MiR-21 promotes osteogenic transformation in ankylosing spondylitis fibroblasts and modulates bone metabolism in a murine arthritis model potentially involving the MAPK NF-κB pathway

Abstract

Background

Ankylosing spondylitis (AS) is a chronic inflammatory disease characterized by pathological bone formation. Synovial fibroblasts are key effector cells in this process, and their dysregulated osteogenic transformation is a critical event. The role of microRNA‑21 (miR‑21) and the mitogen-activated protein kinase (MAPK)/nuclear factor kappa-B‌ (NF‑κB) pathway in inflammation and bone metabolism is established; however, whether miR‑21 promotes AS fibroblast osteogenesis specifically via this pathway is unclear. This study aimed to determine whether miR-21 promotes osteogenic transformation and bone metabolism in AS synovial fibroblasts via the MAPK/NF‑κB pathway.

Methods

Gain‑ and loss‑of‑function of miR‑21 was achieved in AS synovial fibroblasts by transfection with miR‑21 mimic or inhibitor, followed by evaluation of proliferation, osteogenic differentiation (alkaline phosphatase/alizarin red staining), and MAPK/NF‑κB pathway activity. In a proteoglycan‑induced arthritis (PGIA) mouse model, miR‑21 was inhibited by Antagomir to assess spinal pathology, serum inflammatory cytokines, bone metabolism markers, and MAPK/NF‑κB signaling.

Results

In vitro, miR-21 overexpression significantly promoted fibroblast proliferation, osteogenic differentiation, and activated the MAPK/NF-κB pathway, while its inhibition had opposite effects. In vivo, AntagomiR-21 treatment ameliorated spinal cartilage damage, reduced serum levels of inflammatory cytokines (IL-6, IL-1β, TNF-α) and bone metabolism markers (RANKL/OPG), and suppressed the MAPK/NF-κB pathway in PGIA mice.

Conclusions

Our findings suggest that miR-21 is associated with synovial fibroblast proliferation and osteogenesis, potentially via the MAPK/NF‑κB pathway, and that its inhibition alleviates inflammatory arthritis phenotypes in mice. These observations indicate that miR‑21 may warrant further investigation as a candidate target in AS-related inflammation and cartilage pathology.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13018-026-06983-1.

Introduction

Ankylosing spondylitis (AS) is a chronic inflammatory condition primarily affecting the spine and sacroiliac joints, with the resulting spinal fusion and ankylosis severely compromising patients’ quality of life [1, 2]. Its pathogenesis involves genetic predisposition, immune dysregulation, microbial factors, and environmental triggers [3], among which immune abnormalities-such as altered lymphocyte apoptosis and T helper 1/T helper 2 imbalance-lay a central role [4]. However, the molecular mechanisms that drive synovial fibroblast‑mediated ectopic ossification remain incompletely understood, and identifying key regulators of this process could reveal new therapeutic targets.

Within the synovial microenvironment of AS, synovial fibroblasts are recognized as pivotal effector cells [5]. These cells not only perpetuate local inflammation but also exhibit an enhanced potential for osteogenic differentiation, directly contributing to the pathological ectopic ossification characteristic of AS progression [6]. Identifying the key molecular drivers that regulate this aberrant osteogenic transformation in synovial fibroblasts is therefore crucial for understanding AS pathogenesis.

MicroRNA-21 (miR-21) is a multifunctional regulator involved in various physiological and pathological processes [7]. It is overexpressed in many solid tumors, where it promotes proliferation and migration [8, 9]. miR-21 promotes the osteogenic differentiation of mesenchymal stem cells in bone metabolism [10]. A growing body of evidence has demonstrated that non‑coding RNAs (ncRNAs) play critical regulatory roles in various musculoskeletal conditions [11–13]. For example, miRNAs have been shown to hold therapeutic potential in tendon injuries [11], serve as diagnostic biomarkers in osteoarthritis [12], and siRNAs have been used to investigate tendon homeostasis [13]. Furthermore, ncRNAs affect skeletal muscle differentiation and are implicated in the pathogenesis of multiple musculoskeletal diseases [14]. Consistent with this, a recent meta‑analysis published in this journal demonstrated that miRNAs are promising non‑invasive diagnostic biomarkers for osteoarthritis with high diagnostic accuracy [15]. Notably, miR-21 is also dysregulated in autoimmune diseases. For example, in AS, elevated miR-21 in synovial fibroblasts activates Janus kinase/Signal Transducer and Activator of Transcription 3 signaling and upregulates pro-inflammatory mediators [16], suggesting a potential role in modulating local fibroblast activity. However, whether and how miR-21 directly drives osteogenic transformation remains unknown. In the context of AS specifically, emerging evidence suggests that aberrant ncRNA expression, influenced by genetic and environmental factors, contributes to the dysregulation of inflammation and osteogenic/osteoclastogenic differentiation [17]. However, whether and how miR-21 directly drives osteogenic transformation in AS synovial fibroblasts remains unknown.

