Dulbecco's modified Eagle's medium (DMEM; cat. no. 11965-092), penicillin-streptomycin-amphotericin B (cat. no. 15240-062) and cell culturing reagents were procured from Lonza Group Ltd. Fetal bovine serum (FBS; cat. no. 10270-106) was sourced from Gibco (Thermo Fisher Scientific, Inc.). Human osteoblastic osteosarcoma cells (MG-63) and human bone marrow stromal cells (HS-5) were obtained from the National Center for Cell Science (Pune, India) and the American Type Culture Collection, respectively. TGF-β1 was obtained from R&D Systems, Inc. Antibodies against CBP (cat. no. 7389; 1:1,000), acetylated-lysine (cat. no. 9441; 1:1,000) and α-Tubulin (cat. no. 2125; 1:1,000) were acquired from Cell Signaling Technology, Inc., and antibodies against Runx2 (cat. no. sc-390715; 1:100) and MMP-13 (cat. no. 18165-1-AP; 1:3,000) were purchased from Cell Signaling Technology, Inc., and Proteintech Group, Inc., respectively. Scrambled control siRNA (cat. no. Sc-37007) and CBP siRNA (cat. no. Sc-29244) were purchased from Santa Cruz Biotechnology, Inc. miR-4327 mimic (GeneGlobe ID: YM00470747) was purchased from Qiagen GmbH.

HS-5 cells were differentiated into primary osteoblasts by culturing in DMEM along with 10% FBS, 50 µM ascorbic acid, 10 nM β-glycerophosphate and 0.1 µM dexamethasone for seven days. DMEM containing 10% FBS was used to maintain MG-63 cells. Penicillin-streptomycin-amphotericin B was used in the culture media and cells were incubated in a humidified chamber with 5% CO₂ at 37˚C. In the present study, TGF-β1 was used at 5 ng/ml.

miRNAs that target CBP 3'-UTR were identified based on miRNA-target prediction databases. Human miRNA sequences were retrieved from miRDB (https://mirdb.org/). Mature miRNA sequences were analyzed and binding sites were predicted using STarMir (https://sfold.wadsworth.org/cgi-bin/starmirWeb.pl). STarMir provides a logistic probability score (LogitProb), signifying the confidence level of binding between miRNA and target mRNA, based on defined interaction parameters such as site type, ΔG_hybrid ≤-14 kcal mol^-1 and ΔG_total ≤-10 kcal mol^-1 (39,40). The highly probable miRNAs were classified using miRmap (https://mirmap.ezlab.org/). The predicted miRNAs were loaded into Venny v.2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny/) to identify the miRNAs common among the three databases. The common miRNAs were shortlisted based on the LogitProb score (cut-off <0.75) and miRDB score (cut-off <50). Finally, the validated miRNAs were eliminated using TarBase (https://dianalab.e-ce.uth.gr/tarbasev9) and a web-based search (https://scholar.google.com) for obtaining unvalidated miRNAs.

Total RNA was isolated using RNAiso Plus (Takara Bio, Inc.) from cells at 80% confluence. Complementary DNA (cDNA) was synthesized using an iScript cDNA synthesis kit (Bio-Rad Laboratories, Inc.). qPCR was performed using SYBR Green (Takara Bio, Inc.) with primers for precursor miRNAs. Expression patterns of mature miRNAs were analyzed using a miRCURY LNA kit (Qiagen GmbH) with mature miRNA primers. The ΔΔCq method was used to determine the relative expression of precursor and mature miRNAs (13,30). U6 served as an endogenous control. Table I shows the primers used to determine precursor miRNA expression in human osteoblasts. The PCR protocol consisted of denaturation at 95˚C for 5 sec, followed by annealing and extension at 60˚C for 34 sec, for 40 cycles. All the experiments were performed in triplicate according to the manufacturers' protocols.

MG-63 cells (60-70% confluence) were transiently transfected with scrambled control (30 nM) or small interfering (si)RNA for CBP (30 nM) or negative control (50 nM) or miR-4327 mimic (50 nM) using X-tremeGene transfection reagent (Roche Diagnostics) or Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.), as previously described (26). After 24 h of transfection at 37˚C, cells were immediately left untreated (control) or subjected to TGF-β1 treatment. Whole-cell lysates were collected for co-immunoprecipitation and western blot analyses and total RNA was used for RT-qPCR analysis.

