Organic di-selenide hydrogel microspheres for multimodal treatment of osteoarthritis

Abstract

Osteoarthritis (OA) involves multiple pathological processes and presents significant clinical challenges in treatment. Traditional therapies focus on individual factors in cartilage, synovium, or subchondral bone, limiting their ability to comprehensively address OA pathogenesis. In this study, a ROS/MMP13 dual-responsive organic selenium hydrogel microsphere (HSPHR) is developed to trigger a localized microenvironmental response specific to early OA by exploiting the disease’s pathological features. Simultaneously, the organic selenium component effectively enhances selenoprotein levels in cartilage, synovium, and subchondral bone, enabling multimodal treatment for osteoarthritis. HSPHR injections into joints reduce cartilage damage, synovial hyperplasia, and bone sclerosis in post-traumatic OA, while promoting new cartilage in defect models. It enhances selenoprotein synthesis and activates the PI3K-AKT-mTOR pathway in key cells, improving mitochondrial function and antioxidant capacity, thus reversing OA-related changes. Here, we present a multimodal therapeutic strategy for OA lesions and reveal shared regulatory pathways among different cell types. This approach offers distinct insights for the multimodal treatment of degenerative joint diseases.

Osteoarthritis encompasses multiple pathological processes making it difficult to treat. Here, the authors develop ROS/MMP13 dual-responsive organic selenium hydrogel microspheres (HSPHR) to modulate intra-articular multi-tissue selenoprotein levels, achieving multimodal treatment.

Introduction

Osteoarthritis (OA) is one of the most prevalent degenerative joint disorders, with increasing incidence and prevalence, particularly among the elderly population. The Global Burden of Disease Study 2021 reports a substantial rise in the global burden of OA from 1990 to 2021, with ~466.3 million new cases and a total prevalence of about 606.9 million cases in 2021^1,2. The pathogenesis of OA involves alterations across multiple joint structures. Cartilage matrix degradation, a hallmark of OA, is often driven by abnormal chondrocyte metabolism and overexpression of matrix-degrading enzymes, such as matrix metalloproteinases (MMPs) and members of the ADAMTS family³. Additionally, subchondral bone hyperplasia and sclerosis are central to OA progression. Changes in the subchondral bone microenvironment, including bone marrow edema and angiogenesis, have been shown to precede cartilage degeneration and are strongly linked to OA pain and progression⁴. Synovial hyperplasia, typically accompanied by low-grade inflammation, is another critical pathological feature of OA. Proliferation and fibrosis in the synovial tissue may be driven by various inflammatory factors and signaling pathways, with the role of CEMIP (KIAA1199) in synovial inflammation and fibrosis well-documented⁵. Moreover, chronic synovial inflammation can impact cartilage and subchondral bone health through the release of inflammatory mediators⁶. The pathomechanisms of OA are multifaceted, involving complex interactions among cartilage, subchondral bone, and synovium. A deeper understanding of these mechanisms is crucial for developing effective therapeutic approaches^7,8.

Mitochondrial dysfunction and oxidative stress are key features of OA. Mitochondria, as the primary source of cellular energy, generate ATP via oxidative phosphorylation (OXPHOS). However, mitochondrial impairment leads to reduced ATP production and excessive reactive oxygen species (ROS) generation. This is associated with increased mitochondrial DNA (mtDNA) damage in chondrocytes from patients with OA compared to healthy individuals⁹. Additionally, mitochondrial dysfunction in synovial tissue correlates with an imbalance between M1 and M2 macrophages in OA. M1 macrophages, in particular, produce excessive ROS, contributing to mitochondrial dysfunction and sustained inflammation¹⁰. ROS also serve as secondary messengers, triggering the differentiation of hematopoietic stem cells or monocyte/macrophage progenitor cells into osteoclasts. Osteoclast activation and abnormal bone remodeling are central to OA pathogenesis¹¹. Therefore, regulating ROS production and clearance while maintaining intracellular redox balance holds significant clinical potential for simultaneously addressing OA across multiple tissue types.

Selenoproteins, crucial antioxidants, are essential for maintaining the body’s redox balance¹². Glutathione peroxidase (GPX) and thioredoxin reductase (TrxR) are two key selenoproteins that play a pivotal role in scavenging hydrogen peroxide and lipid peroxides, thereby protecting cells from oxidative damage¹³. In recent years, there has been growing interest in the use of selenium nanoparticles in orthopedics and OA. Exploiting their antioxidant properties, bioactive materials derived from selenium have been shown to enhance selenoprotein expression, reduce excess ROS in OA cartilage, and improve the local microenvironment, thereby promoting cartilage regeneration^14–16. Despite these promising properties, the application of selenium nanoparticles in OA treatment remains limited, as their mechanisms of action are not fully understood, and effective strategies for coordinated management of multiple pathological tissues in OA are still lacking.

This study chemically conjugated selenocysteine using a thio-alkene click reaction to create a ROS/MMP13 dual-responsive organic bis-selenium hydrogel microsphere (HSPHR). This approach, designed to leverage the pathological characteristics of OA, enables early-stage, localized microenvironmental responses to the disease. At the same time, the organic selenium component replenishes selenoprotein levels in cartilage, synovium, and subperiosteal tissues. This not only provides multidimensional precision therapy for joint lesions but also offers deeper insights into the critical role of selenoproteins in OA pathogenesis. Intra-articular injection of HSPHR alleviated cartilage degeneration, improved synovial hyperplasia, and delayed subperiosteal bone sclerosis in traumatic OA models, while accelerating new cartilage formation in cartilage defect models. Compared to conventional selenium nanoparticles, this microsphere demonstrated enhanced antioxidant activity and prolonged retention in the joint cavity. Through multidimensional targeted intervention, HSPHR promotes cartilage repair and functional restoration. Mechanistically, HSPHR significantly upregulates TXNRD1 expression, activates the PI3K-AKT-mTOR signaling pathway, and shifts cellular energy metabolism, thereby restoring chondrocyte metabolic homeostasis. Notably, activation of the PI3K-AKT-mTOR pathway was also observed in synovial macrophages and osteoclasts treated with HSPHR. In summary, this study proposes a therapeutic strategy based on OA’s pathophysiological features to achieve synergistic treatment of multiple joint lesions, providing distinct insights for precision therapy in degenerative joint diseases.

Result and discussion

Role of oxidative stress imbalance mediated by selenoprotein in osteoarthritic cartilage damage

The selenoprotein family consists of proteins containing selenocysteine (Sec) residues, which play vital roles in antioxidant defense, immunomodulation, and toxin antagonism in vivo. Knee tissues from six patients, including both normal individuals and patients with OA, were analyzed for cartilage appearance and imaging. Patients with OA exhibited significant cartilage degradation and prominent osteophyte growth around the knee joint (Fig. 1A, B). Selenoprotein gene expression was measured in both normal and damaged cartilage via PCR, revealing the presence of Gpx1, Txnrd1, Txnrd2, Txnrd3, Sellnow, Sellnot, and Msrb1 in human chondrocytes. Notably, Gpx1 and Txnrd1 exhibited the most significant expression differences between normal and OA cartilage (Fig. 1C). Similar trends were observed in protein tissue levels from both normal and OA cartilage (Fig. 1D–F). Immunofluorescence analysis further demonstrated significantly lower protein levels of GPX1 and TXNRD1 in normal chondrocytes compared to OA chondrocytes, with a negative correlation to the pathological markers COL2 and MMP13 in OA (Fig. 1G, H). Histological assessments of patients with OA identified regions of relatively normal and damaged cartilage. In damaged regions, there was considerable loss of the cartilage matrix, and immunohistochemical analyses showed a marked decrease in GPX1 and TXNRD1 expression at these damaged sites. Similar observations were made in 14-week-old DMM rats (Fig. 1I–N). These results highlight a significant reduction in selenoproteins, particularly antioxidant proteins like GPX1 and TXNRD1, in OA-associated chondrocytes. The antioxidant function of selenoproteins is particularly crucial, as they maintain cellular redox balance by participating in redox reactions, thereby protecting cells from oxidative stress-induced damage. Moreover, selenoproteins play a key role in immune regulation, enhancing immune responses. A deficiency in selenoproteins is closely linked to the development of various metabolic disorders^17,18. GPX1 serves as an antioxidant enzyme, protecting articular cartilage and slowing the progression of OA by scavenging ROS. TXNRD1, involved in cellular redox homeostasis and antioxidant defense, contributes to OA progression. Inhibition of TXNRD1 has been shown to prevent cartilage extracellular matrix degradation by activating the Nrf2 signaling pathway, thereby decelerating OA progression. The synthesis of selenoproteins in the body relies on selenium participation.

