Overexpression of MicroRNA-148b-3p stimulates osteogenesis of human bone marrow-derived mesenchymal stem cells: the role of MicroRNA-148b-3p in osteogenesis

Bone marrow aspirates of healthy donors undergoing therapeutic surgery were used to isolate MSCs as previously described [20]. The entire procedure was approved by the Ethics Committee of Mashhad University of Medical Sciences and carried out with written informed consent of the donors. Briefly, bone marrow samples were aspirated in a sterile condition during surgery. Then, they were aliquoted into 75-cm² culture flask by adding Dulbecco’s modified Eagle’s medium-low glucose (DMEM-LG) (Gibco, Paisley, Scotland) consisting 20% fetal bovine serum (FBS; Gibco, Paisley, Scotland) as well as 1× penicillin/streptomycin (Pen/Strep; Gibco, Paisley, Scotland). Cell suspension was maintained at 37 °C in a humidified incubator with 5% CO₂. After 3 days, cultures were rinsed with 1× phosphate-buffered saline (PBS) twice to eliminate non-adherent cells and allowed to reach confluency. Then, confluent cells were trypsinized and subcultured at a density of 5–10 × 10³ cells/cm² in new flasks. The characterized BM-MSCs in terms of cell morphology, cell markers and osteogenic and adipogenic capability were used for this study at passage three.

Naturally, primary miRNAs are produced after transcription of miRNAs by RNA polymerase II. Then, primary miRNAs are cleaved to hairpin pre-miRNAs which are transferred to cytoplasm followed by processing into RNA duplex (~ 22-nucleotide) [21]. The mechanism of RNA splicing can be more efficient when pre-miRNA are flanked by specific sequences [22, 23]. In an attempt to improve our expression strategy, we designed our miRNA based on the pattern as indicated below. The sequences of the mature hsa-miR-148b-3p (5′-UCAGUGCAUCACAGAACUUUGU-3′) as well as hsa-miR148b-5p (5′-AAGUUCUGUUAUACACUCAGGC-3′) were obtained from miRBase (http://​www.​mirbase.​org/​). Besides, anti-mature sequences were designed as complementary sequences of matures. According to this, oligonucleotides pairs (forward and reversed) were annealed for both miRNAs. After ligation, the whole sequences were flanked by XhoI recognition site, which is essential for constructing hairpin structures. Also, the two restriction sites were developed at the both 5′ and 3′ ending sites followed by a sequences of 5 T residues (RNA polymerase III transcription termination signal) (Fig. 1a). Then, hybridized products of these oligonucleotids and their reverse stretches were cloned in between the unique SgrAI and EcoRI recognition site of the LV shuttle construct SHC007-hEEF1a1.EGFP. SHC007-hEEF1a1.EGFP is derived from the shRNA plasmid DNA control vectors SHC007 (Sigma-Aldrich, St. Louis, MO), in which the human phosphoglycerate kinase 1 gene promoter as well as the Streptomyces alboniger puromycin N-acetyl-transferase-coding sequence are replaced by the human eukaryotic translation elongation factor 1 alpha 1 gene promoter and the coding sequence of the enhanced green fluorescent protein (EGFP) of Aequorea Victoria. The consequential plasmids, pLV.hU6.miR-148b-3p.hEEF1a1.EGFP (hereinafter called pLV-miR-148b-3p) and pLV.hU6.miR-148b-5p.hEEF1a1.EGFP (hereinafter referred to as pLV-miR-148b-5p) interceded the expression of miR-148b-3p and miR-148b-5p, respectively. Both plasmids also encode the EGFP expression. The correctness of DNA constructs were approved by restriction endonuclease digestions (using AccI, HaeII, PVuI, and XhoI) as well as nucleotide sequence analysis. Sequencing was accomplished through Leiden Genome Technology Center (http://​www.​lgtc.​nl/​) using a BigDye Terminator v3.1 Cycle Sequencing Kit and a 3730xl DNA Analyzer (both from Thermo Fisher Scientific, Waltham, MA). Enzymes used in this study were purchased from New England Biolabs (BIOKÉ, Leiden, the Netherlands) and Fermentas (Thermo Fisher Scientific, USA) and applied according to the standard protocols. Detailed genetic maps of pLV-miR-148b-3p, pLV-miR-148b-5p as well as SHC007-hEEF1a1.EGFP (hereinafter designated pLV-Ctrl) are provided in Fig. 1b, c.

