The exosome extraction kit was purchased from Shanghai Gefan Biology (Shanghai, China). Wnt/β-Catenin inhibitor XAV939 was purchased from R & D systems (Minneapolis, Minnesota, USA). CXCR4 antibody (11073-2-ap) was provided by Proteintech (Wuhan, Hubei, China). CXCR4 siRNA (sc-35421) was provided by Santa Cruz Biotechnology (Dallas, Texas, USA). Abcam (Cambridge, United Kingdom) supplied all other primary antibodies, including anti-N-cadherin and anti-E-cadherin. Other reagents were provided by Sigma Aldrich (Darmstadt, Germany).

Clean specimens of OCCC were selected for CAF isolation and culture. After two generations, purified third generation cells were used for preliminary identification. NFs cells were cultured in the same way from normal ovarian clean specimens removed during hysteromyoma operation. The cells were identified by morphological observation and immunocytochemical staining. The separation and purification process has been described in our previous study [12].Written informed consent was provided by the patients with OCCC, and the study was approved by the Research Ethics Committee of Shandong Cancer Hospital and Institute.

ATCC (Virginia, USA) supplied the ovarian cancer cell line ES2. According to the relevant instructions, the cells were propagated 1:2 when they were 70% confluent. The ES-2 cells were cultured in McCoy's5A medium containing 10% fetal bovine serum and 1% streptomycin. CAFs and NFs were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 mg/ml streptomycin.

CXCR-4 siRNA (sc-35421) was used to transfect CAFs with Lipofectamine 3000 (Invitrogen, CA). qRT-PCR and WB were used to verify the results.

Exosomes were extracted using the exosome extraction kit (Shanghai Gefan Biology, Shanghai, China). After successful cell culture, cells were grown in media without exocrine secretions. Cells were centrifuged to isolate supernatants which were in turn centrifuged for 20 min × 3000 g at 4°C. Supernatants were harvested and filtered using a 0.22μm filter. To 1ml of cell culture supernatant, 0.2ml of exosome separation and purification solution were added, then mixed well with a pipette gun. The mixture was placed in the refrigerator overnight at 4°C and subsequently centrifuged at 1500g×30 min at 4°C. After the supernatant was removed, the pellet was centrifuged at 1500g×5min and 4°C to remove residual supernatant.

Cell invasion analyses were performed using a 24-well plate housing a transwell chamber. The top section of the chamber was supplemented with serum-free medium and cell suspension (400μl, 5×10⁴ cells/L), while 10% fetal bovine serum was added to the bottom section in 500μl of culture medium. Cells were cultured under standard conditions for 24 hours, then migrated cells were isolated, fixed with formaldehyde and stained using crystal violet (0.1%). The mean number of migrated cells was estimated by counting cells from 10 random microscopic visual fields.

Standard Radio immunoprecipitation Assay (RIPA) extraction methodology was followed to extract total cell proteins using RIPA solution. The bicinchoninic acid assay (BCA) was used to establish the concentration of protein in cell lysates. Membrane blocking to avoid non-specific antibody binding was carried out for 1 hour using 5% defatted milk. This was followed by primary antibody incubation overnight at 4°C. TBS-Tween buffer was used for 3 × 5-minute membrane washes, then stably membrane-bound primary antibodies were probed with secondary antibodies for an hour at room temperature. Chemiluminescence was used to detect antibody signals, which were scanned and detected by a gel imager (Thermo Fisher Scientific).

Exosomes were purified from the fibroblast supernatant. One milliliter of purified exosome suspension and DiO (1 μM) were mixed for 10 min to enable staining. The mixture was sedimented by two cycles of ultra-high speed centrifugation at 100000 g/min at 4°C for 40 min. Labeled exosome pellets were re-suspended with DMEM after removing the supernatant. The labeled exosomes were co-incubated with OCCC ES2 cells for 24 hours. After removing the culture medium, 3 washes with PBS were performed on the cells, followed by ten minutes of paraformaldehyde (4%) fixation and 3 further PBS washes. The nuclei were stained under room temperature conditions for five minutes with DAPI, and the excess stain was removed with 3 PBS washes. ES2 cell internalization of DiO-labelled exosomes was monitored by confocal microscopy.

We manipulated datasets and performed statistical analyses using version 22 of the Statistical Package for Social Sciences (SPSS; IBM, USA) software. Experimental results are represented as means ± standard deviation. Differences between the two groups were explored using the student’s t test, while one-way ANOVA was used for comparison among larger groups. Statistical significance of results was taken at probability (p) values of P<0.05.

After the exosomes were extracted using the exosome extraction kit, CAF exosomes were found to be half cup-shaped (Fig. 1A) by electron microscopy, and NF exosomes were also half cup-shaped (Fig. 1B). We used a Malvern nanoparticle tracking analyzer to record 10s videos of diluted exosome particle trajectories. Additionally, we used NTA software to analyze particle diameters and concentrations and thereby established that the mean CAF exosome diameter was 176.0nm (Fig. 1C). The mean NFs exosome diameter was 104.8nm (Fig. 1D). There is a slight difference in the average particle size between them, which needs to be further studied. The exosomes were purified from the supernatant of fibroblasts, and the labeled exosomes were incubated with OCCC ES2 cells for 24 hours. The results of confocal microscopy showed that DIO-labeled exosomes were internalized by ES2 cells Fig. (2).