The MAPK/NF‑κB pathway, which modulates bone formation, is activated by pro‑inflammatory cytokines and contributes to bone remodeling dysfunction [18, 19]. Given its pivotal role in these processes, targeting the MAPK/NF‑κB axis represents a promising therapeutic strategy for ankylosing spondylitis. Furthermore, mechanistically, miR-21 can post-transcriptionally suppress target genes such as phosphatase and tensin homolog (PTEN), a negative regulator of the MAPK cascade [20–22]. Downregulation of PTEN by miR-21 may lead to MAPK/NF-κB activation, thereby promoting inflammatory signaling [15]. However, whether miR-21 influences AS progression through this specific axis remains unexplored.

A direct causal link between miR-21, the MAPK/NF-κB pathway, and the osteogenic transformation of AS synovial fibroblasts has yet to be established. Therefore, this study investigates whether miR-21 is associated with these processes and whether its inhibition modulates arthritis-related pathology in a PGIA mouse model. We performed gain- and loss-of-function experiments in vitro and evaluated the effects of miR-21 silencing in vivo.

Materials and methods

Cell culture and transfection

Cell culture: Human AS synovial fibroblasts (Cat No.: HUM‑iCell‑s010RA‑1) were obtained from iCell Bioscience Inc (Shanghai) Biotechnology Co., Ltd. Under standard conditions in a Thermo Fisher CO₂ incubator, cells were cultured using iCell primary fibroblast growth medium (iCell Bioscience Inc, PriMed‑iCell‑003).

Cell transfection: For transfection, cells in good growing state were seeded in 6-well plates. Lipofectamine™ 3000 reagent (Cat No.: L3000-008; Invitrogen) and plasmid constructs (miR-21 mimic or miR-21 inhibitor; Gema) were diluted separately according to the manufacturer’s instructions.

The sequences used were:

hsa‑miR‑21 mimic: 5′‑UAGCUUAUCAGACUGAUGUUGA‑3′ (sense); 5′‑AACAUCAGUCUGAUAAGCUAUU‑3′ (antisense).

hsa‑miR‑21 inhibitor: 5′‑UCAACAUCAGUCUGAUAAGCUA‑3′.

Complexes of Lipofectamine™ 3000 and plasmid (equal volumes, 15 min room temperature incubation) were added dropwise to cells. Transfection efficiency was analyzed by qPCR 48 h post-transfection.

Cell processing and grouping

Cells in the exponential growth phase were divided into five groups: a Control group, a miR-21 mimic group (transfected with a plasmid construct for miR-21 overexpression), a mimic negative control group (transfected with a scrambled control plasmid), a miR-21 inhibitor group (transfected with a plasmid to inhibit miR-21), and an inhibitor negative control group (transfected with a corresponding control plasmid).

Cell proliferation capacity assay

CCK8 assay: Following group-specific treatments for 24 h, CCK-8 working solution (Cat No.:C0039; Beyotime) was added to each well as recommended. After incubation, absorbance at 450 nm was measured and cell viability was calculated.

Colony formation assays: For grouping and treatment, cells were seeded into 12-well plates. The medium was refreshed every three days with regular monitoring. When clones reached a microscopically appropriate size, they were imaged and then returned to the incubator until most contained > 50 cells. Cells were sequentially washed with PBS, fixed with 4% paraformaldehyde (Cat No.: A500684, Sangon), washed, and stained with crystal violet (Cat No.: 548-62-9; Shanghai Qiangshun; 2 min).

Assessment of osteogenic differentiation

Alkaline phosphatase (ALP) staining and activity assay: For ALP activity assay, cells were lysed with 400 µL of lysis buffer (3 min on ice; Cat No.: P0013J; Beyotime), and centrifuged. Protein concentration was determined with a micro-volume spectrophotometer. Following the kit protocol (Cat No.: P0321M; Beyotime), samples, standards, and blanks were incubated with substrate (37 °C, 10 min), stopped, and absorbance read at 405 nm. ALP activity (U) was derived from a para‑nitrophenol standard curve, and specific activity (U/mg) was normalized to protein content. All assays were performed in triplicate.

For ALP staining, cells were fixed with 4% paraformaldehyde (Cat No.: A500684; Sangon) for 15 min, washed, and incubated with BCIP/NBT working solution (Cat No.: C3206; Beyotime) for 24 h. After washing with distilled water to stop the reaction, ALP‑positive cells were observed and photographed under a microscope.