Fig. 1. The integrated role of selenoproteins in osteoarthritis regulation.

A Illustrative images depicting the affected and unaffected sides in osteoarthritis (OA) patients. B X-ray imaging of OA patients versus individuals with relatively normal joint conditions. C Heatmap illustrating the expression of genes related to selenoprotein synthesis in human chondrocytes. D–F Schematic representations of selenoproteins in human chondrocytes, alongside Western Blot analyses of proteins involved in cartilage synthesis and degradation, with corresponding quantification. G, H Immunofluorescence analysis of COL2, MMP13, TXNRD1, and GPX1 in human chondrocytes, including quantitative assessment. I, J S.O. and T.B. staining in both human and rat OA and normal tissues. K–N Comparative analysis of GPX1 and TXNRD1 Immunohistochemistry staining in both human and rat OA and normal tissues, accompanied by quantification. (n = 3 independent experiments per group, The data were represented as mean ± SD. Statistical significance was determined by one-way ANOVA. Source data are provided as a Source Data file).

SeNPs counteract OA cartilage degeneration by reactivating selenoprotein-driven antioxidant pathways

The imbalance in the antioxidant capacity of chondrocytes is a key mechanism in the pathogenesis of OA. Selenoproteins, such as GPX and TXNRD1, play essential roles in cellular antioxidant pathways. A clear quantitative and qualitative association has been established between selenoprotein deficiency in cartilage tissues and OA progression. Selenium nanoparticles (SeNPs), developed based on prior studies¹⁹, were designed to directly replenish selenium in chondrocytes. Transmission electron microscopy revealed the structural morphology of SeNPs (Fig. 2A). The particle size was approximately 250 nm, consistent with dynamic light scattering measurements (Fig. 2C). Energy-dispersive X-ray (EDX) confirmed the uniform distribution of selenium on the SeNPs surface (Fig. 2B). SeNPs exhibited a Zeta potential of −11.7 mV (Fig. 2D). Cytotoxicity evaluations revealed that rat chondrocytes exposed to SeNPs at concentrations of 20 ng/mL and 50 ng/mL maintained over 97% viability, as indicated by live-dead cell staining (Fig. 2E). SeNPs at these concentrations were then administered to IL-1β-treated rat chondrocytes, resulting in upregulation of cartilage synthesis-related genes and modulation of chondrocyte degradation-related genes. This effect appeared to be linked to the concentration-dependent upregulation of Gpx1, Txnrd1, and Sephs1 genes in response to SeNPs (Fig. 2G). Immunofluorescence analysis further revealed a significant downregulation of selenoproteins in IL-1β-treated osteoarthritic chondrocytes, which was reversed by SeNPs treatment, restoring cartilage matrix metabolism (Fig. 2F, H, Supplementary Fig. 1). Additionally, SeNPs treatment increased the expression of phosphorylated selenoprotein synthase, with a concomitant dose-dependent increase in the expression of its downstream targets, GPX1 and TXNRD1 (Fig. 2I–K).

Fig. 2. Selenium nanoparticles restore selenoprotein synthesis to maintain chondrocyte matrix metabolic homeostasis.

A Transmission electron microscopy image of SeNPs. B Surface EDX spectrum of SeNPs. C Particle size distribution of SeNPs, D Zeta potential of SeNPs. E Live-dead staining images for Ctrl, 20 ng/mL selenium nanoparticles, and 50 ng/mL selenium nanoparticles. F Immunofluorescence staining of MMP13, COL2 and ROS after selenium nanoparticles treatment of inflammatory chondrocytes. G Heatmap of gene expression associated with matrix metabolism and selenoprotein synthesis in osteoarthritic chondrocytes. H Immunofluorescence staining of TXNRD1, GPX1, SEPHS1 after selenium nanoparticles treatment of inflammatory chondrocytes. I–K Western blot images of matrix metabolism and selenoprotein expression in OA chondrocytes treated with selenium nanoparticles and their quantitative statistical analysis. (n = 3 independent experiments per group, The data were represented as mean ± SD. Statistical significance was determined by one-way ANOVA. Source data are provided as a Source Data file).

To further assess whether SeNPs exert therapeutic effects in OA by restoring intra-articular microenvironmental homeostasis, the impact of SeNPs on DMM was evaluated following intra-articular injection (Supplementary Fig. 2A). Micro-CT analysis revealed that SeNPs treatment increased the number of trabecular bones in the coronal plane and expanded the volume of subchondral bone in the tibial region in the sagittal plane. Trabecular parameters, including BV/TV, Tb/Th, and Tb/Sp, indicated improved subchondral bone heterotopia after the restoration of selenoprotein synthesis (Supplementary Fig. 2B–D). H&E staining showed that the inhomogeneous articular surface was improved in the SeNPs-treated group compared to the saline group. Furthermore, the number of arthroplasts was reduced, as demonstrated by S.O. staining, and the superficial cartilage layer was thicker, with a corresponding decrease in OARSI scores in the SeNPs-treated group (Supplementary Fig. 2E–H). Further analysis of cartilage matrix composition revealed that COL2 levels were restored to normal after SeNPs treatment, while the cartilage degradation marker MMP13 was significantly downregulated (Supplementary Fig. 2I–K). Additionally, this study determined whether SeNPs rescued cartilage loss by restoring selenoprotein synthesis in chondrocytes. The number of GPX1 and TXNRD1 positive cells in joint tissues treated with SeNPs was higher than in saline-treated DMM rats. This suggests that SeNPs supplementation effectively upregulated the selenoprotein levels in chondrocytes Supplementary Fig. 2L–N). SeNPs were employed for the treatment of DMM mice to further evaluate their impact on post-traumatic OA locomotor function and willingness to move (Supplementary Fig. 3). Spontaneous activity levels were recorded during a 3-min open field test (OFT). Compared to the sham-operated group, DMM operated mice exhibited significantly reduced activity levels, active time, active distance, and average speed. Treatment with 50 ng/mL SeNPs improved the mice’s willingness to move (Supplementary Fig. 3A–E). Furthermore, pain and gait status were assessed through footprint experiments, showing that SeNPs alleviated pain in DMM mice, improving stride and step length while shortening forefoot and hindfoot footprint lengths (Supplementary Fig. 3F–H). These results suggest that SeNPs may reverse the OA process by enhancing selenoprotein synthesis in OA chondrocytes, restoring metabolic homeostasis in the intra-articular microenvironment, slowing cartilage matrix degeneration, and ultimately restoring joint function. SeNPs have demonstrated efficacy in treating lumbar disc herniation, where SeNP supplementation promoted GPX1 protein synthesis in the nucleus pulposus cells (NPCs), maintained ECM homeostasis in NPCs, and significantly enhanced mitochondrial activity¹⁹. Despite the promising attributes of SeNPs, their long-term safety and biometabolism pose certain limitations²⁰. Although these nanoparticles exhibit low toxicity in vivo, their narrow therapeutic window and precise toxicity thresholds limit their clinical applicability. As a result, several researchers have explored combining SeNPs with cartilage tissue engineering. For example, Hu W et al. developed OHA/HA-ADH@SeNPs hydrogels for the sustained release of SeNPs, targeting GPX1 activation¹⁵. Gao W et al. created Se/PRP-OGel hydrogels, utilizing zero-valent selenium nanoenzymes for radical scavenging while activating cartilage regeneration through PRP growth factors¹⁴. While these studies have broadened the scope of SeNPs applications, most have focused on the radical scavenging effects of the nanoparticles themselves, with limited exploration of the mechanisms through which selenoproteins regulate disease progression.

