MSCs and basal culture medium were purchased from Cyagen Biosciences, Inc., (Santa Clara, CA, USA). Recombinant human HMGB1 protein and fetal bovine serum were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ras inhibitor (Selleckchem, Houston, TX, USA), which was a transferase inhibitor for H-Ras and K-Ras, was used at a concentration of 5 µM.

Adherent MSCs were trypsinized and passaged once cell confluence reached ~80%. Cells at passage 3–5 were used in the present experiments.

To induce osteogenic differentiation, MSCs were cultured in basal culture medium supplemented with fetal bovine serum (FBS).

To observe MSC differentiation following exposure to HMGB1, MSCs were cultured in basal culture medium or 25 ng/ml HMGB-1 for 5 days. Total RNA was extracted with TRIzol reagent (Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's protocol. RT-qPCR was performed to observe the expression of osteoblastic markers using a StepOne Plus Real-Time PCR system with SYBR Green (Roche Diagnostics, Basel, Switzerland) as a double-strand DNA-specific binding dye. Primer sequences were as follows: Osteocalcin (OCN), forward 5′-AAGCAGGAGGGCAATAAGGT and reverse, CAAGCAGGGTTAAGCTCACA; and GAPDH, forward 5′-CGTCCCGTAGACAAAATGGT and reverse 5′-GGCTGGTGGTCCAGGGGTCT (Sangon Biotech Co., Ltd., Shanghai, China). According to the manufacturer's protocol, DNA hydrolase was used to remove genomic DNA. The reaction mixture included buffer (5 µl), dNTP (4 µl), primer (4 µl), Taq (1 µl), sample (1 µl), SYBR Green (1 µl) and ddH₂O to make a total volume of 50 µl. Thermal cycling conditions were as follows: 95°C for 30 sec and 40 cycles of 95°C for 5 sec and 60°C for 35 sec. Relative target gene expression levels were calculated according to the 2-^ΔΔCq method (20). All PCR reactions were performed in triplicate. mRNA quantification of the target genes and the GAPDH housekeeping gene was performed in separate tubes.

MSCs (10⁴ cells/cm2) were seeded into 24-well plates and cultured in basal culture medium. After one week, ALP activity was assessed using a BCIP/NBT ALP color development kit (Beyotime Institute of Biotechnology, Haimen, China), according to the manufacturer's protocol. For the quantitative determination of ALP activity, the MSCs were incubated with p-nitrophenol phosphate (Beyotime Institute of Biotechnology) as a substrate, washed with PBS buffer and lysed with 0.1% Triton X-100 in 10 mM Tris HCL, (pH 9.0). Absorbance was determined at 405 nm using a microplate reader and compared with p-nitrophenol standard titration curve. Data were presented as the mean ± standard deviation (n=3).

Cells were harvested and lysed in RIPA buffer (Cyagen Biosciences, Inc.) supplemented with protease inhibitors. Following concentration determination (5.534 µg/µl), protein samples (60 µg) were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane by western blotting. Membranes were blocked in 5% skim milk for 1 h and incubated with antibodies against GTP-Ras (cat. no. 16117; 1:500; Active Ras Detection kit; Thermo Fisher Scientific, Inc.), Ras (cat. no. ab52939; Abcam, Cambridge, UK), extracellular-signal-regulated kinases (ERK; cat. no. ab196883; 1:500; Abcam), phosphorylated (p-)ERK (cat. no. ab65142; 1:500; Abcam), p38 (cat. no. 9212; 1:1,000; Abcam), p-p38 (cat. no. 9215; 1:1,000; Abcam) and β-actin (cat. no. dc-130301; 1:2,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), respectively, at 4°C overnight. Following incubation with horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (cat. no. 31210; 1:5,000; Thermo Fisher Scientific, Inc.), immunoreactive proteins were visualized using enhanced chemiluminescence reagent (Thermo Fisher Scientific, Inc.). Relative quantification of the bands was performed using ImageJ software (version 5.0; National Institutes of Health, Bethesda, MA, USA).

Soluble proteins in the medium of the stromal cell lines were measured using the Human Cytokine Array G1000 (AAH-CYT-G1000; RayBiotech Inc., Norcross, GA, USA), according to the manufacturer's protocol. These arrays can detect 120 proteins, respectively. Stromal cells were plated n Dulbecco's modified Eagle medium (DMEM) supplemented with 10% FBS 3 days prior to the experiment and were 75–90% confluent when the media were collected and filtered. DMEM containing 10% FBS was also hybridized to the arrays and used for normalization. Ten technical and biological replicates were performed, and both exhibited a high correlation (correlation coefficient, >0.9) (data not shown). Hybridization was performed overnight at 4°C. All slides were scanned using a Gene Pix 4000B Microarray Scanner (Molecular Devices, LLC, Sunnyvale, CA, USA) and analyzed using Gene Pix Pro 6.0 software. The median signal score was averaged across the triplicates on each array. Results were subsequently normalized using internal controls by subtracting the values for the control DMEM supplemented with 10% FBS.

