PSMB9 exacerbates chondrocyte injury in osteoarthritis via activation of the NF-κB pathway

Abstract

Background

Osteoarthritis (OA) is a degenerative disease with incompletely understood mechanisms. The proteasome subunit PSMB9 has been implicated in immune regulation, but its specific role in OA pathogenesis remains unclear. This study aimed to investigate whether PSMB9 mediates IL-1β–induced chondrocyte injury by activating the NF-κB pathway through promoting IκBα degradation, and to explore the regulatory relationship between IL-6 and PSMB9 in OA progression.

Methods

Differentially expressed genes were identified by integrating human and mouse OA datasets. A mouse destabilization of the medial meniscus (DMM) model was established. The function of PSMB9 in OA chondrocytes and its effect on the NF-κB pathway were analyzed using hematoxylin–eosin (H&E) staining, immunohistochemistry, Western blot, CCK-8, Edu, flow cytometry, and immunofluorescence to observe p65 nuclear translocation.

Results

(1) PSMB9 was significantly upregulated in multiple OA datasets and models; (2) PSMB9 expression increased in the cartilage of OA patients and mice; (3) PSMB9 overexpression exacerbated IL-1β-induced chondrocyte apoptosis, inhibited proliferation, upregulated the expression of the inflammatory factor IL-6 and the matrix-degrading enzyme MMP13, promoted extracellular matrix (ECM) degradation, and decreased COL2A1 expression; (4) PSMB9 activated the NF-κB pathway by promoting IκBα degradation, and inhibition of this pathway alleviated cell injury; (5) silencing IL-6 reduced PSMB9 expression.

Conclusion

PSMB9 may participate in the activation of the NF-κB pathway and potentially contribute to chondrocyte injury in OA, making it a promising target for future research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-026-06796-2.

Introduction

Osteoarthritis (OA) is the most common degenerative joint disease, characterized by progressive destruction of articular cartilage, subchondral bone sclerosis, osteophyte formation, and synovial inflammation. Globally, OA affects millions of people and is a leading cause of pain, functional impairment, and disability in middle-aged and elderly individuals, imposing substantial economic and healthcare burdens on society [1]. Although the precise etiology and pathogenesis of OA are not yet fully understood, it is widely believed to be closely associated with metabolic imbalance in joint tissues and chronic low-grade inflammation triggered by factors such as mechanical stress, aging, obesity, and genetic predisposition [2–5]. Within the complex pathological process of OA, disruption of chondrocyte homeostasis is a central element. This involves not only an imbalance in anabolic and catabolic pathways but also dysregulation of processes such as apoptosis, autophagy, and other related cellular functions, ultimately leading to degradation of the extracellular matrix (ECM) and accelerated cartilage destruction [6].

The proteasome is a crucial multi-subunit protein complex within cells responsible for degrading unnecessary or damaged proteins, thereby maintaining cellular protein homeostasis [7, 8]. The 26 S proteasome is a multi-protein complex composed of one 20 S core particle and one or two 19 S regulatory particles. The core 20 S particle has a barrel-shaped structure formed by four stacked heptameric rings (αββα). The two outer rings consist of α-subunits (α1–7), which serve as binding sites for regulatory complexes (including the 19 S, 11 S, and PA200 families), facilitating peptidase activity and controlling the entry of cytoplasmic proteins into the proteasome chamber. The two inner rings are composed of β-subunits (β1–7), three of which (β1, β2, and β5) contain catalytic sites that hydrolyze peptide bonds [9]. In cells stimulated by interferon-gamma (IFN-γ) or tumor necrosis factor-alpha (TNF-α), these proteolytically active subunits can be replaced by β1i (LMP2/PSMB9), β2i (MECL-1/PSMB10), and β5i (LMP7/PSMB8), thereby forming the immunoproteasome [10, 11]. Beyond the standard proteasome, immunoproteasomes—variant proteasomes with inducible β-subunits—play significant roles in various pathological conditions, including autoimmune diseases, inflammatory disorders, and cancer [12–15]. Traditionally, PSMB9 was thought to be primarily expressed in immune cells and involved in antigen presentation. However, recent studies indicate that PSMB9 expression is also markedly increased in various non-immune cells (including tumor cells and endothelial cells) and may participate in regulating apoptosis, proliferation, and inflammatory responses [16–18]. Notably, some studies have found PSMB9 to be highly expressed in the synovial tissue of rheumatoid arthritis and potentially linked to the activation of the NF-κB pathway [16]. Nevertheless, the expression, biological function, and specific molecular mechanisms of PSMB9 in osteoarthritic chondrocytes remain unclear.

The nuclear factor kappa B (NF-κB) signaling pathway serves as a critical link between inflammatory responses and cartilage degradation, playing a central regulatory role in the defense networks cells deploy against infection, stress, or injury [19]. Its activity is precisely regulated by the inhibitor of NF-κB alpha (IκBα) [19]. Under resting conditions, IκBα acts as a “molecular brake,” binding to and sequestering NF-κB (typically the p50/p65 heterodimer) in the cytoplasm [20]. In the OA milieu, inflammatory cytokines (such as IL-1β and TNF-α) or mechanical stress can activate this pathway [21, 22]. These stimuli activate the upstream IKK complex, which phosphorylates IκBα, targeting it for degradation via the ubiquitin-proteasome pathway. This process releases NF-κB, allowing its translocation into the nucleus [20]. Activated NF-κB (primarily the p65 subunit) enters the nucleus and initiates the transcription of various downstream target genes, including matrix metalloproteinases (MMPs) and interleukin-6 (IL-6) [21, 22]. These factors collectively promote ECM degradation, inhibit the synthesis of cartilage matrix components such as type II collagen and aggrecan, and further amplify the inflammatory cascade, creating a vicious cycle that exacerbates chondrocyte damage [23, 24]. Previous research suggests that PSMB9 deficiency may delay IκBα degradation, thereby inhibiting NF-κB activation [25]. However, whether this regulatory relationship exists in osteoarthritic chondrocytes remains unknown.

