microRNA-214-3p Suppresses Ankylosing Spondylitis Fibroblast Osteogenesis via BMP–TGFβ Axis and BMP2

Recent investigations suggest microRNAs (miRs) exert functions in fibroblast osteogenesis in ankylosing spondylitis (AS), an inflammatory rheumatic disease. But the mechanism of miR-214-3p in osteogenic differentiation in AS is not clearly understood yet. In this study, fibroblasts were obtained from the capsular ligament of patients with AS and femoral neck fracture and cultured for osteogenic induction and identified. The roles of miR-214-3p and bone morphogenic protein 2 (BMP2) in AS fibroblast osteogenesis were assessed via gain- and loss-of-function, alizarin red S staining, and alkaline phosphatase (ALP) detection. Levels of miR-214-3p, BMP2, osteogenic differentiation-related proteins, and BMP–TGFβ axis-related proteins were further measured. Consequently, miR-214-3p was downregulated in AS fibroblasts, with enhanced ALP activity and calcium nodules, which were reversed by miR-214-3p overexpression. BMP2 was a target gene of miR-214-3p and promoted AS fibroblast osteogenesis by activating BMP–TGFβ axis, while miR-214-3p inhibited AS fibroblast osteogenesis by targeting BMP2. Together, miR-214-3p could prevent AS fibroblast osteogenic differentiation by targeting BMP2 and blocking BMP–TGFβ axis. This study may offer a novel insight for AS treatment.


INTRODUCTION
As a highly heritable and prototypic spondyloarthropathy, ankylosing spondylitis (AS) is closely associated with inflammatory response, progressive rigidity, and peripheral arthritis and mainly affects the spine and pelvis (1,2). AS begins at the early stage of life and always results in physical dysfunction and reduced health-related quality of life (3). With AS progression, chronic inflammation and new bone formation persist in all segments of the spine, and the calcification of vertebral body and paravertebral ligament will aggravate (4). New bone formation is based on increased differentiation of osteoblasts (5). AS patients are at a high risk of spinal fracture, spinal cord injury, osteoporosis, hypertension, cardiovascular diseases, pulmonary complication, and metabolic syndrome (6,7). As is reported, current anti-rheumatic drug therapies improve back pain, peripheral arthritis acute phase responses, disturbed sleep, and overall quality of life; however, the major contributing factor of AS-new bone formation-is not affected (5). In light of this, the search for novel targets for AS treatment should focus on osteogenic differentiation to prevent new bone formation. microRNAs (miRs) are small non-coding RNA molecules, involved in many biological processes with interaction with their mRNAs (8). Aberrant miR expression is related to the pathogenesis of rheumatoid arthritis and osteoarthritis (9). A recent study points out that miR-214 is consistently decreased in AS and potentially serves as a non-invasive biomarker for AS diagnosis (10). Importantly, miR-214-3p is involved in osteogenesis of maxillary sinus membrane stem cells (MSMSCs) (11). In this study, we found miR-214-3p could target bone morphogenetic protein 2 (BMP2). BMPs, a subset of the transforming growth factor b (TGFb) superfamily, exert their roles in bone formation via BMP receptors (BMPRs), which in turn activate Smad1/5/8 axis and facilitate osteoblastic gene transcription (12). In particular, high localized expression of BMP-2 at the site of ankylosing enthesitis is involved in ankylosis development in a mouse model of spondyloarthritis (13). Importantly, Shea Carter et al. put forward the notion that activation of BMP pathways may be a very early or even initial event in AS (14). A former research has demonstrated that TGFb1 is involved in AS development and is elevated upon complement attack in osteoblasts and osteoclasts (15). TGFb1 and BMP2 have significant effects on adhesion, proliferation, and differentiation of primary human osteoblasts (16). It is recognized that the TGFb-BMP axis is essential for osteogenesis and bone formation in mammalian development (17). From all of the above, it is reasonable to hypothesize that there may be an interaction among miR-214-3p, BMP2, and TGFb-BMP signaling pathway in AS. Thus, a series of experiments were performed in this study to justify the hypothesis.

