1 (34, 35) or with rabbit pAbs to FLOT1 (Abcam), GM130 (BD Transduction Laboratories), calnexin (CANX, Santa Cruz Biotechnology), and α_v C-terminal domain.

Sucrose Gradient

Exosome analysis in a linear sucrose gradient was performed as described previously by Théry et al. (8).

Analysis of Exosome Uptake

After exosome purification, the resulting exosome pellet was labeled with PKH26 Red fluorescent dye (Sigma) as recommended by the manufacturer. Subconfluent DU145 cells were serum-starved for 24 h followed by incubation with 5 μg/ml labeled exosomes or with PKH26 Red dye alone for 60 min. Subsequently, cells were fixed with 4% paraformaldehyde, permeabilized with 1% Triton X-100, blocked with 5% BSA, and stained with Alexa Fluor Phalloidin-488 as well as DAPI, followed by a confocal microscope analysis using a Zeiss LSM510. 75 cells were counted in each experiment to determine the percentage of cells that have internalized exosomes. A z-stack image analysis was also performed to evaluate exosome internalization into the recipient cells.

Analysis of Exosome-mediated α_vβ₆ Transfer

FACS analysis was employed to detect α_vβ₆ on the surface of non-permeabilized viable cells and was carried out either with PC3 cells that express these receptors endogenously or with DU145 cells that have acquired α_vβ₆ expression upon internalization of exosomes. Typically, DU145 cells were serum-starved for 18 h and then treated with 20 μg/ml exosomes obtained from PC3 cells for 24 h. The cells treated with PC3-derived exosomes were trypsinized, extensively washed with PBS, and subsequently stained for 45 min at 4 °C with the mAb 4B4 (Biogen Idec) specific to β₆. Samples were incubated with Alexa Fluor 488 goat anti-mouse secondary antibody (Molecular Probes) for 30 min at 4 °C, washed with PBS, and analyzed. The data were acquired using FACSCalibur flow cytometer (BD Biosciences). In parallel, an IB analysis was performed to detect β₆ integrin subunit expression in the β₆-negative recipient cell lines, DU145 and C4-2B, upon exosome incubation as described above. IB analysis of β₆ integrin uptake was also performed after acid wash treatment. For this purpose, upon exosome incubation, DU145 cells were washed with or without a sodium acetate buffer (0.2 m acetic acid/0.5 m NaCl, pH 2.8) (36), lysed, and then analyzed by SDS-PAGE.

Quantitative Real Time PCR

Quantitative real time (qRT) PCR analysis was performed as described earlier (37). Each reaction was carried out, at least in triplicate; standard deviations were calculated using Excel (Microsoft) software. The sequences of oligonucleotides used are as follows: GAPDH (sense, GGGAAGGTGAAGGTCGGAGT; antisense, GTTCTCAGCCTTGACGGTGC); β₆ (sense, GGTCTCATCTGGAAGCTACTGGTGTCA; antisense, GGTCTCCCAGATGCACAGTAGGACAACC).

Cell Transfection and Exosome Purification

Exosomes were purified from cells stably transfected with a β₆ shRNA in pLKO.1 (Dharmacon catalog number: RHS3979-9624888; TTCAGCATCATTTGTCAATGG, targeting sequence nt 893–913 (26)). Alternatively, cells were transiently transfected with 100 nm of either non-silencing (NS) siRNA (Dharmacon catalog number D-001810-01-20) or a β₆ siRNA smart pool (Dharmacon catalog number: M-008012-01-0010) containing four siRNAs. The targeting sequences are: 1) GCUAAAGGAUGUCAAUUAA, nt 449–467; 2) GAACGGCUCUUUCCAGUGU, nt 1660–1678; 3) CAUCUCAGCUUAUGAAGAA, nt 1342–1360; and 4) GCCAACCCUUGCAGUAGUA, nt 830–848 (NCBI Reference Sequence: NM_000888.4). After two rounds of transfection, cells were allowed to propagate, and then exosome isolation was performed as described above.