In this study, we identified proteasome family members associated with OA by analyzing multiple human datasets (e.g., GSE215039) and mouse data, and found that PSMB9 expression is elevated in chondrocytes from OA patients. The increased expression of PSMB9 in human osteoarthritic cartilage tissues and a mouse OA model was further verified by immunohistochemistry. Subsequently, using an IL-1β-induced chondrocyte model, we investigated the role of PSMB9 in OA and its regulatory effect on the NF-κB pathway. This research aims to provide new insights into the pathology of osteoarthritis and identify potential therapeutic strategies.

Materials and methods

Dataset selection and processing

We obtained the high-throughput transcriptomic dataset GSE215039 of human primary chondrocytes from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/). This dataset includes a calibrated and quality-controlled count matrix, which we first subjected to quantile normalization and log2 transformation to ensure data accuracy and comparability. The transcriptomic data were derived from primary chondrocyte samples of five patients with osteoarthritis (OA) undergoing total knee replacement and five non-OA patients undergoing scoliosis surgery, with all samples untreated, making them suitable for investigating molecular differences between normal and diseased chondrocytes. Furthermore, this study incorporated multiple published joint tissue expression profile datasets, including GSE169077 from knee cartilage of healthy individuals and OA patients; integrated and batch-corrected normal and OA synovial tissue datasets GSE55235 and GSE1919; and mouse OA model datasets GSE53857 and GSE41342 covering different postoperative time points. These datasets provided essential references for investigating pathological differences in knee cartilage, molecular alterations in synovial tissue, and the dynamic progression of disease in mouse models. All public datasets were sourced from GEO, with their links provided in the Data Availability section.

Clinical samples

Cartilage tissues were obtained from end-stage knee osteoarthritis (OA) patients undergoing total knee arthroplasty at The First Affiliated Hospital of Anhui Medical University. Given limited sample availability, we adopted an intra-individual region-based control strategy: within each patient, areas with relatively intact morphology and a smooth surface (ICRS grade 0–I) on the tibial plateau were defined as controls, and severely worn areas as the OA group. All patients (n = 3; mean age 65.7 ± 8.2 years; M/F 2:1) were diagnosed with primary knee OA. The study was approved by the hospital ethics committee (PJ-YX2024-022), and written informed consent was obtained. Cartilage was collected intraoperatively and immediately separated into the two regions. To accommodate both histological and cellular experiments, multiple macroscopically normal tissue blocks (n = 3 per donor, total of 9 blocks) were harvested from the control region of each donor. These blocks were divided into two portions: one portion was processed together with OA tissues for histological analysis to compare with the damaged region, and the remaining portion was used for primary chondrocyte isolation for subsequent cellular experiments.

Animal experiments

C57BL/6J mice were obtained from the Experimental Animal Center of Anhui Medical University and housed under specific pathogen-free conditions (23–25 °C, 45–65% humidity, 12 h light/dark cycle). Three independent animal experiments were conducted, each comprising 5 mice per group. Eight-week-old male mice were anesthetized with 2% pentobarbital sodium (MedChemExpress, USA). Following shaving and disinfection, a midline incision was made over the right posterior knee; the patellar ligament and surrounding soft tissues were bluntly dissected to expose the joint cavity. The medial meniscotibial ligament was transected to establish a post-traumatic osteoarthritis model [26]. The wound was closed in layers, and postoperative analgesic and anti-infection treatments were administered. Mice were divided into two groups: the experimental group underwent destabilization of the medial meniscus (DMM) surgery on the right knee, while the control group received no surgery. For control mice, cartilage from the left limb was collected for primary chondrocyte isolation and subsequent cellular assays (unstimulated control), whereas the right limb was processed alongside the operated knees from the DMM group for histological analyses (including immunohistochemistry, H&E staining, and Safranin O/Fast Green staining). For DMM mice, cartilage from the non-operated left limb was also harvested for primary chondrocyte isolation and stimulated with IL-1β (10 ng/mL) for 24 h to serve as the in vitro OA group, while the operated right limb was used for histological analyses. All knee joint tissues were collected at 8 weeks post-surgery. All animal procedures were approved by the Animal Ethics Committee of Anhui Medical University (Approval No. LLSC20253028).

Isolation of chondrocytes

Primary mouse chondrocytes were isolated as previously described [27]. Articular cartilage was harvested from the knee joints of C57BL/6J mice, washed three times with PBS, and minced into approximately 0.25 mm³ pieces using a sterile scalpel. The tissue fragments were pre-digested twice with high-concentration type II collagenase (25 mg/mL in DMEM) for 15 min each at 37 °C under 5% CO₂, followed by further digestion with low-concentration type II collagenase (5 mg/mL) for 6–8 h under gentle agitation. The cell suspension was then filtered through a 100 μm cell strainer and centrifuged at 300 × g for 5 min. After washing and resuspension in PBS, cells were seeded at a density of 2–2.5 × 10⁵ cells/cm² and cultured in DMEM supplemented with 15% fetal bovine serum and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ incubator. To simulate osteoarthritis conditions, cells were treated with 10 ng/mL IL-1β for 24 h upon reaching 70% confluence.

