Abstract

Exosomes are small nanovesicles that carry bioactive molecules which can be delivered to neighbouring cells to modify their biological functions. Studies have showed that exosomes from ovarian cancer (OVCA) cells can alter the cell migration and proliferation of cells within the tumour microenvironment, an effect modulated by the invasiveness capacity of their originating cells. Using an OVCA cell line xenograph mouse model, we showed that exosomes derived from a high invasiveness capacity cell line (exo-SKOV-3) promoted metastasis in vivo compared with exosomes from a low invasiveness capacity cell line (exo-OVCAR-3). Analysis from anin vivo imaging system (IVIS) revealed that exo-SKOV-3 formed metastatic niches, whereas exo-OVCAR-3 formed colonies of clustered cells close to the site of injection. Interestingly, kinetic parameters showed that the half-maximal stimulatory time (ST50) of tumour growth with exo-OVCAR-3 (4.0 ± 0.31 weeks) was significantly lower compared with the ST50 in mice injected with exo-SKOV-3 (4.5 ± 0.32 weeks). However, the number of metastic nodes in mice injected with exo-SKOV-3 was higher compared with exo-OVCAR-3. Using a quantitative mass spectrometry approach (SWATH MS/MS) followed by bioinformatics analysis using the Ingenuity Pathway Analysis (IPA), we identified a total of 771 proteins. Furthermore, 40 of these proteins were differentially expressed in tumour tissues from mice injected with exo-SKOV-3 compared with exo-OVCAR-3, and associated with Wnt canonical pathway (β-catenin). Finally, we identified a set of proteins which had elevated expression in the circulating exosomes in association with tumour metastasis. These observations suggest that exosomal signalling plays an important role in OVCA metastasis.

Introduction

Ovarian cancer (OVCA) is the fifth leading cause of deaths among women worldwide, with metastasis attributed to most OVCA mortalities [1]. Specifically, intra-abdominal metastasis is a common observation in OVCA patients, and is a major cause of unfavourable outcomes [2,3].

OVCA, unlike most other solid tumours, has a high tendency to metastasise through the peritoneal cavity and only rarely by the vasculature [4]. OVCA disseminates by dissociation of epithelial cancer cells from the primary site of origin [5]. The cancer cells then float within the peritoneal cavity and attach to secondary sites of implantation. In fact, the organs in the peritoneal cavity become covered with a single layer of mesothelial cells that coat an underlying stroma composed of extracellular matrix and stromal cells [6]. Once the dissociated cancer cells attach to the mesothelial cell layer, they begin to invade the targeted organs [5]. Significant evidence now indicates that cells within the tumour microenvironment can communicate with each other by extracellular vesicles (EVs), mainly through a specific type of EVs called exosomes [7,8].

Exosomes have been identified as key mediators in cell–cell communication, as they enable the degradation of the extracellular matrix and dissemination of cancer cells to distant sites [9]. Exosomes also promote the formation of a favourable microenvironment for metastatic outgrowth of disseminated cancer cells [10]. Exosomes are small vesicles approximately 100 nm in size that originate from the endosomal compartment of the cells. Exosomes are formed by inward budding of the cell membrane, which captures the protein receptors of the cell membrane to form early endosomes, which then mature to late endosomes [11]. Multivascular bodies (MVBs) are subsequently formed by invaginations of the late endosomes, and fuse with the cell membrane to release exosomes into the extracellular environment [12].

Exosomes have been likened to a ‘fingerprint’ of their originating cells, as they contain several bioactive molecules, including proteins, lipids, mRNAs, and miRNAs, which can be delivered to the target cells at either local or distant locations [12,13]. The delivery of these biomolecules can cause changes in gene expression and signalling in the target cells. Thus, communication between cells within the tumour microenvironment by exosomes might play a key role in both the development and progression of cancer [14]. For example, human peritoneal mesothelial cells (HPMCs) from the normal omentum acquired the spindle phenotype characteristic of mesenchymal cells when inoculated with exosomes derived from OVCA cells. Exosomes transferred CD44 to the HPMCs, thereby increasing the expression of matrix metalloproteinase-9 and subsequent target cell invasion [15].

Exosomes may have an influenence on the cellular microenvironment and thereby contribute to the progression and malignancy of tumour cells. Recent studies have highlighted the importance of exosomes in cancer progression and metastasis [16–18], and we have recently obtained in vitro evidence that exosomes from OVCA cells induce the migration and proliferation of mesenchymal stem cells and endothelial cells within the tumour microenvironment [19]. The effect of exosomes on the target cell depended on the invasiveness capacity of the originating cells, suggesting that highly invasive OVCA cells secrete exosomes with a specific package of bioactive molecules that promote cancer progression and metastasis [19]. However, the effect and mechanisms associated with the functions of exosomes on tumour growth and metastasis in vivo remain to be established.

The aims of the present study were to investigate the effect of exosomes derived from a high and low invasiveness capacity cell lines on tumour growth and metastasis in vivo, and to establish the potential singling pathways associated with the effect of OVCA exosomes. Using an in vivo imaging system (IVIS) to monitorng tumour growth and a liquid chromatography–mass spectrometry (LC-MS/MS) with SWATH acquisition to identify signalling molecules, we demonstrated that exosomes from highly invasiveness capacity OVCA cells promote peritoneal metastasis in vivo and that this phenomenon was associated with changes in the Wnt/β-catenin pathway signalling pathways. Finally, we identified a set of proteins within exosomes, where their abundance increases in association with tumour metastasis, suggesting that the circulating exosomes might be used as a real-time monitoring system to evaluate tumour progression.

Materials and methods

Cells culture and reagents

SKOV-3, and OVCAR-3 OVCA cell lines were obtained from the American Type Cell Collection. All cells were maintained in Phenol Red-free RPMI 1640 medium supplemented with 10% heat-inactivated foetal bovine serum (FBS) (Bovogen, Interpath Services Pty Ltd) and 1% antibiotic–antimycotic (Life Technology). Cell cultures were incubated at 37°C in 8% O2 and 5% CO2. All experimental procedures were conducted within an ISO17025 accredited (National Association of Testing Authorities, Australia) research facility. All data were recorded within 21 Code of Federal Regulation (CFR) part 11 compliant electronic laboratory notebooks (Lab Archives, Carlsbad, CA 92008, U.S.A.). The present study was approved by the Ethics Committee of the University of Queensland (approval number 2016000300). The schematic in Supplementary Figure S1 summarises the experimental design used in the present study. Authentication of human cell lines was performed by Short Tandem Repeat (STR) analysis and is presented in the Supplementary Figure S2.

Exosome isolation and characterisation

Prior to exosome isolation, the OVCA cells were grown to >70% confluence before being washed with phosphate-buffered saline (PBS) and incubated in FBS-free medium supplemented with 1% antibiotic–antimycotic (Life Technologies, U.S.A.) for 24 h at 37°C in 8% O2 and 5% CO2. Exosomes were then isolated from the cell-conditioned medium by differential ultracentrifugation and ultrafiltration. Briefly, cell-conditioned medium was centrifuged at increasing speeds: (i) 300×g for 10 min at 4°C, (ii) 2000×g for 10 min at 4°C, (iii) 12000×g for 10 min at 4°C, and (iv) 100000×g for 120 min at 4°C. The resulting pellet was re-suspended in 15 ml PBS and filtered using a 0.22-μm filter (Millipore, Massachusetts, U.S.A.). The filtered solution was concentrated using a 100000 Dalton Nominal Molecular Weight Limit (NMWL) Amicon Ultra-15 Centrifugal Filter Unit (Merck Millipore) by centrifugation at 4000×g for 30 mins at 4°C. Exosomes were characterised by size distribution, detection of proteins associated with exosomes and morphology according to recommendations by the International Society of Extracellular Vesicles [20] using nanoparticle tracking analysis (NTA), Western blot analysis, and electron microscopy. The exosomal protein concentration was measured using DC™ Protein Assay kit (Bio-Rad, Australia). The purity of the exosomes isolations were assessed by comparing the ratio of vesicle counts to protein concentration.

Linear correlation was observed between the number of vesciles and protein per isolation across all preparations from SKOV-3 and OVCAR-3 cells.

Inoculation of SKOV-3 and OVCAR-3 exosomes in a mouse xenograft model

Experiments using mice were approved by the University of Queensland Animal Ethics Committee (MRI-UQ/TRI/348/15/CCQ), and conducted in accordance with the Australian Code for the Care and Use of Animals for Scientific Purposes, and performed at Mater Research Institute-University of Queensland, Translational Research Institute (Woolloongabba, Australia). On day 0, luciferase SKOV-3 cells (5 × 106) were injected intraperitoneally into 25 female NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice aged 6–8 weeks. Mice were age-matched across the groups, to discard variability in results associated with the mice rather than the treatment (i.e. exosomes) used. On day 4, the same 25 NSG mice were divided into three groups that included control mice injected with PBS (n=8), and mice injected with exo-OVCAR-3 (n=8) and exo-SKOV-3 (n=9). For the injections, 10 μg of exosomal proteins containing 1.5 × 109 vesicles in 500 μl final volume or 500 μl of PBS were injected intraperitoneally twice a week for 6 weeks.

Tumour development was assessed weekly by monitoring luciferase signal using an in vivo bioluminescent imaging system (PerkinElmer, Glen Waverley, VIC, Australia) as described previously [21]. Animal health was monitored through daily observations and weekly assessment of weight. After 6 weeks, mice were killed by CO2 gas one animal at a time and immediately blood was collected by cardiac puncture (to obtain the maximum volume of blood) into EDTA tubes (BD Vacutainer K2E (EDTA)). Mice were then dissected and tumour burden determined by counting the number of tumour nodules within the peritoneal cavity of each mouse. Tumour tissues were stored at −80°C for the Sequential Windowed Acquisition of All Theoretical Mass Spectra-MS (SWATH-MS) analysis.

Isolation of circulating exosomes (mice-exo)

Exosomes were isolated from plasma by differential centrifugation coupled to ultrafiltration and were characterised as previously described in the ‘exosome isolation and characterisation’ section.

Tissue collection, processing, and protein extraction

Tumour tissues and metastasis nodes were collected from the killed animals and stored in 1.5 ml Eppendorf tubes at −80°C. Protein was extracted from the frozen tumour tissues by homogenising the tissue in Radioimmunoprecipitation Assay Buffer (RIPA buffer) (Sigma–Aldrich) using a Qiagen TissueLyser II at 30 Hz for 6 min. The homogenised tissue was transferred to a pre-chilled microcentrifuge tube and centrifuged at 15000×g for 15 min at 4°C. The supernatant containing total cellular proteins was collected and quantified using DC™ Protein Assay kit (Bio-Rad, Australia).

