BM-MSCs were isolated from fresh bone marrow by sterile puncture of a healthy donor. The present study was approved by the Anhui Medical University Ethics Committee, and informed written consent was obtained from the donor. The isolation, expansion and identification of BM-MSCs were performed as described previously (10). Briefly, following density gradient centrifugation over Ficoll-Hypaque (1.077 g/ml; Beijing Solarbio Science &Technology Co., Ltd., Beijing, China), BM-MSCs were cultured (1×10⁶ cells/ml) in Dulbecco's High Glucose Modified Eagles Medium (GE Healthcare Life Sciences, Logan, UT, USA), supplemented with 10% fetal calf serum (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and containing 5% CO₂ at 37°C. Non-adherent cells were removed after 24 h and the culture medium was changed twice a week. Adherent cells were cultured until 80–90% confluence was reached, after which trypsin (0.25%; Gibco) digestion was conducted on passage three cells. BM-MSCs were analyzed by flow cytometry (FACSCalibur™; BD Biosciences, Franklin Lakes, NJ, USA) following staining with monoclonal antibodies, as follows: Fluorescein isothiocyanate (FITC)-conjugated anti-cluster of differentiation (CD) 90 (1:10; cat. no. 555595); FITC-anti-CD14 (1:10; cat. no. 555397), phycoerythrin (PE)-anti-CD29 (cat. no. 555443), cyanine5-PE-anti-CD34 (cat. no. 555823), PE-anti-CD166 (cat. no. 559263), FITC-anti-CD44 (1:10; cat. no. 555478), PE-anti-CD31 (1:10; cat. no. 560983), FITC-anti-CD45 (1:10; cat. no. 555482), PE-anti-CD13 (1:10; cat. no. 560998) and FITC-anti-CD105 (1:10; cat. no. 561443; all: BD Biosciences). BM-MSCs were confirmed by negative staining of hematopoietic and endothelial lineage markers (CD14, CD34, CD31 and CD45) and positive staining of CD90, CD105, CD166, CD29, CD44 and CD13.

LPS (100 ng/ml) and PM (100 ng/ml) were purchased from R&D Systems, Inc. (Minneapolis, Minnesota, USA). The BM-MSC feeder layers in 12-well plates were stimulated with (experimental group) or without (control) TLR2 agonist, PM, or TLR4 agonist, LPS, for 24 h. The supernatants were removed and the feeder layers were washed twice with fresh culture medium.

Total RNA was isolated from confluent MSCs with TRIzol (Promega Corporation, Madison, WI, USA). The purity and concentration of total RNA samples were determined with NanoDrop ND-1000 (Thermo Fisher Scientific, Inc., Wilmington, DE, USA). Total RNA of each sample was used to prepare the miRNA sequencing library using the TruSeq Small RNA Library Prep kit (RS-200-0012; Illumina, Inc., San Diego, CA, USA), according to the manufacturer's protocol. This included the following steps: i) 3′-Adapter ligation with truncated T4 RNA ligase 2 (New England BioLabs, Inc., Ipswich, MA, USA); ii) 5′-adapter ligation with T4 RNA ligase (Ambion; Thermo Fisher Scientific, Inc.); iii) cDNA synthesis with reverse transcription primers (Bioligo Biological Technology, Co., Ltd., Shanghai, China); iv) polymerase chain reaction (PCR) amplification; and v) extraction and purification of PCR amplified fragments [length,~138–158 bp; corresponding to small RNAs (length,~15–35 nt)] from the agarose gel. Subsequently, the completed libraries were quantified with an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA).

The samples were diluted to a final concentration of 8 pM and cluster generation was performed on the Illumina cBot using TruSeq Rapid SR cluster kit (Illumina, Inc.), following the manufacturer's protocols. Sequencing was performed on an Illumina HiSeq 2000 using TruSeq Rapid SBS kits (Illumina, Inc.), according to the manufacturer's protocols. The DNA fragments in the libraries were denatured with 0.1 M NaOH to generate single-stranded DNA molecules, captured on flow cells (Illumina, Inc.), amplified in situ and finally sequenced for 36 cycles on Illumina HiSeq 2000.

Using high throughput sequencing with an Illumina Hiseq 2000, total clean reads from the control and trial libraries were obtained, and the length distribution of the clean reads was summarized. Image analysis and base calling were performed using Off-Line Basecaller software (v1.8.0; Illumina, San Diego, CA, USA). Subsequently, the 3′-adapter sequence was trimmed from the clean reads [that had passed the Solexa CHASTITY quality filter (Illumina)] and the reads of length <15 nt were discarded. Remaining reads (length, ≥15 nt) were aligned to the latest known human reference miRNA precursor set [Sanger miRBase 19 (http://www.mirbase.org/)] using Novoalign software (v2.07.11; http://www.novocraft.com/products/novoalign/). Reads (<2 counts) were discarded when calculating the miRNA expression. In order to characterize the isomiR variability, any sequence that matched the miRNA precursors in the mature miRNA region ±4 nt (with ≤1 mismatch) were accepted as mature miRNA isomiRs, which were grouped according to the 5-prime (5p) or 3-prime (3p) arm of the precursor hairpin.

miRDeep2 (http://www.mdc-berlin.de/en/research/research_teams/systems_biology_of_gene_regulatory_elements/projects/miRDeep) was used to predict novel miRNAs. For novel miRNA prediction, all sequence data was pooled from the following 3′-adapter trimmed files: LPS, PM and contrimmed_tags.fa, all adapter trimmed sequences of length <17 bp and mismatch >1 were excluded from the prediction pipeline. The higher the novel miRNA score of the miRDeep2, the more reliable the novel miRNA was considered to be.