Statistical significance was determined using the two-tailed Student's t-test, assuming equal variances. The χ² test was used to compare rates. SPSS version 17.0 (SPSS, Inc., Chicago, IL, USA) was used to analyze the data. P<0.05 was considered to indicate a statistically significant difference.

To determine the influence of HMGB1 stimulation on human MSCs at the protein level, antibody arrays were performed. The same three groups were stimulated with 25 ng/ml HMGB1 and compared with non-stimulated cultures (Fig. 1A). In line with the search criteria for differential secretion, which were >1.5-fold induction or repression and P<0.05, bioinformatic data analysis detected 10 cytokines which were differentially secreted between the groups. Increased cytokine secretion was detected for insulin-like growth factor binding protein 4, macrophage colony stimulating factor (M-CSF), eotaxin-3, epidermal growth factor receptor (EGFR), vascular endothelial growth factor (VEGF), placental growth factor, angiopoetin-2, chemokine C-C motif ligand 5 (CCL-5), urokinase plasminogen activator receptor and macrophage migration inhibitory factor following induction with HMGB1 (Figs. 1B and C). These results suggest that HMGB1 promotes the secretion of multiple cytokines from MSCs.

Hierarchical clustering of these 10 cytokines was performed with normalized cytokines secretion values for the two distinct groups: the HMGB1-treated group and the non-stimulated control group (Fig. 2). Six differentially secreted cytokines were visualized as induced, as indicated by red, and four cytokines were repressed, as indicated by green.

Notably, among the six induced cytokines, three cytokines, EGFR, M-CSF and VEGF, were associated with Ras signaling, and three cytokines were involved in signal transduction: EGF-R, CCL-5 and VEGF. These 10 differentially secreted cytokines have also been associated with cell development (5), the regulation of growth (5) and cell migration (8), response to stress (12) and immune system processes (9) according to the Gene Ontology database (http://amigo.geneontology.org/amigo) for annotation, visualization and integrated discovery. Various differentially secreted cytokines were associated with more than one biological process.

Some of the differentially secreted cytokines detected in the present study also have a role in other molecular and cellular functions, such as cellular growth (such as M-CSF) and proliferation (such as macrophage migration inhibitory factor), cell morphology (such as RANTES) and cellular development (such as placental growth factor). Based on the 10 differentially secreted cytokines, the cytokine-cytokine receptor interaction pathway (0460hsa in KEGG) was demonstrated to be the dominant pathway that was influenced by HMGB1. Here, the secretion of CCL5, CCL26, VEGFA and CSF1 was induced (Fig. 3). The results demonstrate that the differentially secreted cytokines serve an important role in the biological function of MSCs.

Antibody array assay analysis demonstrated that the expression of EGFR was markedly increased in the MSCs following stimulation with HMGB1. It has previously been reported that the interaction between EGF and EGFR activates Ras/MAPK signaling pathway (21). To investigate whether Ras/MAPK signaling pathway was activated in MSCs, the protein levels of GTP-Ras, p-p38 and p-ERK were determined in the control and HMGB1-treated MSCs by western blotting. HMGB1 treatment induced a marked increase in the protein expression levels of GTP-Ras and this effect was sensitive to incubation with the Ras inhibitor. Notably, HMGB1 treatment increased the expression levels of p-p38 and p-ERK and this effect was sensitive to the incubation with the Ras inhibitor. These results suggests that HMGB1 activates the Ras signaling pathway and subsequently stimulates the p38 and ERK pathways of MSCs (Fig. 4A).

ALP staining and quantification of the expression of the late OCN was performed in MSCs. ALP activity and OCN gene expression levels were significantly higher in the HMGB1-stimulated MSCs, as compared with the control cells (P<0.01). Furthermore, as compared with the results in the HMGB1-induced group, the upregulation of ALP activity and OCN expression in MSCs was significantly reduced following Ras inhibitor treatment (P<0.01; Figs. 4B-D). These results suggest that HMGB1 potentiates the osteogenic differentiation of MSCs through the Ras/MAPK signaling pathway.