The cell line used in the present study was wild-type drug-sensitive MCF-7 breast cancer cells (MCF-7/S) purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). In selected experiments, MCF-7/S expressing green fluorescent protein (GFP-S) was generated and characterized as previously reported (14). Cells were cultured in a 5% CO₂ atmosphere at 37°C in Dulbecco's modified Eagle's medium with a high glucose content (HyClone; GE Healthcare Life Sciences, Logan, UT, USA), supplemented with 10% fetal bovine serum (FBS) (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), 100 U/ml penicillin and 100 µg/ml streptomycin. For routine passage, cultures were split 1:3 when they reached between 70 and 80% confluence every 2–3 days. To minimize the influence of exosomes in FBS, FBS was ultracentrifuged at 100,000 × g at 4°C overnight to spin down any potential vesicular content (Avanti J-30I; Beckman Coulter, Inc., Brea, CA, USA) and used for all studies. The protocols followed for the extraction, purification and analysis of DRβ-H are described in a previous study (4).

The inhibitory effect of DRβ-H on breast cancer cells was determined usingan MTT assay as previously described with minor modifications (5). Briefly, 5×10³ cells were cultured in 96-well plates in triplicate. After 8 h, DRβ-H was added at 0, 5, 10, 20, 40 or 80 µg/ml. As a negative control, cells were treated with PBS. The cells were cultured for 0, 24, 48 and 72 h and then 20 µl MTT was added into each well at 5 mg/ml final concentration for 4 h. Following incubation, 150 µl dimethyl sulfoxide was added to dissolve the formazan crystals for 20 min at room temperature. The optical density value for each well was detected at a wavelength of 490 nm with a microplate reader (5082 Grodig; Tecan Group, Ltd., Mannedorf, Switzerland).

Breast cancer cells exposed to the indicated concentrations of DRβ-H (0, 10, 20 or 40 µg/ml) for 24 h were harvested. Cells were washed twice with cold PBS, incubated with Annexin-V-FITC (BDBiosciences, Franklin Lakes, NJ, USA) and propidium iodide for 15 min at room temperature in the dark, and analyzed using a flow cytometer (BD FACS Calibur; BD Biosciences, Franklin Lakes, NJ, USA).

Exosome isolation and characterization was performed using protocols as previously described (12). Exosomes were lysed for protein/RNA extraction, or diluted in PBS for incubation assay or labeled for confocal observation (12). The morphology and size of exosomes were examined by transmission electron microscopy (JEM-1010; JEOL, Ltd., Tokyo, Japan) as previously reported (15).

Exosomes were stained with a PKH26 Red Fluorescent Cell Linker kit (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) according to a previously published protocol (12). As the negative control, exosomes without PKH26 labeling were also prepared. For observation of exosome uptake, confocal laser scanning microscopy (LSM710; Zeiss GmbH, Jena, Germany) was used (150,000 magnification).

Exosomal RNA was isolated using the Total Exosome RNA and Protein Isolation kit (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) in accordance with the manufacturer's protocol. Cellular RNA was prepared using TRIzol^®reagent (Thermo Fisher Scientific, Inc.). miRNA profiles were analyzed using an Affymetrix Gene Chip miRNA Array 3.0 (Capital Bio Corporation, Beijing, China) according to the manufacturer's protocol. The levels of miRNAs between exosomes and MCF-7/S were calculated as previously reported (16).

miRNA inhibitors and the negative controls were synthesized by GenePharma Co., Ltd. (Shanghai, China). Transfection of miRNA inhibitors was performed using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Cells and exosomes were harvested 48 h after transfection for subsequent analysis viapolymerase chain reaction (PCR).

Proteins were extracted from exosomes using the Total Exosome RNA and Protein Isolation kit (Invitrogen; Thermo Fisher Scientific, Inc.) in accordance with the manufacturer's protocols. Equal amounts of proteins were subjected to SDS-PAGE and transferred to polyvinylidenedi fluoride membranes. The CD63 protein level was measured by western blotting analysis using anantibody against CD63 (Santa Cruz, USA). β-actin (Sigma, Germany) was used as an internal control for normalization. Then, bound proteins were visualized using the ECL Plus kit (EMD Millipore, Billerica, MA, USA) with Image Lab Software version 5.2.1 (Bio-Rad, USA).

The RT-qPCR was run in triplicate using the SYBR green (Biouniquer Technology, Nanjing, China) technique. Briefly, cDNA for miRNA was synthesized using the BU-Script RT kit (Biouniquer Technology, Nanjing, China). Specific stem-loop primers (Springen Biotechnology, Nanjing, China) were designed for the selected miRNAs, and U6 was used as an internal control. Expressions of miRNAs and U6 were evaluated using following primers: miR-130a forward, 5′-TTCACATTGTGCTACTGTCTGC-3′; miR-183 forward, 5′-TATGGCACTGGTAGAATTCACT-3′; miR-20b forward, 5′-TGTCAACGATACGCTACGA-3′; miR-25 forward, 5′-TCTGGTCTCCCTCACAGGAC-3′; and miR-452 forward, 5′-GCGAACTGTTTGCAGAGG-3′. Gene amplifications were performed on a Light Cycler 480 (Roche, Australia) as follows: 91°C for 5 min followed by 45 cycles (91°Cf or 15 sec, and 60°C for 30 sec), followed by melting curve detection. The relative miRNA expression was calculated using the 2^−ΔΔCq method.