Cell Culture and Osteogenic Induction
Twenty patients with AS involving both hips and requiring joint replacement from January 2017 to January 2019 were selected as the experimental group (16 males and four females), with an average age of 32.8 ± 1.8 years. These patients met the modified New York criteria (1984) for AS (Raychaudhuri and Deodhar, 2014), free of other immune diseases such as wind dampness. Eighteen patients with femoral neck fracture (free of AS and other immune diseases) who needed open surgery or joint replacement were selected as the control group (12 males and six females), with an average age of 40.7 ± 2.4 years. The specimen was the tissue of the hip joint capsule, which was the waste during the surgery. The tissue of the cut hip joint capsule ligament was collected as the specimen. The ligaments were cut into 0.5 mm 3 pieces and washed twice with phosphate buffered saline (PBS). Then the ligament pieces were preserved in Dulbecco's modified Eagle medium (DMEM) (Thermo fisher, Carlsbad, California, USA) containing 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin in a 5% CO 2 incubator at 37°C. After cell adhesion, the ligament pieces were discarded, and the culture medium was changed every 3 days. Cells were separated at 1:3, and cells in the third passage were used for further study. When fibroblast confluence reached 80%, osteogenic differentiation was induced by adding 0.1 ml/L dexamethasone + 10 mmol/L b-glycerophosphate + 50 ml/L ascorbic acid (18).

HEMATOXYLIN AND EOSIN STAINING
The third passage cells were washed with precooled PBS for 10 min, and fixed with 4% paraformaldehyde for 30 min. According to the standard scheme, cells were stained with HE (Beijing Solarbio Science & Technology Co., Ltd, Beijing, China), and observed under the optical inverted microscope (Leica, Solms, Germany).

Immunocytochemistry
Cleaned, acidic, and sterilized cover slides were placed in a cell culture dish, added with cell suspension and cultured for 24-48 h. When the cover slides were paved with cells, the culture medium was removed, and the slides were washed in PBS, fixed with 4% formaldehyde for 15 min, and incubated with 0.3% Triton X-100 for 20 min. After PBS washing, cells were blocked for 30 min with 1% bovine serum albumin (BSA), and incubated for 1 h with primary antibody anti-Vimentin (1: 100, ab8978, Abcam, Cambridge, MA, USA) at 37°C. Subsequently, cells were incubated for 30 min with goat anti-mouse IgG H&L (HRP) (1:2000, ab205719, Abcam). Finally, cells were conventionally stained with 2, 4-diaminobutyric acid (DAB) and counterstained in hematoxylin. The positive expression of vimentin in fibroblasts was visualized under a light microscope at 40× magnification.

Reverse Transcription Quantitative Polymerase Chain Reaction
The one-step method of TRIzol (Invitrogen, Carlsbad, CA, USA) was employed to extract total RNA, and the extracted highquality RNA was confirmed using ultraviolet analysis and formaldehyde denaturation electrophoresis. RT-qPCR was conducted based on instructions of RT-qPCR kit (Thermo Fisher Scientific, Shanghai, China) with glyceraldehyde-3phosphate dehydrogenase (GAPDH) as an internal reference for BMP2 and U6 for miR-214-3p. PCR primers were provided by Shanghai Sangon Biotechnology Co., Ltd. (Shanghai, China) ( Table 1). The amplification and dissolution curves were confirmed after the reaction, and data were analyzed by 2 -DDCt method (19).

Dual-Luciferase Reporter Gene Assay
The database (http://starbase.sysu.edu.cn/) (20) was used to predict the target binding site of miR-214-3p with BMP2. BMP2 fragment containing the binding site of miR-214-3p was cloned into the pmirGLO oligosaccharide enzyme vector (Promega, Madison, WI, USA), and the pmirGLO-BMP2-wild type (Wt) reporting vector was constructed. The pmirGLO-BMP2-mutant type (Mut) was constructed with the mutant binding site of miR-214-3p based on pmirGLO-BMP2-Wt. After that, the constructed vectors were transfected into fibroblasts and then co-transfected with miR-214-3p mimic and miR-NC respectively. After 48 h, the luciferase activity was assessed using dual luciferase reporter gene assay system (Promega, Madison, WI, USA), and the relative activity was calculated as the ratio of firefly luciferase activity to renilla luciferase activity.