Alizarin red staining and quantitative analysis: Cells were cultured to the logarithmic growth phase and treated according to experimental groups. After treatment, the supernatant was removed and cells were fixed with 4% paraformaldehyde. Cells were then stained with alizarin red staining solution (Cat No.: G1452; Solarbio) at room temperature for 20 min. Following staining, the solution was aspirated, and cells were washed thoroughly with distilled water. Calcium deposition was examined and imaged under an optical microscope (SOPTOP ICX41; Ningbo Shunyu Instruments Co., Ltd.).

In vivo arthritis model (PGIA) for AS-related pathology evaluation

Ten-week-old female specific pathogen free BALB/c mice (Shanghai SLAK Laboratory Animal Co., Ltd.; license no. SCXK(Shanghai)2022-0004) were acclimatized for one week prior to modeling (housed in the barrier facility of Hunter Biolnsight Co., Ltd. (License No.: SYXK(Zhejiang) 2024-0003).

The PGIA model recapitulates several inflammatory features of AS but does not fully reproduce ectopic ossification; it is used here to assess the effect of miR-21 on cartilage pathology and bone metabolism markers. Spinal cartilage was selected for subsequent analyses because it is the primary site of cartilage damage and inflammation in the PGIA model, allowing assessment of local MAPK/NF‑κB activation in relation to histological changes. PGIA model was induced as previously described Hu et al. [23]. At week 0, model group mice were intraperitoneally injected with 0.1 mL of bovine proteoglycan (Cat No.: D8428; Sigma) emulsified in Freund’s complete adjuvant (Cat No.: F5881; Sigma). A booster injection of the same antigen emulsified in Freund’s incomplete adjuvant (Cat No.: F5506; Sigma) was administered at week 3. The normal control mice received injections of an equal volume of sterile saline emulsified in the corresponding adjuvant at the same time points.

Group processing

At week 8 post-modeling, mice in the treatment group were randomly subdivided into three subgroups: the PGIA group (n = 8) received 200 µL saline via tail vein injection; the PGIA + AntagomiR-21 group (n = 6) received AntagomiR-21 (50 mM per injection [16], dissolved in 200 µL saline; mmu-AntagomiR-21: sense (5’-3’): UCAACAUCAGUCUGAUAAGCUA); and the PGIA + AntagomiR-NC group (n = 6) received a scrambled control Antagomir at an equivalent dose. Injections were administered twice per week. Mice in the sham-operated group (n = 8) received 200 µL saline on the same schedule.

Sample collection

At week 8, 2 mice each from the Sham and PGIA groups were euthanized, and spinal cords were collected to verify successful establishment of the in vivo model.

The remaining mice were euthanized at week 16. Serum was obtained via abdominal aorta blood collection, followed by centrifugation (3000 rpm, 15 min) and storage at − 80 °C. The spine was then dissected; intervertebral discs were carefully excised, adjacent cartilaginous endplates were removed, and tissues were rinsed with PBS and stored at − 80 °C.

For each group, 6 serum samples were analyzed by ELISA. Cartilage tissues were allocated for histology and immunohistochemistry (IHC) as well as for molecular analyses (qPCR and Western blot), with 3 samples per group used for each assay.

Pathological staining

The spinal cartilage tissue was decalcified, then routinely embedded and sectioned. Sections from each group were processed through standard dewaxing and rehydration procedures and subsequently stored for further use. The detailed protocol followed the method described by Feng et al. [1]. All chemical reagents, including anhydrous ethanol (Cat No.: 100092683) and xylene (Cat No.: 10023418), were supplied by Sinopharm Chemical Reagent Co., Ltd.

Hematoxylin-eosin (H&E) staining: Sections were hydrated, stained with hematoxylin for about 5 min, rinsed in tap water, and counterstained with eosin. All staining reagents were obtained from Shanghai Titan Technology Co., Ltd. (hematoxylin, Cat No.: 61753 C; eosin, Cat No.: 30086E).

Safranin O and Fast Green staining: The hydrated sections were stained with 1% safranin O solution (Cat No.: S2255; Sigma) for approximately 1 h. After discarding the excess stain and rinsing, the sections were counterstained with 0.5% fast green (Cat No.: F7252; Sigma) for 10 s and then differentiated in anhydrous ethanol for 1 min.

The stained sections were mounted with neutral resin and examined under a Nikon Eclipse Ci-L upright optical microscope.