Three online search algorithms, TargetScan version 6.2 (http://www.targetscan.org/vert_60/) and miRanda (http://www.Microrna.org/microrna/home.do) and MicroCosm Targets version 5 (http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5/) were used to predict the target genes of differentially expressed (DE) miRNAs among the BM-MSCs activated with the TLR2 and TLR4 agonists, and BM-MSCs in the absence of agonists. The annotated miRNA target genes that were selected from all the algorithms were considered to be the target genes.

In order to further realize the functions, the predicted target genes were subjected to the analysis of GO project (http://www.geneontology.org). Fisher's exact test was used to find whether there was increased overlap between the DE list and the GO annotation list than would be expected by chance. The P-value denotes the significance of GO term enrichment in the DE genes; the lower the P-value, the more significant the GO term (P<0.05 is recommended). Furthermore, pathway analysis was performed for these target genes. Pathway analysis is a functional analysis that maps genes to the KEGG pathways. The P-value (EASE-score, Fisher's method P-value or hypergeometric P-value) indicates the significance of the pathway correlated to the conditions. A lower P-value, indicates a more significant pathway (P<0.05 is recommended).

A random selection of DE miRNAs between the experimental and control groups from the sequencing data was validated by qPCR. Corroboration of the six novel miRNAs was performed according to the previously described conditions using qPCR assays. Total RNA was isolated from each sample using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). The miRNA quantification was performed by qPCR using Applied Biosystems StepOne Real-Time PCR system (Thermo Fisher Scientific, Inc.) and SYBR premix Ex Taq II kit (Takara Bio, Inc.) according to the manufacturer's protocols. U6 served to normalize the levels of all miRNA transcripts, the primer sequences were forward, 5′-GCT TCG GCA GCA CAT ATA CTA AAAT-3′ and reverse, 5′-CGC TTC ACG AAT TTG CGT GTCAT-3′ for U6. Relative expression of miRNA was evaluated by the 2^−ΔΔCq method (18). All miRNA samples and the internal reference U6 were run in a qPCR reaction. Thermal cycling conditions of the qPCR were 95°C for 10 min; 40 cycles of 95°C for 10 sec; and 40 cycles of 60°C for 60 sec.

To obtain a comprehensive view of the miRNA profile of BM-MSCs stimulated with LPS and PM, Illumine HiSeq 2000 technology was used to detect the global expression level annotated in Sanger miRBase 19.0. A total of 67 miRNAs demonstrated different expression patterns between the LPS and control groups (BM-MSCs without stimulation), with 32 downregulated miRNAs (Table I)and 35 upregulated miRNAs (Table II). In PM-treated cells, the downregulation (Table III) and upregulation (Table IV) were observed for 51 and 46 miRNAs, respectively. To assess the sequencing quality, the length distribution of the clean reads was summarized. Few differences were observed in the length distribution of the sequences from the control and trial libraries (Fig. 1). The most abundant sequence reads were 22, 23, 21 and 24 nt in length, which is consistent with the typical small RNA distribution of mammals.

For identification of the authenticity of the miRNAs detected by Illumine HiSeq 2000 technology, 40 known miRNAs from Tables I–IV were randomly selected for further validation with additional samples by qPCR. As presented in Fig. 2, the qPCR data were highly consistent with Illumine HiSeq 2000 technology, confirming the reliability of the HiSeq 2000 sequencing data.

To identify the novel, unannotated miRNAs in human BM-MSCs, the unlabeled reads were analyzed by miRDeep2. The present study focused on six potential novel miRNAs with the most frequent appearance among all the samples and the highest miRDeep2 score. The expression levels of these six miRNAs were further validated by qPCR. Compared with Illumine HiSeq 2000 sequencing, qPCR indicated the same trend of expression levels for the six novel miRNA candidates (Fig. 3).

To characterize the DE miRNAs between LPS, PM and the controls, the target genes of the DE miRNAs were predicted using three different miRNA target prediction algorithms (TargetScan, MicroCosm and miRanda). Subsequently, the target genes that were identified in all three databases underwent GO analysis. In the LPS group, the targets genes of downregulated miRNAs were significantly enriched in transcripts from the RNA polymerase II promoter and included regulation of macromolecule metabolic processes, gene expression and nucleobase-containing compound metabolic processes; the most enriched GO terms of targets genes of upregulated miRNAs included anatomical structure morphogenesis, regulation of cellular processes, biological processes and cell development. In the PM group, significantly enriched GO terms of targets genes of down-regulated miRNAs included regulation of biological process, system development, transcription from RNA polymerase II promoter and cell differentiation; the most enriched GO terms of targets genes of downregulated miRNAs included regulation of signal transduction, cellular process, cell communication and biological process (Fig. 4). KEGG pathway analysis indicated that the predicted target genes were significantly enriched in a wide range of pathways. LPS-regulated miRNAs predominantly target mitogen activated protein kinases (MAPKs), estrogen, T cell receptors, phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3)-Akt, the mammalian target of rapamycin signaling pathway and cancer-associated signaling pathways, while PM-regulated miRNAs target the P13-Akt, ras-related protein 1, neurotrophin, MAPK and Ras signaling pathways (Fig. 5).