Human primary chondrocytes were isolated as previously described [28]. Control chondrocytes collected using the method described above were washed three times with PBS and then finely minced using ophthalmic scissors. The minced tissue was digested with trypsin and collagenase type II to release the chondrocytes. After digestion, the cell suspension was filtered through a 150-mesh stainless steel sieve and centrifuged at 1000 rpm for 3–5 min. The pelleted cells were washed three times with PBS and subsequently cultured in DMEM supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific, USA). Cells were maintained in a humidified incubator at 37 °C with 5% CO₂. To establish an in vitro osteoarthritis model, the chondrocytes were treated with 10 ng/mL IL-1β (Thermo Fisher Scientific, USA) for 24 h as reported previously [29].

Cell culture

Primary human chondrocytes, primary mouse chondrocytes, and the C28/I2 cell line (obtained from the First Affiliated Hospital of Anhui Medical University) were cultured in high-glucose DMEM medium (Thermo Fisher Scientific, Massachusetts, USA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Massachusetts, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Massachusetts, USA). All cells were cultured at 37 °C under 5% CO₂.

Establishment of cell models

For mechanistic studies in the C28/I2 human chondrocyte cell line, cells were seeded at appropriate densities in 6-well plates. To modulate PSMB9 expression, cells were infected with lentivirus for PSMB9 overexpression or transfected with PSMB9-specific siRNA prior to subsequent treatments. Following genetic manipulation, cells were treated with 10 ng/mL recombinant human IL-1β (Thermo Fisher Scientific, USA) for 24 h to establish an in vitro osteoarthritis (OA) model [29]. To further elucidate whether PSMB9 exerts its effects through the NF-κB signaling pathway, C28/I2 cells were treated with the NF-κB inhibitor JSH-23 (10 µM, KKL MED) for 24 h [30]. For primary chondrocytes, both mouse and human primary chondrocytes were seeded at appropriate densities in 6-well plates and treated with 10 ng/mL recombinant human IL-1β (Thermo Fisher Scientific, USA) for 24 h to simulate OA-like conditions in vitro [29]. Controls: For siRNA knockdown experiments, cells transfected with scrambled siRNA (non-targeting control) were used as the negative control. For IL-1β stimulation experiments, untreated cells cultured under identical conditions with vehicle (PBS) only served as the baseline control. For JSH-23 inhibition experiments, cells treated with vehicle (DMSO) alone were used as the control.

Lentivirus infection

To achieve long-term and stable modulation of PSMB9 expression in chondrocytes for in-depth investigation into the progression mechanisms of osteoarthritis, this study employed a lentiviral vector system for genetic manipulation. Lentiviruses can efficiently infect both dividing and non-dividing cells and stably integrate the target gene into the host genome, making them suitable for mechanistic studies requiring sustained observation. All lentiviral stocks were stored at − 80 °C for subsequent use. The specific experimental procedure was as follows: C28/I2 cells were seeded in 6-well plates at a density of 2 × 10⁵ cells per well. When cell confluence reached 40%–50%, lentiviral infection was performed. The infection system consisted of lentivirus (GeneChem, Shanghai, China) at a multiplicity of infection (MOI) of 80, together with 5 µg/mL Polybrene (MedChemExpress, USA) to enhance viral infection efficiency. After 12 h of infection, the virus-containing medium was removed and replaced with fresh complete medium for continued culture.

siRNA transfection

To functionally validate the role of PSMB9 in osteoarthritis, siRNA-mediated transient knockdown was performed in C28/I2 chondrocytes to suppress PSMB9 expression and assess its effects on chondrocyte phenotypes. Cells were seeded in 6-well plates at 2 × 10⁵ cells per well and cultured in DMEM supplemented with 10% fetal bovine serum at 37 °C and 5% CO₂. Upon reaching 60–70% confluence, transfection was carried out according to the manufacturer’s instructions (OBiO Technology, Shanghai, China). Briefly, 100 nM siRNA and 10 µL Lipofectamine™ 2000 were each diluted in 250 µL Opti-MEM^® Reduced Serum Medium, incubated at room temperature for 5 min, then combined and incubated for another 25 min to form complexes. The mixture was added dropwise to the culture wells containing 1.5 mL medium. Six hours post-transfection, the medium was replaced with fresh complete medium, and cells were harvested 48 h later.

Cell proliferation assay

The proliferative capacity of C28/I2 cells treated with 10 ng/mL IL-1β for 24 h was evaluated using the CCK-8 assay and a 5-ethynyl-2′-deoxyuridine (EdU) immunofluorescence detection kit (Beyotime Biotechnology, Shanghai, China). Cells were digested with trypsin (Beyotime Biotechnology) to prepare a single-cell suspension at a density of 2 × 10⁴ cells/mL, which was seeded into 96-well plates. CCK-8 solution was added to each well, and after 2 h of incubation, the optical density (OD) was measured at 24 and 48 h to plot cell growth curves. The EdU assay was performed according to the manufacturer’s instructions. C28/I2 cells were incubated with EdU working solution, fixed for 15 min, and permeabilized at room temperature. Then, 500 µL of Click reaction solution was added to each well, followed by incubation in the dark at room temperature. Finally, nuclei were stained with Hoechst 33,342, and images were captured under a fluorescence microscope.