Western Blot Analysis
The proteins were extracted to determine the concentration as per the instructions of bicinchoninic acid kit (Thermo Scientific Pierce, Rockford, IL, USA). The extracted proteins were boiled and run on sodium dodecyl sulfate polyacrylamide gel electrophoresis from 80 V to 120 V. Afterward, the proteins were transformed into the polyvinylidene fluoride membranes (Millipore Corp., Billerica, MA, USA Statistical Analysis SPSS 21.0 (IBM Corp., Armonk, NY, USA) was employed for data analysis. Kolmogorov-Smirnov test showed whether the data were in normal distribution. The measurement data were exhibited as mean ± standard deviation. Comparisons between two groups were analyzed with t test; among multiple groups were assessed with one-way analysis of variance (ANOVA) or two-way ANOVA, and pairwise comparisons after ANOVA were conducted by Tukey's multiple comparisons test. p value was obtained by two-tailed test and p < 0.05 inferred a statistical difference.

Isolation and Identification of Fibroblasts
Fibroblasts obtained from patients with AS and femoral neck fracture were observed and identified with HE staining ( Figure  1A). Immunocytochemistry of cells was performed with vimentin, a specific marker of fibroblasts ( Figure 1B). There was no significant difference in morphology between the AS fibroblasts and normal fibroblasts. Cells were in long spindles with oval and regular nuclei. AS fibroblasts and normal fibroblasts were positive for vimentin, confirming that all the cultured cells were fibroblasts from mesoderm.

miR-214-3p Inhibits AS Fibroblast Osteogenesis
miR is a major regulator in many diseases. miR-214 has been reported to be lowly expressed in AS patients and used as a noninvasive biomarker for AS diagnosis (10). Through RT-qPCR, we found that miR-214 expression was lower in AS fibroblasts than in normal fibroblasts before osteogenic induction and further decreased in AS fibroblasts after osteogenic induction ( Figure  2A). Subsequently, miR-214-3p mimic and miR-214-3p inhibitors were transfected into fibroblasts after induction ( Figure 2B). ALP activity was measured and the formation of calcium nodules was detected with ARS staining 21 days later. ALP activity of AS fibroblasts was detected at 7 th , 14 th , and 21 st days after transfection. It was found that ALP activity of AS fibroblasts increased with time after osteogenic induction; miR-214-3p mimic could significantly inhibit ALP activity, while miR-214-3p inhibitor increased ALP activity of AS fibroblasts ( Figure 2C). As was shown in Figure 2D, compared with normal fibroblasts, calcified nodule in AS fibroblasts was markedly increased. miR-214-3p mimic inhibited the formation of calcium nodules in AS fibroblasts, while AS fibroblasts with miR-214-3p inhibitor produced abundant calcium nodules. Additionally, levels of COLI and Runx2 related to osteogenic differentiation were detected in AS fibroblasts ( Figure 2E). As expected, COLI and Runx2 were increased in AS fibroblasts, but decreased noticeably in AS fibroblasts with miR-214-3p mimic, and reached highest levels in AS fibroblasts with miR-214-3p inhibitor. Runx2 expression was almost not detected in normal fibroblasts. Therefore, we concluded that miR-214-3p could prevent AS fibroblast osteogenesis.

miR-214-3p Targets BMP2
BMP2 is one of the proteins expressed extensively during the development of AS and plays an important role in osteogenesis (21). Through bioinformatics analysis and database screening, we found that there were targeted binding sites between miR-214-3p and BMP2, and the complementary sequence of 3′UTR is shown in Figure 3A. In order to confirm the targeting relationship between miR-214-3p and BMP2, we constructed a dual luciferase reporter gene vector and confirmed the targeting relationship between miR-214-3p and BMP2 through luciferase activity ( Figure 3B). Next, BMP2 levels were measured using RT-qPCR and western blot analysis. The obtained results indicated that BMP2 levels in fibroblasts transfected with miR-214-3p mimic were substantially decreased, which could be  reversed by miR-214-3p inhibitor ( Figures 3C, D). This suggested that BMP2 is indeed the downstream target of miR-214-3p in AS fibroblast osteogenesis and is negatively regulated by miR-214-3p.