MG-63 cells were washed with 1X phosphate-buffered saline and lysed with immunoprecipitation lysis buffer [25 mM Tris (pH 8.0), 1% Nonidet P-40, 1 mM ethylenediaminetetraacetic acid, 150 mM NaCl and protease/phosphatase inhibitors] for 10 min at 4˚C. Subsequently, the whole-cell lysate was centrifuged at 12,000 x g for 10 min at 4˚C. The collected supernatant (1 ml/reaction) was incubated at 4˚C for overnight with 10 µl of antibodies against immunoglobulin G (IgG; cat. no. Sc-2025) or Runx2 (cat. no. sc-390715) purchased from Santa Cruz Biotechnology, Inc. The immune complex was pulled down using protein A/G magnetic beads (Bio-Rad Laboratories, Inc.), with magnetic stacker according to the manufacturer's instructions. 1X Phosphate-buffered saline (HiMedia Laboratories Pvt. Ltd.) was used for washing. Eluted proteins were analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (17,41).

Protein samples were extracted using 1X radioimmunoprecipitation assay buffer (Bio Basic, Inc.) with protease and phosphatase inhibitors (MedChemExpress). Protein concentration was determined by the Bradford assay. Protein (50 µg) was loaded into each lane and separated using SDS-PAGE (8% gel) before being transferred onto polyvinylidene difluoride membranes. Membranes were then blocked with 5% (w/v) bovine serum albumin (Sisco Research Laboratories Pvt Ltd.) for 1 h at room temperature and washed with Tris-buffered saline containing 0.1% Tween 20. The membranes were incubated with primary antibodies (1:1,000) against CBP, MMP-13, acetylated-lysine, or Runx2 overnight at 4˚C. α-Tubulin was used as an endogenous control. The membranes were then incubated with horseradish peroxidase-conjugated secondary antibody (1:2,000) 1 h at room temperature, and immunoreactive bands were visualized using an Enhanced Chemiluminescence Substrate (Takara Bio, Inc.). Image Lab 6.1 (Bio-Rad Laboratories, Inc.) was used to quantify and observe band intensities (12).

A dual-luciferase gene reporter assay was performed as previously described (42,43). The forward and reverse primers containing the wild-type (W) or mutant (M) miRNA response elements (MREs) of the 3'-UTR of CBP were synthesized by Eurofins Genomics LLC (Table II) and cloned into an expression vector pmirGLO (Promega, Madison, WI, USA). Negative control miRNA (nc-miRNA) or miR-4327 mimics were transiently co-transfected into MG-63 cells along with the W or M constructs of CBP 3'-UTR using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). After 24 h of transfection the lysates were collected and luciferase assay was performed using DLR™ Assay System (Promega Corp.). The data were normalized using Renilla luciferase activity. The ratio of Firefly luciferase activities to Renilla luciferase was calculated to determine the relative luciferase activity.

All experiments were carried out using biological triplicate and subjected to one-way analysis of variance (ANOVA) using Statistics Kingdom (https://www.statskingdom.com/180Anova1way.html) to verify statistical significance. Tukey's post hoc analysis was conducted to confirm the significance. P≤0.05 was considered to indicate a statistically significant difference.

Western blot analysis was used primarily to check the expression patterns of CBP and MMP13 under TGF-β1 treatment in human primary osteoblasts cells. Results showed a significant upregulation of CBP levels at 1, 2 and 8 h following TGF-β1 treatment compared with the control group, with the maximum expression at 2 h (Fig. 1A and B). Tukey's post hoc analysis indicated that the 2 h TGF-β1 treatment group showed a statistically significant difference in CBP expression compared with the control group. There was also an upregulation of CBP at 4 h of TGF-β1 treatment, but it was not significant. At 2, 4, 8 and 24 h after TGF-β1 treatment, the level of MMP-13 increased significantly. The highest level of MMP-13 expression was seen at 8 h after TGF-β1 treatment in HS-5 cells (Fig. 1A and C).

Since TGF-β1 treatment increased the expression of CBP and MMP-13 in human primary osteoblasts, the functional role of CBP was assessed in TGF-β1-stimulated MMP-13 expression in osteoblastic cells. The results revealed that CBP knockdown decreased both CBP and MMP-13 protein levels when compared with scrambled control (Fig. 2).

Based on the results from Figs. 1 and 2, CBP is necessary for TGF-β1 to stimulate MMP-13 expression in human osteoblasts; it was next investigated whether miRNAs play a role in regulating CBP expression induced by TGF-β1. Through in silico analysis, miRNAs that potentially target the 3'-UTR of CBP were obtained. A total of 77 miRNAs were identified, from which 17 were unvalidated. After scrutinizing the LogitProb score (>0.75) and miRDB score (>50), eight unvalidated miRNAs were shortlisted for further studies (Fig. 3).

