Exosomal ANXA2 derived from ovarian cancer cells regulates epithelial‐mesenchymal plasticity of human peritoneal mesothelial cells

Abstract Ovarian cancer, one of the malignant gynaecological tumours with the highest mortality rate among female reproductive system, is prone to metastasis, recurrence and chemotherapy resistance, causing a poor prognosis. Exosomes can regulate the epithelial‐mesenchymal plasticity of tumour cells, remodel surrounding tumour microenvironment, and affect tumour cell proliferation, invasion and metastasis. However, the function and mechanism of exosomes in the intraperitoneal implantation of ovarian cancer remain unclear. In this study, exosomal annexin A2 (ANXA2) derived from ovarian cancer cells was co‐cultured with human peritoneal mesothelial (HMrSV5) cells; functional experiments were conducted to explore the effects of exosomal ANXA2 on the biological behaviour of HMrSV5 and the related mechanisms. This study showed that ANXA2 in ovarian cancer cells can be transferred to HMrSV5 cells through exosomes, exosomal ANXA2 can not only promote the migration, invasion and apoptosis of HMrSV5 cells, but also regulates morphological changes and fibrosis of HMrSV5 cells. Furthermore, ANXA2 promotes the mesothelial‐mesenchymal transition (MMT) and degradation of the extracellular matrix of HMrSV5 cells through PI3K/AKT/mTOR pathway, finally affects pre‐metastasis microenvironment of ovarian cancer, which provides a new theoretical basis for the mechanism of intraperitoneal implantation and metastasis of ovarian cancer.

In above process, peritoneal mesothelial cells, as the main components of the human peritoneum, 6,7 could create a suitable environment for tumour intraperitoneal implantation and metastasis through the change of biological properties. 8,9 Accumulating studies have suggested that various tumour cells could modulate the characteristics of peritoneal mesothelial cells by secreting various cytokines and bioactive substances, thus leading to intraperitoneal implantation. [10][11][12] Therefore, a better understanding of the underlying molecular mechanism of malignant progression and peritoneal metastasis of ovarian cancer are essential to improve the survival and prognosis of patients and explore new therapeutic targets with ovarian cancer.
Exosomes, with a lipid bilayer membrane, were regarded as microvesicles with a diameter of about 30-150 nm, and widely present in a variety of body fluids, such as blood, urine, ascites, milk, bile and saliva. 13 Exosomes could carry various biologically active substances, including nucleic acids (e.g., mRNA, miRNA, lncRNA, circRNA and DNA), lipids and proteins, [14][15][16] which were involved in numbers of physiological activities and pathological processes, such as signal communication, molecular transport, immune response and antigen presentation. 17 In recent years, many studies have shown that exosomes participated in the regulation of proliferation, invasion, metastasis and epithelial-mesenchymal transition (MMT) of tumour cells, remodelling tumour environment, chemotherapy resistance and angiogenesis, 18,19 including gastric cancer, lung cancer and ovarian cancer. Especially, exosomes derived from cancer cells have potential abilities to target their parent cancer cells through delivering anti-tumour drugs. 20 Moreover, studies have shown that exosomal non-coding RNA were frequently regarded as potential prognostic biomarkers, including miRNA-200 family, miRNA-141 and miRNA-205, 21,22 while investigations of the function and mechanism of proteins in exosomes have been limited.
Annexin A2 (ANXA2), as calcium-dependent membrane phospholipid-binding protein, participates in cell migration, inflammation, fibrinolysis, exocytosis and endocytosis. As an extracellular fibrinolytic receptor, ANXA2 can not only promote proteolysis, neovascularization, invasion and metastasis of tumour cells, but remodel the extracellular matrix23-24. 23,24 Our previous research has confirmed that ANXA2 was significantly associated with FIGO stages, pathological grade, lymph node metastasis and poor prognosis of ovarian cancer. 25,26 What is more, overexpression of ANXA2 promoted the invasion and metastasis of ovarian cancer cells and the tumorigenesis ability of transplanted tumours in nude mice. In recent years, proteomic studies also documented that ANXA2 was regarded as an important functional protein in exosomes secreted by bladder, colorectal and ovarian cancer cells. 27 ANXA2 can regulate the formation and transport of extracellular vesicles, 28 and miRNA can regulate the level of ANXA2 in exosome. 29 However, the function and mechanism of exosomal ANXA2 on invasion, metastasis and abdominal implantation of ovarian cancer remain unclear.
In this study, we focused on the effect and potential mechanism of exosomal ANXA2 derived from ovarian cancer cells on peritoneal implantation and metastasis of tumours. The results revealed that exosomal ANXA2 derived from ovarian cancer could confer tumour characteristics to HMrSV5 cells and regulate MMT and extracellular matrix degradation of HMrSV5 cells, finally forming pre-metastasis microenvironment suitable for intraperitoneal implantation and metastasis of ovarian cancer, which not only provide new insights into the molecular mechanism of intraperitoneal implantation and metastasis of ovarian cancer, but may also guide the development of therapeutic treatments to delay or inhibit tumour metastasis.