BMP2 Activates the BMP-TGFb Axis to Promote Osteogenic Differentiation of AS Fibroblasts
Considering BMP2 is indeed the downstream target of miR-214-3p, now we turned to investigate the mechanism of BMP2 in osteogenic differentiation of AS fibroblasts and its potential downstream signaling pathway. Through RT-qPCR and Western blot analysis, we found BMP2 was significantly expressed in AS fibroblasts compared with normal fibroblasts after osteogenic induction. We further transfected the plasmid overexpressing BMP2 or silenced BMP2 into AS fibroblasts ( Figure 4A). ALP detection ( Figure 4B) and ARS staining ( Figure 4C) showed that excessive BMP2 expression promoted AS fibroblast osteogenesis, while silenced BMP2 inhibited AS fibroblast osteogenesis. BMP-TGFb is involved in regulating the osteogenic differentiation of AS fibroblasts (22). Through the KEGG signaling pathway, BMPR2 and Smad5 are downstream pathway proteins of BMP2 ( Figure 4D). Western blot analysis found that BMPR2, p-Smad5, Smad5, p-Smad5/Smad5, and  OCN were significantly increased in AS fibroblasts with excessive BMP2 expression, while silenced BMP2 declined the pathwayrelated factors above ( Figure 4E), indicating that BMP2 promoted the osteogenic differentiation of AS fibroblasts by activating the BMP-TGFb axis.

miR-214-3p Inhibits BMP2 Expression to Further Repress AS Fibroblast Osteogenesis
To further confirm the mechanism of miR-214-3p in osteogenic differentiation of AS fibroblasts, miR-214-3p mimic or miR-214-3p inhibitor was transfected into AS fibroblasts with overexpressing BMP2. The results exhibited that the overexpressed BMP2 was inhibited by miR-214-3p mimic, but the highest BMP2 expression was achieved by miR-214-3p inhibitor ( Figure 5A). Similarly, we recorded ALP activity ( Figure 5B) and ARS staining ( Figure 5C) within 21 days. It indicated that miR-214-3p inhibited BMP2 expression and reduced ALP activity and calcium nodule formation. Western blot analysis also found that levels of BMPR2, p-Smad5, Smad5, p-Smad5/Smad5, and OCN were decreased after BMP2 inhibition ( Figure 5D), suggesting miR-214-3p prevented osteogenic differentiation of AS fibroblasts by targeting BMP2 and inhibiting the BMP-TGFb axis.