Analysis of Exosome Functions

DU145 cells were incubated with exosomes from PC3 cells as described above. Cell adhesion and migration assays either on LAP-TGFβ (R&D Systems) or on fibronectin (FN)-coated wells were performed as described previously (33).

Statistical Analysis

Statistical significance (p value, χ2 tests, and Student's t test) between datasets was calculated using Excel (Microsoft) software. A two-sided p value of ≤ 0.05 was considered statistically significant.

RESULTS

The α_vβ₆ Integrin Is Expressed and Enriched in Exosomes from PrCa Cells

Heterogeneous expression of α_vβ₆ in human (19) and murine (20) PrCa suggests that different cancer cell subsets could release α_vβ₆-containing exosomes into the intercellular environment. We asked whether α_vβ₆ could be transferred in this manner from PrCa cells expressing this integrin to α_vβ₆-negative cells. PC3 and RWPE cell lines were chosen as a model for this study as they have been well characterized for α_vβ₆ expression both in vitro and in vivo (26) (Fig. 1A). We investigated whether these PrCa cell lines release exosomes enriched in α_vβ₆. To evaluate the fidelity of the exosome preparations, we performed electron microscopy and IB analysis. Electron microscopy of negatively stained PC3 exosomes revealed a “cup shape” and a size distribution of 40–100 nm that is typical of exosomes (Fig. 1B, upper panel). IB analysis shows that the exosomal markers, CD63 and CD81, are enriched in exosomes purified from PC3 cells as compared with the parental total cell lysates (TCL; Fig. 1B, bottom panel). The endoplasmic reticulum protein CANX and the Golgi-specific marker GM130, which are expressed in the TCL, are not detected in exosomes (Fig. 1B, bottom panel). IB analysis reveals an enrichment of α_v, β₆, and β₁ integrin subunits in exosomes from PC3 cells and of α_v and β₆ in exosomes from RWPE cells as compared with the parental TCL (Fig. 1, C and D). β₁ integrins are used as exosomal loading control (29) under reducing conditions. We performed a more in-depth characterization of the exosomes with a linear sucrose gradient (Fig. 1E). Most integrin subunits (α_v, β₆, and β₁) are recovered in fractions 3–6 (corresponding to the typical exosome density of 1.13–1.17 g/ml) together with the FLOT1 and CD81 exosomal marker-enriched fractions. Finally, we investigated the expression of α_vβ₆ in exosomes isolated from the α_vβ₆-negative DU145 and C4-2B PrCa cells by IB. As expected, the β₆ integrin subunit is not expressed in DU145 and C4-2B exosomes, whereas the β₁ integrin subunit is expressed and highly enriched in these exosomes (Fig. 1F). In general, our results demonstrate that PC3 and RWPE cells secrete exosomes carrying the α_vβ₆ integrin.

FIGURE 1.

The α_v, β₆, and β₁ integrin subunits are enriched in exosomes from PrCa cells. A, cell surface expression of α_v, β₆, and β₁ integrin subunits in PC3 and RWPE cell lines was evaluated by FACS using Abs to α_v (dashed lines), β₆ (gray continuous lines), β₁ (black continuous lines), or a non-immune mouse IgG (dotted lines) used as negative control. B, upper panel, electron microscopy of negatively stained PC3-derived exosomes. Scale bar = 100 nm. Bottom panel, IB analysis of PC3 TCL and exosome (Exo) lysates. CD63 and CD81 were used as markers enriched in exosomes, whereas CANX and GM130 were used as markers not found in exosomes and tested in non-reducing (bottom left) and reducing (bottom right) conditions. β₁ integrin subunit was used as loading control under reducing conditions. C and D, IB analysis of α_v, β₆, and β₁ integrin subunits in PC3 exosomes and of α_v and β₆ integrin subunits in RWPE exosomes. E, IB analysis of α_v, β₆, and β₁ integrin subunits in density gradient fractions. FLOT1 and CD81 were used as markers enriched in exosomes, whereas CANX was used as a marker not found in exosomes. PC3 lysate (TCL) and the Exo lysate used as starting material were also loaded. F, IB analysis of β₆ and β₁ in PC3, DU145, and C4-2B TCL and Exo.

