NIH-3T3 fibroblast cell lines established from the primary Swiss albino mouse embryonic fibroblast cells were grown in Dulbecco’s modified Eagle’s medium (DMEM; Gibco BRL, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (FBS; Gibco BRL) and 1% penicillin-streptomycin (Gibco BRL). When the cell confluency reached about 50%, cells were treated with various concentrations of 5-aza-CN (Sigma-Aldrich Co., St. Louis, MO, USA) for 24 h.

Global DNA-methylation analysis was conducted using a MethylFlash Methylated DNA Quantification kit (Epigentek, Farming-dale, NY, USA), according to the manufacturer’s instructions. Briefly, Genomic DNA was isolated from cells treated with 5-aza-CN for 24 h by a genomic DNA isolation kit (Qiagen, Valencia, CA, USA); 100–200 ng of genomic DNA was used to quantify methylated DNA colorimetrically. The relative percentage of methylated DNA was quantified using a formula provided by the manufacturer.

For cellular differentiation, cells were treated with or without 5-aza-CN for 24 h, then washed with growth media and replaced in a proper differentiation medium. For osteoblastic induction, we replaced them into DMEM supplemented 2% FBS containing 170 μM ascorbic acid 2-phosphate (AA; Sigma-Aldrich), 5 mM β-Glycerophosphate (β-GP; Sigma-Aldrich) and 200 ng/ml BMP2 (R&D Systems, Minneapolis, MN, USA). For adipogenic differentiation, the media were changed to DMEM supplemented 2% FBS with 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) (Sigma-Aldrich), 1 μM dexamethasone (DEX) (Sigma-Aldrich), 100 μM Indomethacin (Sigma-Aldrich), and 10 μg/ml Insulin (Sigma-Aldrich). For chondrogenic differentiation, the media were changed to chondrogenic differentiation media (Gibco BRL). For neurogenic differentiation, the cells were induced by culturing in DMEM supplemented 2% FBS, 2% dimethyl sulfoxide (Sigma-Aldrich), 100 μM butylated hydroxyanisole (Sigma-Aldrich). All cell differentiation media were changed every other day.

The total RNA was extracted using Trizol Reagent (Gibco BRL). RT was conducted with a RT system containing Moloney Murine Leukemia Virus reverse transcriptase (Promega, Madison, WI, USA). Conventional PCR was performed on a Palm Cycler thermocycler (Corbett Life Science, Sydney, Australia) and products were resolved on a 1.5% agarose gel. Real time RT-PCR was run on a Rotor-Gene RG-3000 (Corbett Life Science) using the SYBR Green PCR Master Mix Reagent kit (Qiagen). Ratios of the intensities of the target genes and β-actin, an internal control, were used as a relative measure of the target gene expression. The mean fold change of expression in the experimental group compared with the control group was calculated using the 2^–∆∆ct method, and the range of the fold changes was calculated from the standard error of the values. To ensure experimental accuracy, real-time RT-PCR assays were conducted in triplicate for each sample. Primer sequences and product sizes are defined in Table 1.

Proteins were extracted with a lysis buffer (Cell Signaling Technology, Danvers, MA, USA) and quantified using the bicinchonic acid protein (BCA) protein-assay reagent (Bio-Rad Laboratories, Hercules, CA, USA). Proteins were resolved by 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane (Whatman GmbH, Dassel, Germany). After blocking in Tris-buffered saline with 5% milk and 0.1% Tween-20, the membrane was incubated with the neurofilament medium (NFM) antibody (Biolegend, San Diego, CA, USA) or Nestin antibody (Abcam, Beverly, MA, USA). The mouse monoclonal primary antibody to β-actin (Sigma-Aldrich) was used as the reference. Signals were visualized using an enhanced chemiluminescence reagent (ECL; Santa Cruz Biotechnology, Santa Cruz, CA, USA) in Image Reader LAS-4000 software (Fujifilm, Tokyo, Japan).

