Sox9-Increased miR-322-5p Facilitates BMP2-Induced Chondrogenic Differentiation by Targeting Smad7 in Mesenchymal Stem Cells

Bone morphogenetic protein 2 (BMP2) induces effective chondrogenesis of mesenchymal stem cells (MSCs) by promoting Sox9 expression. However, BMP2 also induces chondrocyte hypertrophy and endochondral ossification by upregulating Smad7 expression, which leads to the disruption of chondrogenesis. In addition, Smad7 can be inhibited by Sox9. Therefore, the underlying mechanism is not clear. Currently, an increasing number of studies have shown that microRNAs play a pivotal role in chondrogenic and pathophysiological processes of cartilage. The purpose of this study was to determine which microRNA is increased by Sox9 and targets Smad7, thus assisting BMP2 in maintaining stable chondrogenesis. We found that miR-322-5p meets the requirement through next-generation sequencing (NGS) and bioinformatic analysis. The targeting relationship between miR-322-5p and Smad7 was confirmed by dual-luciferase reporter assays, qPCR, and western blotting (WB). The in vitro study indicated that overexpression of miR-322-5p significantly inhibited Smad7 expression, thus causing increased chondrogenic differentiation and decreased hypertrophic differentiation, while silencing of miR-322-5p led to the opposite results. Flow cytometry (FCM) analysis indicated that overexpression of miR-322-5p significantly decreased the rate of early apoptosis in BMP2-stimulated MSCs, while silencing of miR-322-5p increased the rate. A mouse limb explant assay revealed that the expression of miR-322-5p was negatively correlated with the length of the BMP2-stimulated hypertrophic zone of the growth plate. An in vivo study also confirmed that miR-322-5p assisted BMP2 in chondrogenic differentiation. Taken together, our results suggested that Sox9-increased miR-322-5p expression can promote BMP2-induced chondrogenesis by targeting Smad7, which can be exploited for effective tissue engineering of cartilage.


Introduction
Traumatic or degenerative cartilage defects are a challenging clinical issue, as cartilage tissue is devoid of vascular, neural, or lymphatic structures [1]. The above characteristics contribute to the poor self-healing capacity of cartilage. Thus, cartilage needs to be reestablished once injured [2]. Mesenchymal stem cells (MSCs) have been identified as ideal seed cells in cartilage tissue engineering due to their chondrogenic differentiation potential [2][3][4].
Bone morphogenetic protein 2 (BMP2), a member of the transforming growth factor beta (TGF-β) superfamily, is a potent growth factor for the induction of MSC chondrogenic differentiation [5][6][7]. However, BMP2 alone cannot achieve stable chondrogenesis, as it stimulates chondrogenic hypertrophic differentiation and endochondral ossification, which destroy the cartilage phenotype [3,8,9]. Sox9, induced significantly by BMP2, is the key transcription factor that maintains the chondrocyte phenotype and cartilage homeostasis. This molecule plays an important role in the production and protection of the extracellular matrix of articular cartilage [10][11][12][13]. A previous study by our team showed that overexpression of Sox9 enhances BMP-2-induced chondrogenic differentiation in MSCs by downregulating Smad7 expression [14][15][16]. However, the underlying mechanism of how Sox9 regulates Smad7 is not clear.
Increasing evidence indicates that microRNAs (miR-NAs) are crucial for the regulatory network in chondrocyte differentiation and cartilage function [17,18]. miRNAs are a class of noncoding, single-stranded, and small-molecule RNAs that are approximately 18-24 nucleotides in length. They play a crucial role in many biological processes through posttranscriptional negative regulation of target gene expression by sequence-specific binding to the 3 ′ untranslated regions (UTRs) of their target messenger RNAs (mRNAs) [17,19,20]. Therefore, we hypothesized that Sox9 could promote the expression of certain miRNAs that target and inhibit Smad7 expression.
In the present study, we investigated the function of miR-322-5p in BMP2-mediated chondrogenic and hypertrophic differentiation in MSCs. Sox9 was found to increase the expression of miR-322-5p, which targeted Smad7. Our experiments revealed that overexpression of miR-322-5p suppressed BMP2-induced MSC early apoptosis and chondrocyte hypertrophy, thus facilitating BMP2-induced chondrogenic differentiation. These findings help elucidate BMP2-mediated chondrogenic and hypertrophic differentiation, which can be exploited for BMP2-mediated cartilage tissue engineering.