Sample preparation method for proteome analysis

Ion library generation

To generate the ion library, the following samples were reduced, alkylated, and trypsinised using an in-gel digestion method: (i) exo-SKOV-3; (ii) exo-OVCAR-3; (iii) pooled proteins from tissue samples; and (iv) pooled proteins from circulating exosomes [22]. Briefly, extracted protein from exosomes or tissues were mixed with Bolt LDS sample buffer (Thermo Fisher), sonicated for 5 min, and heated at 72°C for 10 min to denature protein complexes. Samples were then separated based on molecular weight using a Bolt 4–12% Bis-Tris plus polyacrylamide gel (Thermo Fisher) at 200 V until full separation was observed (approximately 40 min). Protein bands were visualised in the gels by staining with SimplyBlue Safe Stain according to manufacturer’s instructions (Invitrogen). For each pooled sample, 12 gel fractions were excised generating 48 fractions (12 fractions each from exo-OVCAR-3, exo-SKOV-3, the tissue samples, and the circulating exosomes). The gel pieces were destained using 50% (v/v) acetonitrile in 50 mM ammonium bicarbonate. Proteins were reduced with 10 mM DTT and alkylated with 50 mM idoacetamide. Trypsin was added to each sample at a ratio of 1:50 enzyme:protein. The resulting trypic digests were desalted using SOLAμ HRP 96-well plates (Thermo Fisher) that were applied to LC-MS/MS.

Filter-aided sample preparation

Tumour tissue samples from mice injected with exo-SKOV-3 or exo-OVCAR-3 were compared by SWATH-MS to identify differentially expressed proteins. Equal amounts of sample (∼20 μg) were processed using the filter-aided sample preparation (FASP) method. Six biological replicates were processed from the control (PBS) and exo-OVCAR-3 groups, while nine biological replicates were processed from the exo-SKOV-3 group.

Proteins extracted from the tissue lysates or exosomes were prepared in the presence of five volumes of 8 M urea. Disulphide bridges were reduced with 0.1 M dithiothreitol (DTT). Protein–detergent complexes were dissociated in the presence of 8 M urea with the detergent removed by 30 kDa centrifugal filter units (Pall) facilitated by centrifugation at 10000×g at 20°C for 5 min. Iodoacetamide (IAA) was used with carboxyamidomethylate thiols with excess reagent removed by centrifugation at 10000×g at 20°C for 5 min. The filters were washed with 8 M urea to deplete any remaining detergent from the proteins. The pH was then increased in the extracted samples using 50 mM ammonium bicarbonate buffer. The resultant protein suspension was digested using trypsin diluted in ABC at a 1:50 enzyme:protein ratio prior to an overnight incubation at 37°C. Peptides were then collected by adding 50 μl ABC to the filter before centrifugation at 10000×g at 20°C for 15 min. Finally, the peptides were desalted using a SOLAμ HRP 96-well plate (Thermo Fisher) before being applied to LC-MS/MS.

Sequential window acquisition of all theoretical (SWATH) mass spectra analysis

Ion library generation

The data-dependent acquisition (DDA)-MS raw file was subjected to database searches using ProteinPilot software version 4.5b (Sciex, Framingham, MA). Raw data files from all fractions were analysed using the ProteinPilot software. For exo-SKOV-3 and exo-OVCAR-3, the raw data were searched against a human SwissProt database from UNIPROT database (http://www.uniprot.org/proteomes/UP000000589). For the xenograft samples (the tissues and circulating exosomes), the raw data were searched against a combined human and mouse SwissProt database on 21 May 2018. A global false discovery rate (FDR) of 1% was used as the threshold for the number of proteins for import to ensure our ion library had the highest quality. Shared peptides between human and mouse were also excluded from the library.

SWATH analysis

For SWATH processing, the SWATH Acquisition Microapp (version 2.0) within PeakView (RRID: SCR_015786; version 2.2) was used. Within the Microapp, a setting of three peptides per protein, four transitions per peptide, peptide confidence threshold corresponding to 1% global FDR. The retention time was then manually realigned with a minimum of five peptides that had consistently high signal intensities and distributed along the time axis. The resulting peak area for each protein after SWATH processing was exported to MarkerView (version 1.3.1; Sciex) with the resulting data normalised using the most likely ratio method.

Immunoblotting

For immunoblotting, 20 μg total protein per sample was separated by SDS/PAGE using a 4–12% Bolt™ gel (Thermo Fisher) before being transferred to Immobilon-FL PVDF membranes (Merck Millipore). Membranes were blocked with Odyssey® blocking buffer (TBS) for an hour at room temperature. The following primary antibodies were diluted in blocking buffer according to the manufacturers’ recommendations and incubated with the blocked membranes overnight at 4°C: anti-Calnexin antibody (abcam- ab22595) (1:1000); anti-P Glycoprotein (abcam- ab170904) (1:1000); STIM1 (D88E10, Cell Signaling) (1:1000); S100A10 (4E7E10, Cell Signaling) (1:1000); β-Catenin (D10A8)XP (Cell Signaling) (1:1000); Phosphor-β-catenin (Ser33/37/Thr41) (Cell Signaling) (1:1000); Phosphor-β-catenin (Thr41/Ser45) (Cell Signaling) (1:1000); Phosphor-β-catenin (Ser552)(D8E11) (Cell Signaling) (1:1000); and Phosphor-β-catenin (Ser675) (D2F1)XP (Cell Signaling) (1:1000). The membranes were then washed three times in TBS-T (TBS and Tween-20) for 10 min before secondary antibodies anti-Rabbit IgG H&L (IRDye® 800CW) (1:10000) were applied for 1 h at room temperature. Visualisation of the immunoreactive proteins was performed using Image Lab™ Software (Bio-Rad) at 800 and 680 nm. The band density was measured using Image lab 6.1 Bio-Rad software and the changes in experimental groups were normalised to total protein.

Bioinformatics analysis of proteomic profiles

Protein classification was performed using FunRich software version 3.0. Pathway enrichment analyses were performed with Ingenuity Pathway Analysis (IPA) that identified canonical pathways, diseases and functions, and protein networks. Significantly enriched pathways were identified with a criterion P-value <0.05.

Statistical and correlation analysis

The data are presented as the mean ± SEM. To assess the distribution of the data, the Shapiro–Wilk normality test was used. For normal distribution, the comparisons between two and more groups were performed by means of unpaired Student’s t test and analysis of variance (ANOVA), respectively. If the ANOVA demonstrated a significant interaction between variables, post hoc analyses were performed by the multiple-comparison Bonferroni correction test. For non-nomality distribution the Kruskal–Wallis test (non-parametric) and multiple comparison using Dunn’s test was used. Statistical significance was defined as at least P<0.05. Correlation and statistical analysis was performed using the Spearman’s method in R, while the heatmaps were plotted using the ggplot2 R package.

Results

Characterisation of exosomes

Tumour-derived exosomes have emerged as critical mediators in tumour progression and metastasis. Recently, we demonstrated in vitro study that exosomes from an OVCA cell line with high migratory capacity (e.g., SKOV-3) increases the migration and proliferation of endothelial and mesenchymal stem cells compared with the effect of exosomes isolated from a cell line with low migratory capacity (e.g., OVCAR-3) [19]. Therefore, we hypothesise that exosomes derived from high migratory cancer cells enhance tumour growth and metastasis in vivo compared with the effect of exosomes from a low migratory capacity. To test this hypothesis, exosomes were isolated from SKOV-3 (high migration capacity, exo-SKOV-3) and OVCAR-3 (low migration capacity, exo-OVCAR-3) and characterised by NTA, transmission electron microscopy (TEM), and Western blot (CD9, CD63, and TSG101 (Figure 1). Particles identified had a mode of approximately 100 nm (Figure 1A,D), and were enriched with protein exosomal markers (Figure 1B). The isolated exosomes were also visualised under TEM demonstrated the presence of exosome vesicles identified by the typical small cup shaped morphology and lipid bilayer (Figure 1C). To further analyse the purity of the exosomes isolation, we determined the ratio between number of vesicles and protein per preparation, and no difference were observed between exosomes from SKOV-3 compared with OVCAR-3 cells across all the preparation of exosomes (Supplementary Figure S3). Moreover, a positive correlation (P<0.05) between the number of exosomes and protein concentration obtained from OVCRA-3 and SKOV-3 was identified (Supplementary Figure S3). We performed quantitative proteomic analysis in exosomes from OVCAR-3 and SKOV-3 and the proteins were analysed usind the IPA (Supplementary Table S1 and Figure 1E,F). Changes in the protein profile between exo-OVCAR-3 compared with exo-SKOV-3 were associated with the canonical pathways such as acute phase response signalling and LXR/RXR activation (Figure 1E). IPA network analysis showed that exo-SKOV-3 was enriched with proteins that modulate cancer signalling by up-regulating the expression of ACTN4, CD44, and Collagen type IV (Figure 1F). Thus, these results suggest that exosomes from tumour cells may engage in interactions between the cell constituents of the tumour microenviroment and target tissues.

Exosome isolation and characterisation

Figure 1
Exosome isolation and characterisation

(A) Size distribution by NTA of exosomes isolated from OVCA cells, SKOV-3 and OVCAR-3. (B) Representative Western blot showing exosomal markers TSG101, CD9, and CD63. (C) TEM demonstrates the presence of exosome vesicles from cell culture supernatants from SKOV-3 and OVCAR-3 cells and identified as small cup-shaped vesicles ∼100 nm in diameter. (D) Mean and mode size of exo-OVCAR-3 and exo-SKOV-3 with the standard deviation of the mean and mode size. (E) The top canonical protein pathways that were significantly differentially expressed in exo-OVCAR-3 compared with exo-SKOV-3. The bar chart represents the percentage of proteins that were mapped to each canonical pathway, showing those that were up-regulated (in red) and down-regulated (in green). (F) IPA analysis: a network was performed on dysregulated proteins of exo-SKOV-3 compared with exo-OVCAR-3. In (C), images are a representation of three independent TEM preprations per group (i.e., exosomes isolated from SKOV-3 or OVCAR-3) with at least five images analysed per sample.

Figure 1
Exosome isolation and characterisation

(A) Size distribution by NTA of exosomes isolated from OVCA cells, SKOV-3 and OVCAR-3. (B) Representative Western blot showing exosomal markers TSG101, CD9, and CD63. (C) TEM demonstrates the presence of exosome vesicles from cell culture supernatants from SKOV-3 and OVCAR-3 cells and identified as small cup-shaped vesicles ∼100 nm in diameter. (D) Mean and mode size of exo-OVCAR-3 and exo-SKOV-3 with the standard deviation of the mean and mode size. (E) The top canonical protein pathways that were significantly differentially expressed in exo-OVCAR-3 compared with exo-SKOV-3. The bar chart represents the percentage of proteins that were mapped to each canonical pathway, showing those that were up-regulated (in red) and down-regulated (in green). (F) IPA analysis: a network was performed on dysregulated proteins of exo-SKOV-3 compared with exo-OVCAR-3. In (C), images are a representation of three independent TEM preprations per group (i.e., exosomes isolated from SKOV-3 or OVCAR-3) with at least five images analysed per sample.