HSPHR facilitates the release of organic selenium in response to ROS/MMP13 stimuli

To enhance the in vivo bioavailability of elemental selenium, selenocysteine was grafted onto the hyaluronic acid (HA) chain via EDC/NHS-activated HA using thiomercapto-ene click chemistry, resulting in the synthesis of a selenocysteine-rich selenium di-selenium bonded HA-Se-Se (HS) hydrogel (Fig. 3A). To improve the environmental responsiveness of HA-Se-Se, a MMP13 responsive short peptide (MRP, KCGPQG) was incorporated into the HS via an amidation reaction, forming the HSM hydrogel. FTIR analysis revealed a broad peak at 3287 cm^–1, corresponding to the NH stretching vibration in HSM. By integrating methacryloyl glycoside (MA) modified HA (HAMA) with RGD peptide (RGDMA) and utilizing classical microfluidic techniques, complex hydrogel microspheres (HSPHR) were efficiently synthesized (Supplementary Fig. 4A), exhibiting uniform porosity with diameters ranging from 100 to 120 µm (Supplementary Fig. 4B). Scanning electron microscopy analysis revealed that the microspheres retained a loose porous structure during lyophilization, with a significant reduction in pore size compared to single-phase HAMA. The surface also exhibited modifications resembling short peptide-like structures (Fig. 3C). 77Se NMR analysis identified a single Se-Se peak at 290.23 ppm (Fig. 3B). Raman spectroscopy revealed a characteristic Se-Se metallic-like vibrational peak at 254 cm^–1 (Fig. 3D), while FTIR analysis confirmed Se-Se peaks (Fig. 3E). XPS further detected characteristic Se peaks (Fig. 3F, G). Based on the distinct characteristics of Se-Se, which responds specifically to the ROS environment, and MRP, which is responsive to MMP13, the in vitro degradation and release of Se elements from HSPHR were evaluated in the presence of H₂O₂, the MMP13 enzyme, and a combination of both. The results indicated that HSPHR exhibited accelerated degradation when exposed to both H₂O₂ and the MMP13 enzyme, with the degradation process further expedited when both agents were present simultaneously (Fig. 3I, Supplementary Fig. 4C). These results suggest the presence of two distinct degradation mechanisms responsive to the microenvironment. The release efficiency of elemental selenium was significantly enhanced within the first 24 h upon the introduction of H₂O₂ and the MMP13 enzyme. This enhancement is likely due to the combined effects of ROS-induced oxidation, which disrupts the Se-Se bond, and MMP13-mediated cleavage of the MRP peptide (Fig. 3H). The environment-responsive burst release of organic selenium not only boosts its antioxidant capacity early on but also contributes to the depletion of MMP13 in the early OA environment, thereby fostering an optimal protochondral microenvironment during the maintenance period. To track the retention of HSPHR microspheres within the joint cavity, Cy5.5-labeled HSPHR microspheres were injected into the joint cavities of both normal and OA-affected rats. A sustained release over a two-week period was achieved in vivo, with significantly high fluorescence intensities confirming the environment-responsive performance of HSPHR (Fig. 3J, K). Fluorescence intensity measurements in various organs suggested that partially absorbed nanoparticles were metabolized through the liver (Supplementary Fig. 4D). In vitro antioxidant capacity testing revealed that HSPHR exhibited superior antioxidant efficacy compared to 50 ng/mL SeNPs alone (Fig. 3L–N). The onset and progression of OA are intricately linked to the dynamics of pathological factors, such as ROS, MMPs, and other components of the joint microenvironment. While moderate selenium supplementation can effectively restore intracellular protein metabolism and intra-articular redox homeostasis, it is important to note that the therapeutic window for selenium is quite narrow²¹. Se-Se bonds have been extensively reported to efficiently respond to ROS and exhibit distinct functional properties^22–25. Building on this, a structure of selenium-diselenylated HA grafted with an MMP13-responsive peptide was developed, capable of achieving a focal dual response to ROS and MMP13, thereby intelligently facilitating selenium release. The release mechanism was achieved through a feedback loop of “environmental signaling-structural dissociation”, enabling adaptive regulation within the biological system. HSPHR demonstrated superior biosafety and enhanced in vitro antioxidant activity compared to conventional selenium supplementation methods.

Fig. 3. Preparation and characterization of ROS/MMP13 dual-responsive hydrogel microspheres.

A Schematic representation of the hydrogel microsphere preparation process. B 77Se NMR imaging of HA-Se-Se. C Electron microscopy images depicting the ultramicroscopic morphology of HAMA and HSPHR hydrogel microspheres. D Raman spectrum comparison between HSPHR and HAMA. E Fourier transform infrared spectroscopy (FTIR) imaging of HAMA and HSPHR hydrogel microspheres. F, G X-ray photoelectron spectroscopy (XPS) analysis of the elemental composition of HAMA and HSPHR hydrogel microspheres. H, I Analysis of the responsive release and degradation profiles of HSPHR hydrogel microspheres. J, K In vivo metabolic infrared imaging and quantitative analysis of HSPHR hydrogel microspheres in Sham and DMM rat models. L–N In vitro evaluation of antioxidant capacity of PBS, SeNPs and HSPHR hydrogel microspheres by ABTS, DPPH and hydroxyl radical assays. Data shown are mean ± SD (n = 3 measurements). Source data are provided as a Source Data file.

HSPHR improve cartilage metabolic homeostasis in OA by restoring mitochondrial function regulation through TXNRD1-mediated PI3K-AKT-mTOR pathway

Following a 3-day incubation period, chondrocytes cultured on HSPHR exhibited significantly elevated levels of COL2 and notably reduced expression of MMP13 compared to those cultured on HAMA hydrogel microspheres (Supplementary Fig. 5A, B, E, F). Furthermore, the expression of GPX1 and TXNRD1 was significantly upregulated (Supplementary Fig. 5C, D, G, H). Following pretreatment with IL-1β, a comparative analysis of HR microspheres, 50 ng/mL SeNPs, and HSPHR microspheres was conducted to evaluate their efficacy in enhancing selenoprotein synthesis and reversing OA in vitro. Consistent with previous results, SeNPs effectively regulated selenoprotein synthesis and maintained chondrogenic metabolic homeostasis. However, chondrocytes co-cultured with HSPHR showed more pronounced improvements, with significantly increased expression of GPX1, TXNRD1, SEPHS1, and COL2 proteins, and marked downregulation of MMP13 protein synthesis (Supplementary Fig. 5I–N).

To explore the molecular mechanisms underlying the protective effects of HSPHR on cartilage extracellular matrix homeostasis, RNA sequencing was performed on chondrocytes treated with HSPHR. Batch analysis identified 343 differentially expressed genes (DEGs), with 114 genes upregulated and 229 genes downregulated in response to HSPHR treatment (Supplementary Fig. 6A). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses revealed alterations in chondrocyte matrix metabolism and the enrichment of osteoblast-associated pathways, which were strongly correlated with the PI3K-AKT signaling pathway. Additionally, pathway enrichment analysis suggested an inflammatory-protective role of HSPHR (Supplementary Fig. 5O, P). Proteomic analysis further confirmed the upregulation of TXNRD1 and GPX1 protein expression after HSPHR treatment (Supplementary Fig. 5Q). Gene Set Enrichment Analysis (GSEA) revealed significant enrichment of glycolysis-related genes in chondrocytes (Supplementary Fig. 6B).