Fig. (1). Detection of exosomes in CAFs and NFs. (A) electron microscopy of exosomes in CAFs; (B) electron microscopy of exosomes in NFs; (C) the average size of exosomes in CAFs; (D) the average size of exosomes in NFs.

CXCR4 siRNA was supplied to CAFs to decrease CXCR4 expression. And the results were confirmed by Western blotting (Fig. 3A). The exosomes of CAFs, NFs and CAFs ^{CXCR4 siRNA} were extracted and re-suspended for Western blotting. The result showed that the CXCR4 expression in exosomes derived from CAF was significantly more abundant than that in exosomes of NFs and CAFs ^{CXCR4 siRNA} (Fig. 3B).

Fig. (2). Co-localization of exosomes and ES2 cells. DAPI stained ES2 cells, DOI stained exosomes. The process of CAFs-derived or NFs-derived exosomes entering ES2 cells was detected under a fluorescence microscope.

Fig. (3). CAFs could transfer CXCR4 to ES2 via exosomes. Western blotting confirmed that CXCR4 siRNA could decrease CXCR4 expression in CAFs (A). The exosomes of CAFs, NFs and CAFs CXCR4 siRNA were extracted and re-suspended for Western blotting. The result showed that the CXCR4 expression in exosomes derived from CAF was significantly more abundant than that in exosomes of NFs and CAFs CXCR4 siRNA (B). CAFs cells were treated with exocrine secretion inhibitor GW4869(20 μM, Sigma). The supernatants of CAFs,NFs and GW4869-treated CAFs cells were used to treat OCCC cells, while OCCC cancer cell supernatant was used as a control group. The supernatant derived from CAFs could enhance the expression of CXCR4 in OCCC cells, which could be reversed by inhibiting the secretion of CAFs-derived exosomes with GW4689, while the supernatant derived from NFs could not show the corresponding trend (C). It was found that the expression of OCCC CXCR4 protein increased by adding the supernatant of CAFs, but there was no significant change in the expression of CXCR4 mRNA in OCCC (D).

CAFs cells were treated with exocrine secretion inhibitor GW4869(20 μM, Sigma). The supernatants of CAFs,NFs and GW4869-treated CAFs cells were used to treat OCCC cells, while OCCC cancer cell supernatant was used as a control group. The supernatant derived from CAFs could enhance the expression of CXCR4 in OCCC cells, which could be reversed by inhibiting the secretion of CAFs-derived exosomes with GW4689, while the supernatant derived from NFs could not show the corresponding trend (Fig. 3C). It was found that the expression of OCCC CXCR4 protein increased by adding the supernatant of CAFs, but there was no significant change in the expression of CXCR4 mRNA in OCCC (Fig. 3D).

In addition, CAF supernatant with exosome removal (CAFs ^supernatant group) was assigned as a control group. Transwell assay results showed that compared with exosomes from NF and CAF^{CXCR4 siRNA} groups, exosomes from CAFs significantly enhanced invasiveness in ES2 cells (Fig. 4A). Western blotting results showed that compared with ES2 cells treated by the NF exosome group or CAF^{CXCR4 siRNA} exosome group, the expression of EMT transcription factor β-catenin and interstitial protein N-cadherin protein increased significantly in ES2 cells treated by CAFs exosome (Fig. 4B). This suggests that CAFs may promote the invasion of OCCC through EMT, which needs to be verified by further experiments.

The exosomes from CAF cells and CAF ^{CXCR4 siRNA} cells were added to ES2 cells, respectively. Then, the proteins related to EMT, N-cadherin and E-cadherin were detected. Western blotting demonstrated that N-cadherin expression significantly increased in ES2 cells treated with CAF exosomes compared to that treated with CAF^{CXCR4 siRNA} exosomes, while E-cadherin expression illustrated the opposite trend, suggesting that exosomes containing CRCX4 contributed to EMT. Wnt/ β-Catenin pathway is one of the classical pathways related to EMT. When we treated ES2 cells using both exosomes from CAF cells and Wnt/β-Catenin inhibitor XAV939 (10 μM; Selleck, USA), western blotting showed that EMT, induced by CAF exosomes, could be inhibited by XAV939 (Fig. 5A). Finally, transwell assay confirmed that CAF exosomes containing CRCX4 could promote the invasion of ES2 cells (Fig. 5B). When EMT was reversed by XAV939, the invasive ability of OCCC cancer cells was inhibited.

Fig. (4). CXCR4-enriched exosome promotes invasion of ES2 cells. (A) A Transwell assay was used to measure the invasive capacity of ES2 cells after adding CAF exosomes or CAF CXCR4 siRNA exosomes.(B) Western blotting was used to detect the protein expression of β-catenin and N-cadherin after adding CAF exosomes or CAFCXCR4 siRNA exosomes.

Fig. (5). Effect of EMT on the invasion of ES2 cells. (A) Western blotting was used to detect the protein expression of β-Catenin and N-cadherin after the Wnt / β-Catenin pathway inhibitor XAV939 was added. (B) A Transwell assay was used to detect the change in the invasion ability of ES2 cells following the addition of the Wnt/β-Catenin pathway inhibitor XAV939. (C) Pattern diagram. CXCR4-containing exosomes derived from cancer associated fibroblasts promote epithelial mesenchymal transition in ovarian clear cell carcinoma.