Exosomes from a highly migratory cancer cell line enhanced metastasis capacity in vivo

To explore the effect of exosomes in vivo, luciferase-expressed SKOV-3 cells (SKOV-3-Luc) were intraperitoneally injected into the left ovary, and subsequently on day 4 injected in the absence or in the presence of exo-SKOV-3 and exo-OVCAR-3 (Figure 2A). To evaluate tumour size and monitor tumour development, we measured the bioluminescent signal intensities in living mice using an IVIS. exo-OVCAR-3 accelerated tumour growth rate over time compared with the effect of exo-SKOV-3 (Figure 2A,B). The size of tumours in mice injected with exo-OVCAR-3 at week 4 was approximately two-fold higher compared with the values observed in mice injected with exo-SKOV-3 (5.5 × 108 vs 1.9 × 108 Luminoscore for mice injected with exo-OVCAR-3 compared with exo-SKOV-3, respectively) (Figure 2B). Analysis of the kinetic parameters showed that the number of weeks required to reach the half-maximal stimulatory time (ST50) was 4.4 ± 0.28, 3.64 ± 0.19s, and 4.39 ± 0.15 weeks for the control (without exosomes) and mice injected with exo-OVCAR-3 and exo-SKOV-3, respectively (Figure 2C). The ST50 in mice injected with exo-OVCAR-3 was significantly lower compared with exo-SKOV-3, indicating that exo-OVCAR-3 conferred a significant growth advantage and enhanced tumour growth of SKOV-3-Luc cancer cells in vivo. Finally, analysis of the number of metastasis lesions showed that exo-SKOV-3 increased the number of nodes compared with the effect of exo-OVAR-3 cells (Figure 2D). Tumour nodules were dispersed throughout the peritoneal cavity in mice injected with exo-SKOV-3, compared with mice injected with exo-OVCAR-3 which formed colonies of clustered cells near, or at the injection site. Interestingly, the effect of exosomes on tumour growth and metastasis were associated with changes in the mice weight during the time of experiments (Figure 2E). These data suggest that exosomes derived from highly migratory SKOV-3 cells promote metastasis compared with the effect of exosomes from OVCAR-3 cells. To further investigate the mechanism and signalling pathways associated with this phenomenon, we performed quantitative proteomic analysis in the tumour tissues isolated from mice injected with exosomes from SKOV-3 and OVCAR-3 cells.

Exosomes from less invasive OVCA cells (exo-OVCAR-3) are more capable than highly invasive OVCA cells derived exosomes (exo-SKOV-3) at inducing tumour growth

Figure 2
Exosomes from less invasive OVCA cells (exo-OVCAR-3) are more capable than highly invasive OVCA cells derived exosomes (exo-SKOV-3) at inducing tumour growth

(A) Tumour growth monitored by bioluminescent IVIS imaging from week 1 to 5 for three animal groups: Left, control, animal treated with (PBS) (n=8); middle, animal treated with exo-OVCAR-3 (n=8); and right, animal treated with exo-SKOV-3 (n=9). (B) Bar graph of SKOV3-luc tumours growth based on quantitative radiance value of IVIS imaging; control (white bars); animals treated with exo-OVCAR-3 (striped bars); and animals treated with exo-SKOV-3 (black bars). (C) Normalised tumour growth curves of the different treatments: control (white circle); animals treated with exo-SKOV-3 (black circle); and animals treated with exo-OVCAR-3 (white square). (D) The number of tumour nodes in the peritoneal cavity of animals that were treated with PBS, exo-OVCAR-3, and exo-SKOV-3; a quantitative analysis showed an increase in the number of nodes in mice that were injected with exo-SKOV-3 compared with mice that were treated with either exo-OVCAR-3 or PBS. (E) Animal weight. Mice were weighed during the exosome treatment. There was no statistical difference in the body weights of either groups throughout the 6-week study. *P< 0.05, ***P< 0.0005.

Figure 2
Exosomes from less invasive OVCA cells (exo-OVCAR-3) are more capable than highly invasive OVCA cells derived exosomes (exo-SKOV-3) at inducing tumour growth

(A) Tumour growth monitored by bioluminescent IVIS imaging from week 1 to 5 for three animal groups: Left, control, animal treated with (PBS) (n=8); middle, animal treated with exo-OVCAR-3 (n=8); and right, animal treated with exo-SKOV-3 (n=9). (B) Bar graph of SKOV3-luc tumours growth based on quantitative radiance value of IVIS imaging; control (white bars); animals treated with exo-OVCAR-3 (striped bars); and animals treated with exo-SKOV-3 (black bars). (C) Normalised tumour growth curves of the different treatments: control (white circle); animals treated with exo-SKOV-3 (black circle); and animals treated with exo-OVCAR-3 (white square). (D) The number of tumour nodes in the peritoneal cavity of animals that were treated with PBS, exo-OVCAR-3, and exo-SKOV-3; a quantitative analysis showed an increase in the number of nodes in mice that were injected with exo-SKOV-3 compared with mice that were treated with either exo-OVCAR-3 or PBS. (E) Animal weight. Mice were weighed during the exosome treatment. There was no statistical difference in the body weights of either groups throughout the 6-week study. *P< 0.05, ***P< 0.0005.

Quantitative proteomics by SWATH-MS suggest that wnt signalling is associated with the effect of exosomes on tumour growth and metastasis in vivo

We proposed that injecting mice with exosomes could influence cancer cell metastasis by changing the gene expression of tumour cells. Therefore, we investigated the protein profiles of tumour tissues that were isolated from mice treated with the exo-SKOV-3 or exo-OVCAR-3 by SWATH analysis (commonly referred to as data independent analysis (DIA)). SWATH is a quantification technique that combines the traditional mass spectrometry workflows of DDA and Multiple Reaction Monitoring (MRM) or targeted workflows. SWATH analysis identified a total of 2130 proteins in the tumour tissues (Figure 3A–C). Of these proteins, 105 were observed to be differentially expressed in mice injected with exo-SKOV-3 compared with exo-OVCAR-3 (Table 1). A total of 87 and 18 proteins were low and high abundant proteins respectively, in tumour tissue from mice injected with exo-SKOV-3 compared with exo-OVCAR-3. The localisation and molecular functions of the high and low abundant proteins were analysed using FunRich to search against a publicly available protein database. The proteomic profile of the 87 low abundant proteins were mostly located in the focal adhesion ribonucleoprotein complex nucleolus, mitochondrion cytosolic, mitochondrial matrix, cell–cell adherens junction, nuclear envelope, small ribosomal subunit, cytoplasm, lysosome, cytosol, centrosome, nucleus, and ribosome (Table 2). The high abundant proteins were mostly located in the nucleus, endoplasmic reticulum membrane, Box C/D snoRNP complex, and exon–exon junction complex (Table 2). For molecular function, the majority of the low abundant proteins were assigned to a catalytic activity, structural constituents of ribosomes, phosphorylase activity, and protein transporter activity. In a broader biological sense, the dominant subcategories included immune response, protein folding, cell organisation and biogenesis, intracellular signaling cascades, protein transport, and mitochondrion organisation and biogenesis (Table 2). For molecular function, the majority of high abundant proteins were assigned to chaperone activity (Table 2).

Quantitative proteomic analysis of animal tissue inoculated with exosomes and IPA data

Figure 3
Quantitative proteomic analysis of animal tissue inoculated with exosomes and IPA data

Volcano plot of 2130 quantified proteins illustrating expression increases and decreases following treatment. Volcano plots show the log 2 fold changes (on the x-axis) and the −log10 adjusted P-values (on the y-axis). Each dot represents a single protein expression. Dots below or above the threshold (P=0.05) indicate significantly altered protein expression (high or low abundant proteins). Data were selected at the cut-off values adj-P<0.05 and fold change > 1.5. (A) Volcano plot of differentially expressed proteins from tumour tissue of mice treated with PBS compared with mice treated with exo-SKOV-3. (B) PBS treatment compared with exo-OVCAR-3 exosome treatment. (C) exo-OVCAR-3 treatment compared with exo-SKOV-3 exosome treatment. (D) IPA analysis network was performed on 105 dysregulated proteins in mice treated with exo-SKOV-3 compared with those treated with exo-OVCAR-3. (E) Western blot of β-catenin expression in cancer tissue post exosome treatment. (FI) Western blot quantitative analysis of total β-catenin and phosphorylated β-catenin; Phosphor-β-catenin (Ser33/Ser37/Thr41), Phosphor-β-catenin (Thr41/Ser45), Phosphor-β-catenin (Ser552), and Phosphor-β-catenin (Ser675). Band intensity was measured by Image lab 6.1 Bio-Rad software, with the background subtracted in all samples. Changes in the protein expression were normalised to the total β-catenin protein using Image lab 6.1 Bio-Rad software. Data are presented as means ± SEM. **P< 0.001.

Figure 3
Quantitative proteomic analysis of animal tissue inoculated with exosomes and IPA data

Volcano plot of 2130 quantified proteins illustrating expression increases and decreases following treatment. Volcano plots show the log 2 fold changes (on the x-axis) and the −log10 adjusted P-values (on the y-axis). Each dot represents a single protein expression. Dots below or above the threshold (P=0.05) indicate significantly altered protein expression (high or low abundant proteins). Data were selected at the cut-off values adj-P<0.05 and fold change > 1.5. (A) Volcano plot of differentially expressed proteins from tumour tissue of mice treated with PBS compared with mice treated with exo-SKOV-3. (B) PBS treatment compared with exo-OVCAR-3 exosome treatment. (C) exo-OVCAR-3 treatment compared with exo-SKOV-3 exosome treatment. (D) IPA analysis network was performed on 105 dysregulated proteins in mice treated with exo-SKOV-3 compared with those treated with exo-OVCAR-3. (E) Western blot of β-catenin expression in cancer tissue post exosome treatment. (FI) Western blot quantitative analysis of total β-catenin and phosphorylated β-catenin; Phosphor-β-catenin (Ser33/Ser37/Thr41), Phosphor-β-catenin (Thr41/Ser45), Phosphor-β-catenin (Ser552), and Phosphor-β-catenin (Ser675). Band intensity was measured by Image lab 6.1 Bio-Rad software, with the background subtracted in all samples. Changes in the protein expression were normalised to the total β-catenin protein using Image lab 6.1 Bio-Rad software. Data are presented as means ± SEM. **P< 0.001.