We investigated the roles of the TXNRD1 and GPX1 proteins in cartilage biology. In the ATDC5 chondrocyte cell line, TXNRD1 and GPX1 were individually overexpressed via transfection with overexpression plasmids or silenced using small interfering RNA (siRNA).

Chondrogenic induction assays revealed that TXNRD1 overexpression significantly enhanced chondrogenic efficiency in the presence of IL-1β (Supplementary Fig. 7A–D). Conversely, TXNRD1 silencing substantially impaired chondrogenic progression (Supplementary Fig. 7E–H). We propose that this observation is likely attributable to TXNRD1 overexpression promoting the scavenging of ROS in OA chondrocytes (Supplementary Fig. 7I–L). Therefore, TXNRD1 may play a more crucial role in cartilage protection.

Consequently, this study examined the restoration of mitochondrial function through TXNRD1, a key protein that enhances cellular antioxidant capacity and regulates the PI3K-AKT-mTOR pathway²⁶. To further investigate the downstream signaling events triggered by TXNRD1, TRi-1, a chemical inhibitor of TXNRD1²⁷, was employed. Under HSPHR treatment, TXNRD1 expression was significantly elevated, leading to the activation of the PI3K-AKT-mTOR pathway and increased phosphorylation of associated proteins. IL-1β treatment, however, significantly downregulated TXNRD1 expression and inhibited the phosphorylation of PI3K, AKT, and mTOR. Following HSPHR treatment, TXNRD1 expression returned to normal, restoring proper regulation of the PI3K-AKT-mTOR pathway. The addition of TRi-1 again suppressed this pathway (Supplementary Fig. 8A), indicating that HSPHR activates TXNRD1, which in turn initiates downstream PI3K-AKT-mTOR signaling. Further analysis of mitochondrial structural proteins, morphology, and function (Supplementary Fig. 8B, G–J) revealed that IL-1β significantly downregulated intrinsic mitochondrial proteins ATP5A, MT-ND4, COX 4, and TOMM20. HSPHR treatment restored these proteins to normal levels, while TRi-1 application reduced their expression. Similar trends were observed for mitochondrial membrane potential (MMP) (Supplementary Fig. 8D) and mitochondrial staining (Supplementary Fig. 8E). Transmission electron microscopy showed that IL-1β induced mitochondrial swelling and cristae loss, which was reversed by HSPHR treatment, while TRi-1 blocked this effect. Additionally, HSPHR restored the decreases in matrix synthetic proteins and reduced matrix degradative proteins induced by IL-1β, while enhancing the production of matrix synthetic proteins in chondrocytes. These results collectively demonstrate that HSPHR exerts its effects by upregulating TXNRD.

The regulatory role of HSPHR in the PI3K-AKT-mTOR signaling pathway was further explored using LY294002 (a PI3K phosphorylation inhibitor²⁸) and 740Y-P (a PI3K phosphorylation activator²⁹) (Fig. 4 and Supplementary Fig. 10). HSPHR-mediated activation of the PI3K-AKT-mTOR pathway restored the expression of mitochondrial-specific proteins in osteoarthritic chondrocytes (Fig. 4A, B, F–I, Supplementary Fig. 9A–F, K, L), reestablished mitochondrial morphology (Fig. 4D), and reversed the abnormal MMP (Fig. 4E; Supplementary Fig. 9M). These protective effects were similarly validated following 740Y-P administration (Supplementary Fig. 10A, B, D–R), with HSPHR demonstrating equivalent activation potential to 740Y-P. LY294002 treatment reversed these effects. Additionally, Seahorse extracellular flux analysis was employed to investigate the metabolic response of inflammatory chondrocytes to HSPHR, evaluating glycolytic stress states by real-time monitoring of the extracellular acidification rate (ECAR) following sequential glucose and 2-deoxy-d-glucose (2-DG) injections. Compared to controls, IL-1β-induced inflammatory chondrocytes exhibited significantly enhanced glycolytic activity and glycolytic reserve. However, in the HSPHR group, chondrocyte glycolytic parameters showed a progressive decline. Intervention with the PI3K inhibitor LY294002 significantly increased glycolytic reserve and accelerated mitochondrial damage (Fig. 4J, Supplementary Fig. 9Q, R). Similarly, sequential treatment effects of oligonucleotides, cyanocarbonamide, trifluoromethylbenzohydrazide (FCCP), and rotenone were recorded via real-time oxygen consumption rate (OCR) measurement. Cartilage cell OXPHOS parameters were assessed based on cellular mitochondrial stress detection (Fig. 4K, Supplementary Fig. 9N–P). Compared to the control group, the IL-1β group exhibited markedly reduced basal respiration, ATP production, and maximal respiration, while OXPHOS parameters progressively increased in the HSPHR group. Correspondingly, cellular OXPHOS significantly decreased after LY294002 intervention. These findings suggest that HSPHR incubation induces metabolic reprogramming in inflammatory chondrocytes, shifting energy production from glycolysis to OXPHOS, potentially linked to the activation of the chondrocyte PI3K-AKT-mTOR signaling pathway. Subsequent analysis of antioxidant capacity revealed that HSPHR treatment enhanced the clearance of excessive O₂^–, •OH, and other free radicals (Fig. 4L–O). Collectively, these results demonstrate that HSPHR therapy restores mitochondrial protein synthesis pathways, reestablishes mitochondrial OXPHOS, and exerts intracellular antioxidant effects, thereby mitigating cartilage damage.

Fig. 4. HSPHR activates the PI3K-AKT-mTOR pathway to restore mitochondrial oxidative phosphorylation and mitigate chondrocyte damage.

A WB bands of PI3K, P-PI3K, AKT, P-AKT, and P-mTOR WB bands. B WB bands of ATP5A, MT-ND4, COX 4, TOMM20, and LC3B WB. C WB bands of ACAN, COL2, MMP13, and ADAMTS5 bands. D TEM of mitochondria. E JC-1 staining. F TOMM20 immunofluorescence staining. G ATP5A immunofluorescence staining. H COX IV immunofluorescence staining. I MT-ND4 immunofluorescence staining. J Real-time ECARs of chondrocytes during glycolytic stress testing. K Real-time OCR in chondrocytes during mitochondrial stress testing. L–O Scavenging efficiency of chondrocytes against three typical free radicals PTIO, ABTS, •OH, and DPPH. (n = 4 independent experiments per group, The data were represented as mean ± SD. Statistical significance was determined by one-way ANOVA. Source data are provided as a Source Data file).

The precise role of the PI3K-AKT signaling pathway in maintaining articular cartilage homeostasis and its involvement in OA progression remain unclear. Various in vitro studies have provided conflicting findings on the function of AKT signaling in cartilage homeostasis^30,31. While sustained AKT activation may compromise articular cartilage integrity, other studies suggest that AKT activation exerts a protective effect on chondrocytes³². Additionally, AKT pathway activation increases intracellular ATP levels and promotes chondrogenic differentiation^33,34. The PI3K/AKT signaling pathway has been shown to protect chondrocytes by regulating survival, proliferation, and extracellular matrix synthesis, while downstream activation of mTOR promotes coenzyme Q10 production through the mTORC1-HMGCR signaling pathway³⁵. In this section, the benefits of HSPHR in selenium supplementation and its potential mechanisms of action were investigated. Co-culture with HSPHR led to the upregulation of intracellular TXNRD1 expression, restoring mitochondrial structure and function by activating the PI3K-AKT-mTOR pathway. This shift in energy metabolism towards OXPHOS exerted a protective effect on chondrocytes. Mechanistically, TXNRD1 accelerates PTEN degradation by disrupting the interaction between Trx1 and phosphatase and tensin homolog (PTEN), thereby activating the PI3K-AKT-mTOR signaling pathway²⁶. Notably, HSPHR’s activation of PI3K was comparable to that of 740Y-P, indicating that the bioactive material developed in this study can replace small-molecule chemotherapeutic agents. This strategy reduces toxicity associated with drug metabolism and holds greater promise for clinical application.