IHC staining assay

Endogenous peroxidase activity was blocked with hydrogen peroxide solution (Sinopharm Chemical Reagent Co., Ltd.; Cat No.: 73113760). After washing, slides were mounted on plastic racks and subjected to antigen retrieval by immersion in boiling Tris-EDTA buffer (1 mmol, pH 9.0; Sigma; Cat No.: 648310, E9884), followed by maintenance for an additional 15 min. After blocking nonspecific sites with bovine serum albumin (BSA; Cat No.: B2064; Sigma), sections were incubated overnight with primary antibodies. The next day, slides were incubated with secondary antibodies, developed with DAB, and counterstained with hematoxylin. Finally, sections were mounted with neutral resin for microscopic observation. Details of all antibodies used are provided in Table S1.

ELISA

After clotting at room temperature for 20 min, blood samples were centrifuged. The resulting serum was collected and stored at − 80 °C. Serum concentrations of IL-6, Interleukin-1β (IL-1β), TNF-α, Osteoclastogenesis inhibitory factor (OPG), and Receptor Activator of Nuclear Factor Kappa B Ligand (RANKL) were determined using commercially available ELISA kits (Cat No.: RX203049M, RX203049M, RX202412M, RX202895M, RX202913M; RUIXIN BIOTECH), according to the manufacturer’s instructions. The detailed procedures followed established protocols described in the cited literature [24, 25].

Western blotting assay

Western blotting was used to assess the expression of phosphorylated and total extracellular regulated protein kinases‌ (ERK), c-Jun N-terminal kinase (JNK), p38 MAPK, and NF‑κB proteins in AS patient‑derived synovial fibroblasts.

The same analysis was applied to evaluate the expression of these proteins, along with ALP and bone gla protein (BGP), in mouse spinal cartilage tissue.

For tissue samples, mouse spinal cartilage was minced finely with sterile scissors in a culture dish on ice. The tissue fragments were homogenized in pre‑chilled lysis buffer using a tissue homogenizer, followed by incubation on ice. Cell samples were washed with PBS after treatment and lysed directly on ice.

Lysis buffer was prepared by supplementing RIPA lysis buffer (Cat No.: P0013B; Beyotime) with 1 mM PMSF (S Cat No.: P0100; Solarbio), a protein phosphatase inhibitor cocktail (Cat No.: P0100; Solarbio), and a protease inhibitor cocktail (Cat No.: P6730; Solarbio). Protein concentration was determined with a BCA assay kit (Cat No.: P0012S; Beyotime) before heat denaturation.

Equal protein quantities were resolved via SDS‑PAGE using a pre‑stained marker (Cat No.: EC1020-B; Shandong Sikojie Biotechnology). Proteins were blotted onto a PVDF membrane (Cat No.: 10600023; GE Healthcare). Membranes were blocked with 5% BSA (Cat No.: 4240GR100; Biofroxx) and then incubated sequentially with primary antibodies (overnight) and HRP-conjugated secondary antibodies. For detection, bands were visualized with ECL Plus reagent (Cat No.: PE0010; Solarbio) and quantified by densitometry. All steps followed established protocols [26], and antibody details are provided in Table S1.

qPCR

qPCR was performed to evaluate transfection efficiency, cellular MAPK (p38) and NF‑κB mRNA, and cartilage miR‑21 expression.

For tissue processing, samples were placed in a mortar pre-chilled with liquid nitrogen and ground into a fine powder. Before the liquid nitrogen had fully evaporated, approximately 30 mg of tissue was rapidly transferred to a new RNase-free 1.5 mL centrifuge tube.

RNA extracted from tissue and cell samples (EZ‑10 Column‑Based Total RNA Extraction Kit, Cat No.: B610583-0100; BBI) was reverse-transcribed into cDNA using TRUEscript RT MasterMix (Cat No.: PC7002; Beijing Edra Bio-Tech Co., Ltd.2). qPCR was carried out with 2× Dual SYBR Green Mix (Cat No.: PC6202; Beijing Edlay Biotechnology Co., Ltd.). All procedures followed established protocols described in the literature [23]. The sequences of primers used are listed in Table S2.

Statistical analysis

Data are expressed as mean ± SD. Analyses were performed with SPSS 20.0, using one-way ANOVA (with post-hoc tests based on variance) or the Kruskal–Wallis H test, depending on data distribution. A p-value < 0.05 was deemed statistically significant.

Results

MiR-21 promoted AS synovial fibroblast proliferation

First, primary human AS synovial fibroblasts were isolated. These cells grew adherently and exhibited characteristic polygonal and spindle-shaped morphologies (Fig. 1A). To assess the function of miR-21 in AS, gain- and loss-of-function models were established through mimics or inhibitors transfection (Fig. 1B), which confirmed the successful and specific modulation of miR-21 expression in each experimental group (p < 0.01).

