IMMT downregulation promotes osteoarthritis development by inducing mitochondrial dysfunction - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Orthop Surg Res . 2026 Mar 8;21:258. doi: 10.1186/s13018-026-06775-7 Search in PMC Search in PubMed View in NLM Catalog Add to search IMMT downregulation promotes osteoarthritis development by inducing mitochondrial dysfunction Lvlin Yang Lvlin Yang 1 Department of Orthopedics, People’s Hospital of Ningxia Hui Autonomous Region, Ningxia Medical University, No.301 Zhengyuan Street, Jinfeng District, Yinchuan, 750001 Ningxia China Find articles by Lvlin Yang 1, ✉ , Binyang Wang Binyang Wang 2 Department of Orthopedics, the Third Clinical Medical School of Ningxia Medical University, Yinchuan, 750001 Ningxia China Find articles by Binyang Wang 2 , Qing Ma Qing Ma 2 Department of Orthopedics, the Third Clinical Medical School of Ningxia Medical University, Yinchuan, 750001 Ningxia China Find articles by Qing Ma 2 , Junjie Wang Junjie Wang 2 Department of Orthopedics, the Third Clinical Medical School of Ningxia Medical University, Yinchuan, 750001 Ningxia China Find articles by Junjie Wang 2 , Yinghao Cheng Yinghao Cheng 2 Department of Orthopedics, the Third Clinical Medical School of Ningxia Medical University, Yinchuan, 750001 Ningxia China Find articles by Yinghao Cheng 2 Author information Article notes Copyright and License information 1 Department of Orthopedics, People’s Hospital of Ningxia Hui Autonomous Region, Ningxia Medical University, No.301 Zhengyuan Street, Jinfeng District, Yinchuan, 750001 Ningxia China 2 Department of Orthopedics, the Third Clinical Medical School of Ningxia Medical University, Yinchuan, 750001 Ningxia China ✉ Corresponding author. Received 2025 Dec 26; Accepted 2026 Feb 17; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13081484 PMID: 41794759 Abstract Objective This study aimed to identify key mitochondria-related genes involved in the pathogenesis of osteoarthritis (OA). Methods Publicly available OA-related gene expression datasets were analyzed. Differential expression analysis, immune infiltration analysis, weighted gene co-expression network analysis (WGCNA), and machine learning were used to identify hub differentially expressed mitochondria-related genes (DE-MRGs) for constructing a molecular signature to predict OA risk. In addition, a nomogram, protein–protein interaction (PPI) network, gene set enrichment analysis (GSEA), and a competing endogenous RNA (ceRNA) network of signature DE-MRGs were generated to assess their predictive performance and regulatory mechanisms in OA. The expression of signature DE-MRGs was validated in interleukin (IL)-1β-treated chondrocytes, and the effects of IMMT knockdown on OA pathogenesis and mitochondrial dysfunction were investigated in vitro. Results A molecular signature comprising seven DE-MRGs ( IMMT , LONP1 , TUFM , SOD2 , CYCS , CAT , and DLD ) demonstrated high predictive performance for OA. The nomogram established using these genes also exhibited high predictive accuracy for OA risk. Furthermore, GSEA revealed that IMMT was implicated in pathways such as Notch signaling. A ceRNA network involving MEG3–miR-370-3p– IMMT was predicted. Quantitative PCR confirmed significant downregulation of IMMT , TUFM , CAT , and DLD in IL-1β-treated chondrocytes. Knockdown of IMMT promoted OA development and induced mitochondrial dysfunction in chondrocytes. Conclusion The identified DE-MRG signature holds significant potential for predicting OA. The downregulation of IMMT may contribute to OA pathogenesis by inducing mitochondrial dysfunction. These findings provide a foundation for developing more effective diagnostic and therapeutic strategies for OA. Supplementary Information The online version contains supplementary material available at 10.1186/s13018-026-06775-7. Keywords: Osteoarthritis, Mitochondrial dysfunction, IMMT, Molecular signature Introduction Osteoarthritis (OA) is a prevalent degenerative musculoskeletal disorder and a leading cause of pain and disability worldwide [ 1 ]. It is characterized by synovial inflammation, subchondral bone remodeling, and progressive cartilage degradation [ 2 , 3 ]. With an estimated 303 million individuals affected globally, OA presents a significant public health burden [ 4 ]. Its prevalence continues to rise due to an aging population and increasing obesity rates [ 5 ]. Despite considerable progress in understanding OA pathophysiology, the mechanisms underlying its onset and progression remain unclear, and effective interventional therapies remain limited in clinical practice. Therefore, elucidating key molecular mechanisms involved in OA is essential for identifying novel diagnostic biomarkers and developing targeted therapeutic strategies. Mitochondria are essential organelles responsible for energy production and the regulation of various cellular metabolic processes. Increasing evidence suggests that mitochondrial dysfunction plays a pivotal role in the pathogenesis of numerous diseases, including OA [ 6 , 7 ]. Mitochondrial dysfunction is primarily characterized by reduced ATP production, increased reactive oxygen species (ROS) generation, disrupted calcium homeostasis, and mitochondrial DNA damage, all of which contribute to chondrocyte injury and promote OA development [ 8 – 10 ]. Multiple mechanisms have been implicated in mitochondrial dysfunction-mediated OA pathogenesis, including heightened oxidative stress and inflammation, disrupted chondrocyte metabolism, and impaired autophagy [ 11 , 12 ]. Several key mitochondrial genes have been linked to OA development. For example, the age-related downregulation of LONP1 has been shown to exacerbate OA by inducing mitochondrial dysfunction [ 13 ]. The AMPKα–SIRT1–PGC-1α signaling pathway, which is crucial for mitochondrial biogenesis, has been proposed as a potential therapeutic target for OA [ 14 ]. Moreover, dysregulated expression of mitophagy-related proteins such as Parkin and P62 has been observed in OA [ 15 ]. Notably, artemisinin has been reported to promote mitophagy by downregulating TNFSF11 and inhibiting PI3K/AKT/mTOR signaling in cartilage, thereby offering therapeutic benefits for OA [ 16 ]. Given these findings, elucidating the molecular mechanisms underlying mitochondrial dysfunction in OA holds significant potential for advancing both prevention and treatment strategies. However, the key mitochondria-related genes (MRGs) involved in OA pathogenesis and diagnosis remain largely unexplored. In this study, we analyzed publicly available OA-related gene expression datasets to identify hub differentially expressed MRGs (DE-MRGs) and construct a molecular signature for predicting OA risk. We further explored the predictive performance and regulatory mechanisms of these DE-MRGs in OA. The expression of signature DE-MRGs was validated in interleukin (IL)-1β-treated chondrocytes, and the functional role of a key DE-MRG was investigated in OA pathogenesis and mitochondrial dysfunction. Our findings provide new insights into the role of mitochondrial dysfunction in OA pathogenesis and identify a promising biomarker for OA diagnosis and risk prediction. Materials and methods Data acquisition and preprocessing The OA-related gene expression datasets GSE57218 , GSE51588 , and GSE117999 were retrieved from the Gene Expression Omnibus (GEO) database [ 17 ]. GSE57218 comprised 33 OA and 7 normal control (CTRL) cartilage tissue samples, GSE51588 included 40 OA and 10 CTRL cartilage tissue samples, and GSE117999 contained 10 OA and 10 CTRL cartilage tissue samples. As these datasets originated from different batches, the sva package (version 3.38.0) [ 18 ] in R (version 3.6.1) was used to correct for batch effects. The corrected datasets were then merged to form the training dataset for subsequent analyses. Additionally, the GSE114007 dataset, which includes 20 OA and 18 CTRL cartilage tissue samples, was obtained from the GEO database and used as an independent validation dataset. This dataset was generated using the GPL11154 