HSPHR modulate the joint microenvironment, restoring bone metabolic homeostasis and macrophage polarization

In OA pathogenesis, the imbalance in the synovial microenvironment and abnormal bone metabolism are central factors. Synovial macrophages directly induce chondrocyte catabolism by secreting pro-inflammatory factors such as IL-1β and TNF-α, which not only upregulate MMP-13 and ADAMTS-5 expression but also activate osteoclast precursors, promoting their differentiation into mature osteoclasts. Overactivation of osteoclasts through the RANKL/RANK signaling pathway leads to abnormal subchondral bone resorption, disrupting the mechanical balance of the bone-cartilage unit and releasing bone-derived factors, thereby establishing a vicious cycle of “inflammation-bone destruction-bone remodeling.” Our study demonstrated that HSPHR treatment significantly impacted the inflammatory signaling pathways in chondrocytes and bone mineralization. To further investigate the regulatory effects of HSPHR on these two cell types, 50 ng/mL RANKL was initially used to induce osteoclast differentiation in RAW264.7 cells. HSPHR treatment effectively inhibited osteoclast activity and prevented further osteoblastic fusion of the cells (Supplementary Fig. 11A, B). Additionally, markers associated with osteoblastic activity, including CTSK, MMP9, and CD40L, were significantly reduced (Supplementary Fig. 11C, D). Evaluation of subchondral bone lesions in rats two weeks after DMM revealed excessive osteoclast activation in the DMM group at the two-week postoperative time point. HSPHR treatment effectively suppressed this process, restoring the bone density that was initially reduced (Supplementary Fig. 11E, F, I–L). The mechanism may involve the upregulation of selenoproteins within osteoclast precursor cells, accelerating the clearance of ROS generated by cartilage tissue, thereby inhibiting osteoclast activation in subchondral bone (Supplementary Fig. 11C, G, H). An alternative explanation is a direct effect of HSPHR on osteoclasts. Studies suggest that increased selenoprotein synthesis can directly influence osteoclast activity and reverse their activation process³⁶. These findings highlight that HSPHR treatment significantly impacts inflammatory signaling pathways, bone mineralization, and metabolism, indicating its potential to restore bone metabolic homeostasis by inhibiting osteoclast formation.

Transcriptome sequencing was performed on osteoclasts treated with HSPHR to elucidate the mechanism by which HSPHR protects against abnormal subchondral bone ossification. The analysis identified 3202 DEGs. Notably, upregulation of the Pik3cg gene indicated activation of the PI3K-AKT-mTOR signaling pathway (Fig. 5A). KEGG and GSEA enrichment analyses further highlighted the critical role of the PI3K-AKT pathway (Fig. 5B, C). Selective activation and inhibition of the PI3K-AKT-mTOR pathway in osteoclasts using 740Y-P and LY294002 demonstrated that pathway inactivation effectively promoted osteoclast activation (Fig. 5D–G, I–L), while pathway activation inhibited osteoclast activation (Fig. 5H, M, Supplementary Fig. 12).

Fig. 5. HSPHR regulates osteoclast activation through the PI3K-AKT-mTOR signaling pathway.

A Volcano plot of differentially expressed genes between HSPHR and CTRL. B KEGG pathways enriched by differentially expressed genes between HSPHR and CTRL. C ESGA plot of KEGG pathways enriched. D–F Representative images and quantitative analysis of Trap staining and β-Actin staining after HSPHR treatment of osteoclasts. G, E, I–L Representative WB images and quantitative analysis of CTSK, CD40L, P-PI3K, P-AKT, and P-mTOR following HSPHR treatment of osteoclasts, along with quantitative statistical analysis. H, M Trap staining and representative Western blot images of CD206, CD86, ARG1, NOS2, P-PI3K, P-AKT, and P-mTOR following HSPHR and 740Y-P treatment of osteoclasts. (n = 4 independent experiments per group, The data were represented as mean ± SD. Statistical significance was determined by one-way ANOVA. Source data are provided as a Source Data file.).

Additionally, the regulatory effect of HSPHR on synovial macrophages was investigated. Flow cytometry (FCA) revealed significant changes in macrophage polarization following HSPHR treatment, with a shift from the M1 to the M2 phenotype (Supplementary Fig. 13A–D). This was further confirmed by immunofluorescence co-staining (Supplementary Fig. 13E). Transcriptome sequencing of HSPHR-treated M1 macrophages revealed significant enrichment of the PI3K-AKT pathway as well (Supplementary Fig. 14A–C). Selective activation and inhibition of the PI3K-AKT-mTOR pathway in macrophages using 740Y-P and LY294002, respectively, showed that inactivation of this pathway exacerbated the inflammatory state of macrophages (Supplementary Fig. 14D–Q), while pathway activation promoted anti-inflammatory differentiation (Supplementary Fig. 15A–N).

Osteoclast hyperactivation significantly contributes to bone diseases linked to increased oxidative stress, which is a hallmark of early osteolytic destruction in OA³⁷. Various selenium compounds exhibit distinct effects and mechanisms in anti-osteopathy treatment. Selenium supplementation boosts selenoproteins, counteracting the high oxidative stress observed during osteoclastogenesis. Macrophage polarization plays a pivotal role in OA progression. Dysregulated M1 macrophages contribute to inflammation and cartilage degradation, whereas M2 macrophages promote cartilage repair by secreting anti-inflammatory cytokines and growth factors, making M2 polarization a promising therapeutic strategy for OA³⁸. The PI3K-AKT-mTOR pathway is essential for regulating cell growth, survival, and metabolism. In osteoclasts, inhibition of this pathway suppresses osteoclast activation by restoring mitochondrial function and depleting excess intracellular ROS. In M1-polarized macrophages, activation of the PI3K-AKT pathway suppresses chemokine production and promotes M2 polarization^39,40. This research investigates the function of HSPHR in osteoclasts and macrophages, demonstrating that HSPHR mitigates oxidative stress in osteoclasts, inhibiting their formation. Simultaneously, HSPHR modulates the immune microenvironment of macrophages, reversing the polarization of inflammatory macrophages. Additionally, evidence suggests a potential interaction between synovial tissue, cartilage, and subchondral bone mediated by selenoproteins in the regulation of OA, warranting further exploration of this interaction.

HSPHR enables full-cycle treatment of OA pathology, promotes cartilage regeneration, and slows the progression of OA

To evaluate the effect of HSPHR on delaying the progression of OA in vivo, a medial meniscus DMM-induced OA model was established (Fig. 6A). Histological analysis showed cartilage erosion and destruction in the saline-treated DMM group eight weeks post-surgery. Injection of HR microspheres alone partially restored glycosaminoglycan levels. The therapeutic efficacy was further enhanced by the precise and efficient delivery of HSPHR (Fig. 6B–E). HSPHR exhibited dual responsiveness to ROS and MMP13 in the early stages of OA. This intelligent function promoted COL2 synthesis while effectively inhibiting MMP13 levels (Fig. 6F–H). Additionally, HSPHR enhanced the endogenous synthesis and expression of GPX1 and TXNRD1 proteins (Fig. 6I–K), which was associated with downstream activation of the PI3K-AKT pathway and induction of mitochondrial autophagy (Fig. 6L–N). The impact of HSPHR on synovial inflammation following DMM surgery was also assessed. HSPHR treatment significantly reduced synovial inflammation, downregulated the immune-regulation marker INOS, and increased IL-10 secretion in DMM rats (Supplementary Fig. 13F–H, K, L). These findings may be attributed to the upregulation of selenoproteins in inflamed synovial tissue (Supplementary Fig. 13I, J, M, N). Motor function in DMM mice treated with HSPHR was evaluated next. The results showed significantly reduced relative activity levels, active time, active distance, and average velocity in DMM mice compared to the sham-operated group. HSPHR treatment reversed these trends (Supplementary Fig. 16A–E), improving stride and step length while reducing forefoot and hindfoot footprint dimensions (Supplementary Fig. 16J, K). Micro-CT analysis demonstrated that HSPHR effectively prevented the development of subchondral bone sclerosis (Supplementary Fig. 16F–H). These findings indicate that HSPHR has the capacity to address OA through a multidimensional approach, restoring metabolic homeostasis within cartilage mechanisms.