Hematoxylin–eosin (H&E) staining

Following euthanasia, mouse knee joints were harvested, and the surrounding musculature was removed. Human cartilage tissues were also collected following the same procedure. Both types of samples were fixed in 4% paraformaldehyde at 4 °C for 48 h, followed by decalcification in 15% EDTA at 37 °C for several weeks. Subsequently, the specimens were embedded in paraffin and sectioned at a thickness of 5 μm. Hematoxylin–eosin staining was performed according to the manufacturer’s instructions (Solarbio, Beijing, China). Briefly, paraffin sections were deparaffinized, rehydrated, and then stained with hematoxylin and eosin.

Safranin O–Fast Green staining

Sections of mouse knee joints and human cartilage tissues were stained using a Safranin O–Fast Green staining kit (Solarbio, Beijing, China) according to the manufacturer’s instructions. After staining, the tissues were dehydrated, cleared, mounted, and observed under a light microscope to evaluate cartilage morphology. The severity of osteoarthritis was assessed histologically using the Osteoarthritis Research Society International (OARSI) scoring system [31].

Immunohistochemical staining

Paraffin sections of mouse knee joints and human knee cartilage were subjected to antigen retrieval with an antigen retrieval solution (Boster Bio, California, USA) at 37 °C for 30 min. Endogenous peroxidase activity was blocked by incubation with 3% hydrogen peroxide at room temperature for 10 min. The sections were then incubated overnight at 4 °C with the following primary antibodies: MMP-13 and IL-6 (Affinity Biosciences, Jiangsu, China); Col2a1 (Abcam plc, Cambridge, UK); PSMB9 (ImmunoWay Biotechnology, Texas, USA). Immunoreactivity was detected using a streptavidin–biotin detection system (ORIGENE, Maryland, USA), followed by color development with DAB (ZSGB-BIO, Beijing, China) and counterstaining with hematoxylin (Solarbio, Beijing, China).

Immunofluorescence staining

To observe NF-κB p65 nuclear translocation, C28/I2 cells were treated with 10 ng/mL IL-1β for 30 min, followed by fixation with 4% paraformaldehyde and permeabilization with 0.5% Triton X-100. Antigen retrieval was performed using antigen retrieval solution (Boster Bio, California, USA). After blocking endogenous peroxidase activity with 3% hydrogen peroxide, cells were incubated overnight at 4 °C with the primary antibody against p65 (ImmunoWay Biotechnology, Texas, USA). Subsequently, cells were incubated with corresponding secondary antibodies (Beyotime Biotechnology, Shanghai, China) for 4 h at room temperature, and nuclei were counterstained with DAPI (Beyotime Biotechnology) for 5 min. Images were captured using a fluorescence microscope.

Detection of cell apoptosis by flow cytometry

Cell apoptosis was quantitatively analyzed by flow cytometry using an Annexin V FITC/PI Apoptosis Detection Kit (KeyGEN BioTECH, Jiangsu, China) according to the manufacturer’s instructions. Briefly, approximately 5 × 10⁵ cells were collected and resuspended in 500 µL binding buffer. The cells were then stained with 5 µL Annexin V FITC and 5 µL propidium iodide (PI), followed by incubation in the dark at room temperature for 15 min. Apoptosis was detected using a FACScan flow cytometer (Beckman Coulter, California, USA), and data were analyzed with FlowJo software.

Western blot

Total protein was extracted using RIPA lysis buffer (Beyotime Biotechnology, Shanghai, China) supplemented with protease and phosphatase inhibitors (Beyotime Biotechnology) at a 1:100 ratio. Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), and 20–30 µg of protein per lane were transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% skim milk or 5% bovine serum albumin at room temperature, followed by overnight incubation at 4 °C with primary antibodies diluted as recommended by the manufacturers. After incubation with corresponding secondary antibodies for 2 h at room temperature, protein bands were visualized using an enhanced chemiluminescence (ECL) reagent (Abbkine Scientific, Wuhan, China). The primary antibodies used in this study were as follows: PSMB9, IκBα, P65, p65 (phospho-Ser536), IκB-α (phospho-Ser36) (all from ImmunoWay Biotechnology, USA); MMP-13, IL-6 (both from Affinity Biosciences, China); Col2a1 (from Abcam plc, UK); β-actin and β-Tubulin (both from Bioworld Technology, USA).

RNA extraction and reverse transcription quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted from cells using TRIzol™ reagent (Thermo Fisher Scientific, USA). RNA concentration and purity were measured with a NanoDrop system, and samples with an A260/A280 ratio between 1.8 and 2.0 were considered acceptable. Reverse transcription was performed using the RevertAid First Strand cDNA Synthesis Kit (Vazyme Biotech, Nanjing, Jiangsu, China) according to the manufacturer’s instructions. Briefly, 1 µg of total RNA was reverse-transcribed in a 20 µL reaction mixture containing 4 µL of 5× Reaction Buffer, 1 µL of Ribolock RNase Inhibitor, 2 µL of 10 mM dNTP Mix, 1 µL of RevertAid M-MuLV RT, and nuclease-free water. The reaction was carried out under the following conditions: 25 °C for 5 min, 42 °C for 60 min, and 70 °C for 5 min.

RT-qPCR was performed on a Thermo Fisher Scientific 7500 Real-Time PCR System using SYBR Premix Ex Taq (Vazyme Biotech). Each 20 µL reaction contained 10 µL of SYBR Premix Ex Taq, 0.8 µL each of forward and reverse primers (10 µM), 2 µL of cDNA template, and nuclease-free water. The thermal cycling protocol consisted of an initial denaturation at 95 °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 5 s and annealing/extension at 60 °C for 30 s. A melt-curve analysis was performed to verify amplification specificity. The relative mRNA expression of target genes was calculated using the 2^(−ΔΔC_t ) method with human and mouse GAPDH as the internal reference genes, respectively. Primer sequences are listed in Table 1.

