The miR‐193a‐3p‐MAP3k3 Signaling Axis Regulates Substrate Topography‐Induced Osteogenesis of Bone Marrow Stem Cells

Abstract Substrate topographical features induce osteogenic differentiation of bone marrow stem cells (BMSCs), but the underlying mechanisms are unclear. As microRNAs (miRNAs) play key roles in osteogenesis and bone regeneration, it would be meaningful to elucidate the roles of miRNAs in the intracellular signaling cascade of topographical cue‐induced osteogenic differentiation. In this study, the miRNA expression profile of the topographical feature‐induced osteogenic differentiation group is different from that of the chemical‐factors‐induced osteogenic differentiation group. miR‐193a‐3p is sensitive to substrate topographical features and its downregulation enhances osteogenic differentiation only in the absence of osteogenesis−inducing medium. Also, substrate topographical features specifically activate a nonclassical osteogenetic pathway—the mitogen‐activated protein kinase (MAPK) pathway. Loss‐ and gain‐of‐function experiments demonstrate that miR‐193a‐3p regulates the MAPK pathway by targeting the MAP3k3 gene. In conclusion, this data indicates that different osteogenic‐lineage‐related intracellular signaling cascades are triggered in BMSCs subjected to biophysical or chemical stimulation. Moreover, the miR‐193a‐3p‐MAP3k3 signaling axis plays a pivotal role in the transduction of biophysical cues from the substrate to regulate the osteogenic lineage specification of BMSCs, and hence may be a promising molecular target for bone regenerative therapies.

. Expression levels of  expression in hBMSCs in the flat (OS+), random, and random (OS+) groups at 14 days. Results are means ± SEM (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test. Significant differences are denoted by *p < 0.05, **p < 0.01, and ***p < 0.001. Supplementary Fig. 6. The promotion of osteogenic differentiation of hBMSCs by low hsa-miR-193a-3p expression was blocked by osteogenic differentiation medium (OS). Effect of agomir-193a-3p, antagomir, or their corresponding scrambled controls on the Runx2 (a) and Ocn (b) mRNA levels in hBMSCs cultured in osteogenic differentiation medium. (c) The expression of miR-193a-3p at 48 hours posttransfection in osteogenic medium. Gross images (d) of ALP staining of hBMSCs treated with agomir-193a-3p, antagomir, or their corresponding scrambled controls for 5 days in osteogenic differentiation medium. Results are means ± SEM (n = 3). Samples were subjected to one-way ANOVA with Tukey post hoc test. Significant differences are denoted by *p < 0.05.

Materials and Methods
Fabrication of electrospun PLLA nanofiber. PLLA powder (0.7 g) was added to 10 mL of trifluoroethanol and stirred overnight. The solution was ejected from a 20 mL syringe with a steel needle (inner diameter: 0.5 mm), using a programmable syringe pump, into a conventional electrospinning apparatus (Elite; Ucalery, Beijing, China) at a rate of 0.7 mL/h. High-voltage equipment was used to provide a constant voltage (16 kV) to the tip of the needle when the fluid was ejected, and a metal plate (20 × 30 cm 2 ) was used as a collector at a distance of 15 cm from the tip of the needle, to obtain randomly arranged PLLA nanofibers (random group). To fabricate flat PLLA films (flat group), the PLLA polymer solution was cast on a flat glass plate at a thickness of 75 μm, and dried at 40°C for 6 h. The samples (PLLA nanofibers and flat films) were kept in a vacuum oven (DZF-6210; Blue Pard, Shanghai, China) at room temperature for 2 weeks to remove residual solvent. The nanotopography of the films was imaged by scanning electron microscopy (SU-8010; Hitachi, Tokyo, Japan), at an accelerating voltage of 15 kV.

Gel-MA hydrogel fabrication and rheological characterization.