Illumina HiSeq 2000 sequencing platform. Differential expression analysis Based on the training dataset, differentially expressed genes (DEGs) between OA and CTRL samples were identified using the limma package (version 3.34.7) [ 19 ] in R (version 3.6.1), with a false discovery rate (FDR) threshold of < 0.05 and an absolute log₂ fold change (FC) > 0.5. Bidirectional hierarchical clustering of DEGs was performed using the pheatmap package (version 1.0.8) [ 20 ] in R (version 3.6.1), and the results were visualized as a heatmap. Immune infiltration analysis The proportions of immune cell types in OA and CTRL samples were estimated using the CIBERSORT algorithm [ 21 ]. Differences in immune cell proportions between the two groups were assessed using the Kruskal–Wallis test. Immune cell types with significantly different proportions were retained, and their correlations were determined using the cor function in R (version 3.6.1). Weighted gene co-expression network analysis (WGCNA) To identify modules associated with disease status and immune cell infiltration, WGCNA (version 1.61) [ 22 ] in R (version 3.6.1) was performed on the DEGs. During WGCNA, the adjacency function and module division were defined. Modules were constructed with the following parameters: minimum module size = 30 genes and cutHeight = 0.995. Genes within each module were then extracted for further analysis. Identification of DE-MRGs Mitochondria-related genes were obtained from the Human MitoCarta3.0 database. DE-MRGs were identified as the overlapping genes among mitochondria-related genes, DEGs, and WGCNA module genes. Protein–protein interaction (PPI) network construction The STRING database (version 11.0) [ 23 ] was used to identify interactions between proteins encoded by DE-MRGs. The PPI network was constructed using Cytoscape (version 3.9.0) [ 24 ]. Functional enrichment analysis of DE-MRGs in the PPI network was conducted using the clusterProfiler package (version 4.4.4) [ 25 ] in R (version 3.6.1), with FDR < 0.05 considered statistically significant. Modules within the PPI network were identified using the Molecular Complex Detection (MCODE) plugin (version 1.4.2) [ 26 ] in Cytoscape, with the following parameters: node score cutoff = 0.2, degree cutoff = 2, and K-core = 2. Additionally, hub genes in the PPI network were identified using the cytoHubba plugin (version 0.1) [ 27 ] in Cytoscape, employing four topological analysis algorithms: MCC, MNC, DEGREE, and EPC. Genes identified as important by all four algorithms were considered hub DE-MRGs in the PPI network. Construction and validation of a molecular signature based on hub DE-MRGs Using the training dataset, three machine learning algorithms—Least Absolute Shrinkage and Selection Operator (LASSO), Recursive Feature Elimination (RFE), and Random Forest (RF)—were employed to select key gene features. Specifically, LASSO regression analysis was performed on the hub DE-MRGs using the lars package (version 1.2) [ 28 ] in R (version 3.6.1). The caret package (version 6.0–76) [ 29 ] was used to apply the RFE algorithm and identify the optimal feature combination of DE-MRGs. The RF algorithm was implemented via the randomForest package (version 4.6–14) [ 30 ] to screen feature DE-MRGs. The overlapping genes identified by all three algorithms were selected as the final DE-MRG combination and designated as signature DE-MRGs. A molecular signature was then constructed based on the expression of these signature DE-MRGs in the training dataset using a multivariate regression algorithm implemented in the rms package (version 6.3-0) [ 31 ]. The model’s performance was assessed via receiver operating characteristic (ROC) curve analysis using the pROC package (version 1.12.1) [ 32 ]. To further validate the model, its performance was evaluated using an independent dataset ( GSE114007 ). Establishment of a nomogram A nomogram was developed based on the signature DE-MRGs using the rms package (version 6.3-0) in R (version 3.6.1). The model’s predictive accuracy was evaluated using a calibration curve. Additionally, decision curve analysis (DCA) was performed using the rmda package (version 1.6) [ 33 ] to assess the net benefit of each signature DE-MRG in predicting OA risk. The nomogram was further validated using the independent dataset GSE114007 . Construction of a PPI network of signature DE-MRGs The GeneMANIA database [ 34 ] was used to identify 20 interacting proteins associated with the signature DE-MRGs to predict relationships based on co-localization, shared protein domains, co-expression, and pathway interactions. A PPI network of signature DE-MRGs was then constructed. Gene set enrichment analysis for signature DE-MRGs Gene set enrichment analysis (GSEA) [ 35 ] was conducted based on whole-genome expression levels to identify KEGG pathways significantly associated with each signature DE-MRG. The enrichment score (ES), normalized enrichment score (NES), and nominal P-value were calculated. An FDR < 0.05 was used as the threshold for significant enrichment. Prediction of small-molecule drugs Small-molecule drugs associated with signature DE-MRGs were identified from the Comparative Toxicogenomics Database (CTD, 2023 update) [ 36 ] using “osteoarthritis” as the search keyword. The relationships between signature DE-MRGs and disease-related small-molecule drugs were analyzed. Additionally, correlations between signature DE-MRGs and immune cell infiltration results were examined. Construction of a competing endogenous RNA (ceRNA) network miRNA–DE-MRG interactions were predicted using miRWalk 3.0 [ 37 ]. miRNAs associated with OA were retrieved from the HMDD v3.2 database [ 38 ], and overlapping OA-related miRNA–signature DE-MRG pairs were identified. To construct the ceRNA network, lncRNAs targeting OA-related miRNAs were identified using the DIANA-LncBase (version 3) database [ 39 ], while OA-associated lncRNAs were retrieved from the LncRNADisease v2.0 database [ 40 ]. Only OA-related lncRNAs with validated connections to OA miRNAs were retained. The final lncRNA–miRNA–signature DE-MRG interaction network was integrated to construct an OA-related ceRNA network. Chondrocyte culture and treatment Human primary cartilage chondrocytes (HUM-iCell-s018; Cellverse Co., Ltd., Shanghai, China) were cultured in RPMI-1640 (Gibco, USA) supplemented with 20% fetal bovine serum (FBS) (Gibco) and 1% penicillin–streptomycin (Gibco) in a 37 °C humidified incubator with 5% CO₂. To induce an OA-like phenotype, chondrocytes were stimulated with varying concentrations of IL-1β (0, 0.5, 1, 5, 10, and 20 ng/mL) for 24 h. Cell viability was assessed using the CCK-8 assay, and the optimal IL-1β concentration for OA induction was determined. CCK-8 assay Chondrocytes were seeded in a 96-well plate at 100 µL/well. After treatment, CCK-8 solution (10 µL; Beyotime, Jiangsu, China) was added to each well and incubated for 1–4 h. Absorbance at 450 nm was measured using a microplate reader (Epoch, BioTek, VT, USA). Quantitative PCR (qPCR) Total RNA was extracted using RNAiso Plus (TRIzol) (TaKaRa, Japan) and reverse-transcribed into cDNA. The expression levels of the seven signature DE-MRGs were quantified via qPCR using SYBR Green PCR Master Mix (Lifeint, Xiamen, China). The primer sequences used in this study are shown in Table 1 . GAPDH served as an internal control, and relative gene expression was calculated using the 2 −ΔΔCt method. Table 1. The primer sequences used in this study Gene name Primer sequence (5′ to 3′) Forward Reverse IMMT CGGGCCTGTCAGTTATCGG CAATGGACGGAGGACAAACTT LONP1 GACGATCCCCGATGTGTTTCC GGGCGAGACGAACTTTCCTT TUFM GGGGCTAAGTTCAAGAAGTACG CACATGAGCCGCATTGATGG SOD2 GCTCCGGTTTTGGGGTATCTG GCGTTGATGTGAGGTTCCAG CYCS CTTTGGGCGGAAGACAGGTC TTATTGGCGGCTGTGTAAGAG CAT TGGAGCTGGTAACCCAGTAGG CCTTTGCCTTGGAGTATTTGGTA DLD CTCATGGCCTACAGGGACTTT GCATGTTCCACCAAGTGTTTCAT GAPDH TGCAACCGGGAAGGAAATGA GCATCACCCGGAGGAGAAAT Open in a new tab Cell transfection Short hairpin RNAs (shRNAs) targeting IMMT (sh1-IMMT, GCACTATCCTATATGCCAAAT; sh2-IMMT, CAGCCTGAGGAATCTTTAA; sh3-IMMT, CCATTTCCGGGAAAGTGTA) were