| Exosome extraction
Ovarian cancer cells were cultured to 70%-80% confluence (10-cm dish) with complete culture medium. Cells were washed with PBS three times after the supernatant was removed, and then routinely cultured in fresh serum-free medium for 48 h. The culture supernatant was collected and centrifuged at 300 g for 10 min at 4°C to remove dead cells and fragments in the culture supernatant. Then, the culture supernatant was centrifuged at 2,000 g for 10 min at 4°C to remove biopolymers and apoptotic bodies. A Millipore aseptic filter (pore size: 0.22 μm; Millipore) was used to remove large vesicles and particles as well as bacteria. Then, the filtered supernatant was centrifugated at 10,000 g for 30 min at 4°C to remove cell fragments, large vesicles and impurities. The precipitation at the bottom of the tube was discarded, and the supernatant was centrifuged at 100,000 g for 90 min at 4°C. The supernatant was removed, and the exosome precipitation

| Transmission electron microscope (TEM)
The morphology and structure of the exosomes in the supernatant of each cell culture were identified by TEM. The exosomes were resuspended in 100 μl PBS and stored at 4°C. Approximately 10-20 μl of exosome suspension was added to a carbon-coated copper mesh matched by the electron microscope. The copper mesh was exposed to room temperature for 30 min to precipitate the exosomes. Filter paper was used to absorb the liquid from the side of the filter. The exosomes were then negatively stained with 3% phosphotungstic acid dye solution (20-50 μl) for 20 min. After the filter paper was dried, the exosomes were washed with PBS three times. After drying at room temperature, the exosomes were observed and photographed under TEM.

| Nanoparticle Tracking Analysis (NTA)
The extracted exosome precipitate was re-suspended with 100 μl PBS and stored at 4°C. The exosome sample was diluted with PBS and fully mixed. The mixed exosome sample (300 μl) was detected by

Electrophoresis & Brownian Motion Video Analysis Laser Scattering
Microscopy, the particle size and purity of the exosomes were analysed by Zetaview visual nanoparticle tracer.

| Exosome fluorescence labelling uptake experiment
To determine whether the exosomes from ovarian cancer cells can be taken up by HMrSV5, 10 mg Dil dye (Solarbio) was adjusted to a concentration of 2 mg/ml with 5 ml dimethyl sulfoxide (DMSO). The purified exosomes were suspended in 200 μl PBS, 2 mg/ml Dil dye was added, and the working concentration was adjusted to 15 μM.
The mixture was incubated at room temperature for 30 min and then centrifuged at 100,000 g for 90 min to remove unbound dye, washed with PBS for three times. The supernatant was discarded, and Dillabelled exosomes were re-suspended in 200 μl PBS, and then coincubated with HMrSV5 for 6 h or 12 h, respectively. HMrSV5 cells co-cultured with Dil-labelled exosomes were fixed with 4% paraformaldehyde (Solarbio) for 20 min at room temperature and washed with PBS three times. Then, 500 μl 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) solution (1:1000 dilution with PBS) was incubated for 5 min and then washed with PBS for three times. The condition of exosomes taken up by cells was observed by laser confocal microscope.

| Real-time PCR
Total RNA was extracted by Trizol Reagent (Invitrogen) and quantified, reverse transcription was conducted according to the instructions of the PrimeScript RT reagent kit with gDNA Eraser (Takara Bio). The following procedure was implemented with SYBR Premix Ex Taq II kit (Takara Bio).
The PCR reaction was carried out on a fluorescence quantitative PCR instrument (7500 Real-Time PCR Detection System; Applied Biosystems).

| Cell migration ability detected by scratch test
The cells in logarithmic phase were routinely digested with trypsin and seeded in a 6-well culture plate. When the cells reached 90% confluence, the cell layer was gently scratched using a 100 μl micropipette tip, washed with PBS to remove floated cells, and then cultured in serum-free medium for 48 h at 37°C. The wound-healing ability of the cells was observed and photographed under a microscope at 0 and 48 h. This experiment was repeated three times.

