The K562 cells (kindly provided by Dr.Jingqiong Hu, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China) were grown in RPMI 1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% MV-free (by ultrafiltration) fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 24 mM sodium bicarbonate, 2 mM L-glutamine, 100 IU/ml penicillin and 100 IU/ml streptomycin in a humidified 5% CO₂ atmosphere at 37°C. Cell viability was evaluated by trypan blue exclusion assay, and all cultures utilized were 95% viable. A method was established using ultrafiltration to remove MVs from fetal bovein serum. Initially, debris in fetal bovein serum was removed by 0.22 µm filter (EMD Millipore, Billerica, MA, USA). Subsequently, the supernatant was purified by stirring ultrafiltration with a 100,000 molecular weight cut off ultrafiltration membrane (EMD Millipore) to remove MVs.

Peripheral blood was collected in ethylenediaminetetraacetic acid-coated tubes from healthy donors at Union Hospital (Tongji Medical College, Huazhong University of Science and Technology) between August and October 2012. The criteria for volunteers selection consisted of no recent illnesses or treatments for a chronic medical condition. No medical history was obtained from the donors. The collection of blood occurred between the morning and early afternoon. The average age for the female and male donors was 35.79 and 31.81 years, respectively. Written informed consent was obtained from all volunteers and this study was approved by the ethics committee of Tongji Medical College, Huazhong University of Science and Technology.

Supernatants were collected from cell culture after 24 h. Cells were centrifuged at 300 × g for 30 min, and the supernatant was used for MVs preparation. Cell debris was removed by centrifugation at 2,500 × g for 30 min. MVs were purified by ultracentrifugation at 16,000 × g for 120 min, and subsequently washed with sterile filtered phosphate-buffered saline (PBS). The samples were resuspended in radioimmunoprecipitation assay (RIPA) buffer supplemented with a protease inhibitor cocktail (Agilent Technologies, Inc., Santa Clara, CA, USA) and stored at −80°C. All centrifugations were accomplished at 4°C. MVs quantity was determined with the Pierce BCA Protein Assay kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA).

Peripheral blood was centrifuged at 400 × g for 30 min, and the plasma was used for MVs preparation. Cell debris and platelets were removed by centrifugation at 2,500 g for 20 min twice. MVs were purified by ultracentrifugation at 16,000 × g for 120 min, and subsequently washed with sterile filtered PBS (32). All centrifugations were accomplished at 4°C. MV pellets were resuspended in RIPA buffer supplemented with a protease inhibitor cocktail (Agilent Technologies, Inc.) and stored at −80°C. MVs quantity was determined with the Pierce BCA Protein Assay kit (Thermo Fisher Scientific, Inc.).

For transmission electron microscopy, the pelleted MVs were fixed in 2.5% (w/v) glutaraldehyde in PBS, dehydrated and embedded in Epon (SPI Supplies, Inc., West Chester, PA, USA). Ultrathin sections (65-nm) were cut and stained with uranyl acetate (SPI Supplies, Inc.) and Reynold's lead citrate (SPI Supplies, Inc.). The sections were examined in a Tecnai G² 12 transmission electron microscope (FEI, Hillsboro, OR, USA).

Total RNA was isolated from K562 cells and MVs using the mirVana microRNA Isolation kit according to the manufacturer's protocol (Ambion; Thermo Fisher Scientific, Inc.). The quality, yield and size of the microRNA fractions were analyzed using a 2100 Bioanalyzer (Agilent Technologies, Inc.). For RNA isolated from mononuclear cells, only an RNA integrity number (RIN) ≥7 was used, along with its matched plasma sample for profiling. Since the intact 18 and 28s ribosomal RNA were variable in the MVs, the RIN was not a constraint for these samples, although a RIN between 6.5 and 6.8 was observed.

RNA from K562 cells, and MVs both from K562 cells and normal human volunteers' peripheral blood cells were used for microRNA microarray. Human microRNA microarrays from Agilent Technologies, Inc., which contain probes for 888 human microRNAs from the miRBase v14.0 (mirbase.org/pub/mirbase/14.0/), were used in the study. In total, 100 ng of total RNA extracted from serum was used as inputs for sample labeling and hybridization preparation, following the manufacturer's protocol.