designed using the Designer of Small Interfering RNA website. These shRNAs and a negative control (sh-NC) were cloned into a U6-ShRNA-zsGreen-Puro vector to generate lentiviral constructs. Lentiviral particles were produced by co-transfecting the lentiviral vectors with the packaging plasmids (pMDLg/pRRE: pVSV-G: pRSV-Rev at a 5:3:2 ratio) into primary chondrocytes using Lipofectamine TM 2000 Transfection Reagent (Invitrogen, USA). After 48 h of transfection, cells were harvested for further analysis. Cell apoptosis detection by flow cytometry Following different treatments, chondrocyte apoptosis was assessed using the Annexin V-FITC Apoptosis Detection Kit (Beyotime). Cells were stained with 5 µL Annexin V-FITC and 5 µL propidium iodide for 15 min in the dark. Apoptotic cells were analyzed using a FACSCanto II flow cytometer (BD Biosciences, USA). Enzyme-linked immunosorbent assay (ELISA) Following treatment, culture supernatants from chondrocytes were collected, and the concentrations of IL-6 and TNF-α were measured using ELISA kits (Boster Immunoleader, Pleasanton, CA) according to the manufacturer’s protocol. Western blot analysis Chondrocytes were lysed in RIPA lysis buffer (Beyotime) containing phenylmethylsulfonyl fluoride (PMSF) at 4 °C for 15 min. The protein supernatant was collected by centrifugation and quantified using the BCA Protein Assay Kit (Thermo, USA). Proteins were separated on 10% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, USA). The membranes were incubated overnight at 4 °C with the following primary antibodies: IMMT (10179-1-AP; 1:5,000, Proteintech), PINK1 (ab216144; 1:1,000, Abcam, MA, USA), Parkin (ab77924; 1:2,000, Abcam), LC3 (ab63817; 1:2,000, Abcam), and GAPDH (SB-AB0037; 1:3,000, Shanghai Sheng’er Biotechnology Co., Ltd., Shanghai, China). After washing, membranes were incubated with HRP-conjugated IgG H&L secondary antibodies (ab97051;1:10,000, Abcam) at 25 °C for 1 h. Protein bands were visualized using an enhanced chemiluminescence (ECL) kit (Solarbio, Beijing, China), and band intensities were quantified using ImageJ software. Mitochondrial membrane potential (Δψm) measurement Chondrocytes were incubated with JC-1 dye at 37 °C in the dark for 15 min following different treatments. After centrifugation, cells were resuspended and analyzed by flow cytometry (BD Biosciences) using a 488 nm laser. FlowJo software was used to calculate the percentage of dye accumulation, representing mitochondrial membrane potential changes. ATP measurement ATP levels in chondrocytes were quantified using an ATP Assay Kit (Beyotime) following treatment. Luminescence intensity was measured using a luminometer according to the manufacturer’s instructions. ROS detection Intracellular ROS levels were measured using a Reactive Oxygen Species Assay Kit (Beyotime). Cells were incubated with 2′,7′-dichlorofluorescein diacetate (DCFH-DA) at 37 °C for 20 min. ROS fluorescence intensity was detected using a confocal microscope at 485 nm. Transmission electron microscopy (TEM) observation Chondrocytes were fixed with 2.5% glutaraldehyde for 3 h and 1% osmium tetroxide for 1 h in the dark. Cells were then dehydrated through a graded ethanol series (30%–100%), immersed in propylene oxide, and embedded in Epon-812 resin. Ultrathin sections were stained sequentially with uranyl acetate and lead citrate and observed using a JEM-1230 transmission electron microscope to examine mitochondrial morphology and autophagosome formation. Statistical analysis All data are presented as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA). Comparisons between groups were performed using one-way ANOVA (for multiple groups) or Student’s t -test (for two-group comparisons). P < 0.05 was considered statistically significant. Results Identification of DEGs The datasets GSE57218 , GSE51588 , and GSE117999 were processed to remove batch effects. The distribution of samples before and after batch effect correction (Fig. 1 A) demonstrated successful batch effect removal, with a uniform distribution of samples across datasets. The corrected datasets were then merged to form the training dataset, comprising 83 OA and 27 CTRL samples. Differential expression analysis identified 673 DEGs, including 309 upregulated and 364 downregulated genes. A volcano plot illustrating the distribution of DEGs (Fig. 1 B), while a heatmap of sample clustering based on DEG expression (Fig. 1 C). The clustering heatmap confirmed that OA and CTRL samples were clearly distinguishable based on DEG expression patterns. Fig. 1. Open in a new tab Identification of differentially expressed genes (DEGs) based on the GSE57218 , GSE51588 , and GSE117999 datasets. A Distribution of samples based on gene expression levels before and after batch effect removal. B Volcano plot of DEGs between OA and control samples in the merged dataset. Blue and red nodes represent significantly downregulated and upregulated DEGs, respectively. C Sample clustering heatmap based on DEG expression levels. DEGs: differentially expressed genes; OA: osteoarthritis Immune infiltration analysis To characterize the immune microenvironment, the proportions of immune cell types in OA and CTRL samples were estimated. A total of 10 immune cell types exhibited significantly different distributions between the groups, including CD8⁺ T cells, naive CD4⁺ T cells, resting memory CD4⁺ T cells, follicular helper T cells, activated NK cells, monocytes, M0 macrophages, M1 macrophages, resting myeloid dendritic cells, and neutrophils (Fig. 2 A). Correlation analysis revealed that the strongest correlation was observed between CD8⁺ T cells and M0 macrophages ( r = -0.445, P < 0.005, Fig. 2 B). Fig. 2. Open in a new tab Immune infiltration analysis, weighted gene co-expression network analysis (WGCNA), and protein–protein interaction (PPI) module analysis. A Proportion of immune cell types in OA and control samples. B Correlation analysis of 10 differentially abundant immune cell types. C Left: Adjacency matrix weight parameter power selection plot by WGCNA. The x-axis represents the weight parameter power, and the y-axis represents the square of the correlation coefficient between log(k) and log(p(k)) in the corresponding network. The red line indicates the threshold where the square of the correlation coefficient reaches 0.9. Right: Schematic diagram of average gene connectivity under different power parameters, where the red line represents the average connectivity of network nodes under the power parameter value from the left plot (equal to 1). D Module dendrogram, with each color representing a distinct module. E Heatmap showing the correlation between modules, disease states, and immune cells. F Four modules identified from the PPI network, with node border colors corresponding to WGCNA modules. WGCNA: weighted gene co-expression network analysis; PPI: protein–protein interaction; OA: osteoarthritis WGCNA WGCNA was performed on DEGs in the training dataset to construct a co-expression network associated with OA status and immune cell infiltration. DEGs exhibiting similar expression patterns were grouped into distinct modules. To achieve a scale-free network distribution, the adjacency matrix weight parameter (power) was determined, with the optimal power value set at 6, where the square of the correlation coefficient first reached 0.9 (Fig. 2 C). At this threshold, the constructed network exhibited an average node connectivity of 1, consistent with the small-world network property. Subsequently, eight modules were identified (Fig. 2 D). Correlation analysis between modules and OA status or immune cells indicated that all modules exhibited significant correlations with OA and multiple immune cell types, with absolute correlation values exceeding 0.3 (Fig. 2 E). Finally, by integrating DEGs, WGCNA module genes, and 1,136 mitochondrial genes, a total of 157 overlapping genes were identified as DE-MRGs. PPI network analysis Using the STRING database, interaction pairs among DE-MRGs with a connection score > 0.4 were identified, resulting in 464 interaction pairs. A PPI network was constructed based on these interactions (Supplementary Fig. 1 A). Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses revealed that DE-MRGs in the PPI network were significantly enriched in 259 biological processes (e.g., mitochondrial transport), 42 cellular components (e.g., mitochondrial matrix), 48 molecular functions (e.g., electron transfer activity), and 58 KEGG pathways (e.g., metabolic pathways) (Supplementary Fig. 1 B). Further analysis using the MCODE plugin identified four functional modules within the PPI network (Fig. 2 F). Genes within the same module were likely involved in similar biological processes or pathways. Finally, hub genes in the PPI network were identified using the cytoHubba plugin based on four topological analysis algorithms (MCC, MNC, DEGREE, and EPC). Comparison of the top 20 candidate hub genes from each algorithm yielded 16 overlapping genes, which were designated as hub DE-MRGs: CAT , CS , CYC1 , CYCS , DLAT , DLD , GRPEL1 , HSPD1 , HSPE1 , IMMT , LONP1 , PDHX , PHB2 , SOD2 , TIMM23 , and TUFM . Construction and validation of a molecular signature based on hub DE-MRGs Using the training dataset, 8, 12, and 15 feature DE-MRGs were identified from the 16 hub DE-MRGs via LASSO, RFE, and RF algorithms, respectively (Fig. 3 A). By comparing the results of the three algorithms, seven overlapping genes ( IMMT , LONP1 , TUFM , SOD2 , CYCS , CAT , and DLD ) were selected as the final DE-MRG combination for constructing a molecular signature. The expression heatmaps of these signature DE-MRGs in the training and validation datasets are shown in Fig. 3 B. Fig. 3. Open in a new tab Construction and validation of a molecular signature based on hub differentially expressed mitochondria-related genes (DE-MRGs). A Feature selection parameter plots using Random Forest (RF), Recursive Feature Elimination (RFE), and Least Absolute Shrinkage and Selection Operator (LASSO) algorithms. B Heatmap of signature DE-MRG expression in the training and validation datasets. C Comparison of predictive scores between OA and control samples in the training and validation datasets. D Receiver operating characteristic (ROC) curve analysis of the molecular signature for OA prediction in the training and validation datasets. E, F Nomogram constructed using seven DE-MRGs, with corresponding calibration and decision curve analysis (DCA) plots for the training and validation datasets. DE-MRGs: differentially expressed mitochondria-related genes; RF: Random Forest; RFE: Recursive Feature Elimination; LASSO: Least Absolute Shrinkage and Selection Operator; OA: osteoarthritis; ROC: receiver operating characteristic; DCA: decision curve analysis A predictive score for each sample was calculated based on the expression levels of the seven signature DE-MRGs and their corresponding coefficients. The predictive scores were significantly higher in OA samples than in CTRL samples in both the training and validation datasets ( P < 0.001, Fig. 3 C). Moreover, ROC analysis demonstrated that the area under the curve (AUC) values of the molecular signature for predicting OA were 0.941 and 0.819 in the training and validation datasets, respectively (Fig. 3 D), confirming its high predictive performance. Furthermore, nomograms were constructed based on the expression levels of the seven DE-MRGs to predict OA risk in the training (Fig. 3 E) and validation (Fig. 3 F) datasets. Calibration and DCA confirmed the high predictive accuracy of these nomograms (Fig. 3 E and F). PPI analysis of signature DE-MRGs Using the GeneMANIA database, 20 interacting proteins associated with the signature DE-MRGs were identified, and the constructed PPI network is shown in Fig. 4 A. Notably, IMMT was associated with mitochondrial protein complex functions. Fig. 4. Open in a new tab PPI network, gene set enrichment analysis (GSEA), small-molecule drug analysis, and competing endogenous RNA (ceRNA) network for signature DE-MRGs. A PPI network of signature DE-MRGs and 20 interacting proteins based on GeneMANIA. B GSEA of pathways significantly associated with IMMT. C Small-molecule drugs associated with signature DE-MRGs. D Correlation analysis of signature DE-MRGs with multiple immune cell types. E ceRNA network. Squares, triangles, and circles represent lncRNAs, miRNAs, and DE-MRGs, respectively. Red and gray lines indicate lncRNA–miRNA and miRNA–DE-MRG interactions. PPI: protein–protein interaction; GSEA: gene set enrichment analysis; DE-MRGs: differentially expressed mitochondria-related genes; ceRNA: competing endogenous RNA GSEA for signature DE-MRGs To better understand the biological pathways of signature DE-MRGs, GSEA was conducted. The results showed that IMMT , CAT , CYCS , DLD , IMMT , LONP1 , SOD2 , and TUFM were significantly enriched in 12, 14, 16, 12, 17, 14, and 8 KEGG pathways. For instance, IMMT was involved in focal adhesion and Notch signaling pathways (Fig. 4 B). Analysis of small-molecule drugs associated with signature DE-MRGs Using CTD, 31 DE-MRG–small molecule drug pairs were identified (Fig. 4 C). Among these, all seven DE-MRGs were predicted to be targeted by acetaminophen. Additionally, significant correlations were observed between signature DE-MRGs and multiple immune cell types (Fig. 4 D). ceRNA network analysis To investigate the regulatory mechanisms of signature DE-MRGs, a ceRNA network was constructed. Using the miRWalk 3.0 and HMDD v3.2 databases, 262 OA-related miRNA–signature DE-MRG pairs were identified. Furthermore, 77 OA-related long non-coding RNA (lncRNA)–miRNA pairs were retrieved from the DIANA-LncBase and LncRNADisease v2.0 databases. An OA-related lncRNA–miRNA–signature DE-MRG ceRNA network was constructed (Fig. 4 E), revealing several potential regulatory axes, including MEG3–miR-370-3p–IMMT. Validation of signature DE-MRGs in OA chondrocytes qPCR was performed to validate the differential expression of the seven signature DE-MRGs ( IMMT , LONP1 , TUFM , SOD2 , CYCS , CAT , and DLD ) in OA chondrocytes. Primary chondrocytes were treated with varying concentrations of IL-1β to induce OA. Cell viability assays indicated a significant reduction in viability following treatment with 10 and 20 ng/mL of IL-1β ( P < 0.05, Fig. 5 A). As the inhibitory effect was more pronounced at 20 ng/mL, this concentration was selected for subsequent experiments. qPCR analysis revealed that, compared to the control group, IMMT , TUFM , CAT , and DLD were significantly downregulated in IL-1β-treated chondrocytes ( P < 0.01, Fig. 5 B), consistent with the bioinformatics findings. However, LONP1 showed no significant difference, while SOD2 and CYCS were significantly upregulated ( P < 0.01, Fig. 5 B). Among IMMT , TUFM , CAT , and DLD , IMMT is a mitochondria-shaping protein crucial for maintaining mitochondrial cristae structure and remodeling [ 41 ]. Disruption of mitochondrial cristae structure can simultaneously affect various biological processes, such as energy metabolism, ROS generation, and apoptosis, all of which may influence the pathogenesis of OA [ 42 ]; and 2) previous studies have revealed that changes in IMMT expression under disease conditions are associated with cellular stress and dysfunction [ 43 ], but its role in OA has not been reported. Therefore, IMMT was selected for subsequent functional experiments. Fig. 5. Open in a new tab Construction of the OA cell model and validation of signature DE-MRGs in OA chondrocytes. A CCK-8 assay showing chondrocyte viability following treatment with different concentrations of IL-1β. B Quantitative PCR analysis of signature DE-MRG expression in OA chondrocytes. OA: osteoarthritis; DE-MRGs: differentially expressed mitochondria-related genes. * P < 0.05, ** P < 0.01, *** P < 0.001 IMMT knockdown in chondrocytes To determine whether IMMT functions as a key regulator in OA, IMMT expression was knocked down in chondrocytes via cell transfection. Quantitative PCR analysis confirmed that transfection with