Cells were split by 0.2% ethylenediaminetetraacetic acid (EDTA) in PBS. For intracellular markers, cells were fixed and permeabilized with 90% ice-cold methanol for 30 min and treated with Fc block (anti-FcgRII/III) to prevent nonspecific binding for 15 min. Indirect staining was done with primary and secondary antibodies on each step. Stained cells were analyzed on FACScan (BD Biosciences, Mountain View, CA, USA). FACS data was quantified by mean fluorescence intensity (MFI). The arithmetic mean was used in the computation of the MFI. The following antibodies and their corresponding isotype controls were used; Antibody against cluster of differentiation 146 (CD146) and octamer-binding transcription factor 4 (OCT4) was purchase from Santa Cruz Biotechnology and antibodies against mouse IgG1 and IgG2b for κ Isotype control from Biolegend.

After 14 days of neuronal differentiation, cells were fixed with 10% neutral formalin for 24 h and immunofluorescence staining was done using the tyramide signal amplification (TSA) Kit (Invitrogen, Carlsbad, CA, USA). Briefly, after blocking endogenous peroxidase, cells were reacted with the NFM antibody (Biolegend) overnight and subsequently with the horseradish peroxidase (HRP)-conjugated secondary antibody (Cell Signaling Technology). Cells were incubated in Alexa Fluor 488 tyramide solution (Thermo Fisher Scientific, Waltham, MA, USA) and then counterstained with propidium iodide (Thermo Fisher Scientific). Cells were photographed using an LSM confocal microscope (Carl Zeiss, Oberkochen, Germany).

For alkaline phosphatase (ALP) enzyme staining, cells were fixed with 4% formaldehyde, rinsed three times with deionized water, and treated with a 5-bromo-4-chloro-3-indolyl-phosphate (BCIP)/nitro blue tetrazolium (NBT) solution (Sigma-Aldrich) for 15 min. The reaction was stopped by the addition of water, and the stains were photographed and quantified by Scion Image software (Scion Corporation, Frederick, MD, USA). To evaluate mineralization, Alizarin Red S staining was applied. Briefly, cells were fixed with 70% ethanol and then stained with 40 mM Alizarin Red S (pH 4.2) (Sigma-Aldrich) solution for 10 min. After extraction with 10% cetylpyridinium chloride (Sigma-Aldrich), the stain was quantified by measuring the absorbance at 570 nm.

Oil Red O staining is performed to evaluate adipogenic differentiation. Cells were fixed with 10% formalin for 30 min and washed with PBS, then were stained with 0.5% Oil Red O solution (Sigma-Aldrich) for 30 min, washed with deionized water, and photographed. The images of stained cells were acquired using an inverted contrast-phase microscope (Olympus Corporation, Tokyo, Japan).

To evaluate chondrogenic differentiation, Alcian blue staining was used. Briefly, cells were fixed with 4% formaldehyde, washed with deionized water, and stained with a 1% Alcian blue (pH 2.5) (Sigma-Aldrich) solution for 30 min. Stained cells were washed with deionized water and photographed.

Data are expressed as means ± standard errors. Statistical analysis was carried out using the analysis of variance (ANOVA) software (Analyse-it Software Ltd., Leeds, WY, UK). Results were considered significant at a probability level of p < 0.05. For all experiments, at least three independent trials were performed.