After shortlisting eight unique miRNAs that were predicted to target the 3'-UTR of CBP, their presence and expression patterns at the precursor level were analyzed in control or TGF-β1-treated MG-63 cells. Under TGF-β1 stimulation, miR-3924 was significantly downregulated at all time points (Fig. 4A); mir-3133 was significantly downregulated at 24 h (Fig. 4B); miR-4327 and miR-4264 were significantly downregulated at 1, 2 and 4 h (Fig. 4C and D); miR-1185-3p (the accession no. of the sequence used to design this primer: NR_031575) was significantly upregulated at 1, 2, 8 and 24 h (Fig. 4E); miR-6083 was significantly downregulated at 1, 2, 4 and 8 h (Fig. 4G); and miR-7-1-3p and miR-600 did not show any significant upregulation or downregulation (Fig. 4E and F). Tukey's post hoc analysis showed that the groups treated with TGF-β1 for 1, 2, 4 and 24 h were statistically significant compared with the control with regard to the expression profile of mir-6083. Also, the 1- and 24-h TGF-β1-treated groups were significantly different compared with the control with regard to mir-1185-3p. Analyses of the LogitProb and miRDB scores (Fig. 3), along with expression patterns assessed by RT-qPCR (Fig. 4C), identified miR-4327 as having the most favorable characteristics for targeting CBP.

Next, the mature expression pattern of miR-4327 was analyzed in MG-63 cells. A similar pattern of significant downregulation of mature miR-4327 expression at 2, 4 and 8 h after TGF-β1 treatment in MG-63 cells was observed (Fig. 5A). Tukey's post hoc analysis revealed a statistical significance in miR-4327 expression at 2 and 8 h of TGF-β1 treatment compared with the control in MG-63 cells. In addition, miR-4327 expression was significantly downregulated at 1, 2 and 4 h after TGF-β1 treatment in HS-5 cells (Fig. 5B). Although a significant upregulation of miR-4327 expression was observed after 8 and 24 h of TGF-β1 treatment, the expression patterns of mature miR-4327 at early time points of TGF-β1 treatment were consistent in both MG-63 and HS-5 cells (Fig. 5). These findings supported an inverse correlation, as observed in the case of CBP expression (Fig. 1).

As miR-4327 expression was downregulated by TGF-β1 stimulation, the present study aimed to analyze its functional role using miRNA overexpression studies. The negative control (nc)-miRNA (5'-UCACCGGGUGUAAAUCAGCUUG-3') or miR-4327 mimic (5'-GGCUUGCAUGGGGGACUGG-3') were transiently transfected into MG-63 cells and they were treated with TGF-β1 for 24 h or left untreated (control). RT-qPCR analysis showed that miR-4327 overexpression caused a substantial elevation of its endogenous expression under TGF-β1 treatment or control conditions (Fig. 6A). To determine the association between CBP and Runx2 and the effect of HAT activity on Runx2 expression, whole-cell lysates were collected after 2 h of TGF-β1 treatment and subjected to coimmunoprecipitation using IgG or Runx2 antibody, followed by immunoblotting using antibodies against Runx2, acetylated-lysine, or CBP (Fig. 6). In the miR-4327 mimic-transfected group, the levels of acetylated Runx2 (Fig. 6B and C), Runx2 (Fig. 6B and D) and CBP (Fig. 6B and E) were significantly downregulated compared with those in the nc-miRNA group. Overexpression of miR-4327 significantly downregulated the expression of CBP in human osteoblasts (Fig. 6F and G). Furthermore, western blot analysis for the aliquots of the aforementioned whole-cell lysates was performed. In the nc-miRNA group, TGF-β1 treatment substantially increased MMP-13 expression, whereas miR-4327 mimic-transfected cells significantly reduced MMP-13 expression (Fig. 6H and I). These results indicated that the interaction between CBP and Runx2 and the acetylation of Runx2 could be due to the HAT activity of CBP in human osteoblasts (Fig. 6B and C). CBP expression decreased by miR-4327 overexpression, which might have altered Runx2 stability for MMP13 expression in these cells (Fig. 6B-E).

As miR-4327 overexpression decreased the extent of CBP-mediated Runx2 acetylation and Runx2 expression, which in turn reduced MMP-13 expression under TGF-β1 treatment, the present study further examined if miR-4327 directly interacted with the 3'-UTR of CBP, using a dual-luciferase reporter assay system as previously described (42,43). In silico analyses identified two distinct MREs in the 3'-UTR of CBP for miR-4327 (Fig. 7A and B). A substantial decrease in luciferase activity was noticed in the samples transfected with the wild (W) CBP 3'-UTR MRE (sites 1 and 2) constructs and miR-4327 mimic, whereas no significant changes were noticed in the samples transfected with the mutant (M) CBP 3'-UTR MREs (sites 1 and 2) constructs and miR-4327 mimic or nc-miRNA (Fig. 7C and D). These results indicated the direct targeting of CBP by miR-4327 in human osteoblasts.