The microarray image information was converted into spot intensity values using Scan Control Software version 7.0 (Agilent Technologies, Inc.). The signal upon background subtraction was exported directly into the GeneSpring GX 11.0 software (Agilent Technologies, Inc.) for quartile normalization and further analysis.

For testing of candidate microRNAs acquired on microarrays, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was performed using miScript SYBR Green PCR kit (Qiagen GmbH, Hilden, Germany). The assays were performed on three samples for six candidates (miR-494, miR-483-5p, miR-26a, miR-223, miR-21 and miR-22).

Data were analyzed using the statistical software package SAS v3.0. (Shanghai Biotechnologies Corporation, Shanghai, China). The statistical significance was calculated by the Student's t test. P≤0.05 was considered to indicate a statistically significant difference. Three-way Venn diagrams were used to indicate the number of significant microRNAs [P<0.05, fold-change (FC)>2 and FC<0.5] identified by the Student's t test (K562-MVs vs. control MVs).

The human TargetScan Release 5.1 (http://www.targetscan.org) was used for prediction of microRNA targets. In addition, potential target gene-associated pathways were analyzed using TargetScan Release 5.1 based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database (http://www.genome.jp/kegg/). The enrichment P-value of the target genes involved in every pathway was calculated, and the regulatory interactions between genes and microRNAs were integrated.

MVs derived from K562 cells and normal human volunteers' peripheral blood cells were collected and observed under a transmission electron microscope, which revealed vesicular structures characteristic of MVs (Fig. 1A and B).

A microarray containing probes for 888 human microRNAs was initially used to screen the significant differential expression levels of microRNAs between K562-MVs and control groups (Table I). The filtered and normalized data were subjected to hierarchical cluster analysis comparing the microRNA expression profile of K562-MVs and control group samples. Fig. 2 illustrates the hierarchical clustering of the differentially expressed microRNAs in the pairwise comparison of the two samples. There were seven microRNAs, including miR-494, miR-1275, miR-484-5p, miR-1308-v15.0, miR-575, miR-1268 and miR-125a-3p, with significantly higher expression levels in the K562-MV group than in the healthy group (FC=7.18–96.49; P<0.05; Table I). By contrast, miR-151-3p, miR-1974-v14.0, miR-26a, miR-24, miR-22, miR-93, miR-223, miR-23b, miR-103, miR-361-5p, miR-21, miR-126*, miR-107, miR-27b, miR-27a and miR-185 displayed a significantly lower expression level in the K562-MV group than in the healthy group (FC<0.5; P<0.05; Table I).

The presence and levels of specific microRNAs from both K562-MVs and their parental cells were determined using microarray analysis probing for 888 human microRNAs. The microRNA profiles of K562 cells confirmed the alterations previously reported (33). Furthermore, the results demonstrated that 30 microRNAs were above the normalized threshold in the 888 microRNAs, which was calculated based on the 95 percentile of the negative control probe signal in both normal cells and MVs (Fig. 3 and Table II). Of the 303 positive microRNAs, 104 were not significantly different between MVs and their parental K562 cells. By comparison, 77 microRNAs were present at elevated levels within MVs, while 122 were present at a higher proportion in the parental cells. This observation may suggest that the compartmentalization of microRNAs from cells into MVs is an active (selective) process, at least for certain microRNAs.

Two upregulated (miR-494 and miR-484-5p) and four downregulated (miR-26a, miR-22, miR-223 and miR-21) microRNAs of K562-MVs (in terms of their expression levels compared with those in the control) were selected for microarray data validation via RT-qPCR, and the results correlated well with the findings of microarray analysis.

To further study the functions of the aberrantly expressed microRNAs in MVs and in their parental cells, a predicted target analysis for these microRNAs was performed. The data revealed that 23 microRNAs exhibited altered expression in K562-derived MVs compared with control MVs. These microRNAs targeted 1,354 regulatory genes, which affect cellular apoptosis, proliferation and molecular signaling pathways. Notably, 43 oncogenes and tumor suppressor genes were identified among these aberrant microRNAs detected in MVs derived from K562 cells (Table III). In these dysregulated microRNAs, 4 microRNAs targeted 2 oncogenes of the B-cell lymphoma family, while 12 microRNAs targeted 16 oncogenes of the rat sarcoma viral oncogene homolog (RAS) family. In addition, 24 microRNAs targeted 22 housekeeping genes in the p53 signaling pathway (Table IV). p53 is a tumor suppressor protein that regulates the expression of a wide variety of genes (34). It has been reported that the p53 signaling pathway has a key role in the induction of apoptosis of CML cells (35). With the exception of breast cancer 1, which was targeted by miR-125a-3p, and insulin-like growth factor 1, which was targeted by miR-1275 and miR-1207-5p, 16 of the above housekeeping genes were targeted by the downregulated microRNAs.