To construct lentiviral vector particles, HEK (293 T) cells were infected with transfection medium composed of molar ratio of 1:1:2 packaging plasmids [PSPAX2 (Addgene; plasmid number; 12,260), pLP/VSVG (Life Technologies Europe)] and lentiviral vector shuttle plasmid pLV-miR-14bb-3p/miR-148b-5p (to generate LV-miR-148b3p/−5p particles) or pLV-Ctrl (to produce control vector [LV-Ctrl] particles) supplemented with NaCl 150 mM, and 25-kDa polyethylenimine (PEI; PolysciencesEurope, Hirschberg an der Bergstrasse, Germany). Following overnight incubation of DNA–PEI complexes, medium was refreshed with DMEM-LG supplemented with 5% FBS and 25 mM HEPES-NaOH (pH 7.4). After 48 h, supernatants of transfected cells were collected and centrifuged to remove cellular debris followed by filtering through 0.45-μm pore-sized, 33-mm diameter polyethersulfone Millex-HP syringe filters (Millipore, Amsterdam, the Netherlands). Viral particles were precipitated by adding a cushion of 20% (wt/vol) sucrose to the cleared supernatants, and then ultracentrifuged at 15000 rpm for 2 h at 4 °C in a SW32 swinging bucket rotor (Beckman Coulter Nederland, Woerden, the Netherlands). Finally, the supernatants were thoroughly removed and PBS (1%) containing 1% bovine serum albumin (Sigma-Aldrich Chemie, Taufkirchen, Germany) were added to viral pellet followed by gentle rocking overnight on the shaker (110 rpm) at 4 °C. As a final point, viral vectors were aliquoted in 1.5 ml microtubes and stocked at − 80 °C.

Human BM-MSCs were seeded at density of 5 × 10³ cells/cm². The following day, the cells were incubated for ±20 h with LV-miR-148b-3p or LV-miR-148b-5p or LV-Ctrl in culture medium supplemented with 2.5 μg/ml DEAE-dextran sulfate (Sigma-Aldrich Chemie, Taufkirchen, Germany). The other day, cells were rinsed with PBS and the transfection medium was changed with osteogenic differentiation medium (ODM) containing DMEM-LG included 15% FBS, and 100 U/ml Pen/Strep supplemented with 10 mM ß-glycerophosphate, 50 μg/ml ascorbic acid, and 0.1 μM dexamethasone (all from Sigma-Aldrich Chemie, Taufkirchen, Germany). To check the transduction efficiency, direct EGFP fluorescence was evaluated at day 7 using an inverse fluorescence microscopy (Nikon Eclipse TE2000-U, Tokyo, Japan).

Twenty-one days after osteogenic induction in order to stain calcified nodules, cells were washed twice with PBS and fixed with 10% formalin (Sigma-Aldrich Chemie, Taufkirchen, Germany) for 45 min at room temperature (RT). Then, fixed cells were incubated with 1% Alizarin red S staining solution [ARS; Sigma-Aldrich Chemie (pH 4.1–4.3)] for 30 min to develop mineralized bone nodules as previously described [24]. Calcium deposition was identified as orange and red bodies under light microscope (Nikon Eclipse TE2000-U) [25, 26]. To quantify absorbed ARS [27], stained cells were detached using 10% acetic acid (Sigma-Aldrich Chemie, Taufkirchen, Germany) and homogenized. Then, the acidity of the supernatants was neutralized by 10% ammonium hydroxide. Finally, the absorbance was determined at the wavelength of 405 nm by plate reader (BioTek, Bad Friedrichshall, Germany).

To immunostain markers of interest in human BM-MSCs, they were osteo-inducted at the density of 1 × 10⁵ cells/well in 48-well plates (BD Biosciences, San Jose, CA). Fourteen days after osteogenic differentiation, induced cells were washed twice with PBS and fixed with PBS (1×) containing 4% paraformaldehyde at RT for 30 min. Cells were permeabilized with 0.1% Triton X-100 in PBS and then incubated overnight at 4 °C in the presence of mouse anti-alkaline phosphatase (ALP; BD Biosciences, San Diego, US) and collagen type I (ColI; Sigma-Aldrich Chemie, Taufkirchen, Germany) primary antibodies diluted 1:200 and 1:1000 in PBS + 0.1% donkey serum (Santa Cruz Biotechnology, Santa Cruz, CA), respectively. The next day, cells were incubated with Alexa Fluor 568-conjugated donkey anti-mouse IgG (H + L) secondary antibody (Life Technologies Europe, Bleiswijk, the Netherlands) diluted 1:400 in PBS + 0.1% donkey serum for 4 h at RT. Finally, the nuclei were stained with Hoechst 33258 staining (Sigma-Aldrich Chemie, Taufkirchen, Germany) for 10 min. Cells were completely washed with PBS after each step. Images were taken with a fluorescence microscope attached to digital color camera. The Mean fluorescence signal intensity was determined using ImageJ (version 4.1; National Institutes of Health, Bethesda, MA).