| Cell apoptosis detected by flow cytometry
Cells in the exponential growth phase were digested with trypsin without EDTA and neutralized in complete medium. The single-cell suspension was collected and centrifuged at 978 g for 5 min, and then

| Cell cytoskeleton detected by Phalloidin staining
The cells in the exponential growth phase are routinely digested and centrifuged. When the cell density grows to 40%-50%, the culture medium is discarded and washed with PBS for three times. After fixed with 4% paraformaldehyde for 20 min and washed with PBS for three times. Cells were incubated with Triton X-100 (Beyotime) for 10 min. Subsequently, the cells were incubated with the Actin-Tracker Green-488 (green fluorescent probe) (Beyotime) diluted in 1% BSA with PBS (1:100) for 1 h at room temperature. Next, the nuclei were stained by DAPI diluted with PBS (1:1000) for 10 min at room temperature. Cell cytoskeleton was observed under a laser confocal microscope. and expressed as mean ± standard deviation (variance ± SD). Comparisons between two groups were analysed by Student's t-test. Single analysis of variance (ANOVA) was used to analyse comparisons among more than two groups. Bilateral p-values < 0.05 were considered statistically significant.   Figure 2D). Exosomal ANXA2 protein derived from the OVCAR3-ANXA2-H-exo group and ES-2-ANXA2-H-exo group were significantly higher than those in OVCAR3-ON-exo group and ES-2-ON-exo group (p < 0.05; Figure 2D).

| Expression of ANXA2 and migration and invasion abilities of HMrSV5 cells co-cultured with exosomes derived from OVCAR3 and ES-2 cells
To investigate whether HMrSV5 cells can take up exosomes derived  Figure S2A). Scratch test showed that migration capacities were significantly higher in HMrSV5 cells co-cultured with OVCAR3-exo or ES-2-exo than that in control group (both p < 0.05; Figure 3C,D). Transwell assay showed that the invasion capacities were significantly higher in HMrSV5 cells co-cultured with OVCAR3-exo or ES-2-exo than that in control group (both p < 0.05; Figure 3E,F).

| Effect of GW4869 on the expression of ANXA2 and migration and invasion abilities of HMrSV5 cells co-cultured with OVCAR3-exo and ES-2-exo
We further added the exosome inhibitor GW4869 to the cul-  Figure 4A). The levels of ANXA2 mRNA in HMrSV5 cells did not change significantly after addition of GW4869 in the culture medium by real-time PCR (p > 0.05, Figure S2B). Scratch test showed that, compared with those in the control groups, migration abilities were inhibited in the OVCAR3+GW4869 and ES-2+GW4869 groups after the addition of GW4869 (all p < 0.05; Figure 4B,C). Transwell assay showed that, compared with those in the control groups, invasion abilities were inhibited in the OVCAR3+GW4869 and ES-2+GW4869 groups after the addition of GW4869 (all p < 0.05; Figure 4D,E).

| Migration and invasion abilities of HMrSV5 cells co-cultured with exosomal ANXA2 derived from ovarian cancer cells
The influence of exosomal ANXA2 on biological characteristics of

| Effect of ANXA2 on migration, invasion, mesothelial-mesenchymal transition (MMT) and fibrosis of HMrSV5 cells
To confirm the effect and mechanism of ANXA2 on the biological behaviour of HMrSV5 cells, HMrSV5 cells with stable overexpression and inhibition of ANXA2 protein were constructed by lentiviral transfection. Western blot showed that the expression level of ANXA2 in the HMrSV5-shANXA2 group was significantly lower than that in the HMrSV5 and HMrSV5-NC groups (both p < 0.05).
The expression level of ANXA2 in the HMrSV5-ANXA2-H group was significantly higher than that in the control group (HMrSV5-ON) (p < 0.05; Figure S3). We explored the effect of ANXA2 on the migration ability of HMrSV5 cells by Scratch test. The results showed that the migration ability of HMrSV5-shANXA2 cells was significantly lower than that of HMrSV5 and HMrSV5-NC cells (both p < 0.05).
The migration ability of HMrSV5-ANXA2-H cells was significantly higher than that of HMrSV5-ON cells (p < 0.05; Figure 7A).
Transwell assay was employed to detect the effect of ANXA2 on the invasion ability of HMrSV5 cells. The results showed that the invasion ability of HMrSV5-shANXA2 cells was significantly decreased compared with that of HMrSV5 and HMrSV5-NC cells (both p < 0.05); the invasion ability of HMrSV5-ANXA2-H cells was significantly increased compared with that of HMrSV5-ON cells (p < 0.05; Figure 7B).  Figure 7D).