α_vβ₆ Integrin Is Transferred via Exosomes between Different PrCa Cells

Our finding that α_vβ₆ is sorted into exosomes motivated further evaluation of the phenotypic and functional impact of PC3-derived exosomes in other PrCa recipient cells. To investigate whether α_vβ₆ could be transferred between different PrCa cells, we first evaluated the efficiency of exosome internalization by the α_vβ₆-negative DU145 cell line. These cells were incubated for 1 h with PKH26 Red-labeled exosomes obtained from PC3 cells, and a confocal analysis was performed. As confirmed by z-stack image analysis, we have found PC3-derived exosomes to be efficiently internalized by DU145 cells cultured in serum-free medium (Fig. 2A). In particular, a total of 75 cells were counted, out of which 43% show internalized PC3-derived exosomes. In addition, we evaluated de novo expression of α_vβ₆ in the α_vβ₆-negative cells, DU145 and C4-2B, after incubation for 24 h with PC3-derived exosomes. First, we observe a significant increase in cell surface expression of α_vβ₆ in DU145 cells by FACS analysis (Fig. 2B). Second, we show by IB analysis that the β₆ integrin subunit is detected in the recipient cells upon incubation with PC3-derived exosomes, whereas no detectable change in β₁ integrin subunit levels is observed (Fig. 2C). Furthermore, to exclude the presence of non-specifically bound/non-internalized exosomes on the surface of the recipient cells, we washed DU145 cells with or without a sodium acetate buffer (36) followed by cell lysis and IB analysis for the β₆ integrin subunit. As shown in Fig. 2D, β₆ integrin was detected in the recipient cells upon incubation with PC3-derived exosomes, and its levels were slightly reduced after acid wash treatment. Finally, to rule out the possibility that exosomes could lead to transfer of PC3-associated exosomal β₆ mRNA or transcription factors that may enhance β₆ mRNA expression, we performed a qRT-PCR assay using specific primers for the β₆ integrin subunit. As shown in Fig. 2E, there is no significant change in β₆ mRNA levels upon incubation with PC3-derived exosomes. Thus, we conclude that PC3-derived exosomes are efficiently internalized by DU145 cells and that the α_vβ₆ integrin is transferred between different PrCa cells and topographically targeted to the correct site, i.e. the plasma membrane of the cancer cells.

FIGURE 2.

Exosomal transfer of α_vβ₆ between PrCa cells. A, DU145 cells were incubated with PKH26 Red dye alone (left) or with 5 μg/ml labeled PC3-derived exosomes (right) for 1 h, and a confocal analysis was carried out to evaluate exosomal internalization. DAPI was used to detect nuclei (blue), Alexa Fluor Phalloidin-488 was used to label actin (green), and PKH26 Red was used to label the exosomes (red). Scale bar = 20 μm. Z-stack analysis shows the presence of exosomes (representative exosomes shown by the arrow) inside the cells. B, FACS analysis of DU145 cells incubated for 24 h with PC3-derived exosomes and analyzed for α_vβ₆ integrin surface expression. Neg. Ctrl, negative control. C, IB analysis of lysates (75 μg per lane) of α_vβ₆-negative DU145 and C4-2B cells incubated with α_vβ₆-positive PC3 exosomes using Abs to β₆ or β₁ integrin subunit; CANX was used as loading control for the cell lysates. D, IB analysis of lysates (75 μg per lane) of DU145 cells incubated with or without PC3-derived exosomes, before and after acid wash. E, qRT-PCR analysis of β₆ mRNA levels in DU145 and C4-2B cells incubated with or without PC3-derived exosomes. Results were normalized to the β₆ mRNA levels in PC3 cells and are shown as average ± S.E.