A predicted target analysis was also performed for the 104 microRNAs co-expressed in MVs and in their parental K562 cells. In total, 97 microRNAs targeted 3,743 regulatory genes. Data regarding miR-17*, miR-191*, miR-1914*, miR-1974_v14.0, miR-1977_v14.0, miR-1979_v15.0 and miR-33b* were not available in TargetScan 5.1.

Gene ontology (GO) analysis revealed that the high-enrichment GOs targeted by the microRNAs co-expressed in MVs and the corresponding K562 cells were involved in various processes, including cell communication, signal transduction, RNA metabolism, transcription and cell differentiation. The pathway analysis revealed that there were 13 different pathways corresponding to the target genes, 12 of which were enriched (P<0.01) (Table V).

To elucidate the mechanisms by which microRNAs co-expressed in MVs and parental K562 cells regulated leukemogenesis, a bioinformatic approach was undertaken to identify signaling pathways and target genes regulated by these microRNAs. Using SAS v3.0, which is designed to integrate microRNA target genes into signaling pathways, mRNAs involved in the MAPK signaling pathway, focal adhesion pathway, insulin signaling pathway and Wnt signaling pathway were observed to be significantly enriched in these microRNA target genes. These target genes were analyzed, and it was noticed that 34 microRNAs targeted 44 genes of the MAKP pathway, suggesting that the MAPK signaling pathway may play a significant role in MVs and their parental K562 cells.

The microRNA-gene interaction networks are represented in Fig. 4. As indicated in Fig. 4, the miR-30 family (with the exception of miR-30a and miR-30d) contained the most targeted mRNAs, with degrees ranging from 6 to 9, while miR-17, miR-16 and miR-15a/b occupied an important position in the MAPK pathway. These microRNAs may be important in the pathogenesis of CML. It was speculated that miR-17, miR-16 and miR-15a/b may be the key regulators of the MAPK signaling pathway in CML cells.

A rearrangement (translocation) of genetic material between chromosomes 9 and 22 is associated with several types of blood cancer, particularly CML (36,37). As a typical CML cell line, K562 cells presented t(9:22) (38); thus, the chromosome location of the differentially expressed microRNAs was further examined. Several differentially regulated microRNAs in K562-derived MVs, including miR-27b, miR-24, miR-23b and miR-126*, were located in chromosome 9. A total of 12 co-expressed microRNAs in MVs and their corresponding parental K562 cells [Homo sapiens (hsa)-let-7a, hsa-let-7f, miR-126, miR-126*, miR-23b, miR-24, miR-27b and miR-7] were mapped to chromosome 9, whereas hsa-let-7b, miR-1249, miR-130b and miR-185 were mapped to chromosome 22. The genomic locations of a number of these microRNAs were distant from the ABL1 and BCR gene locus (Fig. 5). Using target prediction programmes, 875 predicted targets were observed to be regulated by these 12 microRNAs, which may participate in leukemia via the above 875 genes. Using TargetScan 5.1 based on the KEGG pathway database, signaling pathways were analyzed for predicted targets involved in the MAPK, transforming growth factor (TGF)β, p53, ErbB, Wnt, vascular endothelial growth factor, Notch and mechanistic target of rapamycin (mTOR) signaling pathways, focal adhesion and cell cycle. Table VI summarizes the predicted targets associated with CML.

Notably, it was observed that 70 of the 104 microRNAs co-expressed in MVs and in the corresponding parental K562 cells targeted 184 genes that were mapped to chromosome 9, while 50 of the 104 microRNAs targeted 66 genes that were mapped to chromosome 22. These genes were correlated with the cell cycle, p53 signaling pathway, MAPK signaling pathway and RAS oncogene family.