| Effect of ANXA2 on the proliferation, apoptosis and PI3K/AKT/mTOR signalling pathway of HMrSV5 cells
Cell Counting Kit-8 was utilized to detect the role of ANXA2 protein on the proliferation ability of HMrSV5 cells. The results showed that, compared with HMrSV5 and HMrSV5-NC groups, the proliferation ability of HMrSV5-shANXA2 cells showed no significant change (p > 0.05; Figure 7E); compared with HMrSV5-ON group, the proliferation ability of HMrSV5-ANXA2-H cells showed no significant change (p > 0.05; Figure 7E).  38,39 It has been reported that ANXA2 can be transported from cell to cell by exosomes in a Ca 2+ -dependent manner during the formation and release of exosomes. 28 However, the function and mechanism of exosomal ANXA2 on peritoneal implantation of ovarian cancer have not been fully elucidated.
Researches have shown that exosomal ANXA2 derived from breast cancer cells can not only promote angiogenesis in a tissue plasminogen activator (tPA) -dependent manner but also activate multiple signalling pathways and regulate the secretion of factors in macrophages. 40 Additionally, exosomal ANXA2 derived from serum was associated with malignant progression and angiogenesis in triple-negative breast cancer. 41 In this study, we firstly confirmed that ANXA2 protein in exosomes derived from ovarian cancer varies with the content of ANXA2 in the host cells, and the expression of ANXA2 was upregulated in HMrSV5 cells after co-cultured with OVCAR3-exo or ES-2-exo, indicating that ANXA2 can not only regulate the formation and release of exosomes secreted by ovarian cancer cells, but also transfer to HMrSV5 cells via exosomes, thus having potential effect on the biological characteristics of target cells.
Researches have confirmed that integrin α5β1/AEP and CD44 mediated by exosomes derived from ovarian cancer cells can induce the destruction of the mesothelial barrier. 10,38 Exosomal miR-99a-5p derived from ovarian cancer cells promotes tumour invasion and metastasis by regulating the expression of fibronectin and vitronectin in HMrSV5 cells, 39 suggesting biological cargos derived from exosome could perform important effects on recipient cells.
We further confirmed that exosomal ANXA2 derived from ovarian cancer can promote the migration, invasion and apoptosis of HMrSV5 cells, indicating that ovarian cancer cells can transfer their tumour characteristics to HMrSV5 cells through exosomal ANXA2, and further to form a pre-metastasis microenvironment suitable for ovarian cancer cell adhesion and colonization. However, during this process, exosomal ANXA2 derived from ovarian cancer cells has no significant effect on the proliferation of HMrSV5 cells, which may act in the early stage of ovarian cancer intraperitoneal implantation, and the proliferation capacity is not sufficient to change, or the process was affected and regulated by other proteins and nucleic acids which warrants further investigation.
A series of studies have indicated that exosomes not only participate in modulating EMT of tumour cells but are also involved in the MMT of mesothelial cells. 42 During abdominal implantation and metastasis of tumours, exosomes have the potential function to reshape the peritoneal microenvironment to promote adhesion and metastasis of cancers. 43 It is reported that gastric cancerderived exosomes can transfer various biological molecules to induce MMT of HMrSV5 cells, such as FasL, MMP2, nicotinamide N-methyltransferase (NNMT), tripartite motif-containing protein 3 (TRIM3), miR-106 and miR-21-5p, which could regulate biological characteristics of HMrSV5 cells, thus facilitating metastasis of gastric cancer cells. [44][45][46][47][48] Studies have shown that PI3K/AKT/mTOR pathway played critical roles in the development of cancer, 49 which can be abnormally activated in a variety of tumours and promote the growth, invasion, metastasis and EMT of tumours, such as ovarian cancer, breast cancer, gastric cancer, lung cancer, and head and neck squamous cell carcinoma. [50][51][52][53][54] In this study, we successfully confirmed that, after exosomal ANXA2 derived from ovarian cancer uptaken by HMrSV5, ANXA2 protein can not only promote migration, invasion, apoptosis and MMT of HMrSV5 through downstream PI3K/AKT/mTOR pathway, but also accelerate the degradation of extracellular matrix, which provided pre-metastatic niche for peritoneal invasion and metastasis of ovarian cancer. The above studies provide sufficient evidences that how the exosomal ANXA2 regulate biological behaviour and MMT of HMrSV5, providing new insights into the underlying mechanism of peritoneal metastasis of ovarian cancer.

| CON CLUS IONS
In conclusion, to the best of our knowledge, we demonstrated for the first time that the expression of ANXA2 in exosomes derived from ovarian cancer cells varies with the content of ANXA2 in these cells. Furthermore, exosomal ANXA2 can not only regulate the migration, invasion and apoptosis abilities of HMrSV5 cells but also induce MMT, degradation of the extracellular matrix and fibrosis through the PI3K/AKT/mTOR signalling pathway, thus remodelling the pre-metastasis microenvironment and facilitating peritoneal metastasis ( Figure 7I). However, further verification by clinical samples and more animal experiments are still required in future.
Taken together, our observations shed light on the function and mechanism by which exosomal ANXA2 derived from ovarian cancer cells promotes the migration, invasion and apoptosis abilities of HMrSV5 cells, providing credible evidence to clarify the mechanism of peritoneal implantation and metastasis of ovarian cancer.

ACK N OWLED G EM ENT
Not applicable.

CO N FLI C T O F I NTE R E S T
The authors have declared that no competing interest exists.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.