sh1-IMMT and sh3-IMMT significantly reduced IMMT expression in chondrocytes ( P < 0.001, Supplementary Fig. 2 A). As sh1-IMMT resulted in a greater reduction in IMMT expression than sh3-IMMT, it was selected for subsequent experiments. Western blot analysis further confirmed a significant decrease in IMMT protein expression following sh1-IMMT transfection ( P < 0.001, Supplementary Fig. 2 B), confirming successful IMMT knockdown in chondrocytes. IMMT knockdown promotes OA development To assess the effect of IMMT knockdown in an OA cell model, chondrocytes were subjected to IL-1β stimulation. Compared with the sh-NC group, the sh-NC + IL-1β group exhibited a significant reduction in cell viability (Fig. 6 A), increased apoptosis (Fig. 6 B), and elevated levels of IL-6 and TNF-α (Fig. 6 C) (all P < 0.05). Similarly, relative to the sh-NC group, the sh-IMMT group showed significantly decreased cell viability, increased apoptosis, and higher concentrations of IL-6 and TNF-α (all P < 0.05, Fig. 6 A–C). Furthermore, in comparison with the sh-NC + IL-1β group, the sh-IMMT + IL-1β group exhibited a further decrease in cell viability, along with a significant increase in apoptosis and IL-6 and TNF-α concentrations (all P < 0.05, Fig. 6 A–C). These findings indicate that IMMT knockdown promotes OA development. Fig. 6. Open in a new tab Knockdown of IMMT promotes OA development. Chondrocytes were transfected with sh-NC or sh-IMMT and/or stimulated with IL-1β. The mock group comprised untreated chondrocytes. A CCK-8 assay showing chondrocyte viability under different conditions. B Flow cytometry analysis of apoptosis following different treatments. C ELISA measurement of IL-6 and TNF-α concentrations under different conditions. * P < 0.05, ** P < 0.01, *** P < 0.001 IMMT knockdown induces mitochondrial dysfunction in chondrocytes Given that IMMT is a key inner mitochondrial membrane protein essential for maintaining mitochondrial function, we further examined the effect of IMMT knockdown on mitochondrial function in chondrocytes. Compared with the sh-NC group, IMMT knockdown significantly increased mitochondrial membrane potential ( P < 0.001, Fig. 7 A), inhibited ATP production ( P < 0.01, Fig. 7 B), and elevated ROS levels (Fig. 7 C). In addition, IMMT knockdown increased the number of mitochondrial autophagosomes (Fig. 7 D) and promoted the expression of autophagy-related proteins (PINK1, Parkin, and LC3) (all P < 0.05, Fig. 7 E). These results indicate that IMMT knockdown induces mitochondrial dysfunction and promotes autophagy in chondrocytes. Fig. 7. Open in a new tab Knockdown of IMMT induces mitochondrial dysfunction in chondrocytes. Chondrocytes were transfected with sh-NC or sh-IMMT. The mock group comprised untreated chondrocytes. A Flow cytometry with JC-1 staining to assess mitochondrial membrane potential under different conditions. B ATP assay for intracellular ATP levels following different treatments. C DCFH-DA staining to determine reactive oxygen species (ROS) concentrations. D Transmission electron microscopy images showing mitochondrial morphological changes and autophagosomes. E Western blot analysis of autophagy-related proteins (PINK1, Parkin, LC3, p62, Beclin1). * P < 0.05, ** P < 0.01, *** P < 0.001 Discussion Mitochondria are essential organelles that maintain chondrocyte homeostasis. OA is increasingly recognized as a mitochondrial disease, with mitochondrial dysfunction contributing to OA pathogenesis [ 44 ]. Therefore, MRGs may serve as promising biomarkers for OA prevention and treatment. In this study, we developed a seven-gene MRG signature, which demonstrated high predictive performance for OA risk. Moreover, a nomogram constructed using these seven DE-MRGs exhibited high predictive accuracy for OA risk. Notably, IMMT was implicated in pathways such as Notch signaling, and a ceRNA network, MEG3–miR-370-3p– IMMT , was identified. Furthermore, quantitative PCR confirmed that IMMT , TUFM , CAT , and DLD were significantly downregulated in IL-1β-treated chondrocytes. Functional analyses revealed that IMMT knockdown promoted OA development and induced mitochondrial dysfunction in chondrocytes. Accumulating evidence has indicated that changes in the immune system play an essential role in the development of OA. Increased levels of various T cell subtypes, such as CD4⁺ and CD8⁺ T cells, have been detected in patients with OA, highlighting their potential involvement in the OA development [ 45 ]. Macrophages are crucial in regulating OA progression by coordinating synovial inflammation and promoting the degradation of the cartilage matrix [ 46 ]. Neutrophils are the first immune cells to infiltrate the synovium following joint injury, and their activity is crucial for OA progression [ 47 ]. Additionally, dendritic cells in OA patients show elevated levels of inflammatory cytokines, indicating their involvement in the OA pathogenesis [ 48 ]. Consistent with these findings, we also found the involvement of various immune cells in OA samples, such as CD8⁺ T cells, naive CD4⁺ T cells, M0 macrophages, M1 macrophages, resting myeloid dendritic cells, and neutrophils, confirming the crucial role of immune cells in OA. Moreover, mitochondria are essential regulators of immune function in bone and cartilage diseases, and mitochondrial metabolism is implicated in the activation of immune cells [ 49 ]. Consequently, identifying MRGs associated with immune cell activation may help to clarify the key mechanisms underlying OA. The rapid advancements in microarray technology and bioinformatics have facilitated the high-throughput analysis of gene expression in various diseases, enhancing our understanding of their genetic underpinnings. In this study, publicly available OA-related gene expression datasets were analyzed using WGCNA and machine learning approaches to identify key MRGs associated with OA and immune cell infiltration. These methods have been widely applied to uncover genes implicated in diverse disease phenotypes [ 50 ]. As a result, seven DE-MRGs ( IMMT , LONP1 , TUFM , SOD2 , CYCS , CAT , and DLD ) were identified and used to establish a molecular signature, which exhibited high predictive performance. Furthermore, the nomograms constructed using these DE-MRGs demonstrated high predictive accuracy for OA risk, as confirmed by calibration and DCA. Given the growing interest in nomograms as effective tools for disease risk prediction [ 51 ], our findings suggest that tnomograms may serve as a valuable tool for predicting OA. Among the identified DE-MRGs, IMMT (also known as Mic60 or Mitofilin) plays a crucial role in mitochondrial integrity [ 52 ]. As a key regulatory subunit of the MICOS complex, IMMT interacts with the Sorting and Assembly Machinery (SAM) to maintain mitochondrial membrane architecture [ 53 , 54 ]. Dysregulation or depletion of IMMT has been implicated in mitochondrial dysfunction and cell death, contributing to various diseases, including degenerative encephalopathy [ 55 ] and cancer [ 56 , 57 ]. In this study, IMMT was found to be downregulated in OA samples based on public data analysis, and this finding was validated in IL-1β-treated chondrocytes. Functionally, IMMT knockdown promoted OA development by increasing cell apoptosis and inflammatory cytokine levels (IL-6 and TNF-α) while inducing mitochondrial dysfunction in chondrocytes. These results suggest that IMMT downregulation may drive OA pathogenesis by impairing mitochondrial function. Furthermore, we observed that IMMT knockdown promoted the expression of autophagy-related proteins (PINK1, Parkin, and LC3). Further co-localization analysis of mitochondria (TOM20) with lysosomes (LAMP1) or LC3 by immunofluorescence is required to detect the effect of IMMT on mitophagy. Apart from IMMT , other DE-MRGs may also play roles in OA pathogenesis. LONP1, a mitochondrial protease, has been reported to decline with age, contributing to mitochondrial dysfunction and OA pathogenesis [ 13 ]. Similarly, SOD2 depletion in chondrocytes results in increased ROS levels and