Gel-MA, an alterable elastic hydrogel, is commonly used as a mechanical model for cell culture. Briefly, Gel powder was added to phosphatebuffered saline (PBS) at 50°C and stirred until fully dissolved at 10% (w/v). Methacrylic anhydride was added to this solution at a rate of 0.3 mL/minute while stirring at 50°C until the target volume was reached. The mixture was next dialyzed against distilled water at 40°C for 2 weeks using 7000 molecularweight (MW) cutoff dialysis tubing. The water was changed every day to remove salts and methacrylic acid. Subsequently, the Gel-MA solution was lyophilized and stored at −80°C for future use. Lyophilized Gel-MA at 3%, 5%, 7%, 10% and 20% (w/v) was dissolved in human mesenchymal stem cell basal medium containing 0.5% (w/v) 2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropiophenone (Aladdin, Shanghai, China) at 50℃ to prepare a pre-polymer solution. Two milliliters of the mixture were added to a 35-mm-diameter dish and exposed to ultraviolet (UV) light for 10 min. The rheological properties of the Gel-MA hydrogel were determined using a HAAKE RheoWin MARS 40 (Thermo Fisher, Waltham, MA, USA). Gel-MA hydrogels were placed within 35-mm-diameter parallel plates separated by 1.00 mm. Ampl sweeps were conducted from τ₀ 0.01000-1,000 Pa (f = 1.000 Hz, T = 37℃). Frequency sweeps were performed at a constant strain from τ₀ 0.03000-4 Pa according to the Gel-MA hydrogel concentration and frequencies of 0.1-14.00 Hz at 37℃.
Cell culture and seeding on PLLA nanofilms, Gel-MA, or six-well plates. The hBMSCs used in this research were supplied by Cyagen Biosciences Inc. (Guangzhou, China); they were obtained as surgical waste material from normal male donors aged 20-30 years. The culture medium was human mesenchymal stem cell-basal medium containing 5% (w/v) mesenchymal stem cell-qualified fetal bovine serum (FBS), 10 μg/mL (w/v) L-glutamine, and 100 IU/mL penicillin-streptomycin (Cyagen Biosciences Inc.). The medium was changed every 2-3 days. At 80-90% confluence, hBMSCs were detached with 0.25% (w/v) trypsin/ ethylenediaminetetraacetic acid (EDTA) (Gibco, Grand Island, NY, USA) and subcultured at a density of 5 × 10 5 cells per T75 flask, or plated onto six-well plates at a density of 1 × 10 5 cells per well. Fourth-passage hBMSCs were used in this study. PLLA nanofiber scaffolds and flat polymer films were cut into 4 × 4 cm sections and attached to the cap of centrifuge tubes (50 mL), as in our previous study. After placement within into six-well plates, these materials were sterilized with radiation for 12 h, and washed three times with PBS. Cells cultured on flat polymer films in OS comprising 50 mg/mL ascorbic acid, 10 mM sodium β-glycerol phosphate, and 10 −8 M dexamethasone were used as flat (OS+) and random (OS+) positive controls. Negative controls were cultured on flat polymer films in normal culture medium without osteoinductive supplements.
Cell morphology and cytoskeletal remodeling. Morphological changes and cytoskeletal remodeling of hBMSCs were imaged by confocal microscopy. hBMSCs cultured on nanofibers for 4 h, and hBMSCs cultured on Gel-MA for 24 h, were washed with PBS, fixed in 4% (w/v) paraformaldehyde for 15 min, washed twice in PBS, permeabilized with 0.2% (w/v) Triton X-100 for 10 min, and blocked with 5% (w/v) bovine serum albumin (BSA) for 60 min at room temperature. The flat or random PLLA membranes were stained with tetramethylrhodamine isothiocyanate (TRITC) phalloidin (100 nM; Solarbio, Shanghai, China) 40 min at room temperature, followed by counterstaining with 1 μg/mL 4',6diamidino-2-phenylindole (DAPI) (blue) for 10 min. The cells were visualized by confocal laser scanning microscopy (TCS SP8; Leica, Wetzlar, Germany). Image Pro Plus software (Media Cybernetics, Silver Spring, MD, USA) was used to evaluate the effect of Gel-MA concentration on cell morphology. The shape of the cell and nucleus of at least 30 cells per group were analyzed.