Fig. 6. Evaluation of the efficacy of HSPHR microspheres in early osteoarthritis management.

A Schematic illustration depicting the intra-articular injection of hydrogel microspheres into the knee joint cavity. B–E Images of S.O., toluidine blue, hematoxylin, and eosin-stained sections of rat knee joint sections from different treatment subgroups and their quantitative statistical analysis. F–K Representative images of COL2, MMP13, TXNRD1, GPX1 immunohistochemical staining of knee joint sections from different treatment subgroups of rats and their quantitative statistical analysis. L–N Representative images of p-mTOR, p-AKT immunofluorescence staining of knee joint sections from different treatment subgroups of rats and their quantitative statistical analysis. (n = 8 rats per group). The data were represented as mean ± SD. Statistical significance was determined by one-way ANOVA. Source data are provided as a Source Data file).

To further investigate the beneficial effects of HSPHR on cartilage regeneration, an in vivo model of full-thickness cartilage defects was developed. Chondrogenic progenitor cells (CPCs) were incorporated into HSPHR microspheres, following our established treatment protocol for advanced cartilage defects⁴¹. Gross examination revealed limited regenerative capacity in both the 4-week and 8-week postoperative saline-treated groups, as evidenced by the presence of cavities and irregular margins (Supplementary Fig. 17B, H). The enhanced protochondral formation potential of HSPHR facilitated accelerated cartilage formation, attributed to its unique cell-loading properties, which allowed controlled selenium release in response to injury localization. This process resulted in optimal tissue integration and the development of hyaline-like cartilage morphology (Supplementary Fig. 17C, I). In contrast, saline treatment failed to regenerate collagen for the extracellular matrix components of the newly formed cartilage. However, HSPHR treatment successfully restored the natural hyaline cartilage-like ECM, as demonstrated by increased expression of COL2 and suppression of the fibrocartilage marker COL1 (Supplementary Fig. 17E–G, K–M). Mechanistically, HSPHR effectively modulates chondrocyte energy metabolism through the PI3K-AKT-mTOR pathway, restoring mitochondrial function and promoting cartilage regeneration. Following HSPHR treatment, no significant abnormalities were observed in vital organs (heart, liver, spleen, lungs, and kidneys) (Supplementary Fig. 18A) or blood biochemical parameters (Supplementary Fig. 18B, C). In the early stages of OA, inflammatory dysregulation leads to an imbalance in the joint microenvironment and excessive osteoclast activation. In advanced stages, irreversible cartilage destruction, periosteal bone remodeling, and abnormal osteoblast differentiation occur. Single-target treatment strategies for arthritis cannot fundamentally slow OA progression. Effective treatment requires multi-tissue, multidimensional intervention. Microfluidic technology is widely used in OA treatment with adaptable hydrogel systems^42–44. It efficiently loads small-molecule drugs, nanoparticles, exosomes, and genetically engineered viruses^43,45. In this study, HSPHR microspheres are tailored to OA pathology, integrating bioactive components organic selenium seamlessly with the hydrogel system. Multi-effect regulation of the cartilage-synovium-subchondral bone interaction offers greater adaptability in arthritis therapy. HSPHR can slow subchondral bone loss in early arthritis by inhibiting osteoclast activity and modulating the joint immune environment. In advanced stages, following irreversible cartilage matrix loss, HSPHR supplementation can reprogram chondrocyte energy metabolism and restore proliferative capacity, enabling redifferentiation into hyaline cartilage. As shown in Fig. 7, this strategy achieves comprehensive, full-cycle, and multi-pathogenic treatment for OA.

Fig. 7. Schematic illustration of HSPHR preparation.

The dual-response design delivers organic selenium to multiple joint tissues, enabling multimodal treatment for osteoarthritis.

Methods

SeNPs preparation

SeNPs were synthesized via microbial reduction using sodium selenite (Na₂SeO₃) as the precursor⁴⁶. Saccharomyces cerevisiae was first inoculated into a yeast extract-peptone-dextrose (YPD) medium, containing 1% yeast extract, 2% peptone, and 2% glucose, and incubated in a shaker at 30 °C for 12 h to activate the culture. Following this, 10% of the culture was transferred to a shaker containing 30 mL of medium supplemented with 100 g/L mother liquor and incubated for another 12 h. Sodium selenite was then added to achieve a final concentration of 2 g/L in the medium. Distilled water was added, and the resulting precipitate was homogenized. The mixture was incubated at 30 °C for 48 h, followed by autolysis at 40 °C for 24 h. Finally, the SeNPs were isolated using ultrasonic disruption and subsequent centrifugation.

Preparation of HSP

To prepare a solution of HA, 100 mg of HA was dissolved in an appropriate volume of water and sonicated for 30 min. The solution was then dialyzed in 0.001 M dilute hydrochloric acid for 4 h to achieve acidification. The acidified HA was subsequently dissolved in dimethyl sulfoxide (DMSO), with the addition of 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxysuccinimide (NHS) for activation, allowing the reaction to proceed for 6 h. Selenocystamine and an MMP13-responsive peptide (KCGPQG) were then dissolved in the activated HA solution, maintaining a weight ratio of selenocystamine, HA, and short peptides at 2:2:1. The mixture was sonicated for 30 min and allowed to react for 12 h to facilitate the formation of diselenide bonds. The resulting DMSO solution was placed in a dialysis bag with an appropriate molecular weight cutoff and subjected to aqueous dialysis, followed by lyophilization for 48 h to obtain the hyaluronan monomer containing diselenide bonds.

Preparation of HAMA

The synthesis of HAMA was conducted following previously established protocols. Briefly, 4 mL of methacrylic anhydride (MA) was gradually added to 100 mL of a 2 wt.% HA solution. Subsequently, 5 M NaOH was incrementally introduced up to 4 mL using a micro-syringe pump, and the reaction was allowed to proceed overnight under ice bath conditions. After completion, the reaction mixture was dialyzed against ultrapure water for three days using a dialysis membrane with a molecular weight cutoff of 3500 Da to remove residual reactants. The purified product was then freeze-dried and stored at −20 °C for future use.

Preparation of microspheres

The three-way connector was interfaced with two independent microfluidic syringe pumps via hoses. The external phase, designated as the continuous phase, was used to carry the internal phase, which consisted of a mixture of HAMA, HSP, and RGD-MA (Engineering For Life, China) solution, including a photoinitiator, serving as the dispersed phase. Two types of microspheres were synthesized: HAMA microspheres and composite HSR microspheres, with the latter comprising HAMA, HSP, and RGD-MA in a mass ratio of 5:5:1. The aqueous and oil phases were sequentially introduced into the microfluidic device inlet at a flow rate ratio of 15:500, facilitating a photocrosslinking reaction under UV irradiation to solidify the microsphere structure. The collected microspheres were then washed repeatedly with 75% ethanol to remove any residues of the continuous phase. Following this, the microspheres were transferred to a freshly prepared PBS buffer for 4 h, a process repeated three times to ensure thorough removal of residual additives. To create a porous microsphere structure, the purified microspheres were first subjected to freeze treatment at −80 °C for 4 h, followed by lyophilization in a freeze-dryer for 48 h. After drying, the microspheres were stored in a freeze-dryer.