reduced collagenase expression, further exacerbating OA [ 58 ]. However, the roles of TUFM , CYCS , CAT , and DLD in OA remain unclear, necessitating further investigation into their specific contributions to mitochondrial dysfunction and disease pathogenesis. To further elucidate the biological mechanisms of IMMT , we conducted GSEA, revealing its significant involvement in Notch signaling. The Notch pathway regulates cell proliferation, differentiation, and fate determination and plays a critical role in chondrocyte differentiation and osteochondral ossification [ 59 ]. Notably, Notch signaling is activated in OA cartilage, where it has been shown to exert both protective and detrimental effects on OA pathogenesis [ 60 ]. Based on our GSEA results, we speculate that IMMT may contribute to OA pathogenesis, potentially through the activation of the Notch signaling pathway; however, it is essential to detect the expression of key Notch pathway markers (e.g., Notch1, Hes1, Hey1) in the IMMT -knockdown cells to confirm the association between IMMT and Notch pathway. Additionally, our ceRNA network analysis identified the MEG3–miR-370-3p–IMMT axis as a potential upstream regulatory mechanism. lncRNA MEG3 has been implicated in OA development [ 61 , 62 ], while miR-370-3p has been identified as a key regulator in OA pathogenesis [ 63 ]. Therefore, we hypothesize that the MEG3–miR-370-3p axis may regulate IMMT expression in OA. However, our study did not include experimental validation of these interactions, necessitating further research to confirm their functional relevance. The strengths of this study include the identification of seven DE-MRGs associated with OA and the experimental validation of IMMT expression and function in chondrocytes, which provide new insights into OA pathogenesis. Furthermore, the constructed molecular signature holds potential as a predictive tool for assessing OA risk. However, this study has several limitations. First, the sample size in the public datasets was relatively small, and additional datasets with larger cohorts are required for validation. Second, there were discrepancies between the bioinformatics analysis and qPCR validation results regarding the expressions of LONP1 , SOD2 , and CYCS in IL-1β-treated chondrocytes. These differences may arise from various factors, including the inherent limitations of bioinformatics analysis that may not fully capture the complexities of gene regulation in biological systems, as well as variations in cellular environments and the specific experimental conditions employed. Despite these inconsistencies, the robustness of our seven-gene signature remains strong, as indicated by the AUC > 0.8 in both training and validation sets. Further validation across larger cohorts and different conditions are still needed to enhance the reliability of our findings. Third, the in vitro experiments were conducted exclusively in chondrocytes, and further in vivo studies are needed to explore the effects and regulatory mechanisms of IMMT in OA. Fourth, the MEG3–miR-370-3p– IMMT regulatory axis was primarily based on database predictions. The absence of direct experimental validation represents a significant limitation that needs to be addressed in future studies. Lastly, while the DE-MRG molecular signature shows good performance in the GEO datasets, its applicability to real-world clinical settings remains unclear. Factors such as potential batch effects and population heterogeneity may influence its clinical predictive value. Future studies should focus on validating the translational relevance of this signature in large-scale clinical trials to assess its real-world applicability and predictive accuracy. In conclusion, the identified DE-MRG signature has significant potential to predict OA. The downregulation of IMMT may contribute to OA pathogenesis by inducing mitochondrial dysfunction. Moreover, IMMT may be involved in OA pathogenesis potentially through Notch signaling or the MEG3–miR-370-3p–IMMT axis; however, these hypotheses require additional experimental validation. These findings provide a foundation for the development of novel diagnostic and therapeutic strategies for OA. Supplementary Information Below is the link to the electronic supplementary material. 13018_2026_6775_MOESM1_ESM.tif (20.3MB, tif) Supplementary Fig. 1: PPI network and functional enrichment analysis of DE-MRGs. (A) PPI network of DE-MRGs, with node border colors indicating WGCNA modules. (B) Significant Gene Ontology (GO) biological processes, cellular components, molecular functions, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enriched by DE-MRGs. PPI: protein–protein interaction; WGCNA: weighted gene co-expression network analysis; DE-MRGs: differentially expressed mitochondria-related genes. 13018_2026_6775_MOESM2_ESM.tif (300.9KB, tif) Supplementary Fig. 2: Knockdown efficiency of IMMT in chondrocytes. (A) Quantitative PCR analysis of IMMT mRNA expression following transfection. (B) Western blot analysis of IMMT protein expression following transfection. *** P < 0.001. Author contributions Lvlin Yang, Qing Ma carried out the conception and design of the research, Junjie Wang, Yinghao Cheng participated in the acquisition of data. Binyang Wang carried out the analysis and interpretation of data. Lvlin Yang, Yinghao Cheng participated in the design of the study and performed the statistical analysis. Lvlin Yang participated in obtaining funding. Lvlin Yang conceived of the study, and participated in its design and coordination and drafted the manuscript and revision of manuscript for important intellectual content. All authors read and approved the final manuscript.Lvlin Yang, Qing Ma carried out the conception and design of the research, Junjie Wang, Yinghao Cheng participated in the acquisition of data. Binyang Wang carried out the analysis and interpretation of data. Lvlin Yang, Yinghao Cheng participated in the design of the study and performed the statistical analysis. Lvlin Yang participated in obtaining funding. Lvlin Yang conceived of the study, and participated in its design and coordination and drafted the manuscript and revision of manuscript for important intellectual content. All authors read and approved the final manuscript. Funding This work was supported by Ningxia Hui Autonomous Region key research and development project (No. 2022BEG03101), Ningxia Natural Science Foundation Project (Youqing Project) (No. 2022AAC05051), Ningxia Hui Autonomous Region Youth Top Talent Training Project (No. 2018), Ningxia Hui Autonomous Region Youth Science and Technology Talent Recruitment Project, and Zhongwei City Health Research and Development Plan Project (No. 2024wsjk013). Data availability A mitochondria-related gene signature accurately predicts osteoarthritis risk. IMMT knockdown accelerates osteoarthritis progression and mitochondrial dysfunction. IMMT may drive osteoarthritis via the Notch signaling pathway. The MEG3–miR-370-3p–IMMT axis may regulate osteoarthritis development. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Andraskar K. Recent Advances in the Understanding and Management of Osteoarthritis: A Comprehensive Review. Journal of Internal Medicine and Pharmacology (JIMP). 2024;1(01):53–62. 10.61920/jimp.v1i01.20. 2. Yao Q, Wu X, Tao C, Gong W, Chen M, Qu M, Zhong Y, He T, Chen S, Xiao G. Osteoarthritis: pathogenic signaling pathways and therapeutic targets. Signal Transduct Target therapy. 2023;8:56. 10.1038/s41392-023-01330-w. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Martel-Pelletier J, Barr AJ, Cicuttini FM, Conaghan PG, Cooper C, Goldring MB, Goldring SR, Jones G, Teichtahl AJ, Pelletier JP, Osteoarthritis. Nat reviews Disease primers. 2016;2:16072. 10.1038/nrdp.2016.72. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Ding Y, Liu X, Chen C, Yin C, Sun X. Global, regional, and national trends in osteoarthritis disability-adjusted life years (DALYs) from 1990 to 2019: a comprehensive analysis of the global burden of disease study. Public Health. 