Exosomal Transfer of α_vβ₆ Integrin Enhances Cell Adhesion and Migration on LAP-TGFβ

We next investigated whether PC3-derived exosomes internalized by DU145 cells might affect the ability of these cells to adhere and migrate on common ECM proteins such as LAP-TGFβ and FN, two major ligands found in the tumor microenvironment and specific for α_vβ₆. We reasoned that in cells that have incorporated α_vβ₆-containing exosomes, the functional consequences of the aforementioned de novo α_vβ₆ expression could confer a gain of function in the recipient cells. To investigate this hypothesis, we incubated α_vβ₆-negative DU145 cells with or without exosomes purified from PC3 cells in which α_vβ₆ was expressed (parental, Ctrl.shRNA, and NS siRNA), stably down-regulated (by β₆ shRNA), or transiently down-regulated (by β₆ siRNA). After confirming down-regulation of β₆ in exosomes, we next performed adhesion and migration assays (Fig. 3) using either LAP-TGFβ or FN as substrates. We find that 24 h of treatment with α_vβ₆-containing exosomes increases cell adhesion and migration on both LAP-TGFβ and FN in DU145 cells in a concentration-dependent manner; the effect was observed at 5 μg/ml exosomal preparation and reached a maximum at 5 μg/ml on FN and 20 μg/ml on LAP-TGFβ (Fig. 3). In contrast, incubation with exosomes from either PC3 β₆ shRNA or β₆ siRNA abrogates the adhesive and migratory properties on LAP-TGFβ, whereas cell adhesion and migration on FN are still increased (Fig. 3). Comparable results for cell adhesion to FN⁵ were obtained with LNCaP, another PrCa cell line that is known to express androgen receptor. Our data indicate that the horizontal transfer of α_vβ₆ via exosomes enhances cell adhesion and migration on LAP-TGFβ, whereas the increased adhesion and migration on FN may be mediated by other integrins in addition to α_vβ₆.

FIGURE 3.

PC3 exosome uptake by DU145 cells enhances adhesion and migration on LAP-TGFβ in an α_vβ₆ integrin-dependent manner. A, adhesion (left) and migration assays (center) of DU145 cells either untreated or treated with 20 μg/ml PC3 exosomes in which β₆ integrin is either expressed (parental and Ctrl.shRNA (Ctrl)) or down-regulated (β₆ shRNA). Left, cells were seeded for 2 h on BSA (1%)-, FN (10 μg/ml)-, or LAP-TGFβ (10 μg/ml)-coated wells, and cell adhesion was quantified by crystal violet staining. Center, cells were seeded on Transwell chambers coated with BSA (1%), FN (10 μg/ml), or LAP-TGFβ (10 μg/ml), and cell migration was evaluated as described under “Experimental Procedures.” Right, IB analysis of β₆ integrin expression in exosomes secreted by PC3 cells [parental (Par.), Ctrl.shRNA (Ctrl), and β₆ shRNA (β₆)]. CD63 was used as marker enriched in exosomes, whereas CANX was used as a marker not found in exosomes. B, migration assay of DU145 cells either untreated or treated with increasing concentrations of PC3 exosomes (5–40 μg/ml). C, left, migration assay of DU145 cells treated with or without 5 μg/ml exosomes secreted by PC3 cells previously incubated either with a non-silencing siRNA (NS siRNA) or with an siRNA specific for β₆ mRNA(β₆ siRNA). Right, IB analysis of β₆ integrin expression in exosomes secreted by PC3 cells. FLOT1 was used as marker enriched in exosomes, whereas CANX was used as a marker not found in exosomes. In A, B, and C, bar graphs show the number of DU145 cells migrated on BSA (white bars), LAP-TGFβ (gray bars), and FN (black bars), expressed as the percentage of total number of cells attached on both filter layers. Values are reported as means ± S.E. *, p ≤ 0.001 as determined by a Student's t test and χ2 tests for adhesion and migration assays, respectively.