Materials and Methods
2.1. Cell Culture and Chemicals. Mouse bone marrow MSC C3H10T1/2 and human embryonic kidney (HEK) 293 cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States). Cell lines were maintained in complete Dulbecco's modified Eagle's medium (DMEM, BioExplorer, USA) supplemented with 10% fetal bovine serum (FBS, PAN Biotech, Germany), 100 mg/ml streptomycin, and 100 U/ml penicillin at 37°C in a humidified atmosphere with 5% carbon dioxide (CO 2 ). Unless mentioned otherwise, all chemicals were purchased from Thermo-Fisher Scientific or Sigma-Aldrich.

Construction and Generation of Recombinant
Adenoviral Vectors AdBMP2, AdSox9, AdGFP, AdshSox9, and AdRFP. AdEasy technology was used to generate recombinant adenoviruses as previously described [21]. AdBMP2, AdSox9, and AdshSox9 were previously characterized [8,14,16,22]; AdGFP and AdRFP were used as mock virus controls. Briefly, the full-length transcript of mousederived Sox9 and the coding region of human-derived BMP2 were PCR amplified and subcloned into an adenoviral shuttle vector to generate recombinant adenoviral vectors; vectors containing Sox9 or BMP2 were subsequently used to generate recombinant adenoviruses in HEK-293 cells. AdshSox9 was purchased from Vigene (Shandong, China). For monitoring infection efficiency, AdBMP2 and AdSox9 were flagged with green fluorescent protein (GFP), and AdshSox9 was labeled with red fluorescent protein (RFP).