Epigenetic changes in DNA methylation affects the differentiation potential of cells. To induce epigenetic modification in NIH-3T3 fibroblasts, we used 5-aza-CN, a DNA methyltransferase inhibitor, which is well known to cause DNA demethylation. First of all, in order to evaluate the capacity of 5-aza-CN to transform cell phenotypes, morphological changes of NIH-3T3 fibroblasts were examined under an inverted microscope. Exposure to 5 μM 5-aza-CN for 24 h led standard elongated fibroblasts to display shortening and shrinking phenotypes, which were restored efficiently to their initial shape by an additional 24 h culture in fresh medium without 5-aza-CN (Fig. 1A). Next, we examined the effect of 5-aza-CN on DNA demethylation of NIH-3T3 fibroblasts, which were treated at the indicated concentrations of 5-aza-CN for 24 h and then global DNA methylation status were assessed. As shown in Fig. 1B, methylated DNA was significantly decreased in a dose-dependent manner. Finally, we assessed the multi-lineage differentiation potential of 5-aza-CN-treated NIH-3T3 fibroblasts by evaluating the relative gene expression levels of the stem cell markers, such as (sex determining region Y[SRY]-box 2) (SOX2), Krüppel-like factor 4 (KLF4), NANOG, and OCT4, which are transcription factors that is essential maintaining in self-renewal, or pluripotency, of undifferentiated embryonic stem cells [21]. Expression of these markers was increased clearly, and the expression changes of CD146, a mesenchymal stem cell marker [22] and OCT4 protein in 5-aza-CN-treated NIH-3T3 fibroblasts were also observed in flow cytometry analysis (Fig. 1C and D). MFI of CD146 has increased significantly from 0.74 to 3.54, while the value of OCT4 has slightly changed.

In order to evaluate the differentiation potential, brief exposure of NIH-3T3 fibroblasts to 5-aza-CN was followed by the procedure illustrated in Fig. 2A. For conversion into mesenchymal lineage cells, 5-aza-CN-treated NIH-3T3 fibroblasts were incubated with specific defined differentiation media containing 200 ng/ml BMP2. As shown in Fig. 2B, BMP2 markedly increased the mRNA of osteogenic key molecules, such as ALP, osteopontin (OPN), osteocalcin (OCN), bone sialoprotein (BSP) which are major non-collagenous proteins in bone matrix, and Runt-related transcription factor 2 (RUNX2) and osterix (OSX), which are essential transcription factors for osteoblast differentiation [23], in 5-aza-CN-treated NIH-3T3 fibroblasts cultured with osteogenic medium. Furthermore, the density of ALP and AR-S staining performed to investigate early osteogenic differentiation and mineral nodule formation, respectively, were increased significantly in the group treated with both 5-aza-CN and BMP2 (Fig. 2C and D). Next, the 5-aza-CN-treated NIH-3T3 fibroblasts expressed effectively adipogenic marker genes, such as adipocyte protein 2 (AP2) which is a carrier protein for fatty acid, and CCAAT-enhancer-binding proteins (C/EBP) and peroxisome proliferator-activated receptor gamma (PPARG) which are key transcription factors for adipocyte differentiation [24], and vigorously formed oil droplets in a specific adipogenic medium containing BMP2 compared to control (Fig. 3A and B). Finally, in chondrogenic media, BMP2 induced chondrogenic differentiation of 5-aza-CN-treated NIH-3T3 fibroblasts, which significantly increased the expression of chondrogenic markers such as aggrecan (ACAN) and type II collagen (COL2) which are major matrix proteins of cartilaginous tissue, and SRY-box 9 (SOX9) which is essential transcription factor for differentiation of precursors cells into chondrocyte [25], as well as proteoglycan stained by Alcian blue was significantly increased (Fig. 3C and D). These results indicate that 5-aza-CN-treated NIH-3T3 fibroblasts can be converted to desired mesenchymal lineage cells capable of specific matrix formation if it is properly induced.

After epigenetic modification, NIH-3T3 fibroblasts were incubated with a defined neurogenic medium to evaluate the neurogenic potential. As shown in Fig. 4A and B, mRNA and protein of Nestin and NFM which are intermediate filament proteins expressed in nerve cells [26], were clearly increased in 5-aza-CN-treated NIH-3T3 fibroblasts. Fig. 4C shows the morphologic changes of neurogenic differentiated NIH-3T3 fibroblasts, which have an expanded cytoplasm and elongated axon-like pod as well as an increase of NFM protein in the cell cytoplasm. In contrast, control cells showed no morphological changes or NFM protein.