Table 1
A list of proteins that showed significant differential expression in the tumour tissue of animals that were injected with exo-SKOV-3 compared with animals that were injected with exo-OVCAR-3
Protein name Fold change P-value 
PSB1_HUMAN 0.384535 0.00118 
TLN1_HUMAN 0.698048 0.00138 
DNM1L_HUMAN 0.45033 0.00159 
EF1G_HUMAN 0.322676 0.00391 
ILK_HUMAN 0.746528 0.00399 
SYRC_HUMAN 0.391929 0.00432 
F8I2_HUMAN 1.738235 0.00443 
SNX3_HUMAN 0.477026 0.00452 
PSB3_HUMAN 0.520496 0.0046 
RAB23_HUMAN 0.323951 0.00502 
DHB8_HUMAN 0.308298 0.00527 
SYDC_HUMAN 0.784374 0.0055 
LIS1_HUMAN 0.77368 0.0065 
PFKAL_HUMAN 0.268027 0.0076 
S10AA_HUMAN 3.469367 0.00793 
BZW2_MOUSE 0.466433 0.00873 
ERP29_HUMAN 0.466458 0.00937 
GNPI1_HUMAN 0.639046 0.00966 
LSM1_MOUSE 0.438112 0.01007 
VINC_HUMAN 0.467516 0.01111 
FN3K_HUMAN 0.363036 0.01129 
TPIS_HUMAN 0.721967 0.01131 
CH60_HUMAN 0.643594 0.01169 
SRG2C_HUMAN 0.241958 0.01209 
DDX3X_HUMAN 0.230422 0.01217 
PYGB_HUMAN 0.330536 0.01272 
PGAM1_MOUSE 0.631488 0.01293 
PEBP1_HUMAN 0.345336 0.01329 
ATPO_HUMAN 0.573383 0.01338 
MYDGF_HUMAN 2.109616 0.01472 
AP3B1_HUMAN 1.732305 0.01579 
THTR_MOUSE 0.42316 0.01599 
RHOC_MOUSE 0.748789 0.01715 
RS12_HUMAN 0.501876 0.01735 
NAKD2_HUMAN 1.855666 0.01876 
IDHC_HUMAN 0.68443 0.0189 
RS2_HUMAN 0.537756 0.01945 
MRP1_HUMAN 1.601333 0.01995 
RAB14_MOUSE 0.532819 0.02155 
KGUA_HUMAN 0.630616 0.022 
PFD4_HUMAN 0.507878 0.02273 
FKB11_MOUSE 0.50857 0.02349 
MUTA_HUMAN 0.492338 0.02359 
PRP8_MOUSE 0.237383 0.02364 
EF2_HUMAN 0.587857 0.02386 
RHOA_MOUSE 0.756545 0.02427 
TBCA_HUMAN 1.788367 0.02447 
RSSA_MOUSE 0.317898 0.02455 
EFR3A_HUMAN 2.413713 0.02522 
GOLP3_MOUSE 0.281531 0.02638 
GUAD_MOUSE 0.397567 0.02658 
RL27A_MOUSE 0.433141 0.02688 
IF5_HUMAN 0.534475 0.02809 
PR40A_MOUSE 0.433666 0.02816 
SEPR_MOUSE 2.475975 0.02874 
RS7_MOUSE 0.606223 0.02905 
PTGIS_MOUSE 0.211376 0.02986 
PCAT1_HUMAN 0.258328 0.03037 
CX7A2_HUMAN 0.548943 0.03059 
NOP56_HUMAN 1.569469 0.03068 
ASNS_MOUSE 0.431024 0.03097 
PARP1_HUMAN 0.403411 0.03105 
RLA0_HUMAN 0.521963 0.03162 
ASNA_HUMAN 0.752813 0.03247 
B2MG_MOUSE 1.570482 0.03248 
ACLY_MOUSE 0.619443 0.03303 
PRDX3_HUMAN 0.608166 0.03328 
CALX_MOUSE 1.385379 0.03336 
PNPH_HUMAN 0.644419 0.03419 
ERF1_MOUSE 0.776206 0.03423 
KAP2_HUMAN 0.571257 0.0343 
RBM8A_HUMAN 1.960519 0.03596 
SDHA_HUMAN 0.647884 0.03706 
FAF2_HUMAN 0.273161 0.03815 
TBA1C_HUMAN 0.341774 0.03959 
MA2C1_HUMAN 2.853969 0.03959 
RL3_HUMAN 0.586006 0.03966 
FA50B_HUMAN 0.524123 0.04014 
LRRF1_HUMAN 0.493559 0.04016 
NHLC2_HUMAN 0.324802 0.0402 
CEL2A_MOUSE 2.130846 0.04027 
ACTA_MOUSE 0.525277 0.04028 
RS16_HUMAN 0.759051 0.04095 
STIM1_HUMAN 2.29191 0.04112 
ISC2A_MOUSE 0.291221 0.04124 
KCC2D_MOUSE 0.494293 0.04131 
FNTA_HUMAN 0.524574 0.04131 
SSDH_HUMAN 1.621418 0.04181 
PRPS2_MOUSE 0.595369 0.04294 
2ABA_MOUSE 0.139626 0.04317 
U5S1_HUMAN 0.413438 0.04342 
EPN1_HUMAN 0.455649 0.04347 
EIF3L_HUMAN 0.349349 0.04436 
DYHC1_MOUSE 2.678517 0.04458 
GBG12_HUMAN 0.680559 0.0448 
KCY_HUMAN 0.127653 0.0451 
RS27_MOUSE 2.634619 0.0454 
AGR2_HUMAN 0.415204 0.04626 
CTNB1_MOUSE 0.194459 0.04763 
HMCS2_MOUSE 0.288194 0.04787 
RAD50_HUMAN 0.533035 0.04835 
TLN1_MOUSE 0.358109 0.0484 
ANXA7_HUMAN 0.857859 0.0491 
DCUP_HUMAN 0.228955 0.0499 
DHB4_HUMAN 0.478737 0.04997 
Protein name Fold change P-value 
PSB1_HUMAN 0.384535 0.00118 
TLN1_HUMAN 0.698048 0.00138 
DNM1L_HUMAN 0.45033 0.00159 
EF1G_HUMAN 0.322676 0.00391 
ILK_HUMAN 0.746528 0.00399 
SYRC_HUMAN 0.391929 0.00432 
F8I2_HUMAN 1.738235 0.00443 
SNX3_HUMAN 0.477026 0.00452 
PSB3_HUMAN 0.520496 0.0046 
RAB23_HUMAN 0.323951 0.00502 
DHB8_HUMAN 0.308298 0.00527 
SYDC_HUMAN 0.784374 0.0055 
LIS1_HUMAN 0.77368 0.0065 
PFKAL_HUMAN 0.268027 0.0076 
S10AA_HUMAN 3.469367 0.00793 
BZW2_MOUSE 0.466433 0.00873 
ERP29_HUMAN 0.466458 0.00937 
GNPI1_HUMAN 0.639046 0.00966 
LSM1_MOUSE 0.438112 0.01007 
VINC_HUMAN 0.467516 0.01111 
FN3K_HUMAN 0.363036 0.01129 
TPIS_HUMAN 0.721967 0.01131 
CH60_HUMAN 0.643594 0.01169 
SRG2C_HUMAN 0.241958 0.01209 
DDX3X_HUMAN 0.230422 0.01217 
PYGB_HUMAN 0.330536 0.01272 
PGAM1_MOUSE 0.631488 0.01293 
PEBP1_HUMAN 0.345336 0.01329 
ATPO_HUMAN 0.573383 0.01338 
MYDGF_HUMAN 2.109616 0.01472 
AP3B1_HUMAN 1.732305 0.01579 
THTR_MOUSE 0.42316 0.01599 
RHOC_MOUSE 0.748789 0.01715 
RS12_HUMAN 0.501876 0.01735 
NAKD2_HUMAN 1.855666 0.01876 
IDHC_HUMAN 0.68443 0.0189 
RS2_HUMAN 0.537756 0.01945 
MRP1_HUMAN 1.601333 0.01995 
RAB14_MOUSE 0.532819 0.02155 
KGUA_HUMAN 0.630616 0.022 
PFD4_HUMAN 0.507878 0.02273 
FKB11_MOUSE 0.50857 0.02349 
MUTA_HUMAN 0.492338 0.02359 
PRP8_MOUSE 0.237383 0.02364 
EF2_HUMAN 0.587857 0.02386 
RHOA_MOUSE 0.756545 0.02427 
TBCA_HUMAN 1.788367 0.02447 
RSSA_MOUSE 0.317898 0.02455 
EFR3A_HUMAN 2.413713 0.02522 
GOLP3_MOUSE 0.281531 0.02638 
GUAD_MOUSE 0.397567 0.02658 
RL27A_MOUSE 0.433141 0.02688 
IF5_HUMAN 0.534475 0.02809 
PR40A_MOUSE 0.433666 0.02816 
SEPR_MOUSE 2.475975 0.02874 
RS7_MOUSE 0.606223 0.02905 
PTGIS_MOUSE 0.211376 0.02986 
PCAT1_HUMAN 0.258328 0.03037 
CX7A2_HUMAN 0.548943 0.03059 
NOP56_HUMAN 1.569469 0.03068 
ASNS_MOUSE 0.431024 0.03097 
PARP1_HUMAN 0.403411 0.03105 
RLA0_HUMAN 0.521963 0.03162 
ASNA_HUMAN 0.752813 0.03247 
B2MG_MOUSE 1.570482 0.03248 
ACLY_MOUSE 0.619443 0.03303 
PRDX3_HUMAN 0.608166 0.03328 
CALX_MOUSE 1.385379 0.03336 
PNPH_HUMAN 0.644419 0.03419 
ERF1_MOUSE 0.776206 0.03423 
KAP2_HUMAN 0.571257 0.0343 
RBM8A_HUMAN 1.960519 0.03596 
SDHA_HUMAN 0.647884 0.03706 
FAF2_HUMAN 0.273161 0.03815 
TBA1C_HUMAN 0.341774 0.03959 
MA2C1_HUMAN 2.853969 0.03959 
RL3_HUMAN 0.586006 0.03966 
FA50B_HUMAN 0.524123 0.04014 
LRRF1_HUMAN 0.493559 0.04016 
NHLC2_HUMAN 0.324802 0.0402 
CEL2A_MOUSE 2.130846 0.04027 
ACTA_MOUSE 0.525277 0.04028 
RS16_HUMAN 0.759051 0.04095 
STIM1_HUMAN 2.29191 0.04112 
ISC2A_MOUSE 0.291221 0.04124 
KCC2D_MOUSE 0.494293 0.04131 
FNTA_HUMAN 0.524574 0.04131 
SSDH_HUMAN 1.621418 0.04181 
PRPS2_MOUSE 0.595369 0.04294 
2ABA_MOUSE 0.139626 0.04317 
U5S1_HUMAN 0.413438 0.04342 
EPN1_HUMAN 0.455649 0.04347 
EIF3L_HUMAN 0.349349 0.04436 
DYHC1_MOUSE 2.678517 0.04458 
GBG12_HUMAN 0.680559 0.0448 
KCY_HUMAN 0.127653 0.0451 
RS27_MOUSE 2.634619 0.0454 
AGR2_HUMAN 0.415204 0.04626 
CTNB1_MOUSE 0.194459 0.04763 
HMCS2_MOUSE 0.288194 0.04787 
RAD50_HUMAN 0.533035 0.04835 
TLN1_MOUSE 0.358109 0.0484 
ANXA7_HUMAN 0.857859 0.0491 
DCUP_HUMAN 0.228955 0.0499 
DHB4_HUMAN 0.478737 0.04997 
Table 2
FunRich analysis of significantly enriched and low abundant proteins from animals injected with exo-SKOV-3 revealing the enriched cellular components and molecular functions with fold enrichment and P-values
Classification Fold enrichment P-value 
High abundant proteins 
Cellular component   
Nucleolus 1.72371907 0.033004 
Integral to endoplasmic reticulum membrane 36.4330172 0.027352 
Box C/D snoRNP complex 281.742609 0.003569 
Small nucleolar ribonucleoprotein complex 141.047174 0.007126 
Pre-snoRNP complex 187.984669 0.005349 
Exon–exon junction complex 281.742609 0.003569 
Molecular function   
Chaperone activity 19.26143 0.004745 
Protein metabolism 2.747116 0.091368 
Metabolism 1.442059 0.41232 
Energy pathways 1.486212 0.397088 
Low abundant proteins 
Cellular component   
Ribonucleoprotein complex 7.89255611 0.026779 
Nucleolus 3.04820202 6.98E-05 
Mitochondrion 2.84060432 0.000272 
Cytosolic small ribosomal subunit 21.3375799 0.000371 
Soluble fraction 4.26846427 0.033453 
Nuclear envelope 7.22567347 0.031521 
Focal adhesion 11.3995795 0.013396 
Mitochondrial matrix 6.49405231 0.038315 
Peroxisome 9.22112813 0.000928 
Coated pit 19.817363 0.049758 
Cell–cell adherens junction 19.817363 0.049758 
Early endosome 11.1341292 0.002485 
Mitochondrial inner membrane 8.84494178 0.021655 
6-phosphofructokinase complex 85.6557782 0.011705 
Cytosolic large ribosomal subunit 14.2486829 0.008718 
IκB kinase complex 32.1877518 0.030915 
Microtubule-associated complex 21.3700572 0.003936 
Peroxisomal matrix 46.6026406 0.00081 
Astral microtubule 255.27118 0.003917 
Mitochondrial proton-transporting ATP synthase, catalytic core 25.7566326 0.038496 
Molecular function   
Structural constituent of ribosome 10.29772 0.000124 
Phosphorylase activity 31.38516 0.001843 
Protein transporter activity 78.69595 0.01274 
Regulation of immune response 39.3949257 0.025321 
Biological process   
Regulation of immune response 39.3949257 0.025321 
Protein folding 39.2243859 0.001173 
Cell organisation and biogenesis 45.0147439 0.022191 
Intracellular signalling cascade 104.835002 0.00957 
Protein transport 28.6606135 0.034653 
Molecular function 156.991719 0.00639 
Classification Fold enrichment P-value 
High abundant proteins 
Cellular component   
Nucleolus 1.72371907 0.033004 
Integral to endoplasmic reticulum membrane 36.4330172 0.027352 
Box C/D snoRNP complex 281.742609 0.003569 
Small nucleolar ribonucleoprotein complex 141.047174 0.007126 
Pre-snoRNP complex 187.984669 0.005349 
Exon–exon junction complex 281.742609 0.003569 
Molecular function   
Chaperone activity 19.26143 0.004745 
Protein metabolism 2.747116 0.091368 
Metabolism 1.442059 0.41232 
Energy pathways 1.486212 0.397088 
Low abundant proteins 
Cellular component   
Ribonucleoprotein complex 7.89255611 0.026779 
Nucleolus 3.04820202 6.98E-05 
Mitochondrion 2.84060432 0.000272 
Cytosolic small ribosomal subunit 21.3375799 0.000371 
Soluble fraction 4.26846427 0.033453 
Nuclear envelope 7.22567347 0.031521 
Focal adhesion 11.3995795 0.013396 
Mitochondrial matrix 6.49405231 0.038315 
Peroxisome 9.22112813 0.000928 
Coated pit 19.817363 0.049758 
Cell–cell adherens junction 19.817363 0.049758 
Early endosome 11.1341292 0.002485 
Mitochondrial inner membrane 8.84494178 0.021655 
6-phosphofructokinase complex 85.6557782 0.011705 
Cytosolic large ribosomal subunit 14.2486829 0.008718 
IκB kinase complex 32.1877518 0.030915 
Microtubule-associated complex 21.3700572 0.003936 
Peroxisomal matrix 46.6026406 0.00081 
Astral microtubule 255.27118 0.003917 
Mitochondrial proton-transporting ATP synthase, catalytic core 25.7566326 0.038496 
Molecular function   
Structural constituent of ribosome 10.29772 0.000124 
Phosphorylase activity 31.38516 0.001843 
Protein transporter activity 78.69595 0.01274 
Regulation of immune response 39.3949257 0.025321 
Biological process   
Regulation of immune response 39.3949257 0.025321 
Protein folding 39.2243859 0.001173 
Cell organisation and biogenesis 45.0147439 0.022191 
Intracellular signalling cascade 104.835002 0.00957 
Protein transport 28.6606135 0.034653 
Molecular function 156.991719 0.00639 