Table 1.

Nucleotide sequences of primers used for qPCR

Gene	Sequence
PSMB9(Human)	Forward: GGAGGTCAGGTATATGGAACCC
	Reverse: CCTGGCTTATATGCTGCATCC
PSMB9(Mouse)	Forward: GAGGACTTGTTAGCGCATCTCA
	Reverse: CATATACCTGTCCCCCCTCACA
GAPDH(Human)	Forward: GGAGCGAGATCCCTCCAAAAT
	Reverse: GGCTGTTGTCATACTTCTCATGG
GAPDH(Mouse)	Forward: CAGTGGCAAAGTGGAGATTG
	Reverse: TGCCGTGAGTGGAGTCATAC

Statistical analysis

All data in this study were analyzed using GraphPad Prism 9.0 software. Quantitative data are presented as mean ± standard deviation (mean ± SD). Comparisons between two groups were performed using the unpaired t-test, while comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA). A P value < 0.05 was considered statistically significant. In the figures, significance is denoted as follows: *P < 0.05, **P < 0.01, ***P < 0.001; ns indicates no significant difference.

Result

Differential expression analysis of PSMB9 in public human datasets

To explore the role of the proteasome family in OA, we analyzed multiple public human datasets. In GSE215039 (primary chondrocytes), PSMB9 was among the most differentially expressed genes (Fig. 1A). Analysis of knee cartilage data (GSE169077) confirmed PSMB9 upregulation in OA (Fig. 1B). Integration of synovial tissue datasets (GSE1919/GSE55235) showed PSMB9 was consistently upregulated across studies (Fig. 1C). A Venn diagram revealed PSMB9 was commonly upregulated in chondrocytes, cartilage, and synovial tissue (Fig. 1D), supporting its potential as an OA biomarker.

Fig. 1.

Differential expression analysis of PSMB9 in public human datasets. A Volcano plot of differentially expressed genes in GSE215039. Colors indicate p-value significance. B Volcano plot for GSE169077, showing log2 fold changes. Significantly altered proteasome family genes (P < 0.05) are highlighted. C Scatter plot comparing log2 fold changes between GSE1919 and GSE55235. D Venn diagram of overlapping differentially expressed genes across GSE215039, GSE1919, and GSE55235. NS, not significant; *P < 0.05, **P < 0.01, ***P < 0.001

Analysis of PSMB9 differential expression in mouse public datasets and comparison with human data

To investigate proteasome family gene expression during OA progression, we analyzed microarray data from the mouse DMM model, using the 2-week time point post-surgery as the baseline. Specific comparisons were made against later stages: 4 weeks versus 2 weeks (dataset GSE53857) and 8/16 weeks versus 2 weeks (dataset GSE41342). This multi-time-point analysis revealed that only PSMB9 exhibited sustained upregulation at the later stages (4, 8, and 16 weeks post-surgery) (Fig. 2A). Subsequently, after correcting for batch effects in human synovial datasets (GSE1919 and GSE55235), box plots and principal component analysis (PCA) confirmed more consistent expression profiles and improved dataset clustering (Fig. 2B, C). An integrated heatmap of five human and mouse datasets further demonstrated that PSMB9 was consistently upregulated in human OA samples and in the later stages (8/16 weeks) of the mouse DMM model (Fig. 2D). Together, these cross-species, multi-time-point analyses suggest that PSMB9 is actively involved in OA progression.

Fig. 2.

Analysis of PSMB9 expression in mouse public datasets and cross-species comparison. A Volcano plots showing differential expression of PSMB9 at 2, 4, 8, and 16 weeks after DMM surgery in datasets GSE38537 and GSE14342. B Box plots of expression profiles from datasets GSE1919 and GSE5235 before (upper) and after (lower) batch effect correction. C Principal component analysis (PCA) of sample clustering before (left) and after (right) batch correction. D Heatmap of PSMB9 expression across multiple datasets. NS, not significant; *P < 0.05, **P < 0.01; ***P < 0.001

Expression of PSMB9 is significantly elevated in osteoarthritis tissues and cells

To investigate the role of PSMB9 in osteoarthritis (OA), we established a DMM mouse model and examined its expression in murine and human OA cartilage and in vitro. Histological analysis (Fig. 3) revealed OA-like changes in DMM mouse cartilage, including reduced chondrocytes and disorganized arrangement on H&E staining, and markedly decreased cartilage area on Safranin O–Fast Green staining, indicating severe degradation. Human OA cartilage displayed similar pathological features, and OARSI scores were significantly higher in both mouse DMM and human OA cartilage than in respective controls (all P < 0.001; Fig. 3A). Immunohistochemistry showed upregulated PSMB9, IL-6, and MMP13, and downregulated COL2A1 in mouse OA cartilage (P < 0.001; Fig. 3C, D), with human OA cartilage showing a consistent PSMB9 upregulation (P < 0.001; Fig. 3B). In vitro, IL-1β significantly increased PSMB9 protein (P < 0.01) and mRNA (P < 0.001) in mouse primary chondrocytes (Fig. 3E, F) and PSMB9 mRNA in human primary chondrocytes (P < 0.001; Fig. 3G). In C28/I2 cells, IL-1β upregulated PSMB9 protein (P < 0.01), IL-6 and MMP13 (P < 0.001 and P < 0.01, respectively), downregulated Col2a1 (P < 0.01; Fig. 3F), and elevated PSMB9 mRNA (P < 0.001; Fig. 3H, I).