Chondrogenic Differentiation of MSCs in Micromass
Culture. To mimic the condensation of MSCs, we used micromass culture to induce chondrogenic differentiation as previously described [23]. C3H10T1/2 cells were infected with AdGFP, AdRFP, AdBMP2, AdSox9, or AdshSox9. To manipulate miR-322-5p expression, we infected miR-322-5P agomir and antagomir, purchased from GenePharma (Shanghai, China), by siRNA-Mate™ (GenePharma) according to the manufacturer's instructions, using agomir-NC or antagomir-NC as normal controls, respectively. Twentyfour hours after infection, the cells were collected, resuspended at a high density (∼10 5 per 50 μl of DMEM), subsequently seeded at the center of each well in 6-well plates, and then incubated in a CO 2 incubator. Two hours after incubation, 2 ml of complete DMEM was added to each well; half of the medium was replaced every 3 days. 2.6. Dual-Luciferase Reporter Assay. For detection of the interaction between Smad7 and miR-322-5p, plasmids containing mutant-type Smad7-3 ′ UTR (Smad7-3 ′ UTR-MT) and wild-type Smad7-3 ′ UTR (Smad7-3 ′ UTR-WT) were constructed. When HEK-293 cells cultured in 24-well plates reached 80% confluency, 50 nM miR-322-5p agomir or NC was cotransfected into cells with 2 μg plasmids mediated by Lipofectamine 2000 (Invitrogen). Moreover, Renilla luciferase-(RL-) loaded pRL-TK was transfected as an internal control. After 48 h, the Dual-Luciferase Reporter Gene Assay Kit (Beyotime) was used to detect the intensity of RL and firefly luciferase (FL) according to the manufacturer's protocol. Consequently, the ratio of FL to RL reflected the suppressive effect of miR-322-5p on Smad7. Each group had five duplicate wells. . After the membranes were washed with TBST, they were incubated with the corresponding secondary antibodies (goat antirabbit IgG, 1 : 10000, Zen Bio) for 1 h at room temperature. Following sequential washing with TBST and TBS, the target proteins were detected by an ECL detection kit (Thermo Fisher Scientific), and ImageJ software was used for quantification of band density.
2.8. Apoptosis Detection by Flow Cytometry. Cells were treated with trypsin (0.25%) and monitored under a microscope throughout the process of digestion, which was immediately stopped by adding DMEM containing FBS when most of the cells became round. After centrifugation, the cells were repeatedly resuspended in PBS gently and centrifuged twice, followed by resuspension in 500 μl of PBS for immediate detection on CytoFLEX.
2.9. Mouse Fetal Limb Explant Culture. Forelimbs of mouse embryos (E18.5) were dissected of skin and most soft tissue, except periosteum, under sterile conditions and incubated in DMEM containing 0.5% bovine serum albumin (BSA, Sigma), 50 μg/ml ascorbic acid, 1 mM β-glycerophosphate, 100 mg/ml streptomycin, and 100 U/ml penicillin at 37°C in a humidified atmosphere with 5% carbon dioxide (CO 2 ) for up to 14 days as previously described [25,26]. Five samples were cultured in each well. Upon the initiation of incubation, the skin-free limbs were infected by adding AdBMP2, AdSox9 or AdshSox9, miR-322-5P agomir, or antagomir to the culture medium. Half of the medium was changed every second day. For monitoring the survival of the cells in the forelimbs, GFP and RFP signals were observed under a microscope. The tissues were fixed for histological evaluation after 14 days of culturing.
2.10. Subcutaneous MSC Implantation. The animal use and care and experimental procedures were approved by the Chongqing Medical University Animal Care and Use Committee. The subcutaneous stem cell implantation procedure was carried out as described [3,14]. Briefly, C3H10T1/2 cells were infected with AdBMP2, AdGFP, AdRFP, AdSox9, AdshSox9, miR-322-5P agomir, and antagomir. Twentyfour hours after transfection, the cells were collected and resuspended in PBS-diluted Matrigel (Corning) for subcutaneous injection into the flanks of athymic nude mice (4 weeks old, female, n = 3/group, 4 * 10 6 cells per injection). Four weeks after injection, the animals were sacrificed for collection of ectopic masses. Following fixation in 4% paraformaldehyde (Servicebio, Wuhan, China) for 24 h at room temperature, the masses were subjected to ethylenediaminetetraacetic acid (EDTA) for decalcification at 4°C for 14 days, followed by embedding in paraffin. Serial 5 μm thick sections were processed for special staining and histological evaluations.
2.11. Histological Evaluation: Hematoxylin and Eosin (H&E), Alcian Blue Staining, and Masson's Trichrome. After sequential deparaffinization with xylene and rehydration with graded ethanol, H&E, Masson's trichrome, and Alcian Blue staining was performed using a standard protocol as described previously [8,14,16]. Briefly, the deparaffinized samples were first subjected to antigen retrieval and fixation, followed by H&E and Masson's trichrome staining. A light microscope (Olympus, Japan) was used for histological evaluation.
2.12. Immunohistochemistry Assay. After sequential deparaffinization with xylene and rehydration with graded ethanol, sections were boiled in 10 mM citrate buffer at 95-100°C for 10 min for antigen retrieval, rinsed in 3% H 2 O 2 at room temperature for 10 min to inhibit endogenous peroxidase activity, and blocked with 10% goat serum at room temperature for 10 min. Then, the sections were incubated with primary antibodies against collagen 2α1 (COL2A1) (Abcam, 1 : 400) and collagen 10α1 (COL10A1) (Abcam, 1 : 200) at 4°C overnight. After washing, the sections were incubated with secondary antibody at 37°C for 30 min, followed by incubation with streptavidin-HRP conjugate for 20 min at room temperature. Staining without primary antibody was used as a negative control. A microscope (Olympus, Japan) was used for imaging.  CCAACAGCGTCCTGGAGT  COL2A1  CAACACAATCCATTGCGAAC  TCTGCCCAGTTCAGGTCTCT  Sox9  AGCTCACCAGACCCTGAGAA  TCCCAGCAATCGTTACCTTC  COL10A1  TGCTGCCCTGGTCTTACTCT  GCCTTGGGATCCTAAACCT  GAPDH CTACACTGAGGACCAGGTTGTCT TTGTCATACCAGGAAATGAGCTT 3 Stem Cells International 2.13. Statistical Analysis. All experiments were performed at least three times independently. Data are expressed as the mean ± standard deviation (SD) and were analyzed with SPSS software (Version 21, IBM). Statistical analyses were conducted using one-way analysis of variance and Student's t-test; p < 0:05 was considered statistically significant.