Fig. 1.

Effect of miR-21 on proliferation in human AS synovial fibroblasts (A) Morphology of synovial fibroblasts from human AS: adherent growth pattern, spindle-shaped. Scale bars: 200 μm and 100 μm. (B) qPCR detection of miR-21 mRNA expression in cells following transfection with miR-21 mimic/mimic-NC and miR-21 inhibitor/inhibitor-NC. n = 3. (C) Cell viability assessed by CCK-8 assay following miR-21 modulation. miR-21 mimic enhanced viability compared to the mimic-NC, while knockdown reduced it. n = 3. (D) Statistical analysis of colony formation assays shows that miR-21 mimic enhances cell proliferation capacity. n = 3. (E) Representative images of colony formation assay for each cell group. Abbreviation: AS, Ankylosing Spondylitis; miR-21 mimic, miR-21-overexpressing cell group; mimic-NC, Negative control group; miR-21 inhibitor, miR-21 low-expression cell group; inhibitor-NC, Negative control group. Significant marker: ##P <0.01 vs. mimic-NC; @@P <0.01 vs. inhibitor-NC

CCK‑8 and colony formation assays confirmed that miR‑21 promotes synovial fibroblast proliferation: overexpression stimulated it, while inhibition had the opposite effect (vs. respective NCs, p < 0.01; Figs. 1C–D).

MiR‑21 mediated osteogenic differentiation in AS synovial fibroblasts

To investigate the role of miR-21 in the osteogenic differentiation of synovial fibroblasts-key effector cells in AS-ALP activity and alizarin red S staining were performed. As shown in Figs. 2A-D, overexpression of miR-21 significantly enhanced both ALP activity and calcium nodule formation compared with the mimic-NC group (p < 0.01). Inhibition of miR‑21 significantly suppressed mineralization activity (ALP and alizarin red staining) (p < 0.01); conversely, miR‑21 promotes osteogenic differentiation in AS‑derived synovial fibroblasts.

Fig. 2.

MiR-21 is associated with osteogenic differentiation in human AS synovial fibroblasts (A) ALP staining showed enhanced activity in the miR-21 mimic group and reduced activity in the inhibitor group. Scale bars: 200 μm and 100 μm. (B) Alizarin red staining revealed markedly increased calcium deposition in the miR-21 mimic group and decreased deposition in the inhibitor group. Scale bars: 200 μm and 100 μm. (C, D) The corresponding bar graph quantifies the results, showing significantly increased values in the miR-21 mimic group and decreased values in the inhibitor group, respectively. n = 3. Abbreviation: AS, Ankylosing Spondylitis; miR-21 mimic, miR-21-overexpressing cell group; mimic-NC, Negative control group; miR-21 inhibitor, miR-21 low-expression cell group; inhibitor-NC, Negative control group; ALP, Alkaline phosphatase. Significant marker: ##P <0.01 vs. mimic-NC; @@P <0.01 vs. inhibitor-NC

Involvement of MAPK/NF-κB signaling in miR-21-regulated osteogenesis

Based on these findings, we next examined whether miR-21 regulates the MAPK/NF-κB signaling pathways in human AS synovial fibroblasts. MiR‑21 overexpression promoted MAPK/NF‑κB pathway activation, increasing its key mRNA levels and the phosphorylation of ERK, JNK, p38 MAPK, and NF‑κB (all p < 0.01 vs. mimic‑NC; Fig. 3A-G). On the contrary, inhibition of miR‑21 significantly reduced the mRNA and protein expression of these molecules (all p < 0.01 vs. inhibitor-NC). These results suggest that miR-21 may activate the MAPK/NF-κB signaling pathway in AS synovial fibroblasts.

Fig. 3.

Regulation of the p38 MAPK/NF-κB signaling pathway by miR-21 (A, B) qPCR detection of p38 MAPK and NF-κB mRNA expression in cells following transfection with miR-21 mimic/mimic-NC and miR-21 inhibitor/inhibitor-NC. n = 3. Western blot analysis (C) and corresponding quantification (D-G) of p-ERK/ERK, p-JNK/JNK, p-p38/p38, and p-NF‑κB/NF‑κB in AS synovial fibroblasts following miR-21 modulation. β-actin was used for normalization. Abbreviation: AS, Ankylosing Spondylitis; miR-21 mimic, miR-21-overexpressing cell group; mimic-NC, Negative control group; miR-21 inhibitor, miR-21 low-expression cell group; inhibitor-NC, Negative control group. Significant marker: ##P <0.01 vs. mimic-NC; @P <0.05, @@P <0.01 vs. inhibitor-NC