Physical characterization of the microspheres

The morphology and dimensions of the microspheres were examined and documented using optical microscopy (CX23, Olympus, Japan). Particle size was measured using ImageJ-win64 software, followed by statistical analysis. The lyophilized microspheres underwent gold sputter coating, and their structures were analyzed using scanning electron microscopy (SEM, SU8100, Hitachi, Japan). Pore sizes on the microsphere surfaces were quantified using ImageJ-win64. Comprehensive analysis of the structural properties of the synthesized microspheres and nanoparticles was conducted using Fourier Transform Infrared Spectroscopy (FTIR, Thermo Fisher Scientific, USA). Samples were prepared with potassium bromide pellets and scanned 128 times at a resolution of 4 cm^–1 across a spectral range of 400–4000 cm^–1. The surface elemental composition of the microspheres was analyzed using Energy Dispersive Spectroscopy (EDS, EDAX, USA). Changes in the surface characteristics of the microspheres and nanoparticles were assessed using X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher Scientific, USA). The experimental parameters were carefully configured: the vacuum within the analysis chamber was maintained at 10^–9 mBar, with Al Kα radiation (hv = 1486.68 eV) used as the excitation source. The X-ray spot diameter was set to 500 μm, and the operational voltage was set at 15 kV, with a filament current of 10 mA. The signal accumulation spanned 5 to 10 cycles, and the energy passband was set to 30 eV with an incremental step width of 0.05 eV. Data analysis was performed through peak fitting using Avantage software, with the C1s binding energy standard value of 284.80 eV as the reference point for charge correction.

Antioxidant capacity of the microsphere

The antioxidant capacity of the microspheres was evaluated by assessing the scavenging activity against DPPH radicals, O^-• radicals, and •OH radicals. Microspheres from each experimental group were homogenized using a tissue grinder. A 100 μM DPPH solution was prepared, and various samples were mixed and dispersed in 2.0 mL of ethanol, with PBS as the blank control. The mixture was incubated for 30 min at 37 °C in the absence of light. Absorbance was measured at 515 nm using a UV-Vis spectrophotometer.

Next, 100 μL of either HSRHR microsphere or SeNPs solution was combined with 100 μL of a 100 μM ABTS solution. After thorough mixing, absorbance was measured at 734 nm to evaluate ABTS radical scavenging activity.

The scavenging ability of the microspheres against hydroxyl radicals (•OH) generated through the Fenton reaction was assessed by mixing 100 mg of the sample with 500 μL of a 1 mM FeSO₄ solution and 500 μL of a 100 mM H₂O₂ solution. The mixture was incubated at 37 °C for 1 h. After cooling to 25 °C, 100 μL of the supernatant was combined with 100 μL of a 10 mM 3,3′,5,5′-tetramethylbenzidine (TMB) solution (dissolved in DMSO). Absorbance was measured at 536 nm using a microplate reader.

Degradation of microspheres

For microsphere degradation, microspheres containing an equivalent mass were weighed and placed into a 6-well plate. These were incubated with 3 mL of a composite solution consisting of MMP13 protein (5 units/mL), 0.1% H₂O₂, and a combination of MMP13 protein with 0.1% H₂O₂. The mixture was incubated at 37 °C with shaking at 120 rpm. At predetermined time intervals, the microspheres were collected, examined, and photographed to assess their morphology using an optical microscope. After observation, the microspheres were washed multiple times with sterile deionized water, lyophilized, and weighed. The degradation rate was calculated using the following formula:

$$DP\left(\%\right)=\left(W_0-W_t\right)/W_0\times 100\%$$

Where W₀ is the weight of the initial lyophilized sample and W_t is the weight of the lyophilized sample at time t.

Drug release experiments

Microspheres containing an equivalent mass of the drug were weighed and placed into a 6-well plate. These were incubated with 3 mL of a composite solution consisting of MMP13 protein (5 units/mL), 0.1% H₂O₂, and a combination of MMP13 protein with 0.1% H₂O₂. The concentration of organic valence selenium was measured at predetermined intervals (6, 12, 18, and 24 h, as well as 1, 3, 5, 7, 14, and 21 days) by extracting 1.5 mL of the sample solution, which was then supplemented with an equal volume of buffer and added back to the plate. Supernatants were analyzed spectrophotometrically at 490 nm using a Bio-Tek spectrophotometer (Winooski, VT, USA). A standard curve for selenium radical concentration was generated under identical conditions, and the drug concentration was determined from the obtained absorbance values.

Human cartilage samples assay

Articular cartilage specimens were obtained from patients undergoing total knee arthroplasty, with informed consent. The study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (Approval No. 2025066). A total of six patients participated, with detailed information available in Supplementary Table 1. OA was diagnosed based on radiographic evaluation of the lower limbs. Cartilage samples were categorized into relatively normal (N) and OA sections according to the extent of tissue damage. For subsequent protein and RNA extraction, the cartilage samples were rapidly frozen in liquid nitrogen and stored at −80 °C.

Histological analysis

Human cartilage tissue and rat knee joints were fixed in 10% neutral formalin for 24 h and decalcified in an EDTA solution (pH 7.4) for 4 weeks. After decalcification, the tissues were embedded in paraffin, and the resulting wax blocks were sectioned into 5 µm serial sections for Safranin O/fast green staining. Sections were collected starting from the point where the medial meniscus began to detach, continuing until the meniscus was no longer visible, yielding approximately 50 sections per sample. Cartilage destruction was evaluated by two independent observers under blinded conditions using the Osteoarthritis Research Society International (OARSI) scoring system (grades 0–6) on Safranin O/fast green-stained sections.

Live/Dead cell staining assay

Chondrocytes were treated with SeNPs at concentrations of 25 and 50 ng/mL in 24-well plates. Following treatment, Live/Dead cell staining was performed using reagents from Thermo Fisher Scientific, USA (Waltham, MA, USA) for 20 min. Viable cells were stained green, while non-viable cells were stained red. Fluorescence microscopy (Zeiss, Oberkochen, Germany) was employed to capture images of the stained cells.

Fluorescence measurement of mitochondrial morphology

Mitochondrial detection in chondrocytes was performed using MitoTracker Red (Beyotime, China). Prior to analysis, cells were stained with 200 nM MitoTracker Red for 15 min. Mitochondrial morphology was then observed using an orthogonal fluorescence microscope.

Mitochondrial membrane potential (MMP) level

The MMP was assessed using the Mitochondrial Membrane Potential Assay Kit with JC-1, obtained from Beyotime, China. Chondrocytes were exposed to 0.5 µM JC-1 working solution for 30 min at 37 °C, in the dark, followed by incubation with cold JC-1 staining buffer. Fluorescent images were captured using a fluorescence microscope, and the MMP value was determined by calculating the ratio of red fluorescence intensity (representing JC-1 aggregates) to green fluorescence intensity (representing JC-1 monomers) using ImageJ-win64 software.

Metabolic assays using the XFe96 extracellular flux analyzer

ECAR and OCR measurements were conducted using a Seahorse XFe96 Extracellular Flux Analyzer and a Seahorse XF Glycolytic Rate Assay Kit (Agilent Technologies), following the manufacturer’s instructions⁴⁷. Briefly, 2 × 10⁵ cells per well were plated in XF96 Cell Culture Microplates coated with Cell-Tak. The cells were then incubated in XF RPMI medium (without phenol red) supplemented with 2 mM glutamine, 10 mM glucose, 1 mM pyruvate, and 5 mM HEPES for 45 min at 37 °C (non-CO₂ incubator) prior to the assay. ECAR and OCR were measured at the basal stage (basal glycolysis + mitochondrial acidification), in response to Rot/AA (mitochondrial electron transport chain inhibitors; compensatory glycolysis) and 2-deoxy-D-glucose (a glucose analog; post-2-DG acidification). The basal and compensatory glycolytic rates were calculated using the Seahorse Glycolytic Rate Assay.