NGS and Bioinformatic Analyses Demonstrated That
Silencing of Sox9 Decreased the Expression of miR-322-5p, Which Was Predicted to Target Smad7. The volcano plot (Figure 2(a)) showed that 41 and 57 miRNAs had downregulated and upregulated expression, respectively, in the BMP2+shSox9 group. After prediction for miRNAs targeting Smad7 in the databases miRanda, StarBase, and TargetScan, Venn diagram analysis revealed that only two miRNAs among the miRNAs with downregulated expression in the BMP2+shSox9 group appeared in all three databases (Figure 2(b)). Consequently, miR-322-5p, but not miR-181-5p, was the focus of the following experiments. Next, the scatter plot showed that the correlation coefficient was 0.9, which suggested a reliable trend between the groups (Figure 2(c)).   Stem Cells International performed on C3H10T1/2 cells infected with AdBMP2+-AdSox9, AdBMP2, and AdBMP2+AdshSox9. Subsequently, the expression of miR-322-5p was detected by qPCR, which suggested an increased response by overexpressing Sox9 and a downregulated response by silencing Sox9 (Figure 3(a)). According to the prediction on TargetScan, miR-322-5p matched the position 69-76 of the Smad7 3′UTR The correlation coefficient between groups was 0.9, and miR-322-5p was marked. FC: fold change.

Smad7-Induced Early Apoptosis Was Negatively
Correlated with the Expression Level of miR-322-5p. To determine the effect of miR-322-5p on Smad7-related early apoptosis, we performed flow cytometry (FCM). As indicated by the results, overexpression of Sox9 obviously decreased the early apoptosis rate (Figures 4(b) and 4(c)), which was partly restored by forced expression of miR-322-5p antagomir (Figures 4(a) and 4(b)); moreover, silencing of Sox9 remarkably upregulated the early apoptosis rate (Figures 4(c) and 4(d)), which was partially reversed by use of the miR-322-5p agomir (Figures 4(d) and 4(e)). The cartogram shows the early apoptosis rate in each group (Figure 4(f)).

BMP2-Induced Chondrocyte Hypertrophy in Fetal Mouse
Forelimb Explants Was Inhibited by miR-322-5p. After culture for 14 days, fetal mouse forelimbs were subjected to sectioning and H&E staining to evaluate the length of the hypertrophic zone. The results demonstrated that overexpressing Sox9 reduced the BMP2-induced hypertrophic zone, which was partly reversed by forced expression of the miR-322-5p antagomir (Figures 5(a)-5(c)); moreover, silencing of Sox9 extended the hypertrophic zone, which was partly reversed by the use of the miR-322-5p agomir (Figures 5(c)-5(e)). The arrows in dark blue and yellow represent the lengths of the prehypertrophic and hypertrophic zones (HZs), respectively ( Figure 5(f)). The cartogram shows the length of the HZ in each group ( Figure 5(g)).
Furthermore, immunohistochemistry showed that overexpression of Sox9 increased the synthesis of COL2A1 and decreased the generation of COL10A1, and both changes were partly reversed by use of the miR-322-5p antagomir (Figures 7(f), 7(g) i-iii, 7(h), and 7(i)); moreover, silencing of Sox9 decreased the production of COL2A1 and increased the formation of COL10A1, and both changes were partly reversed by use of miR-322-5p agomir (Figures 7(f), 7(g) iii-v, 7(h), and 7(i)).