Quantitative reverse transcriptase-polymerase chain reaction. The PLLA nanofilms or Gel-MA loaded with cells for in vitro culture were prepared as described previously. Total RNA of cells cultured on PLLA nanofilms was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Total RNA of the cells loaded on Gel-MA was extracted using the RNAprep Pure Cell Kit (DP430; Tiangen, Beijing, China). mRNA was reverse-transcribed into cDNA using the PrimeScript ® RT Reagent Kit (TaKaRa, Tokyo, Japan) at 37°C for 15 min and 85°C for 5 s according to the manufacturer's protocol. The synthesized cDNA samples were subjected to qRT-PCR to determine the expression levels of Runx2, Ocn, Map3k3, Bsp, Erk1, Jnk, and Erk5 using SYBR Green Master Mix (Roche Diagnostics Ltd., Mannheim, Germany) and a 7500 ABI Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Relative gene expression levels were calculated using the 2 (-ΔΔCT) method. Briefly, the mRNA and miR-193a-3p data were normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and U6 (U6 spliceosomal RNA), respectively. The following primer sequences were used for qRT-PCR: Gapdh (5'-CCTCTGACTTCAACAGCGAC-3', 5'-TCCTCTTGTGCTCTTGCTGG-3'), Runx2 miRNA array, gene array and bioinformatics analysis. Total RNA, including small RNAs, from hBMSCs cultured for 14 days on PLLA nanofibers (random) and flat polymer films with (flat OS+) or without OS (flat), was extracted using TRIzol reagent (Invitrogen) and purified using the mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA) according to the manufacturer's protocol. miRNA profiling was performed using an Agilent miRNA array (CapitalBio Corp, Beijing, China). The Agilent array has eight identical arrays per slide (8 × 60 K format), with each array containing probes interrogating 2,549 human mature miRNAs from miRBase R21.0 (www.mirbase.org). Each miRNA was detected by probes repeated 30 times. The array also contained 2,164 Agilent control probes. The miRNA profiling was performed using the Agilent miRNA array. The Agilent array was designed with eight identical arrays per slide (8 × 60 K format), with each array containing probes interrogating 2,549 human mature miRNAs from miRBase R21.0. Each miRNA was detected by probes repeated 30 times. The array also contained 2,164 Agilent control probes.
Briefly, 200 ng of total RNA were used to synthesize first-strand cDNA followed by double-stranded cDNA using a Message Amp Premier RNA Amplification Kit and PCR apparatus (MJ, PTC-225; Ambion). Biotin-labeled cRNA was synthesized using a MessageAmp Premier RNA Amplification Kit (Ambion). The concentration of cRNA was measured using a NanoDrop ND-1000 instrument, and 15 μg of fragmented cRNA was hybridized to each GeneChip Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara, CA, USA), which contains more than 54,000 probe sets to cover over 47,000 transcripts and variants, at 45°C for 16 h (Affymetrix Hybridization Oven 640) according to the manufacturer's instructions. After hybridization, the arrays were washed, stained with streptavidin phycoerythrinonan using an Affymetrix Fluidics Station 450, and scanned using an Affymetrix Scanner 3000 7G. Three replicates of the microarray experiment were performed until high reproducibility was achieved. Each replicate comprised pooled cell lysates from three wells.
Bioinformatics analysis. The miRNA array data were , normalized, summarized and subjected to quality control using GeneSpring software v. 12 (Agilent). The default 90 th percentile normalization method was applied for data pre-processing. To identify differentially expressed genes, we used threshold values of ≥ 2.5 and ≤ −2.5-fold changes, and a t-test p-value of 0.05. The raw data were log2-transformed and median-centered by genes using the Adjust Data function of CLUSTER 3.0 software (Stanford University School of Medicine, Stanford, CA, USA), and further analyzed by hierarchical clustering and heatmaps with average linkage. We performed tree visualization in using Java TreeView (Stanford University School of Medicine). In the heatmap, red denotes high relative expression, and green low relative expression. A PCA was performed to determine the gene expression trends of the samples.
Ontology and pathway analysis. The number of genes potentially regulated by one miRNA ranges from several hundreds to a couple of thousand. The predicted mRNA targets of downregulated miRNAs were picked up by no less than 7 out of 12 commonly used predicting databases: miRWalk, microT v4, miRanda, mirbridge, miRDB, miRMap, miRNAMap, Pictar2, PITA, RNA22, RNAhybrid, and Targetscan. The predicted mRNA targets that overlapped with upregulated mRNAs in gene chips subjected to ontology analysis with KOBAS 3.0 software were identified to determine the enriched biological pathways regulated by the candidate downregulated miRNAs. miRNA transfection. hBMSCs were cultured to 70% confluence in six-well plates in basal medium with 5% (v/v) FBS. Next, 50 nM agomir or 100 nM antagomir (GenePharma) was transfected into cells using