MiR-21 promoted spinal cartilage damage in PGIA mice

Following the in vitro results, a PGIA mouse model was used to assess the in vivo function of miR-21 (Fig. 4A). Histological evaluation via H&E and Safranin O and Fast Green staining revealed pronounced spinal pathology in PGIA mice, characterized by uniform cartilage layer narrowing, a significant loss of chondrocytes, articular cartilage damage, and reduced proteoglycan staining in the matrix (Fig. 4B). Treatment with an AntagmiR-21 effectively suppressed miR-21 overexpression in spinal cartilage tissue (p < 0.01; Fig. 4C) and concurrently alleviated the afore-mentioned pathological changes (Fig. 4D). These findings demonstrate that inhibition of miR-21 mitigates cartilage lesions in PGIA mice.

Fig. 4.

Effect of miR-21 on spinal cartilage damage in PGIA mice (A) In vivo experimental procedure. (B) H&E staining of spinal cartilage in PGIA mice. Representative images show successful induction of the PGIA model, characterized by cartilage damage, marked depletion of chondrocytes, and diminished matrix staining compared to the control group. Scale bars: 200 μm and 100 μm. (C) qPCR detection of miR-21 mRNA expression in spinal cartilage tissue from PGIA-graded mice across each group. n = 3. (D) H&E staining showing the therapeutic effect of AntagomiR-21 on spinal cartilage in PGIA mice. Cartilage damage was ameliorated following treatment. Scale bars: 200 μm and 100 μm. Abbreviation: AS, Ankylosing Spondylitis; AntagomiR-21, miR-21-overexpressing cell group; AntagomiR-NC, Negative control group; PGIA, Proteoglycan-induced arthritis. Significant marker: ##P <0.01 vs. Sham; @@P <0.01 vs. PGIA + AntagomiR-NC

Effect of miR-21 suppression on bone metabolism in PGIA mice

Figures 5A-B demonstrate a marked increase in RANKL expression within the spinal cartilage of PGIA mice (vs. Sham group; p < 0.01), indicating aberrant bone metabolism activation during AS progression. Notably, this increase was significantly attenuated by treatment with AntagomiR-21, as evidenced by a marked reduction in the average optical density values for RANKL compared with the PGIA + AntagomiR-NC group (p < 0.01). This indicates that miR-21 inhibition can alleviate aberrant bone metabolism. As shown in Figs. 5C-E, AntagomiR-21 significantly reduced the protein levels of RANKL and increased the OPG/RANKL ratio in spinal cartilage, indicating a shift towards reduced bone resorption. ELISA assays revealed significant differences in serum OPG and RANKL levels among the groups (p < 0.05; Fig. 5C-D). Western blot results further confirmed that miR-21 inhibition promoted the balance between bone formation and resorption (p < 0.01; Fig. 5E-G).

Fig. 5.

Association of miR-21 with bone metabolism markers in PGIA spinal cartilage (A) Immunohistochemical staining of RANKL in mouse spinal cartilage across experimental groups. Scale bars: 200 μm and 100 μm. (B) The quantitative analysis shows that AntagomiR-21 treatment significantly downregulated RANKL expression. n = 3. (C, D) ELISA quantification of serum OPG and RANKL levels across experimental groups shows that AntagomiR-21 significantly reduces both markers in PGIA mice. n = 6. Western blot analysis (E) and corresponding quantification (F, G) of ALP, and BGP in mouse spinal cartilage following miR-21 modulation. β-actin was used for normalization. n = 3. Abbreviation: AS, Ankylosing Spondylitis; AntagomiR-21, miR-21-overexpressing cell group; AntagomiR-NC, Negative control group; PGIA, Proteoglycan-induced arthritis. Significant marker: ##P <0.01 vs. Sham; @P <0.05, @@P <0.01 vs. PGIA + AntagomiR-NC

Downregulation of miR-21 alleviated spinal cartilage inflammation and inhibited the MAPK/NF-κB pathway in PGIA mice

Compared to control, AntagmiR-21 significantly decreased serum IL-6, IL-1β, and TNF-α in PGIA mice (all p < 0.01; Fig. 6A-C), suggesting a suppression of systemic inflammation. Abnormal activation of the p38 MAPK/NF-κB signal pathway was also observed in the spinal cartilage of PGIA mice (Fig. 6D-H). This activation was effectively suppressed by AntagmiR-21 intervention, reducing pathway activity to levels comparable with those in the Sham group (all p < 0.01).

Fig. 6.