2024;226:261–72. 10.1016/j.puhe.2023.10.030. [ DOI ] [ PubMed ] [ Google Scholar ] 5. He Y, Jiang W, Wang W. Global burden of osteoarthritis in adults aged 30 to 44 years, 1990 to 2019: results from the Global Burden of Disease Study 2019. BMC Musculoskelet Disord. 2024;25:303. 10.1186/s12891-024-07442-w. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Zhang M, Wu J, Cai K, Liu Y, Lu B, Zhang J, Xu J, Gu C, Chen T. From dysfunction to healing: advances in mitochondrial therapy for Osteoarthritis. J translational Med. 2024;22:1013. 10.1186/s12967-024-05799-z. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Liu D, Cai ZJ, Yang YT, Lu WH, Pan LY, Xiao WF, Li YS. Mitochondrial quality control in cartilage damage and osteoarthritis: new insights and potential therapeutic targets. Osteoarthr Cartil. 2022;30:395–405. 10.1016/j.joca.2021.10.009. [ DOI ] [ PubMed ] [ Google Scholar ] 8. Mao X, Fu P, Wang L, Xiang C, Mitochondria. Potential Targets for Osteoarthritis. Front Med. 2020;7:581402. 10.3389/fmed.2020.581402. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Yi D, Yu H, Lu K, Ruan C, Ding C, Tong L, Zhao X, Chen D. AMPK Signaling in Energy Control, Cartilage Biology, and Osteoarthritis. Front cell Dev biology. 2021;9:696602. 10.3389/fcell.2021.696602. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Sun K, Jing X, Guo J, Yao X, Guo F. Mitophagy in degenerative joint diseases. Autophagy. 2021;17:2082–92. 10.1080/15548627.2020.1822097. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Qi Z, Zhu J, Cai W, Lou C, Li Z. The role and intervention of mitochondrial metabolism in osteoarthritis. Mol Cell Biochem. 2024;479:1513–24. 10.1007/s11010-023-04818-9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Cheung C, Tu S, Feng Y, Wan C, Ai H, Chen Z. Mitochondrial quality control dysfunction in osteoarthritis: Mechanisms, therapeutic strategies & future prospects. Arch Gerontol Geriatr. 2024;125:105522. 10.1016/j.archger.2024.105522. [ DOI ] [ PubMed ] [ Google Scholar ] 13. He Y, Ding Q, Chen W, Lin C, Ge L, Ying C, Xu K, Wu Z, Xu L, Ran J, Chen W, Wu L. LONP1 downregulation with ageing contributes to osteoarthritis via mitochondrial dysfunction. Free Radic Biol Med. 2022;191:176–90. 10.1016/j.freeradbiomed.2022.08.038. [ DOI ] [ PubMed ] [ Google Scholar ] 14. Yang Q, Shi Y, Jin T, Duan B, Wu S. Advanced Glycation End Products Induced Mitochondrial Dysfunction of Chondrocytes through Repression of AMPKα-SIRT1-PGC-1α Pathway. Pharmacology. 2022;107:298–307. 10.1159/000521720. [ DOI ] [ PubMed ] [ Google Scholar ] 15. D’Amico D, Olmer M, Fouassier AM, Valdés P, Andreux PA, Rinsch C, Lotz M. Urolithin A improves mitochondrial health, reduces cartilage degeneration, and alleviates pain in osteoarthritis. Aging Cell. 2022;21:e13662. 10.1111/acel.13662. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Li J, Jiang M, Yu Z, Xiong C, Pan J, Cai Z, Xu N, Zhou X, Huang Y, Yang Z. Artemisinin relieves osteoarthritis by activating mitochondrial autophagy through reducing TNFSF11 expression and inhibiting PI3K/AKT/mTOR signaling in cartilage. Cell Mol Biol Lett. 2022;27:62. 10.1186/s11658-022-00365-1. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30:207–10. 10.1093/nar/30.1.207. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinf (Oxford England). 2012;28:882–3. 10.1093/bioinformatics/bts034. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43:e47. 10.1093/nar/gkv007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Wang L, Cao C, Ma Q, Zeng Q, Wang H, Cheng Z, Zhu G, Qi J, Ma H, Nian H, Wang Y. RNA-seq analyses of multiple meristems of soybean: novel and alternative transcripts, evolutionary and functional implications. BMC Plant Biol. 2014;14:169. 10.1186/1471-2229-14-169. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. 2018 Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods in molecular biology (Clifton, NJ). 1711:243–59. 10.1007/978-1-4939-7493-1_12 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. 10.1186/1471-2105-9-559. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Szklarczyk D, Morris JH, Cook H, Kuhn M, Wyder S, Simonovic M, Santos A, Doncheva NT, Roth A, Bork P, Jensen LJ, von Mering C. The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible. Nucleic Acids Res. 2017;45:D362–8. 10.1093/nar/gkw937. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498–504. 10.1101/gr.1239303. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16:284–7. 10.1089/omi.2011.0118. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Ehsani Ardakani MJ, Safaei A, Arefi Oskouie A, Haghparast H, Haghazali M, Mohaghegh Shalmani H, Peyvandi H, Naderi N, Zali MR. 2016 Evaluation of liver cirrhosis and hepatocellular carcinoma using Protein-Protein Interaction Networks. Gastroenterology and hepatology from bed to bench. 9:S14–22. [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Chin CH, Chen SH, Wu HH, Ho CW, Ko MT, Lin CY. cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol. 2014;8(Suppl 4):S11. 10.1186/1752-0509-8-s4-s11. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Goeman JJ. L1 penalized estimation in the Cox proportional hazards model. Biometrical J Biometrische Z. 2010;52:70–84. 10.1002/bimj.200900028. [ DOI ] [ PubMed ] [ Google Scholar ] 29. Deist TM, Dankers F, Valdes G, Wijsman R, Hsu IC, Oberije C, Lustberg T, van Soest J, Hoebers F, Jochems A, El Naqa I, Wee L, Morin O, Raleigh DR, Bots W, Kaanders JH, Belderbos J, Kwint M, Solberg T, Monshouwer R, Bussink J, Dekker A, Lambin P. Machine learning algorithms for outcome prediction in (chemo)radiotherapy: An empirical comparison of classifiers. Med Phys. 2018;45:3449–59. 10.1002/mp.12967. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Tolosi L, Lengauer T. Classification with correlated features: unreliability of feature ranking and solutions. Bioinf (Oxford England). 2011;27:1986–94. 10.1093/bioinformatics/btr300. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Tibshirani R. The lasso method for variable selection in the Cox model. Stat Med. 1997;16:385–95. 10.1002/(sici)1097-0258(19970228)16:4<385::aid-sim380>3.0.co;2-3. [ DOI ] [ PubMed ] [ Google Scholar ] 32. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, Müller M. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77. 10.1186/1471-2105-12-77. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Shan S, Chen W, Jia JD. Transcriptome Analysis Revealed a Highly Connected Gene Module Associated With Cirrhosis to Hepatocellular Carcinoma Development. Front Genet. 2019;10:305. 10.3389/fgene.2019.00305. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. Warde-Farley D, Donaldson SL, Comes O, Zuberi K, Badrawi R, Chao P, Franz M, Grouios C, Kazi F, Lopes CT, Maitland A, Mostafavi S, Montojo J, Shao Q, Wright G, Bader GD, Morris Q. The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function. Nucleic Acids Res. 2010;38:W214–20. 10.1093/nar/gkq537. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA. 2005;102:15545–50. 10.1073/pnas.0506580102. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Davis AP, Wiegers TC, Johnson RJ, Sciaky D, Wiegers J, Mattingly CJ. Comparative Toxicogenomics Database (CTD): update 2023. Nucleic Acids Res. 2023;51:D1257–62. 10.1093/nar/gkac833. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Hemmat N, Mokhtarzadeh A, Aghazadeh M, Jadidi-Niaragh F, Baradaran B, Bannazadeh Baghi H. Role of microRNAs in epidermal growth factor receptor signaling pathway in cervical cancer. Mol Biol Rep. 2020;47:4553–68. 10.1007/s11033-020-05494-4. [ DOI ] [ PubMed ] [ Google Scholar ] 38. Yu DL, Yu ZG, Han GS, Li J, Anh V Heterogeneous Types of miRNA-Disease Associations Stratified by Multi-Layer Network Embedding and Prediction. Biomedicines. 2021. 10.3390/biomedicines9091152. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Karagkouni D, Paraskevopoulou MD, Tastsoglou S, Skoufos G, Karavangeli A, Pierros V, Zacharopoulou E, Hatzigeorgiou AG. DIANA-LncBase v3: indexing experimentally supported miRNA targets on non-coding transcripts. Nucleic Acids Res. 2020;48:D101–10. 