DISCUSSION
Although infliximab (a monoclonal antibody against TNF-a) has been commonly shown to be effective in AS treatment, long-term follow-up studies demonstrated the association between infliximab therapy and increased risk of non-Hodgkin's lymphoma (23). Surprisingly, Md Shaifur Rahman et al. highlighted that BMP plays major roles in regulating osteogenesis and bone formation and implied the clinical applications of TGFb-BMP axis for bone diseases (24). It is commonly recognized that miRs have critical roles in regulating TGFb-BMP and osteoblast differentiation (25). Based on this, we hypothesized prior to the experiments that miR-214-3p, BMP2, and the TGFb/BMP axis may be interacting in the osteogenic differentiation of AS fibroblasts. Collectively, we claimed that miR-214-3p could inhibit osteogenic differentiation of AS fibroblasts by targeting BMP2 and blocking BMP-TGFb axis.
The first important observation in this study was that miR-214 expression was lower in AS fibroblasts and further decreased after The targeting relationship between miR-214-3p and BMP2 confirmed by dual luciferase reporter gene assay, relative to the NC group, ***p < 0.001; (C) mRNA expression of BMP2 measured by RT-qPCR; (D) Protein level of BMP2 measured by western blot analysis. Relative to the normal group, ***p < 0.001. ns represents no statistical differences; relative to the AS+miR-NC group, ### p < 0.001. osteogenic induction, and miR-214-3p overexpression inhibited ALP activity, calcified nodules, COLI, and Runx2 levels. Consistently, miR-214-3p was downregulated in osteogenic differentiation of MSMSCs, and miR-214-3p overexpression substantially diminished levels of osteogenic markers Runx2 and ALP in osteogenic differentiated MSMSCs (11). As recently reported, upregulated osteoclastic miR-214-3p led to inhibited osteoblast activity and bone formation in aged women with fractures (26). Runx2 is a major regulator of osteogenic differentiation by regulating key osteogenic genes, and Runx2 knockout in mice results in complete depletion of bone formation (27). Runx2 and ALP are early markers of osteogenesis; OCN is a late osteogenic marker, and calcified nodules are signs of the final stage of osteogenesis (28,29). Kihara T previously indicated that exogenous COLI increased calcified nodules, osteogenesis, and matrix mineralization of mesenchymal stem cells (30). Importantly, enhanced expression of ALP and Runx2 induced osteogenic differentiation, which stimulated AS progression (31). Interestingly, miR-214 level was positively associated with the number of white blood cells, platelets, C-reactive protein (CRP), and AS Disease Activity Score with CRP (10). In summary, miR-214-3p could prevent osteogenic differentiation of AS fibroblasts.
Furthermore, the data confirmed that BMP2 is a target gene of miR-214-3p. As Ramazzotti G et al. demonstrated, overexpressed miR-214 inhibited osteogenic differentiation of C2C12 cells by suppressing phosphoinositide-phospholipase Cb1, which was crucial to promote BMP-2-induced osteogenesis (32). BMPs could exert potent function in regulating osteoblastic stem cells, among which recombinant human BMP2 is applicable in clinical adjuvant therapy for bone formation (33). Besides, we also found  and speed up new bone formation in vivo (16). Strikingly, evidence revealed that TGFb could induce fibroblast synthesis of COLI, OCN, and ALP, block extracellular matrix digestion, promote the amount of bone and cartilage tissues and bone formation in AS (22). Similarly, miR-214-5p overexpression diminished levels of ALP, Runx2, OCN and COLI, and TGFb/Smad2 in bone marrow stem cells of postmenopausal osteoporosis (18). Taken together, the inhibitory effects of miR-214-3p on osteogenic differentiation of AS fibroblasts were achieved via BMP2 and the BMP-TGFb axis. However, the change of Smad5 protein itself may be caused by other target genes downstream of miR-214-3p, which will be further studied in the future.
To sum up, our study supported the notion that miR-214-3p could inhibit osteogenic differentiation of AS fibroblasts by targeting BMP2 and blocking the BMP-TGFb axis, highlighting promising novel approaches for AS treatment. There are many studies on osteoblast differentiation in normal medium (35,36). On the basis of these studies, we will carry out relevant research on whether miR-214-3p can promote osteogenesis in normal medium in the future. More studies are still needed to further validate our results and deeply analyze the molecular mechanism of AS so as to find out applicable methods for AS patients in the clinic.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Clinical Ethical Committee of Beijing Shijitan Hospital. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Clinical Ethical Committee of Beijing Shijitan Hospital.

AUTHOR CONTRIBUTIONS
YY contributed to the study design and manuscript preparation. LD contributed to the definition of intellectual content and manuscript editing. YH is the guarantor of integrity of the entire study. HJ took charge of the literature research. JZ contributed to experimental studies. ZD contributed to the acquisition and analysis of data. GZ contributed to the study concepts. All authors contributed to the article and approved the submitted version.