To further elucidate the pathways of tumour metastasis in the abdominal cavity, IPA was performed on 105 differentially expressed proteins in mice treated with exo-SKOV-3 and compared with those treated with exo-OVCAR-3. Our network analysis showed that the canonical pathway β-catenin/CTNNB1 gene coding protein was the main target for the dysregulated proteins (Figure 3D). In order to determine the potential role of β-catenin in the effect of exosomes on tumour growth and metastasis in vivo, the protein abundance of total β-catenin and β-catenin phosphorylated at various residues (β-catenin Ser33/Ser37/Thr41, β-catenin Thr41/Ser45, β-catenin Ser552, and β-catenin Ser675) were evaluated by Western blot. Significant differences (ANOVA, P<0.05) on β-catenin Thr41/Ser45, β-catenin Ser552, and β-catenin Ser675 across the groups were identified (Figure 3F–I). The protein abundance of β-catenin Ser675 and β-catenin Ser552 was higher in tumour tissues in mice injected with exo-OVCAR-3 compared with exo-SKOV-3 and control (Figure 3G,I). On the other hand, the protein abundance of β-catenin Thr41/Ser45 was higher in tumour tissues in mice injected with exo-SKOV-3 compared with exo-OVCAR-3 and control (Figure 3H). Taken together, our data suggest that the effect of exosomes on tumour growth and mestatasis in vivo were associated with changes on the phosphorylation of β-catenin.

Analysis of circulating exosomes in mice injected with exosomes from SKOV-3 and OVCAR-3 cells

Due to the fact that most tumour exosomes can be found in biological fluids including plasma, we proposed that the changes in the OVCA tumour growth and metastasis might be reflected in circulating exosome protein profile. Exosomes were isolated from plasma and the protein content determined and quantified (see ‘Materials and methods’ section). No significant difference between the number of circulating exosomes across the groups (exo-OVACR-3, exo-SKOV-3, and control) were identified (Supplementary Figure S4). Next we investigated the protein cargo of circulating exosomes by MS/MS SWATH analysis, and a total of 771 proteins were identified across all the groups. We identified proteins involved in the endosomal pathway (annexins, GTP-binding proteins), vesicles trafficking (Rabs) and lysosomal proteins (lysosome-associated membrane glycoprotein), cytoskeletal (actin, tubulin), cytoskeletal-binding protein (moesin), and motor proteins (myosins). Furthermore, proteins associated with OVCA progression including; apolipoprotein E (APOE), epidermal growth factor receptor (EGFR), fatty acid synthase (FASN), and detoxification enzymes such as a glutathione S-transferase involved in the defense mechanism against toxic molecules (chemodrugs) were identified (Supplementary Table S2). The protein profiles of the circulating exosomes were compared with the control (PBS) versus exo-OVCAR-3 (Figure 4A); control versus exo-SKOV-3 (Figure 4B); and exo-OVCAR-3 versus exo-SKOV-3 (Figure 4C); the data have been presented as a volcano plot (Supplementary Table S3). Among 771 proteins, 40 proteins were significantly differentially expressed in mice injected with exo-SKOV-3 compared with mice injected with exo-OVCAR-3. Of the 40 differentially expressed proteins, 18 were overexpressed while 29 were underexpressed (Figure 4C). The up-regulated proteins were analysed using IPA and a high number of proteins were associated with translation, transcription, transcriptional modification, free radical scavengers, protein synthesis, and cellular movement. Interestingly, circulating exosomes from mice injected with exo-SKOV-3 were significantly enriched with proteins known to be regulators of invasive responses such as adapter molecule CrK (CRK)), cell cycle progression (protein phosphatase 1G (PPM1G)); and negative regulators of cell stress response pathways (PPM1G and Cytochrome c oxidase). Finally, we performed a correlation analysis between the individual protein quantification and the number of metastasis lesions (Figure 4D) or tumour growth (Figure 4E). We identified a total of 33 proteins that significantly correlates (P<0.05) with the number of metastasis lesions such as high mobility group box 3 (HMGB3), EF hand domain containing 2 (EFHD2), CD9, and CUL3 (Figure 4D). Similarly, we identified 32 proteins that significantly correlate (P<0.05) with tumour growth such as transmembrane 4 superfamily protein (TM9S4), proteasome subunit β type, and lactate dehydrogenase A (LDHA) (Figure 4E and Table 3). These data suggest that condition-specific changes associated with tumour growth and mestastasis in cancer cells might be reflected in the circulating exosomes, thus, specific proteins within exosomes may be of clinical utility in the early identification of women with OVCA and cancer progression or metastasis.

Quantitative proteomic analysis from mice-exo

Figure 4
Quantitative proteomic analysis from mice-exo

(A) Volcano plot of 771 quantified proteins illustrating expression increases and decreases following treatment with exosomes. Data were selected at the cut-off values a fold change (FC) ≥ ±1.5 in relative abundance and a P-value <0.05. (A) Volcano plot of differentially expressed protein from circulating exosomes of mice treated with PBS compared with mice treated with exo-SKOV-3. (B) PBS treatment compared with exo-OVCAR-3 exosome treatment. (C) exo-OVCAR-3 treatment compared with exo-SKOV-3 exosome treatment. (D) A heat map representation of 33 proteins where expression correlated with the number of tumour nodes in the peritoneal cavity of animals that were treated with PBS, exo-OVCAR-3 and exo-SKOV-3 (in red, with the scale shown on the left below). (E) A heat map showing the correlation between tumour growth/size and expression of 32 proteins.

