C3H10T1/2 cells were obtained from the Chinese Academy of Science Cell Bank (Shanghai, China) and were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS; both Hyclone Laboratories, Inc., Logan, UT, USA) at 37°C in a humidified atmosphere of 5% CO₂ in air. To induce osteoblast differentiation, the medium was replaced with low-serum medium, consisting of DMEM supplemented with 5% FBS and 200 ng/ml rhBMP-2 (R&D Systems, Minneapolis, MN, USA) and the medium was changed every 2–3 days.

Levels of osteoblast differentiation in C3H10T1/2 cells were determined using ALP staining. For ALP staining, cells were washed with PBS twice, fixed with 70% ethanol for 20 min, rinsed three times with deionized water and then incubated with the BCIP/NBT liquid substrate system (Sigma-Aldrich, St. Louis, MO, USA), an ALP substrate solution, for 30 min. Images of the stained cells were then captured. For quantitative analysis, the ALP stain was extracted with 10% cetylpyridinium chloride for 15 min and quantified by measuring its absorbance at 540 nm. Relative ALP staining was then calculated as a fold change of the control.

Following incubation with 200 ng/ml BMP-2 for 1 or 4 days, total RNA was extracted using TRIzol (Invitrogen Life Technologies, Carlsbad, CA, USA). RNA was also extracted from BMP-2 untreated cells. Total RNA from each sample was quantified using the NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA)and RNA integrity was assessed using standard denaturing agarose gel electrophoresis.

For microarray analysis, the Agilent Array platform (Agilent Technologies, Santa Clara, CA, USA) was employed. Each sample was amplified and transcribed into fluorescent cRNA along the entire length of the transcripts without 3′ bias utilizing a random priming method (Quick Amp Labeling kit, One-Color, Agilent Technologies, p/n 5190-0442). The labeled cRNAs were hybridized onto the Mouse lncRNA Array v2.0 (8×60K; Arraystar Inc., Rockville, MD, USA) which was designed for the global profiling of 31,423 mouse lncRNAs and 25,376 protein-coding transcripts. The slides were washed and the arrays were scanned using the Agilent G2505C microarray scanner. Agilent Feature Extraction software (version 11.0.1.1) was used to analyze the acquired array images. Quantile normalization and subsequent data processing were performed using the GeneSpring GX version 11.5.1 software package (Agilent Technologies). Differentially expressed lncRNAs and mRNAs were identified through fold change filtering.

According to lncRNA genomic locations relative to protein coding genes, lncRNAs are categorized as: i) sense, ii) antisense, iii) bidirectional, iv) intronic and v) intergenic (17). The differentially expressed lncRNAs identified in this study were categorized based on these groupings.

One important function of lncRNAs is to regulate the expression of nearby protein-coding genes. Therefore, protein-coding genes were searched for differentially expressed lncRNAs using the UCSC Genome Browser (http://genome.ucsc.edu/cgi-bin/hgGateway). Genes transcribed within 300 kb were considered to represent nearby coding genes and predicted lncRNAs nearby these coding genes were integrated with differentially expressed mRNA in the microarray (fold change ≥2.0 at day 4) (18). The regulatory network between lncRNAs and nearby coding genes was visually presented using the Cytoscape program (http://www.cytoscape.org/).

Experiments were repeated three times with the exception of microarray experiments. Two group comparisons were performed using the Student’s t test. P<0.05 was considered to indicate a statistically significant difference.

Osteoblast differentiation from MSCs is regulated by various cytokines and growth factors. BMP-2, a transforming growth factor β superfamily member, is well known as one of the most powerful osteoblast promoting factors (19). ALP activity is an early osteoblast differentiation marker. In the present study, ALP activity in C3H10T1/2 cells with and without BMP-2 treatment was analyzed for 7 days to investigate BMP-2-induced C3H10T1/2 cell osteoblast differentiation. A significant increase in ALP acitivity in BMP-2-induced C3H10T1/2 cells was observed (Fig. 1), indicating that BMP-2 stimulates C3H10T1/2 cells into early osteoblast differentiation.

To investigate the expression profiles of lncRNAs during MSC osteoblast differentiation, total RNA was extracted from BMP-2 treated and untreated cells at day 1 and 4. OD_260/280 ratios were close to 2.0 and OD_260/230 ratios were >2.0 for all the samples. The quality of total RNA was checked by gel electrophoresis (Fig. 2A), confirming that the RNA was of good quality. Scatter plot analysis of the microarray data is shown in Fig. 2B. Microarray expression analysis of lncRNAs was then performed. Firstly, lncRNAs upregulated or downregulated by >1.5-fold in BMP-2 treated or untreated C3H10T1/2 cells for 1 day were screened, revealing 886 upregulated and 825 downregulated lncRNAs in the BMP-2 treated group. Secondly, 595 upregulated and 548 downregulated lncRNAs by >2-fold were identified at day 4. Continuously upregulated or downregulated lncRNAs following prolonged BMP-2 treatment may be more important for the control of osteoblast differentiation. We also identified 59 upregulated (Table I) and 57 downregulated lncRNAs (Table II) following 1 and 4 days of BMP-2 treatment in accordance with the described thresholds.

Based on a previous study by Ponting et al(17), of the 116 differentially expressed lncRNAs identified in this study, 60.3% were categorized as intergenic. In addition, 20.7% were catagorized as sense. Details of the categories are presented in Table III.

Unlike proteins or microRNAs, lncRNA function cannot be inferred from sequence or structure. Studies have hypothesized a number of regulatory paradigms to explain the mechanism by which lncRNAs function. In the present study, the genomic context of lncRNAs was highlighted. Differentially expressed nearby mRNAs were combined and 24 differentially expressed lncRNAs and nearby coding genes pairs were identified for 13 differentially expressed lncRNAs and 20 differentially expressed mRNAs. Using the Cytoscape program, a regulatory network was constructed between differentially expressed lncRNA and nearby coding genes (Fig. 3).