The TargetScan (http://www.targetscan.org/) and miRDB (http://www.mirdb.org/miRDB/) databases were employed to predict miRNA targets (17,18). Only the genes predicted by these two independent tools were considered. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were analyzed by the online DAVID program (http://david.abcc.ncifcrf.gov/) (19,20).

Statistical analysis was performed using SPSS 20.0 statistical software (IBM Corp., Armonk, NY, USA). All experiments were performed in triplicate and the representative data are presented as the mean ± standard deviation. Differences were determined using the Student's t-test or using a one-way ANOVA with Student-Newman-Keuls post-hoc test. P<0.05 was considered to indicate a statistically significant difference.

The chemical structure of DRβ-H is presented in Fig. 1A. MCF-7/S cells were treated with different concentrations of DRβ-H (0, 5, 10, 20, 40 or 80 µg/ml) or for different times (0, 24, 48 or 72 h) and the rate of inhibition of growth was measured using an MTT assay. DRβ-H exerted a significant inhibitory effect on MCF-7/S cells in a dose- and time-dependent manner compared with the negative control (Fig. 1B and C).

To explore whether the anti-tumor effect of DRβ-H was also associated with apoptosis, MCF-7/S cells were treated with different concentrations of DRβ-H (0, 10, 20 or 40 µg/ml) for 24 h and the rate of apoptosis was determined by flow cytometry. Incubation of MCF-7/S cells with DRβ-H increased the apoptotic rate compared with control cells treated without DRβ-H (Fig. 1D). Furthermore, it was identified that increased DRβ-H concentrations increased the apoptotic effect, demonstrating a dose-dependent association.

The protocol of exosome isolation was based on differential sedimentation properties, and used a series of centrifugation and ultracentrifugation steps as previously published (12). Exosomes were homogeneous in morphology, and exhibited round vesicles measuring between 50 and 100 nm in diameter as determined by transmission electron microscopy (Fig. 2A). The exosome quantity was evaluated by measuring protein content. The authors' previous study demonstrated that exosomes derived from cancer cells were responsible for drug efficacy during toxic insult (13). Thus, it was hypothesized exosome release is involved in DRβ-H-mediated growth inhibition. MCF-7/S cells were treated with different concentrations of DRβ-H and it was revealed that the quantity of exosomes was decreased in a dose-dependent manner (Fig. 2B). Additionally, treatment of cells with high concentration of DRβ-H reduced exosome marker CD63 compared with the control cells treated with lower concentration of DRβ-H (Fig. 2C), indicating that DRβ-H may inhibit exosome secretion.

To visualize exosome uptake by recipient cells, exosomes were labeled by PKH26 red dye and incubated with GFP-S for 24 h, after which confocal laser scanning microscopy was performed. The internalization of exosomes was indicated by several red fluorescent punctuated signals on the cell membranes and inside the cytoplasm of GFP-S (Fig. 3A; white arrows indicate red fluorescent punctuated signals). The effects of exosomes on cell growth were assessed in DRβ-H-treated MCF-7/S cells. As presented in Fig. 3, the proliferation rate modulated by DRβ-H was relatively lower in MCF-7/S cells without exosomes, whereas a significant increase in proliferation rate was identified in MCF-7/S cells treated with exosomes (Fig. 3B). Furthermore, incubation of MCF-7/S cells with exosomes significantly decreased apoptosis when compared with control cells incubated without exosomes (Fig. 3C). These data, along with the observations in the aforementioned section, collectively suggested that exosomes were able to promote recipient MCF-7/S cells growth and DRβ-H suppressed MCF-7/S cells growth by inhibiting exosome release.

Given that exosomes contain cellular information, an attempt was made to explore whether exosomal miRNAs affect MCF-7/S proliferation. By comparing the miRNA profiles of exosomes and their cells of origin, it was identified that there are numerous miRNAs shared by exosomes and MCF-7/S. Furthermore, a number of miRNAs were detected exclusively in exosomes compared with those in the parental cells, whereas a group of miRNAs expressed by cells were absent in the corresponding exosomes (Fig. 4A). Next, the functions of the top five miRNAs (miR-130a, miR-183, miR-20b, miR-25 and miR-425) more localized in exosomes compared with that incells were examined by transfecting with miRNA inhibitors. The successful knockdown of the five exosomal miRNAs was subsequently validated by RT-qPCR (Fig. 4B). miRNA knockdown exosomes were added into recipient MCF-7/S cells and cell viability was detected by analyzing the proliferation rate. As presented in Fig. 4, of the five miRNAs tested, only miR-130a and miR-425 knockdown exosomes were able to significantly decrease cell proliferation (Fig. 4C), indicating that exosomal miR-130a and miR-425 could enhance MCF-7/s cell viability.

To evaluate the biological processes regulated by miR-130a and miR-425, their targets were investigated using TargetScan and miRDB. Only the genes listed by these two independent algorithms were chosen for further study. A total of 443 genes were detected and a KEGG analysis was performed using the DAVID program. As presented in Table I, the predicted genes were suggested to participate in several signaling pathways, including mammalian target of rapamycin (mTOR), ErbB, mitogen activated protein kinase (MAPK) and transforming growth factor (TGF)-β signaling pathways. Furthermore, a number of pathways associated with the tumor metabolic process were also detected, in particular choline metabolism and proteoglycansin cancer.