DISCUSSION

The present study demonstrates that the α_vβ₆ integrin is transferred by exosomes between different PrCa cells and that exosome-mediated transfer of this integrin promotes cell adhesion as well as cell migration. Thus, we propose that exosomes containing the α_vβ₆ integrin impact neighboring tumor cells by enhancing their ability to migrate.

We have previously shown that the α_vβ₆ integrin plays a dynamic role in regulating the osteolytic lesion forming process in metastatic PrCa (26). Thereby, these findings implicate α_vβ₆ integrin as a master regulator of the metastatic phenotype in PrCa. Until now, it was thought that α_vβ₆ functions only in cells that synthesize it. However, the current finding that α_vβ₆ is enriched in exosomes raises the possibility that this integrin is delivered to neighboring cells through exosomes. Because of the known role of α_vβ₆ in promoting metastasis (18), delivery of this integrin via exosomes may play a role in transferring its related function to different tumor cells.

The novel findings reported in this study are as follows. First, we show that exosomes from PC3 and RWPE PrCa cell lines express high levels of the α_vβ₆ integrin. Second, we demonstrate that PC3-derived exosomes are efficiently internalized by DU145 cells and that the internalized α_vβ₆ integrin is expressed on the surface of these recipient cells. We also show that de novo α_vβ₆ expression occurs as a consequence of an exosome-mediated transfer of this protein between these two PrCa cells and is not due to a change at the mRNA level. Conversely, β₁ integrin subunit levels do not show any detectable change upon exosome uptake in the recipient cells; the reasons why this occurs remain to be investigated. Third, this is the only study showing surface re-expression of a transmembrane receptor of the integrin family; two previous studies have shown that two different transmembrane receptors, MET and KIT, are transferred and re-expressed on the surface of recipient cells (9, 38). Finally, our study is the first to show that the α_vβ₆ integrin that is transferred between different subsets of PrCa cells is functional and promotes cell adhesion and migration on LAP-TGFβ in an α_vβ₆ integrin-dependent manner. Previous studies have suggested but not proven, or excluded, that exosomal transfer of integrins, as well as of other molecules such as heparanase, promotes cell motility on ECM in recipient cells (29, 39). In contrast, Lee et al. (40) demonstrate that exosomes released from human macrophages negatively regulate endothelial cell migration, promoting the internalization and degradation of the β₁ integrin subunit in the recipient cells. The reason for these contradictory findings remains to be fully investigated. We further show that exosomes from PC3 cells promote cell adhesion and migration on FN in an α_vβ₆ integrin-independent manner and believe that β₁ or other integrins may contribute to this increase. In addition, these findings do not exclude the possibility that an uptake of exosomes containing α_vβ₆ by the recipient target cells may modulate their transcriptomic/proteomic profiles, promoting several cell functions such as proliferation, apoptosis, or invasion. Among these, we have ruled out the possibility that exosomes may lead to transfer of PC3-associated exosomal β₆ mRNA or transcription factors that may enhance β₆ mRNA expression (Fig. 2D). On the basis of these findings, it would be interesting to investigate whether the transfer of α_vβ₆ may lead to enhanced secretion of MMP2 and parathyroid hormone-related protein, two known α_vβ₆ downstream targets that are required for tumor cell invasion and bone metastasis, in target recipient cells (26).

Although detection of exosomal integrins offers, in addition to mouse and cell line models (41), a new tool to study aberrations occurring in cancer progression, it also underlines a great potential for diagnosis and risk stratification (42). The exosome-mediated transfer mechanism described here may have wider significance for various types of cancer. It remains to be established whether surrounding normal cells, such as stromal or endothelial cells, may also uptake exosomal integrins or growth factor receptors. Overall, these data suggest that therapeutics interfering with the production, transfer, or uptake of α_vβ₆-containing exosomes may attenuate tumor progression and metastasis.