Figure 4
Quantitative proteomic analysis from mice-exo

(A) Volcano plot of 771 quantified proteins illustrating expression increases and decreases following treatment with exosomes. Data were selected at the cut-off values a fold change (FC) ≥ ±1.5 in relative abundance and a P-value <0.05. (A) Volcano plot of differentially expressed protein from circulating exosomes of mice treated with PBS compared with mice treated with exo-SKOV-3. (B) PBS treatment compared with exo-OVCAR-3 exosome treatment. (C) exo-OVCAR-3 treatment compared with exo-SKOV-3 exosome treatment. (D) A heat map representation of 33 proteins where expression correlated with the number of tumour nodes in the peritoneal cavity of animals that were treated with PBS, exo-OVCAR-3 and exo-SKOV-3 (in red, with the scale shown on the left below). (E) A heat map showing the correlation between tumour growth/size and expression of 32 proteins.

Table 3
A list of exosomal proteins that showed significant differential expression, which correlated with increasing numbers of metastatic nodes in mice peritoneal cavities
Protein name P-value 
AT1A1_HUMAN 0.013784 
X1433E_MOUSE 0.003738 
ACADV_HUMAN 0.032414 
SEPT9_HUMAN 0.04965 
IDH3A_MOUSE 0.045952 
APT_HUMAN 0.011121 
PCBP1_HUMAN 0.038881 
SYHC_HUMAN 0.003464 
HORN_HUMAN 0.043493 
SPCS2_HUMAN 0.01906372 
HMGB3_HUMAN 0.013646 
CO8B_MOUSE 0.022845 
EFHD2_HUMAN 0.015829 
AP3S1_MOUSE 0.004505 
HMCS2_MOUSE 0.044885 
BIG2_HUMAN 0.033804 
TSPO_HUMAN 0.046312 
MA2A1_MOUSE 0.045594 
ARF4_MOUSE 0.008421 
CELF1_HUMAN 0.028023 
SC24A_HUMAN 0.024332 
MIC10_HUMAN 0.04927 
CUL3_HUMAN 0.027538 
MPPA_HUMAN 0.014345 
CD9_MOUSE 0.012714 
MSS4_MOUSE 0.040158 
DMBT1_MOUSE 0.014488 
RRP1_HUMAN 0.002207 
GBG1_HUMAN 0.023686 
JAK1_HUMAN 0.042134 
WIPI3_MOUSE 0.00527 
CP4F8_HUMAN 0.046312 
Protein name P-value 
AT1A1_HUMAN 0.013784 
X1433E_MOUSE 0.003738 
ACADV_HUMAN 0.032414 
SEPT9_HUMAN 0.04965 
IDH3A_MOUSE 0.045952 
APT_HUMAN 0.011121 
PCBP1_HUMAN 0.038881 
SYHC_HUMAN 0.003464 
HORN_HUMAN 0.043493 
SPCS2_HUMAN 0.01906372 
HMGB3_HUMAN 0.013646 
CO8B_MOUSE 0.022845 
EFHD2_HUMAN 0.015829 
AP3S1_MOUSE 0.004505 
HMCS2_MOUSE 0.044885 
BIG2_HUMAN 0.033804 
TSPO_HUMAN 0.046312 
MA2A1_MOUSE 0.045594 
ARF4_MOUSE 0.008421 
CELF1_HUMAN 0.028023 
SC24A_HUMAN 0.024332 
MIC10_HUMAN 0.04927 
CUL3_HUMAN 0.027538 
MPPA_HUMAN 0.014345 
CD9_MOUSE 0.012714 
MSS4_MOUSE 0.040158 
DMBT1_MOUSE 0.014488 
RRP1_HUMAN 0.002207 
GBG1_HUMAN 0.023686 
JAK1_HUMAN 0.042134 
WIPI3_MOUSE 0.00527 
CP4F8_HUMAN 0.046312 

Discussion

Late diagnosis and tumour metastasis are the main causes of poor prognosis in OVCA [23]. Recent in vitro studies have suggested that exosomes have a role in OVCAprogression and metastasis via transfering bioactive molecules to cells within the tumour microenvironment [10,19]. In the present study, we identified that exosomes isolated from a highly invasive OVCA cell line (SKOV-3) promote metastasis in vivo compared with exosomes from a cell line with low invasion capacity (OVCAR-3).

In the present study, we found that exosomes from SKOV-3 were enriched with proteins that modulate cancer signalling by such as ACTN4, CD44, and Collagen type IV. The ACTN4 actin binding protein particularly plays a key role in cancer cells’ motility [24]. ACTN4 participates in migration and focal adhesion of different cancer types, including OVCA [25–27].

When comparing the SKOV-3 exosomes with the OVCAR-3 exosomes in vivo, the SKOV-3 exosomes increased the number of tumour nodes in the abdominal cavity of mice. The bioinformatics analysis predicated that these dysregulated proteins are associated with the β-catenin pathway. However, as hypo-phosphorylated β-catenin forms enhance the translocation into the nucleus and interact with transcription factors, the phosphorylated β-catenin forms were assessed. The β-catenin, which translocates into the nucleus, increases the expression of migration-related genes in the cells [28]. In most cancer cases, β-catenin accumulate in both the cytoplasm and the nucleus. Cytoplasmic β-catenin translocates into the nucleus and binds to the transcription factors of the lymphoid-enhancing factor-1 (LEF-1)/T-cell factor (TCF) increasing expression of cancer progression -related genes, such as c-Myc (MYC) and cyclin D1 (CCND1) [29]. However, in the absence of a Wnt ligand, β-catenin is maintained at an extremely low level. In this scenario, β-catenin forms a complex with casein kinase-1 (CK-1), glycogen synthase kinase-3 (GSK-3), and axin. This complex allows phosphorylation of β-catenin at positions Ser45, Thr41, Ser37, and Ser33. This heavy phosphorylated β-catenin becomes earmarked for ubiquitination and proteasomal targeting for degradation in the cytosol [30]. It was determined that the expressions of the total and non-active forms of β-catenin (phosphorylated β-catenin at positions Ser45, Thr41, Ser37 and Ser33) were not significantly affected by different treatments; however, the phosphorylated β-catenins were significantly influenced by treatment. Specifically, when metastatic tumour tissues were injected with exo-SKOV-3, the expressions of β-catenin Ser45 and Thr41 were increased. In the literature, transfecting endothelial cells with β-catenin Ser45 and Thr41 attenuated E-cadherin and modulated endothelial cell–endothelial cell junction integrity increasing endothelial permeability [31]. Another study reported that a reduction in phosphorylation of inactive β-catenins (Ser45, Thr41, Ser37 and Ser33) at residues of Ser33 and Ser37 by Gamide and Ggly, allowed β-catenin phosphorylated at residues Ser45 and Thr41 to translocate into the nucleus and mediate nuclear signalling by canonical pathways [32].

On the other hand, phosphorylated β-catenin at residue Ser552 and β-catenin at residue Ser675 increased in mice injected with ex-OVCAR-3 in contrast with mice injected with exo-SKOV-3. Phosphorylation of β-catenin Ser552 by protein kinase A (PKA) has been reported to promote the transcriptional activity of β-catenin [33]. Another study demonstrated that phosphorylation of β-catenin at Ser552 by PKA-dependent activation promotes interaction with TCF-4. This resulted in an increased transcriptional activity of TCF-4 and other genes implicated in the cell cycle progression [34]. Moreover, we found that the expression of phosphorylated β-catenin at residue Ser675 increased in mice injected with ex-OVCAR-3. Phosphorylation of β-catenin Ser675 by PKA has been reported to promote the transcriptional activity of β-catenin through a non-canonical pathway that does not involve the stabilisation of β-catenin [33]. Taurin et al., showed that phosphorylation of β-catenin at Ser675 site promoted the interaction between β-catenin and its co-activator CREB-binding protein (CBP) [33] that has been reported to participate in many biological processes including growth control [33,35]. In our study we found that exosomes obtained from OVCA cell lines with different migration capacities may possibly activate phosphorylation of β-catenin at different sites and promote different migration/metastasis rates in vivo and in vitro. Thus, our results provide new insights into how exosomes trigger the expression of phosphorylated β-catenin that could play a role in cancer progression and metastasis.

In the present study, it was hypothesised that the protein profile in the circulating exosomes might reflect the changes that occur in tumour cells, hence, circulating exosomes were analysed. We found 33 proteins which significantly correlated with the number of metastasis lesions, such as HMGB3, EFHD2, CD9, and Cullin3. HMGB3 has been reported to be over-expressed in colorectal cancer [36], gastric carcinoma [37], and bladder cancer [38]. In a study involving lung cancer, HMGB3 gene expression was found to increase in cancer tissue throughout advanced stagesOverexpression of HMGB3 can promote colorectal cancer cell proliferation and migration by activating WNT/β-catenin pathways [36]. Conversely, knockdown of HMGB3 expression in vitro revealed the dependence of this protein in cancer cell growth and invasion [39]. We also identified an increase in the expression of EFHD2. Increased EFHD2 expression has been reported in various human cancer tissues, particularly at highly invasive stages of malignant melanoma [40], and correlated with postsurgical recurrence of stage-I lung adenocarcinoma [41]. Ectopic expression of EFHD2 promoted expression of lamellipodia and membrane ruffles [42] and stimulated epithelial-to-mesenchymal transition [41]. In this study, the tetraspanin family protein, CD9, significantly correlated with the number of metastatic lesions. CD9 has been shown to participate in diverse functions associated with cancer progression and metastasis in multiple cancer types in vitro and in vivo [43–45]. In the OVCA context, increased expression of CD9 has been found in ascites-derived stromal cells from patients, promoting angiogenesis and tumorigenicity [46]. Finally, it was determined that the Cullin 3 expression is positively associated with metastatic lesions in mice. The expression of Cullin3 protein levels has been linked to bladder cancer aggressiveness and liver tumorigenesis [47,48]. Likewise, in breast cancer, Cullin3 expression is significantly enriched in breast cancer lesions, compared with normal breast tissue, and its expression is positively associated with metastasis [49]. Overexpression of Cullin3 protein in breast cancer cells promotes cell invasion, metastasis, and epithelial–mesenchymal transition (EMT) by targeting breast cancer metastasis-suppressor 1 (BRMS1) for degradation. Conversely, knocking down Cullin3 expression in breast cancer cells inhibits cell invasion, metastasis, and EMT [49].

On the other hand, we found that the expression of TM9S4 and LDHA were correlated with tumour growth/size. TM9S4 interacts with a variety of transmembrane and cytosolic proteins, including integrins, growth factor receptors, G-protein-coupled receptors and their intracellular proteins. Overexpression of TM4SF4 in hepatocellular carcinoma promotes cancer cell proliferation; however, the knockdown of TM4SF4 expression leads to the inhibition of cancer cell growth, decreasing cell adhesion to fibronectin [50]. The expression of LDHA is dysregulated in multiple cancer types [51–53]. For example, LDHA is overexpressed in cancer tissue in the context of oesophageal squamous cell carcinoma, and it is associated with augmented tumour cell survival [52]. Indeed, LDHA has been reported to induce cancer cell growth and migration in gastric cancer cells [53].