Suppression of the p38 MAPK/NF‑κB pathway following miR-21 downregulation (A-C) ELISA quantification of serum IL-6, IL-1β and TNF-α levels across experimental groups shows that AntagomiR-21 significantly reduces both markers in PGIA mice. n = 6. Western blot analysis (D) and corresponding quantification (E-H) of of p-ERK/ERK, p-JNK/JNK, p-p38/p38, and p-NF‑κB/NF‑κB in mouse spinal cartilage following miR-21 modulation. β-actin was used for normalization. n = 3. Abbreviation: AS, Ankylosing Spondylitis; AntagomiR-21, miR-21-overexpressing cell group; AntagomiR-NC, Negative control group; PGIA, Proteoglycan-induced arthritis. Significant marker: ##P <0.01 vs. Sham; @@P <0.01 vs. PGIA + AntagomiR-NC

Discussion

This study investigated the role of miR-21 and its link to the MAPK/NF‑κB pathway in AS-an autoimmune-driven inflammatory bone disorder where this relationship remains unclear [27, 28]. Within this process, synovial fibroblasts are considered key effector cells and potential targets for pathological ectopic ossification [29]. Consequently, regulating the biological behavior of these fibroblasts represents a promising therapeutic strategy for modulating cartilage lesions. To explore this mechanism comprehensively, we utilized primary human AS synovial fibroblasts as an in vitro model and the PGIA mouse model for in vivo validation. The PGIA model is well-established in AS research and recapitulates critical disease features, including the structural alterations in spinal cartilage-such as uniform tissue narrowing and significant chondrocyte loss-observed in this study [30]. It should be noted, however, that the PGIA model primarily mimics inflammatory arthritis and does not fully reproduce the ectopic ossification characteristic of AS. Therefore, our in vivo findings relate mainly to cartilage pathology and inflammatory bone metabolism markers, not to AS-specific ossification.

Previous reports indicate that miR-21 plays a significant role in disorders associated with abnormal bone metabolism. Consistent with our findings, Hu et al. reported that miR-21 silencing alleviated bone loss in a different model, supporting a broader role of miR-21 in bone homeostasis [31]. Notably, emerging evidence suggests that miR-21 expression is significantly altered in experimental AS, implicating it in cartilage lesions [32]. However, the precise molecular mechanisms linking miR-21 to AS pathogenesis remain unclear. Therefore, the present study employed both gain- and loss-of-function experiments to define the functional role of miR-21 and its underlying pathway in AS.

Our findings indicate that miR-21 promotes synovial fibroblast proliferation and osteogenic differentiation, contributing to spinal cartilage damage and dysregulated bone metabolism in AS, at least partially potentially through activation of the MAPK/NF‑κB signaling pathway. Inhibition of miR-21 alleviated these cellular phenotypes in both cellular and PGIA mouse models, highlighting its therapeutic potential.

Our observations that miR-21 overexpression enhanced the proliferative capacity of AS synovial fibroblasts align with its established role as an oncogenic miRNA in various cancers, where it promotes cell growth and survival [33]. In the context of AS, this pro‑proliferative effect may fuel the hyperplasia of synovial tissue and contribute to pannus formation, a hallmark of cartilage lesions. Conversely, the significant suppression of proliferation upon miR-21 inhibition underscores its functional importance in maintaining the activated state of these fibroblasts. Consistent with our findings, small interfering RNAs (siRNAs) have been demonstrated to effectively target inflammatory pathways in human rheumatoid arthritis, another chronic inflammatory joint disease, further supporting the therapeutic potential of RNA‑based interventions in arthritis [34]. More importantly, we provide novel evidence linking miR-21 to osteogenic differentiation in AS synovial fibroblasts. The significant upregulation of ALP activity and calcium deposition upon miR-21 overexpression, coupled with their marked downregulation following inhibition, strongly suggests that miR-21 drives a pro‑osteogenic phenotype. This is particularly relevant to the pathogenesis of AS, in which ectopic ossification of spinal structures is a characteristic feature [35]. As shown in Figs. 5C-E, AntagomiR-21 significantly reduced the protein levels of RANKL and increased the OPG/RANKL ratio in spinal cartilage, indicating a shift towards reduced bone resorption. Our data position miR-21 as a key molecular switch that may push synovial fibroblasts toward an aberrant bone‑forming lineage. The regulatory role of miR-21 in bone metabolism is consistent with a broader paradigm in which both miRNAs and lncRNAs function as key regulators of genes critical for bone homeostasis and disease [36]. Moreover, siRNA‑based approaches have shown promise in the management of osteoporosis, particularly by targeting genes involved in metabolic bone processes [37]. Additionally, recent investigations have identified lncRNAs that regulate osteoblast viability and apoptosis during bone fracture healing, further highlighting the diverse regulatory roles of ncRNAs in bone biology [38].