Fig. 3.

Upregulation of PSMB9 in osteoarthritic tissues and cells. (A) Hematoxylin-eosin (H&E) and Safranin O–Fast Green (S&F) staining of articular cartilage from mouse control and DMM groups, and from human control and OA cartilage. (B) Immunohistochemical (IHC) analysis of PSMB9 expression in human control and OA cartilage. (C, D) IHC analysis of PSMB9, IL-6, Col2a1, and MMP13 in sham control and DMM groups. (E) PSMB9 protein in mouse primary chondrocytes stimulated with IL-1β. (F) PSMB9 mRNA in mouse primary chondrocytes stimulated with IL-1β. (G) PSMB9 mRNA in human primary chondrocytes stimulated with IL-1β. (H) Protein levels of PSMB9, IL-6, Col2a1, and MMP13 in IL-1β-treated C28/I2 cells. (I) PSMB9 mRNA levels in C28/I2 chondrocytes following IL-1β stimulation. NS, not significant; *P < 0.05, **P < 0.01, ***P < 0.001

PSMB9 promotes apoptosis, inflammatory response, and extracellular matrix degradation in IL-1β-induced C28/I2 chondrocytes

To investigate the role of PSMB9 in OA, we overexpressed PSMB9 in C28/I2 chondrocytes via lentiviral transduction and induced an OA-like phenotype with IL-1β. The results are summarized in Fig. 4. Western blotting confirmed higher PSMB9 protein levels in the oe-PSMB9 + IL-1β group than in the oe-NC + IL-1β group (P < 0.01). PSMB9 overexpression further increased IL-6 and MMP13 (P < 0.01) and decreased Col2a1 (P < 0.001) (Fig. 4A, B). CCK-8 and EdU assays showed that IL-1β suppressed proliferation, and this inhibitory effect was further enhanced by PSMB9 overexpression (P < 0.01) (Fig. 4C–E). Flow cytometry revealed that IL-1β increased apoptosis, which was further elevated upon PSMB9 overexpression (P < 0.001) (Fig. 4F, G).

Fig. 4.

PSMB9 promotes IL-1β–induced apoptosis, inflammation, and extracellular matrix degradation in C28/I2 chondrocytes. A, B Western blot analysis of extracellular matrix degradation–related proteins (MMP13, Col2a1) and the cytokine IL-6. C, D Cell proliferation assessed by EdU assay. EdU-positive cells are shown in green; nuclei are stained with Hoechst 33,342 (blue). E Cell proliferation measured by the CCK-8 assay (OD450). F, G Apoptosis rate detected by flow cytometry. NS, not significant; *P < 0.05, **P < 0.01, ***P < 0.001

Knockdown of PSMB9 alleviates IL-1β-induced chondrocyte degradation and inflammation

To verify the role of PSMB9 in OA, we knocked it down with siRNA in IL-1β-stimulated C28/I2 cells. Western blot showed a reduced PSMB9 protein level in the si-PSMB9 + IL-1β group compared with the si-NC + IL-1β group (P < 0.001). PSMB9 knockdown also decreased IL-6 and MMP13 levels (P < 0.001) and restored Col2a1 expression (P < 0.05) (Fig. 5A, B). CCK-8 and Edu assays indicated that IL-1β inhibited proliferation, an effect reversed by PSMB9 knockdown (P < 0.01) (Fig. 5C–E). Flow cytometry showed that IL-1β increased apoptosis, and PSMB9 knockdown attenuated this effect (P < 0.001) (Fig. 5F, G).

Fig. 5.

Knockdown of PSMB9 suppresses IL-1β–induced apoptosis, inflammation, and extracellular matrix degradation in C28/I2 chondrocytes. A, B Western blot analysis of extracellular matrix–related proteins (MMP13, Col2a1) and the cytokine IL-6. C, D Cell proliferation assessed by the EdU assay. EdU-positive cells (green) and nuclei (Hoechst 33342, blue) are shown. E Cell proliferation measured by the CCK-8 assay (OD450). F, G Apoptosis rate detected by flow cytometry. NS, not significant; *P < 0.05, **P < 0.01; ***P < 0.001

PSMB9 aggravates IL-1β-induced chondrocyte injury by activating the NF-κB signaling pathway

To determine how PSMB9 affects the NF-κB pathway, we analyzed key protein expression. IL-1β treatment decreased IκBα while increasing p-IκBα, p-P65, and the p-P65/P65 ratio. PSMB9 overexpression (oe-PSMB9 + IL-1β) further reduced IκBα and elevated p-P65 levels compared with IL-1β treatment alone (Fig. 6A, B). Conversely, PSMB9 knockdown (si-PSMB9 + IL-1β) reversed these effects, restoring IκBα/p-IκBα and lowering p-P65 (Fig. 6D, E). Immunofluorescence confirmed that IL-1β increased nuclear accumulation of P65, which was enhanced by PSMB9 overexpression and reversed by its knockdown (Fig. 6C, F). Inhibition of NF-κB with JSH-23 counteracted the damaging effects of PSMB9 overexpression. The oe-PSMB9 + IL-1β + JSH-23 group showed reduced apoptosis (Fig. 7A), enhanced proliferation (Fig. 7C), lower expression of PSMB9, IL-6, and MMP13, and higher Col2a1 levels compared with the oe-PSMB9 + IL-1β group (Fig. 7B, D).