Strengths and limitations of the study

In the present study, we examined the role of exosomes in cancer metastasis using an in vivo model. Our findings provide valuable information about the involvement of cancer-derived exosomes in inducing cancer progression and metastasis in the peritoneal cavity. We also identified a set of oncogenic protein-containing exosomes that correlated with increased metastatic nodes; which could be used as a minimally invasive clinical test to monitor cancer progression with further validation studies. Although this design is imperfect due to a few limitations, it is sufficient to demonstrate the effect of highly invasive OVCA-derived exosomes in promoting cancer progression and metastasis. There remains a number of limitations that may affect the interpretation of the results. First, we used animals with SKOV-3-luc cancer cells and treated mice with exo-SKOV-3, exo-OVCAR-3, or PBS. It is possible that there is interference between the SKOV-3 injected exosomes and the SKOV-3-luc tumour cells releasing exosomes. It would be interesting to compare exo-OVCAR-3 and a more aggressive cell line than exo-SKOV-3. In addition, exosome bio-distribution experiments to identify the targeting preferences of exo-SKOV-3 and exo-OVCAR-3 in the peritoneal cavity may provide further insight for the observations in the present study. Further controls such as heat inactivation or sonication of the exosomes might be used to determine the importance of intact vesicles on the tumour growth. Several questions still need to be addressed before these research findings become translationally relevant. Furthermore, the clinical value of the circulating oncogenic protein-containing exosomes requires further investigation using ascites fluids and human plasma from ovarian-cancer patients. Based on the fact that most peritoneal fluid contents in the abdominal cavity are cleared by the peritoneal membrane and are directly transported into the blood, further studies involving circulating exosomes and ascites-derived exosome content from OVCA patients are warranted.

Conclusions

Exosomes play roles in cell–cell communication under pathological conditions that might be important in clinical applications such as diagnostic tools. Our findings provide valuable information for the direction of future research into the pathological function of exosomes in cancer metastasis. Our results show that exosomes can promote cancer metastasis, an effect associated with changes in the protein content of tumour tissues and modulation of wnt/bcatenin signalling pathway. Interestingly, circulating exosomes from highly metastatic animal models are enriched in proteins involved in cancer metastasis. The present study identified a set of exosome-based biomarkers that might be developed as diagnostic biomarkers for OVCA progression and metastasis.

Clinical prespectives

  • The effect and mechanisms associated with the functions of exosomes on tumour growth and metastasis is poorly understood.

  • The present study demonstrated that exosomes derived from high migratory OVCA cells induce metastasis in vivo while exosomes derived from low migratory OVCA cells promte tumour growth.

  • The findings of the present study suggests that the effect of exosomes on tumour growth and mestatasis in vivo were associated with changes in the phosphorylation of β-catenin and these changes might be reflected in the circulating exosomes.

Acknowledgments

The authors acknowledge the editorial assistance of Debbie Bullock (UQ Centre for Clinical Research, The University of Queensland).

Author Contribution

C.S. and M.A. conceived and designed the study. M.A., Y.H., and C.P. performed the experiments. C.S., M.A., A.L., F.Z., and D.G. performed data analysis. M.A. and C.S. wrote the initial draft of the manuscript. C.S., M.A., D.G., A.L., L.P., Y.H., and J.D.H. edited the manuscript. All authors reviewed/edited the manuscript and approved the final version.

Competing Interests

The authors declare that there are no competing interests associated with the manuscript.

Funding

This work was supported by the Lions Medical Research Foundation (to C.S.); the UQ-Ochsner Seed Fund for Collaborative Research; The University of Queensland; the Faculty of Medicine M+BS Emerging Leaders Medical Research Grant; the Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT) [grant number 1170809]; the Ovarian Cancer Research Foundation; and the scholarship from King Saud University, Ryiadh, K.S.A. (to M.A.).