The mechanistic exploration revealed that the effects of miR-21 are mediated through the MAPK/NF-κB pathway. We observed that miR-21 upregulation significantly increased the phosphorylation (activation) of key components of this pathway, including ERK, JNK, p38 MAPK, and NF-κB, while inhibition had the opposite effect. This pathway is a well‑known central hub integrating inflammatory and stress signals [39], and its aberrant activation is implicated in both inflammation and abnormal bone remodeling [40]. The involvement of ncRNAs in regulating inflammatory responses is well documented, with multiple studies demonstrating that ncRNAs modulate the expression of key cytokines such as TNF‑α, IL‑6, and IL‑1β through pathways including NF‑κB [41]. Our results therefore delineate a plausible axis in which miR-21 enhances synovial fibroblast proliferation and osteogenic differentiation potentially through activation of the MAPK/NF‑κB pathway. This extends previous observations of altered miR-21 expression in AS by defining a specific downstream signaling mechanism.

In the PGIA model, suppression of miR-21 reduces its expression in cartilage and ameliorates associated pathological alterations including cartilage destruction, elevated inflammatory mediators (IL-6, IL-1β, TNF-α), and OPG/RANKL imbalance. The concomitant suppression of the MAPK/NF-κB pathway in treated mice further supports the association established in vitro. These results collectively suggest that targeting miR-21 can modulate multiple aspects of arthritis pathology-inflammation, tissue damage, and bone metabolism markers-through a common signaling cascade. Collectively, these observations align with a growing literature implicating miRNAs in every step of osteogenesis and bone homeostasis, from embryonic skeletal development to the maintenance of adult bone tissue [42].

Several limitations of this study should be acknowledged. First, although we observed that miR-21 modulates the MAPK/NF‑κB pathway, direct gain‑ or loss‑of‑function interventions targeting this pathway-such as pharmacological activation or inhibition-were not performed. Thus, a causal link between miR‑21‑driven osteogenesis and MAPK/NF‑κB signaling remains to be fully established. Second, the PGIA model used in this study recapitulates inflammatory arthritis and cartilage damage but does not fully reproduce the ectopic ossification characteristic of AS. Therefore, our in vivo findings primarily relate to cartilage pathology and bone metabolism markers (e.g., RANKL/OPG), not to AS‑specific aberrant bone formation. Future studies using AS‑relevant models (e.g., proteoglycan‑induced spondylitis or transgenic models) are needed to assess the effect of miR‑21 on spinal ossification. Third, consistent with our findings, Hu et al. reported that targeted silencing of miR‑21 alleviated bone loss in a different model (sympathetic neurostress‑induced osteopenia) [31], supporting a broader role of miR‑21 in bone homeostasis. However, direct comparisons between studies are limited by differences in disease models and endpoints. Fourth, validation in additional animal models would strengthen the conclusions. Finally, the long‑term safety and efficacy of miR‑21 inhibition as a therapeutic strategy require further investigation.

Conclusions

In conclusion, our work suggests that miR-21 is a potential regulator in AS-related synovial fibroblast proliferation and osteogenic differentiation, with evidence implicating the MAPK/NF-κB pathway. In the PGIA model, miR-21 inhibition reduced cartilage damage and altered bone metabolism markers. These findings indicate that miR-21 warrants further investigation as a candidate target for modulating inflammation and cartilage pathology in AS. Future studies are needed to establish causality and to test miR-21 blockade in models that recapitulate ectopic ossification.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AS

Ankylosing spondylitis

miR-21

microRNA‑21

MAPK

Mitogen-activated protein kinase

NF‑κB

Nuclear factor kappa-B

PGIA

Proteoglycan‑induced arthritis

PTEN

Phosphatase and tensin homolog

ALP

Alkaline phosphatase

Author contributions

**Weiwei Pan: ** Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing-original draft. **Wenbiao Zheng: ** Data curation, Investigation, Methodology, Validation, Visualization. **Fanghu Chen: ** Data curation, Formal analysis, Project administration, Writing-review & editing. **Weiqian Wu: ** Conceptualization, Project administration, Supervision, Writing-review & editing.

Funding

This work was supported by the Taizhou Science and Technology Plan Project [grant number 24ywb82].

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All experimental procedures involving the animals adhered to the guidelines of the Institutional Animal Care and Use Committee and were approved by the Laboratory Animal Management and Ethics Committee of Hunter Biolnsight Co., Ltd. (License No.: SYXK(Zhejiang) 2024-0003).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.