Fig. 6.

PSMB9 activates the NF-κB signaling pathway. A, B Western blot analysis of key NF-κB pathway proteins (IκBα, p-IκBα, P65, and p-P65) in C28/I2 cells overexpressing PSMB9. C Immunofluorescence detection of P65 nuclear translocation in C28/I2 cells overexpressing PSMB9. D, E Western blot analysis of key NF-κB pathway proteins (IκBα, p-IκBα, P65, and p-P65) in C28/I2 cells with PSMB9 knockdown. F Immunofluorescence detection of P65 nuclear translocation in C28/I2 cells with PSMB9 knockdown. NS, not significant; *P < 0.05, **P < 0.01; ***P < 0.001

Fig. 7.

JSH-23 alleviates chondrocyte injury by inhibiting the NF-κB pathway. A Apoptosis of C28/I2 cells treated with IL-1β and JSH-23, detected and quantified by flow cytometry. B, C Protein expression levels of cartilage matrix–related and cytokine markers in C28/I2 cells treated with IL-1β and JSH-23, with relative quantification. D Proliferation of C28/I2 cells measured by the CCK-8 assay after treatment with IL-1β and JSH-23. E, F Expression levels of PSMB9 and IL-6 in IL-1β-stimulated C28/I2 cells after IL-6 silencing, with relative quantification. NS, not significant; *P < 0.05, **P < 0.01; ***P < 0.001

Inhibition of IL-6 reduces PSMB9 expression in IL-1β-stimulated chondrocytes

To elucidate the regulatory relationship between IL-6 and PSMB9, IL-6 was knocked down with siRNA in C28/I2 cells under IL-1β stimulation to establish an in vitro OA model. Western blot analysis (Fig. 6E, F) demonstrated that, compared with the IL-1β-alone treatment group, IL-6 expression was effectively suppressed in the si-IL-6 + IL-1β group (P < 0.001), concomitant with a significant decrease in PSMB9 protein expression (P < 0.001).

Discussion

Through both in vivo and in vitro experiments, this study investigated the role and mechanisms of the immunoproteasome subunit PSMB9 in osteoarthritis (OA). Key findings show that PSMB9 is upregulated in OA cartilage. Its overexpression enhanced IL-1β-triggered inflammation, matrix degradation, apoptosis, and NF-κB activation, whereas its knockdown produced opposite effects. Mechanistically, PSMB9 partially mediates its pro-OA effects by activating the NF-κB pathway and its downstream target IL-6. These results identify PSMB9 as a novel regulator in OA, functionally linked to the NF-κB/IL-6 axis. Based on IL-6 knockdown results, we hypothesize that IL-6 may positively feedback to regulate PSMB9 expression—a potential loop requiring future validation via recombinant IL-6 stimulation and NF-κB dependency tests.

PSMB9 expression is significantly elevated in OA cartilage, consistent with its upregulation in chronic inflammatory diseases. As a key inducible immunoproteasome subunit, PSMB9 is central to the pathogenesis of various chronic inflammatory conditions [32–35]. For example, it is expressed in the cholesteatoma matrix in middle ear cholesteatoma [32], localizes to Mallory-Denk bodies in steatohepatitis [36], promotes colitis-associated carcinogenesis [34], and its inhibition attenuates inflammation in TNF-α-stimulated intestinal epithelium [35]. Notably, PSMB9 has been identified as a biomarker associated with RA fibroblasts [37], independently supporting its potential central role in joint inflammation. Although rheumatoid arthritis (RA) and osteoarthritis (OA) are distinct—RA is a chronic autoimmune-driven inflammation where the immune system attacks the synovium, whereas OA is primarily degenerative—both share key signaling pathways (e.g., NF-κB, MAPK, JAK-STAT) and inflammatory mediators (e.g., IL-6, TNF-α, MMPs) [38–42]. Thus, PSMB9 upregulation in OA chondrocytes and RA fibroblasts may reflect this “shared inflammatory pathway,” suggesting that targeting such common hubs could offer broad-spectrum therapeutic strategies for osteoarthritis.

This study delineates the positive regulatory role of PSMB9 in the NF-κB pathway. Data suggest that PSMB9 overexpression is associated with promoted phosphorylation and degradation of IκBα, along with enhanced p65 phosphorylation and its nuclear translocation. Gain- and loss-of-function experiments in OA chondrocytes further indicated that PSMB9 overexpression correlated with increased degradation of both IκBα and phosphorylated IκBα, whereas PSMB9 knockdown was associated with accumulation of IκBα and phosphorylated IκBα. These findings support the perspective that PSMB9 may contribute to NF-κB activation by accelerating IκBα degradation. Notably, prior research has suggested that the ubiquitin-like protein FAT10 can upregulate the expression of the immunoproteasome subunit LMP2 (i.e., PSMB9), thereby potentially influencing IκBα degradation efficiency and modulating NF-κB activation in renal cells [43]. As a catalytic subunit of the immunoproteasome, PSMB9 appears to selectively target IκBα stability rather than broadly increasing overall protein degradation, which hints at its potential signal-regulatory function in non-immune cells [25, 44]. Given that NF-κB overactivation is considered a hallmark of OA cartilage degeneration [45–47], PSMB9 might act as a possible upstream contributor to this process. This observation aligns with existing evidence indicating that PSMB9/LMP2 deficiency is linked to impaired ubiquitinated IκBα degradation in other contexts (e.g., in renal cells [43] and in lymphocytes of diabetic mice [48–50]), and that targeting PSMB9/LMP2 has been shown to inhibit NF-κB activation and may help overcome drug resistance in myeloma [51]. Therefore, targeting PSMB9 could represent a potential strategy for intervening in OA by exploring the inhibition of excessive NF-κB activation.