Two weeks after osteogenesis, induced cells were lysed using Trizol reagent (Invitrogen, Massachusetts, USA) followed by total protein extraction based on the supplier protocol. Then, the concentrations of purified proteins were evaluated by Bradford protein assay kit (Thermo Scientific, Waltham, MA) in which bovine serum albumin (BSA) was used as the standard protein. To determine alkaline phosphatase (ALP) protein expression by Western blotting, 15 μg/ml denatured protein samples were loaded onto sodium dodecyl sulfate (SDS)-containing sample buffer, boiled for 5 min and applied to a 12% SDS-polyacrylamide gel and electrophoresed in Tris-glysine buffer system (Bio-Rad, Hercules, California, USA), which was conducted at 120 V for 60 min. The proteins were then transferred onto polyvinylidene difluoride (PVDF; Bio-Rad, California, USA) membrane in 1× transfer buffer for 1 h at 250 mA in a transblot electrophoretic transfer cell (Bio-Rad California, USA). Membrane was blocked with Tris-buffered saline containing 5% BSA diluted in 0.1% Tween-20 (TBS-T solution) for 3 h at RT and then rinsed with TBS-T solution. For immunodetection of ALP, the blocked membrane was incubated with mouse anti-ALP primary antibody diluted 1:500 in TBS-T plus BSA 2.5% for 1 h at RT, washed four times with TBS-T solution for 20–30 min, and treated with 1:1000 diluted anti-mouse horseradish peroxidase (HRP)-conjugated secondary antibody (Santa Cruz Biotechnology, Texas, USA) for 1 h at RT. Finally, the signals were visualized using chemiluminescence ECL detection kit (Bio-Rad, California, USA) and exposed to X-ray film (Eastman Kodak, Rochester, New York, USA) for 30 s to 10 min. β-actin [mouse anti- β-actin diluted 1:800 (Santa Cruz Biotechnology; Texas, USA)] was used as the standard protein.

Human BM-MSCs could adhere to plastic surface based on their typical properties. The phenotype of primary human BM-MSCs was verified by MSC immunophenotyping, the capacity to form single spindle-fibroblastic cells, and the adipogenesis and osteogenesis potentials. As previously demonstrated, these cells were positive for CD44, CD90, and CD105 (MSC cell surface markers), whilst they were negative for CD34, CD45 (hematopoietic cell markers), and CD11b (pan-macrophage marker) [20]. Alizarin red S staining showed the ability of the primary human BM-MSCs to form mineralized nodules. Moreover, Oil Red O staining displayed that these cells could accumulate lipid droplets under adipogenic conditions. Taken together, findings of osteogenesis and adipogenesis tests displayed that human BM-MSCs have multipotential differentiation properties.

One day after transfection of HEK 293 T cells with LV shuttle and packaging plasmids, almost all cells were positive for EGFP, which indicated the transfection efficiency was high. To check the gene transfer activity of the LV-miR-148b-3p, LV-miR-148b-5p, and LV-Ctrl particles, 5 × 10⁴/cm² HEK 293 T indicator cells were treated with 3 μl of a 1000-fold dilution of all three vector stocks. The results of EGFP expression confirmed all vectors were functional with a similar titer. We used empty LVs as a negative control to ignore the probable effects of plasmid backbones in order to study the osteogenic effect of LV-miRs (miR-3p or miR-5p).

To investigate the osteogenic effects of permanent expression of hsa-miR-148b, human BM-MSCs were infected with LV-miR-148b-3p, which codes both miR-148b-3p and EGFP, or LV-miR-148b-5p, which express both miR-148b-5p and EGFP or LV-Ctrl, that codes only EGFP. The transduction efficiency of the human BM-MSCs 3 days post-transduction was about 90% as indicated by direct EGFP fluorescence microscopy. Then, transduced cells were incubated in ODM for 21 days. Although, LV functionality were similar in all groups (Fig. 2, left column), the amounts of osteogenic differentiation was significantly more in the cells transduced with LV-miR-148b-3p as shown by white calcified matrix (Fig. 2, right column). In the following, osteo-induction was assayed by alizarin red staining (ARS) as well as alkaline phosphatase (ALP) assay.

As demonstrated in Fig. 3a (Additional file 1 Figure S1), miR-148b-3p overexpression resulted in more bright red bone nodular formation in human BM-MSCs in a time dependent manner. Consistently, ARS quantification displayed 1.5-fold increase in calcification in LV-miR-148b-3p-transduced cells in comparison with LV-miR-148b-5p- and LV-Ctrl-transduced cells (Fig. 3b). Similar data was achieved from ALP staining (Fig. 3c) in which miR-148b-3p-overexpressed human BM-MSCs revealed dark purple staining and 2.5-fold enhancement in ALP production compared to LV-miR-148b-5p- and LV-Ctrl-transduced cells (Fig. 3d). Also, the results of immunostaining (Fig. 4) implicated that expression of miR-148b-3p in human BM-MSCs resulted in an enhanced expression of about 2.5-fold in ALP and ColI compared to LV-miR-148b-5p- and LV-Ctrl–transduced cells.

Western blot analysis, on day 14th of post differentiation displayed immunoreactive bands with a weight of 89 kDa in response to the anti-ALP antibody (Fig. 5). Although Western blot analysis revealed that ALP was expressed in all three groups, LV-miR-148b-3p-treated cells exhibited 1.2-fold enhancement in ALP expression compared to LV-miR-148b-5p-transduced cells. Altogether, these data demonstrated stimulatory effects of miR-148b-3p overexpression on osteogenic differentiation in human BM-MSCs.