10.1093/nar/gkz1036. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Bao Z, Yang Z, Huang Z, Zhou Y, Cui Q, Dong D. LncRNADisease 2.0: an updated database of long non-coding RNA-associated diseases. Nucleic Acids Res. 2019;47:D1034–7. 10.1093/nar/gky905. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Tombo N, Imam Aliagan AD, Feng Y, Singh H, Bopassa JC. Cardiac ischemia/reperfusion stress reduces inner mitochondrial membrane protein (mitofilin) levels during early reperfusion. Free Radic Biol Med. 2020;158:181–94. 10.1016/j.freeradbiomed.2020.06.039. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Kan S, Duan M, Liu Y, Wang C, Xie J. Role of Mitochondria in Physiology of Chondrocytes and Diseases of Osteoarthritis and Rheumatoid Arthritis. Cartilage. 2021;13:s1102–21. 10.1177/19476035211063858. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Althurwi SI, Yu JQ, Beale P, Huq F Sequenced Combinations of Cisplatin and Selected Phytochemicals towards Overcoming Drug Resistance in Ovarian Tumour Models. Int J Mol Sci. 2020. . 10.3390/ijms21207500. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. Fernández-Moreno M, Rego-Pérez I, Blanco FJ. Is osteoarthritis a mitochondrial disease? What is the evidence. Curr Opin Rheumatol. 2022;34:46–53. 10.1097/bor.0000000000000855. [ DOI ] [ PubMed ] [ Google Scholar ] 45. Zhu W, Zhang X, Jiang Y, Liu X, Huang L, Wei Q, Huang Y, Wu W, Gu J. Alterations in peripheral T cell and B cell subsets in patients with osteoarthritis. Clin Rheumatol. 2020;39:523–32. 10.1007/s10067-019-04768-y. [ DOI ] [ PubMed ] [ Google Scholar ] 46. Hu J, Wang J. Macrophages: pivotal orchestrators and emerging therapeutic targets in osteoarthritis pathogenesis. J translational Med. 2025;23:1250. 10.1186/s12967-025-07340-2. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 47. Chaney S, Vergara R, Qiryaqoz Z, Suggs K, Akkouch A The Involvement of Neutrophils in the Pathophysiology and Treatment of Osteoarthritis. Biomedicines. 2022. . 10.3390/biomedicines10071604. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 48. Alahdal M, Zhang H, Huang R, Sun W, Deng Z, Duan L, Ouyang H, Wang D. Potential efficacy of dendritic cell immunomodulation in the treatment of osteoarthritis. Rheumatology (Oxford). 2021;60:507–17. 10.1093/rheumatology/keaa745. [ DOI ] [ PubMed ] [ Google Scholar ] 49. Mendelsohn DH, Walter N, Cheung WH, Wong RMY, Schönmehl R, Winter L, El Khassawna T, Heiss C, Brochhausen C, Rupp M. Targeting mitochondria in bone and cartilage diseases: A narrative review. Redox Biol. 2025;83:103667. 10.1016/j.redox.2025.103667. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Zhou J, Huang J, Li Z, Song Q, Yang Z, Wang L, Meng Q. Identification of aging-related biomarkers and immune infiltration characteristics in osteoarthritis based on bioinformatics analysis and machine learning. Front Immunol. 2023;14:1168780. 10.3389/fimmu.2023.1168780. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Huang S, Xie X, Sun Y, Zhang T, Cai Y, Xu X, Li H, Wu S. Development of a nomogram that predicts the risk for coronary atherosclerotic heart disease. Aging. 2020;12:9427–39. 10.18632/aging.103216. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Kozjak-Pavlovic V. The MICOS complex of human mitochondria. Cell Tissue Res. 2017;367:83–93. 10.1007/s00441-016-2433-7. [ DOI ] [ PubMed ] [ Google Scholar ] 53. Ott C, Ross K, Straub S, Thiede B, Götz M, Goosmann C, Krischke M, Mueller MJ, Krohne G, Rudel T, Kozjak-Pavlovic V. Sam50 functions in mitochondrial intermembrane space bridging and biogenesis of respiratory complexes. Mol Cell Biol. 2012;32:1173–88. 10.1128/mcb.06388-11. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Stephan T, Brüser C, Deckers M, Steyer AM, Balzarotti F, Barbot M, Behr TS, Heim G, Hübner W, Ilgen P, Lange F, Pacheu-Grau D, Pape JK, Stoldt S, Huser T, Hell SW, Möbius W, Rehling P, Riedel D, Jakobs S. MICOS assembly controls mitochondrial inner membrane remodeling and crista junction redistribution to mediate cristae formation. EMBO J. 2020;39:e104105. 10.15252/embj.2019104105. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Van Laar VS, Otero PA, Hastings TG, Berman SB. Potential Role of Mic60/Mitofilin in Parkinson’s Disease. Front NeuroSci. 2018;12:898. 10.3389/fnins.2018.00898. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Liu L, Zhao Q, Xiong D, Li D, Du J, Huang Y, Yang Y, Chen R. Suppressing mitochondrial inner membrane protein (IMMT) inhibits the proliferation of breast cancer cells through mitochondrial remodeling and metabolic regulation. Sci Rep. 2024;14:12766. 10.1038/s41598-024-63427-8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Lin HY, Wu HJ, Chu PY. Multi-omics and experimental analysis unveil theragnostic value and immunological roles of inner membrane mitochondrial protein (IMMT) in breast cancer. J translational Med. 2023;21:189. 10.1186/s12967-023-04035-4. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Scott JL, Gabrielides C, Davidson RK, Swingler TE, Clark IM, Wallis GA, Boot-Handford RP, Kirkwood TB, Taylor RW, Young DA. Superoxide dismutase downregulation in osteoarthritis progression and end-stage disease. Ann Rheum Dis. 2010;69:1502–10. 10.1136/ard.2009.119966. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 59. Wong SA, Rivera KO, Miclau T 3rd, Alsberg E, Marcucio RS, Bahney CS. Microenvironmental Regulation of Chondrocyte Plasticity in Endochondral Repair-A New Frontier for Developmental Engineering. Front Bioeng Biotechnol. 2018;6:58. 10.3389/fbioe.2018.00058. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 60. Fazio A, Di Martino A, Brunello M, Traina F, Marvi MV, Mazzotti A, Faldini C, Manzoli L, Evangelisti C, Ratti S. The involvement of signaling pathways in the pathogenesis of osteoarthritis: An update. J Orthop translation. 2024;47:116–24. 10.1016/j.jot.2024.06.002. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 61. Xiong G, Wang S, Pan Z, Liu N, Zhao D, Zha Z, Ning R. Long non-coding RNA MEG3 regulates the progress of osteoarthritis by regulating the miR-34a/Klotho axis. Annals translational Med. 2022;10:454. 10.21037/atm-22-894. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 62. Chen K, Zhu H, Zheng MQ, Dong QR. LncRNA MEG3 Inhibits the Degradation of the Extracellular Matrix of Chondrocytes in Osteoarthritis via Targeting miR-93/TGFBR2 Axis. Cartilage. 2021;13:s1274–84. 10.1177/1947603519855759. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 63. Zhou JL, Deng S, Fang HS, Peng H, Hu QJ. CircSPI1_005 ameliorates osteoarthritis by sponging miR-370-3p to regulate the expression of MAP3K9. Int Immunopharmacol. 2022;110:109064. 10.1016/j.intimp.2022.109064. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials 13018_2026_6775_MOESM1_ESM.tif (20.3MB, tif) Supplementary Fig. 1: PPI network and functional enrichment analysis of DE-MRGs. (A) PPI network of DE-MRGs, with node border colors indicating WGCNA modules. (B) Significant Gene Ontology (GO) biological processes, cellular components, molecular functions, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enriched by DE-MRGs. PPI: protein–protein interaction; WGCNA: weighted gene co-expression network analysis; DE-MRGs: differentially expressed mitochondria-related genes. 13018_2026_6775_MOESM2_ESM.tif (300.9KB, tif) Supplementary Fig. 2: Knockdown efficiency of IMMT in chondrocytes. (A) Quantitative PCR analysis of IMMT mRNA expression following transfection. (B) Western blot analysis of IMMT protein expression following transfection. *** P < 0.001. Data Availability Statement A mitochondria-related gene signature accurately predicts osteoarthritis risk. IMMT knockdown accelerates osteoarthritis progression and mitochondrial dysfunction. IMMT may drive osteoarthritis via the Notch signaling pathway. The MEG3–miR-370-3p–IMMT axis may regulate osteoarthritis development. Articles from Journal of Orthopaedic Surgery and Research are provided here courtesy of BMC ACTIONS View on publisher site PDF (6.6 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top