It should be noted that NF-κB pathway function is highly context-dependent. In this OA model, PSMB9-driven NF-κB activation exacerbated inflammatory damage, consistent with its classic destructive role in chronic degeneration. However, in acute infection models this relationship differs. For example, in CVB3-induced myocarditis, immunoproteasome (including PSMB9) deficiency impaired NF-κB activation but worsened inflammation and injury, as timely NF-κB activation is crucial for cell survival under proteotoxic stress [52]. This indicates that the PSMB9–NF-κB axis may have opposing effects depending on pathology, microenvironment, and disease stage, with outcomes determined by downstream signaling profiles and cellular fate decisions.

Using the NF-κB inhibitor JSH-23, we confirmed that blocking this pathway reverses deleterious phenotypes caused by PSMB9 overexpression, including reduced MMP13 and IL-6, increased COL2A1, and decreased apoptosis. Interestingly, IL-6 knockdown suppressed IL-1β-induced PSMB9 upregulation, suggesting a potential positive feedback loop: PSMB9 activates NF-κB to promote IL-6 production, and IL-6 may in turn enhance PSMB9 expression, forming a self-amplifying inflammatory circuit. Similar loops occur in chronic osteoarthritis, where synovial fibroblasts (FLS) are central drivers [53]. In RA, TNF-α and IL-17 activate FLS via NF-κB, leading to MMP secretion and further cytokine production that reactivates NF-κB, sustaining inflammation [54–60]. Notably, IL-6/STAT3 signaling can induce PSMB9 expression, as shown in HCV infection studies [60]. Our study is the first to place PSMB9 within such a regulatory network in chondrocytes, offering new insight into inflammatory signal persistence in OA.

This study demonstrates that the immunoproteasome subunit PSMB9 promotes cartilage degeneration in OA by activating the NF-κB pathway and upregulating inflammatory factors like IL-6. Notably, IL-6 knockdown reduces PSMB9 expression, suggesting a potential positive feedback loop that requires further validation (Fig. 8). While the immunoproteasome may be protective in other contexts (e.g., neuronal oxidative stress or adaptive immunity) [61, 62], this work clarifies its pathogenic role in the OA cartilage microenvironment.

Fig. 8.

Proposed mechanism: PSMB9 activates the NF-κB pathway in IL-1β-stimulated chondrocytes. Overexpression of PSMB9 may accelerate the degradation of phospho-IκBα (p-IκBα), leading to activation of the NF-κB pathway. This enhances the expression of inflammatory mediators and degradation of the cartilage matrix, thereby exacerbating osteoarthritis. IL-6 may further amplify this process through a positive feedback loop

Conclusion

This study indicates that PSMB9 is upregulated in OA cartilage and may exacerbate IL-1β-induced chondrocyte inflammation, matrix degradation, and apoptosis via activation of the NF-κB/IL-6 axis. PSMB9 could thus represent a novel regulatory factor in OA pathogenesis. Furthermore, the IL-6 knockdown data suggest a potential positive feedback loop wherein IL-6 may regulate PSMB9 expression, a hypothesis that warrants further experimental validation. These findings offer new insights into OA mechanisms and potential therapeutic targets.

Study limitations

This study has several limitations. First, the main conclusions are based on limited sample sizes and primarily cellular evidence, warranting further validation in more complex in vivo and clinical settings. Second, the animal experiments used 8-week-old young adult mice. Although this model is widely adopted for OA induction, the young and metabolically active cartilage may not fully recapitulate the biology of aged human OA, thus limiting the generalizability of our findings to aging or natural degeneration models. Accordingly, our results primarily reflect early-stage molecular events, and future studies using aged or spontaneous OA models are needed. Finally, although NF-κB pathway inhibition preliminarily confirmed the regulatory role of PSMB9, the PSMB9–IL-6 feedback loop and its precise position within the inflammatory network remain to be systematically elucidated through larger sample sizes, multi-omics approaches, and more pathophysiologically relevant models. Addressing these questions will be critical for evaluating the translational potential of PSMB9 as a therapeutic target.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. All authors had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. L.H.Z, J.L.O, and J.C. designed the study, carried out data analyses, interpreted the results, and drafted the manuscript. Z.W.W, S.J.B, X.W, S.S, X.L, K.D.B, D.Z.S were involved in collecting the data, helping with data analyses, interpreting the results, and revising the manuscript. All the authors took part in the experiment.

Funding

This work was supported by the general project of Anhui Province Outstanding Young Talents Support Program for Universities (Grant No. gxyq2022009); the Anhui Institute of translational medicine (Grant No. 2022zhyx-C90); the Foundation of Anhui Medical University (Grant No. 2020xkj209) and the Open Project of Anhui Province Key Laboratory of Occupational Health (Grant No. 2024ZYJKB003).

Data availability

All public datasets used in this study are openly available from the GEO repository under the following accession numbers:- GSE215039: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE215039] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE215039)- GSE169077: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE169077] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE169077)- GSE55235: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE55235] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE55235)- GSE1919: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE1919] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE1919)- GSE53857: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE53857] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE53857)- GSE41342: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE41342] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE41342).

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University, and written consent was obtained from all individuals participating in the study. The Declaration of Helsinki was followed for all experiments.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.