Discussion
Articular cartilage defects caused by trauma or degeneration are increasing yearly with the development of society and changes in people's living habits. However, the repair of defective cartilage is still a challenging issue worldwide. BMP2-induced chondrogenesis of MSCs has been widely accepted as a chondrogenic model in tissue engineering. However, BMP2 induces not only chondrogenesis but also endochondral ossification [8,14]. Previous studies by our team have shown that overexpression of Sox9 promotes BMP2-induced chondrogenesis and may function through the suppressive effect of Sox9 on Smad7 [14,16]. The current study clarified that silencing Sox9 weakened BMP2-induced chondrogenesis by upregulating Smad7 expression. The mechanism may be that Sox9 induces miR-322-5p, which binds to the 3′UTR of Smad7 and inhibits its function.
Sox9 is highly increased by BMP2, and it facilitates BMP2-induced chondrogenic differentiation [11,16,28]. This molecule directly regulates the production of COL2A1, which is chondrocyte specific [29,30]. Smad7 is an inhibitor in the Smad family; it plays an inhibitory role in the TGFβ/BMP pathway, which is crucial in the processes of chondrogenesis [31][32][33]. Previous studies have found that BMP2-induced high expression of Smad7 serves as the key inhibitor of chondrogenic differentiation [32,34]. In addition, the hypertrophic differentiation of chondrocytes, marked with COL10A1 [15,35], was caused by Smad7, at least in part, during chondrogenic differentiation [15]. Further research revealed that Smad7 inhibited the formation of cartilaginous tissue induced by BMP2 by suppressing the P38 and Smad1/5/8 pathways [32,36].
Increasing research has revealed the important roles of miRNAs in the processes of chondrocyte formation and the pathophysiological function of cartilage. In this research, NGS and bioinformatic analysis found that miR-322-5p expression was downregulated by silencing Sox9, and the results were further clarified by qPCR. In addition, overexpression of Sox9 upregulated miR-322-5p expression. These results confirmed that Sox9 upregulates miR-322-5p expression. Moreover, the dual-luciferase reporter assay confirmed the targeting relationship between miR-322-5p and Smad7. Subsequently, in vitro experiments also showed that miR-322-5p could indeed inhibit the expression of Smad7 at the transcriptional and translational levels.
Currently, there are only a few reports on the biological functions of miR-322-5p. A previous report showed that miR-322-5p was involved in cardiac hypertrophy in rats with pulmonary hypertension by targeting IGF-1 [37]. In addition, miR-322-5p is involved in FAM3B-mediated hyperglycemic vascular smooth muscle proliferation and migration [38]. RNA sequencing based on cartilage-derived progenitor and stem cells identified miR-322-5p as one of the core regulatory molecules during the progression of OA [39]. In another RNA sequencing analysis performed by our team, high expression of miR-322-5p was positively correlated with BMP-2-induced chondrogenesis (data not shown). It was reported that overexpression of miR-322-5p activated the TGF-β pathway, which is important in maintaining the homeostasis of articular cartilage [37,38,40]. Given that previous reports and data have shown the correlation of miR-322-5p with cartilage homeostasis and chondrogenesis, this molecule was predicted to be more associated with Smad7 [41]. Thus, when the final two miR-NAs were screened out, we focused on miR-322-5p for subsequent experiments. To the best of our knowledge, this is the first study to reveal the regulatory effect of miR-322-5p on Smad7 and chondrogenesis.

Data Availability
Data will be available on request.