Abbreviations

     
  • ANOVA

    analysis of variance

  •  
  • DDA

    data-dependent acquisition

  •  
  • EFHD2

    EF hand domain containing 2

  •  
  • EMT

    epithelial–mesenchymal transition

  •  
  • EV

    extracellular vesicle

  •  
  • FBS

    foetal bovine serum

  •  
  • HMGB3

    high mobility group box 3

  •  
  • IPA

    Ingenuity Pathway Analysis

  •  
  • IVIS

    in vivo imaging system

  •  
  • LC-MS/MS

    liquid chromatography–mass spectrometry

  •  
  • LDHA

    proteasome subunit β type-and lactate dehydrogenase A

  •  
  • NSG

    NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ

  •  
  • NTA

    nanoparticle tracking analysis

  •  
  • OVCA

    ovarian cancer

  •  
  • PBS

    phosphate-buffered saline

  •  
  • PKA

    protein kinase A

  •  
  • PPM1G

    protein phosphatase 1G

  •  
  • SKOV-3-Luc

    luciferase-expressed SKOV-3 cells

  •  
  • ST50

    half-maximal stimulatory time

  •  
  • SWATH-MS

    Sequential Windowed Acquisition of All Theoretical Mass Spectra-MS

  •  
  • TCF

    T-cell factor

  •  
  • TEM

    transmission electron microscopy

  •  
  • TM9S4

    transmembrane 4 superfamily protein

References

References
1.
Alharbi
M.
,
Zuñiga
F.
,
Elfeky
O.
,
Guanzon
D.
,
Lai
A.
,
Rice
G.
et al. .
(
2018
)
The potential role of miRNAs and exosomes in chemotherapy in ovarian canc.
Endocr Relat Cancer.
25
,
R663
R685
2.
Lengyel
E.
(
2010
)
Ovarian cancer development and metastasis
.
Am. J. Pathol.
177
,
1053
1064
[PubMed]
3.
Bast
R.C.
Jr
,
Hennessy
B.
and
Mills
G.B.
(
2009
)
The biology of ovarian cancer: new opportunities for translation
.
Nat. Rev. Cancer
9
,
415
[PubMed]
4.
Cheng
L.
,
Wu
S.
,
Zhang
K.
and
Xu
T.
(
2017
)
A comprehensive overview of exosomes in ovarian cancer: emerging biomarkers and therapeutic strategies
.
J. Ovarian Res.
10
,
73
[PubMed]
5.
Halkia
E.
,
Spiliotis
J.
and
Sugarbaker
P.
(
2012
)
Diagnosis and management of peritoneal metastases from ovarian cancer
.
Gastroenterol. Res. Prac.
2012
,
[PubMed]
6.
Mutsaers
S.E.
(
2004
)
The mesothelial cell
.
Int. J. Biochem. Cell Biol.
36
,
9
16
7.
Webber
J.
,
Steadman
R.
,
Mason
M.D.
,
Tabi
Z.
and
Clayton
A.
(
2010
)
Cancer exosomes trigger fibroblast to myofibroblast differentiation
.
Cancer Res.
70
,
9621
9630
[PubMed]
8.
Webber
J.P.
,
Spary
L.K.
,
Sanders
A.J.
,
Chowdhury
R.
,
Jiang
W.G.
,
Steadman
R.
et al. .
(
2014
)
Differentiation of tumour-promoting stromal myofibroblasts by cancer exosomes
.
Oncogene
34
,
290
[PubMed]
9.
Yokoi
A.
,
Yoshioka
Y.
,
Yamamoto
Y.
,
Ishikawa
M.
,
Ikeda
S.-I.
,
Kato
T.
et al. .
(
2017
)
Malignant extracellular vesicles carrying MMP1 mRNA facilitate peritoneal dissemination in ovarian cancer
.
Nat. Commun.
8
,
14470
[PubMed]
10.
Lobb
R.J.
,
Lima
L.G.
and
Möller
A.
(
2017
)
Exosomes: key mediators of metastasis and pre-metastatic niche formation
.
Semin. Cell Dev. Biol.
67
,
3
10
11.
Hessvik
N.P.
and
Llorente
A.
(
2017
)
Current knowledge on exosome biogenesis and release
.
Cell. Mol. Life Sci.
1
16
12.
Kobayashi
M.
,
Rice
G.E.
,
Tapia
J.
,
Mitchell
M.D.
and
Salomon
C.
(
2015
)
Exosomes are fingerprints of originating cells: potential biomarkers for ovarian cancer
.
Res. Rep. Biochem.
5
,
101
109
13.
Properzi
F.
,
Logozzi
M.
and
Fais
S.
(
2013
)
Exosomes: the future of biomarkers in medicine
.
Biomark. Medicine
7
,
769
778
[PubMed]
14.
Steinbichler
T.B.
,
Dudás
J.
,
Riechelmann
H.
and
Skvortsova
I.I.
(
2017
)
The role of exosomes in cancer metastasis
.
Semin. Cancer Biol
44
,
170
181
15.
Nakamura
K.
,
Sawada
K.
,
Kinose
Y.
,
Yoshimura
A.
,
Toda
A.
,
Nakatsuka
E.
et al. .
(
2016
)
Exosomes promote ovarian cancer cell invasion through transfer of CD44 to peritoneal mesothelial cells
.
Mol. Cancer Res.
15
,
[PubMed]
16.
Cho
J.A.
,
Park
H.
,
Lim
E.H.
,
Kim
K.H.
,
Choi
J.S.
,
Lee
J.H.
et al. .
(
2011
)
Exosomes from ovarian cancer cells induce adipose tissue-derived mesenchymal stem cells to acquire the physical and functional characteristics of tumor-supporting myofibroblasts
.
Gynecol. Oncol.
123
,
379
386
[PubMed]
17.
Reza
A.M.M.T.
,
Choi
Y.-J.
,
Yasuda
H.
and
Kim
J.-H.
(
2016
)
Human adipose mesenchymal stem cell-derived exosomal-miRNAs are critical factors for inducing anti-proliferation signalling to A2780 and SKOV-3 ovarian cancer cells
.
Sci. Rep.
6
,
38498
[PubMed]
18.
Cappellesso
R.
,
Tinazzi
A.
,
Giurici
T.
,
Simonato
F.
,
Guzzardo
V.
,
Ventura
L.
et al. .
(
2014
)
Programmed cell death 4 and micro RNA 21 inverse expression is maintained in cells and exosomes from ovarian serous carcinoma effusions
.
Cancer Cytopathol.
122
,
685
693
[PubMed]
19.
Sharma
S.
,
Alharbi
M.
,
Kobayashi
M.
,
Lai
A.
,
Guanzon
D.
,
Zuñiga
F.
et al. .
(
2018
)
Proteomic analysis of exosomes reveals an association between cell invasiveness and exosomal bioactivity on endothelial and mesenchymal cell migration in-vitro
.
Clin. Sci.
132
,
CS20180425
20.
Théry
C.
,
Witwer
K.W.
,
Aikawa
E.
,
Alcaraz
M.J.
,
Anderson
J.D.
,
Andriantsitohaina
R.
et al. .
(
2018
)
Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines
.
J. Extracell. Vesicles
7
,
1535750
[PubMed]
21.
He
Y.
,
Wu
A.
,
Harrington
B.
,
Davies
C.
,
Wallace
S.
,
Adams
M.
et al. .
(
2016
)
Elevated CDCP1 predicts poor patient outcome and mediates ovarian clear cell carcinoma by promoting tumor spheroid formation, cell migration and chemoresistance
.
Oncogene
35
,
468
[PubMed]
22.
Shevchenko
A.
,
Tomas
H.
,
Havli
J.
,
Olsen
J.V.
and
Mann
M.
(
2006
)
In-gel digestion for mass spectrometric characterization of proteins and proteomes
.
Nat. Protoc.
1
,
2856
[PubMed]
23.
Yeung
T.-L.
,
Leung
C.S.
,
Yip
K.-P.
,
Au Yeung
C.L.
,
Wong
S.T.C.
and
Mok
S.C.
(
2015
)
Cellular and molecular processes in ovarian cancer metastasis. A review in the theme: cell and molecular processes in cancer metastasis.
.
Am. J. Physiol. Cell Physiol.
309
,
C444
C456
24.
Shao
H.
,
Wang
J.H.-C.
,
Pollak
M.R.
and
Wells
A.
(
2010
)
α-actinin-4 is essential for maintaining the spreading, motility and contractility of fibroblasts
.
PLoS ONE
5
,
e13921
[PubMed]
25.
Yamamoto
S.
,
Tsuda
H.
,
Honda
K.
,
Kita
T.
,
Takano
M.
,
Tamai
S.
et al. .
(
2007
)
Actinin-4 expression in ovarian cancer: a novel prognostic indicator independent of clinical stage and histological type
.
Mod. Pathol.
20
,
1278
[PubMed]
26.
Gao
Y.
,
Li
G.
,
Sun
L.
,
He
Y.
,
Li
X.
,
Sun
Z.
et al. .
(
2015
)
ACTN4 and the pathways associated with cell motility and adhesion contribute to the process of lung cancer metastasis to the brain
.
BMC Cancer
15
,
277
[PubMed]
27.
Kikuchi
S.
,
Honda
K.
,
Tsuda
H.
,
Hiraoka
N.
,
Imoto
I.
,
Kosuge
T.
et al. .
(
2008
)
Expression and gene amplification of actinin-4 in invasive ductal carcinoma of the pancreas
.
Clin. Cancer Res.
14
,
5348
5356
[PubMed]
28.
Valenta
T.
,
Hausmann
G.
and
Basler
K.
(
2012
)
The many faces and functions of β‐catenin
.
EMBO J.
31
,
2714
2736
[PubMed]
29.
Polakis
P.
(
2000
)
Wnt signaling and cancer
.
Genes Dev.
14
,
1837
1851
30.
Wu
G.
and
He
X.
(
2006
)
Threonine 41 in β-catenin serves as a key phosphorylation relay residue in β-catenin degradation
.
Biochemistry
45
,
5319
5323
[PubMed]
31.
Chang
S.-F.
,
Chen
L.-J.
,
Lee
P.-L.
,
Lee
D.-Y.
,
Chien
S.
and
Chiu
J.-J.
(
2014
)
Different modes of endothelial–smooth muscle cell interaction elicit differential β-catenin phosphorylations and endothelial functions
.
Proc. Natl. Acad. Sci. U.S.A.
111
,
1855
1860
32.
He
H.
,
Shulkes
A.
and
Baldwin
G.S.
(
2008
)
PAK1 interacts with β-catenin and is required for the regulation of the β-catenin signalling pathway by gastrins
.
Biochim. Biophys. Acta
1783
,
1943
1954
33.
Taurin
S.
,
Sandbo
N.
,
Qin
Y.
,
Browning
D.
and
Dulin
N.O.
(
2006
)
Phosphorylation of β-catenin by cyclic AMP-dependent protein kinase
.
J. Biol. Chem.
281
,
9971
9976
[PubMed]
34.
Taurin
S.
,
Sandbo
N.
,
Yau
D.M.
,
Sethakorn
N.
and
Dulin
N.O.
(
2008
)
Phosphorylation of β-catenin by PKA promotes ATP-induced proliferation of vascular smooth muscle cells
.
Am. J. Physiol. Cell Physiol.
294
,
C1169
C1174
[PubMed]
35.
Tang
Z.
,
Yu
W.
,
Zhang
C.
,
Zhao
S.
,
Yu
Z.
,
Xiao
X.
et al. .
(
2016
)
CREB‐binding protein regulates lung cancer growth by targeting MAPK and CPSF4 signaling pathway
.
Mol. Oncol.
10
,
317
329
[PubMed]
36.
Zhang
Z.
,
Chang
Y.
,
Zhang
J.
,
Lu
Y.
,
Zheng
L.
,
Hu
Y.
et al. .
(
2017
)
HMGB3 promotes growth and migration in colorectal cancer by regulating WNT/β-catenin pathway
.
PLoS ONE
12
,
e0179741
[PubMed]
37.
Tang
H.-R.
,
Luo
X.-Q.
,
Xu
G.
,
Wang
Y.
,
Feng
Z.-J.
,
Xu
H.
et al. .
(
2012
)
High mobility group-box 3 overexpression is associated with poor prognosis of resected gastric adenocarcinoma
.
World J. Gastroenterol.
18
,
7319
[PubMed]
38.
Li
M.
,
Cai
Y.
,
Zhao
H.
,
Xu
Z.
,
Sun
Q.
,
Luo
M.
et al. .
(
2015
)
Overexpression of HMGB3 protein promotes cell proliferation, migration and is associated with poor prognosis in urinary bladder cancer patients
.
Tumour Biol.
36
,
4785
4792
[PubMed]
39.
Guo
S.
,
Wang
Y.
,
Gao
Y.
,
Zhang
Y.
,
Chen
M.
,
Xu
M.
et al. .
(
2016
)
Knockdown of high mobility group-box 3 (HMGB3) expression inhibits proliferation, reduces migration, and affects chemosensitivity in gastric cancer cells
.
Med. Sci. Monitor
22
,
3951
[PubMed]
40.
Huh
Y.H.
,
Oh
S.
,
Yeo
Y.R.
,
Chae
I.H.
,
Kim
S.H.
,
Lee
J.S.
et al. .
(
2015
)
Swiprosin-1 stimulates cancer invasion and metastasis by increasing the Rho family of GTPase signaling
.
Oncotarget
6
,
13060
[PubMed]
41.
Fan
C.-C.
,
Cheng
W.-C.
,
Huang
Y.-C.
,
Sher
Y.-P.
,
Liou
N.-J.
,
Chien
Y.-C.
et al. .
(
2017
)
EFHD2 promotes epithelial-to-mesenchymal transition and correlates with postsurgical recurrence of stage I lung adenocarcinoma
.
Sci. Rep.
7
,
14617
[PubMed]
42.
Huh
Y.H.
,
Kim
S.H.
,
Chung
K.-H.
,
Oh
S.
,
Kwon
M.-S.
,
Choi
H.-W.
et al. .
(
2013
)
Swiprosin-1 modulates actin dynamics by regulating the F-actin accessibility to cofilin
.
Cell. Mol. Life Sci.
70
,
4841
4854
[PubMed]
43.
Rappa
G.
,
Green
T.M.
,
Karbanová
J.
,
Corbeil
D.
and
Lorico
A.
(
2015
)
Tetraspanin CD9 determines invasiveness and tumorigenicity of human breast cancer cells
.
Oncotarget
6
,
7970
[PubMed]
44.
Huan
J.
,
Gao
Y.
,
Xu
J.
,
Sheng
W.
,
Zhu
W.
,
Zhang
S.
et al. .
(
2015
)
Overexpression of CD9 correlates with tumor stage and lymph node metastasis in esophageal squamous cell carcinoma
.
Int. J. Clin. Exp. Pathol.
8
,
3054
[PubMed]
45.
Kischel
P.
,
Bellahcene
A.
,
Deux
B.
,
Lamour
V.
,
Dobson
R.
,
De Pauw
E.
et al. .
(
2012
)
Overexpression of CD9 in human breast cancer cells promotes the development of bone metastases
.
Anticancer Res.
32
,
5211
5220
[PubMed]
46.
Pasquet
M.
,
Golzio
M.
,
Mery
E.
,
Rafii
A.
,
Benabbou
N.
,
Mirshahi
P.
et al. .
(
2010
)
Hospicells (ascites‐derived stromal cells) promote tumorigenicity and angiogenesis
.
Int. J. Cancer
126
,
2090
2101
[PubMed]
47.
Yuan
W.-C.
,
Lee
Y.-R.
,
Huang
S.-F.
,
Lin
Y.-M.
,
Chen
T.-Y.
,
Chung
H.-C.
et al. .
(
2011
)
A Cullin3-KLHL20 Ubiquitin ligase-dependent pathway targets PML to potentiate HIF-1 signaling and prostate cancer progression
.
Cancer Cell
20
,
214
228
[PubMed]
48.
Grau
L.
,
Luque-Garcia
J.L.
,
González-Peramato
P.
,
Theodorescu
D.
,
Palou
J.
,
Fernandez-Gomez
J.M.
et al. .
(
2013
)
A quantitative proteomic analysis uncovers the relevance of CUL3 in bladder cancer aggressiveness
.
PLoS ONE
8
,
e53328
[PubMed]
49.
Huo
X.
,
Li
S.
,
Shi
T.
,
Suo
A.
,
Ruan
Z.
,
Guo
H.
et al. .
(
2015
)
Cullin3 promotes breast cancer cells metastasis and epithelial-mesenchymal transition by targeting BRMS1 for degradation
.
Oncotarget
6
,
41959
[PubMed]
50.
Paolillo
R.
,
Spinello
I.
,
Quaranta
M.T.
,
Pasquini
L.
,
Pelosi
E.
,
Coco
F.L.
et al. .
(
2015
)
Human TM9SF4 is a new gene down-regulated by hypoxia and involved in cell adhesion of leukemic cells
.
PLoS ONE
10
,
e0126968
[PubMed]
51.
Shim
H.
,
Dolde
C.
,
Lewis
B.C.
,
Wu
C.-S.
,
Dang
G.
,
Jungmann
R.A.
et al. .
(
1997
)
c-Myc transactivation of LDH-A: implications for tumor metabolism and growth
.
Proc. Natl. Acad. Sci. U.S.A.
94
,
6658
6663
52.
Yao
F.
,
Zhao
T.
,
Zhong
C.
,
Zhu
J.
and
Zhao
H.
(
2013
)
LDHA is necessary for the tumorigenicity of esophageal squamous cell carcinoma
.
Tumour Biol.
34
,
25
31
[PubMed]
53.
Cai
Z.
,
Zhao
J.-S.
,
Li
J.-J.
,
Peng
D.-N.
,
Wang
X.-Y.
,
Chen
T.-L.
et al. .
(
2010
)
A combined proteomics and metabolomics profiling of gastric cardia cancer reveals characteristic dysregulations in glucose metabolism
.
Mol. Cell. Proteomics
9
,
2617
2628