TEM images of mitochondria in chondrocytes

Cell samples were fixed with a 2.5% glutaraldehyde solution (Sigma-Aldrich, USA) in phosphate buffer for 4 h. After rinsing with phosphate buffer, the samples were treated with 1% osmium tetroxide solution (OsO4, Sigma-Aldrich, USA) in phosphate buffer for 2 h. The samples were then immersed in ethanol at varying concentrations for 15 min. Following this, the samples underwent treatment with a mixture of embedding agent and acetone (Sigma-Aldrich, USA) for 3 h. The permeated samples were embedded and subjected to overnight heating at 70 °C to facilitate embedding. The embedded samples were sectioned using an ultra-thin microtome, and the sections were stained with uranyl acetate and alkaline lead citrate (Sigma-Aldrich, USA) for 10 min each. Finally, the sections were examined using an H-7650 transmission electron microscope (TEM) (Hitachi, Japan).

Cell culture

To isolate and culture articular cartilage cells, rats were euthanized and sterilized by immersion in 75% ethanol. The skin over the knee joint was incised using sterile instruments to aseptically extract the articular cartilage tissue. The tissue blocks were minced and repeatedly rinsed with a 10% penicillin-streptomycin solution (Gibco, USA). Subsequently, the tissue was washed with sterile phosphate-buffered saline (PBS, RG-CE-10, KETU, China) and digested overnight with 0.2% type II collagenase (Sigma-Aldrich, USA) at 37 °C in an incubator. The next day, undigested tissue was removed by filtration through a cell sieve, and the digested solution was centrifuged, resuspended, and cultured in 100 mm cell culture dishes (704202, NEST Biotechnology, China) to obtain P0 generation rat chondrocytes. These cells were cultured in complete medium (DMEM/F-12, Keygen BioTECH, China) supplemented with 10% fetal bovine serum (FBS, F103, Vazyme Biotech Co., Ltd, China), penicillin (100 units/mL), and streptomycin (100 units/mL) at 37 °C in a 5% CO₂ incubator. Subsequent experiments were conducted using P1 generation rat chondrocytes. For the isolation and culture of bone marrow-derived macrophages (BMMs), six-week-old male SD rats were euthanized by cervical dislocation following anesthesia. After euthanasia, the rats were sterilized with 75% ethanol and placed on a sterile glass plate. The femur and tibia were meticulously isolated, and the bones, immersed in PBS, were transferred to a sterile workbench. The metaphyses of the femur and tibia were excised to expose the medullary cavities. The diaphyseal cavities were then flushed with a complete medium containing 30 ng/mL macrophage colony-stimulating factor (M-CSF, CB34, Novoprotein, China) using a 10 mL sterile syringe and a 20-gauge needle. This process was repeated until the medullary cavities appeared white. The resulting cell suspension was treated with red blood cell lysate, followed by cell lysis and centrifugation. The isolated cells were cultured in a 10 cm petri dish for 16 h. The supernatant was then transferred and treated with M-CSF (30 ng/mL) for three days. For further experimentation, BMMs were cultured with M-CSF (30 ng/mL) and receptor activator of nuclear factor kappa-B ligand (RANKL, 50 ng/mL).

Cellular intervention

Chondrocytes were inoculated at an initial density of 3000 cells/cm². To simulate the arthritic microenvironment in vitro, interleukin-1 beta (IL-1β) at a concentration of 10 ng/mL was used for co-culturing with the cells. Each group of microspheres was thoroughly washed and sterilized using 75% ethanol before being co-cultured with the chondrocytes. A control group was also established, with the culture medium replaced every two days for a total of seven days. To inhibit TXNRD1 activity, TRi-1 is present at a concentration of 12 nM in the cell culture medium⁴⁸. To inhibit the phosphorylation of PI3K, LY294002 was maintained at a concentration of 0.5 μM in the culture media of chondrocytes, macrophages, and osteoclasts⁴⁹.To activate the phosphorylation process of PI3K, the concentration of 740Y-P in the culture medium for chondrocytes, macrophages, and osteoclasts was set to 3 μM²⁹.

Cell transfection

ATDC-5 cells were seeded into 6-well plates (5 × 10⁵ cells/well) and cultured until 80–90% confluency. Interventions were performed as follows: TXNRD1 overexpression, GPX1 overexpression, si-TXNRD1, and si-GPX1 (Anzhenbio, China) corresponding plasmids and siRNA sequences are listed in Supplementary Tables 4, 5. For overexpression groups, following Lipofectamine 3000 instructions, dissolve 2 μg of corresponding overexpression plasmid and 5 μL reagent in serum-free medium, incubate, and mix to form a complex. For knockdown groups, transfect 50 nM corresponding siRNA using Lipofectamine RNAiMAX. The complex was added dropwise to the cell culture system and incubated at 37 °C with 5% CO₂. Transfection efficiency was validated by Western blot analysis 48 h post-transfection.

Induction of Cartilage Formation

ATDC-5 mouse chondrocytes (CL-0856, Pronase Life Technology, China). A 100 μL suspension containing 400,000 cells was pipetted into the center of each well of a 24-well cell culture plate. Cells were cultured at 37°C in a 5% CO₂ environment for 2h, then transferred to medium containing 1% insulin-transferrin-selenium, 10ng/ml transforming growth factor-β3, 100nM dexamethasone, 50 μg/ml vitamin C, 1mM sodium pyruvate, and 40μg/ml proline. Cultures continued for 7 days with concurrent IL-1β and HSPHR treatment. The culture supernatant was discarded, cells were washed three times with PBS, fixed with 4% paraformaldehyde for 20 min, stained with A.B. and S.O., observed under an optical microscope, and images were captured. Quantitative analysis was performed using ImageJ-win64 software.

Trap stain assay

To facilitate in vitro osteoclastogenesis, Raw 264.7 cells were seeded at a density of 2 × 10⁴ cells per well in 48-well plates. The cells were then incubated for five days in the presence or absence of macrophage colony-stimulating factor (M-CSF) at a concentration of 30 ng/mL and receptor activator of nuclear factor kappa-Β ligand (RANKL) at 50 ng/mL. Tartrate-resistant acid phosphatase (TRAP) staining was performed following the manufacturer’s protocol. After fixation with 4% paraformaldehyde for 20 min, staining solution I was applied, and the cells were incubated at 37 °C for 30 min in the dark. Images were captured and stored using an inverted microscope (Carl Zeiss, Germany).

Macrophage polarization detection

Raw 264.7 cells (CL-0190, Pronase Life Technology, China), at a concentration of 2 × 10⁴ cells/mL, were cultured in the presence of microspheres. Following a 24-h treatment with lipopolysaccharide (LPS) at 10 ng/mL and interferon-gamma (IFN-γ) at 20 ng/mL, the cells were harvested. The cells were then stained with antibodies specific to F4/80, CD86, and the mannose receptor (CD206), sourced from eBioscience, USA. After washing with PBS, the expression levels of F4/80, CD86, and CD206 were analyzed using FCA with the Guava® easyCyte system (Merck Millipore, Germany).

Western blot assay

For protein extraction, cells treated with the respective interventions were lysed using RIPA buffer (Beyotime, China). The supernatant was collected after ultracentrifugation to yield the protein sample, and the concentration was determined using a BCA kit (Beyotime, China). Protein samples were mixed with protein loading buffer (Beyotime, China) at a 4:1 ratio and boiled at 100 °C for 5 min. SDS-PAGE electrophoresis was performed using a 4–12% gel (GenScript, China), and the transferred nitrocellulose membrane was incubated in blocking solution (Beyotime, China) for 30 min at room temperature. After blocking, the membrane was incubated overnight at 4 °C with the appropriate primary antibodies (diluted 1:2000). The antibodies used included anti-COL2 (ab188570, Abcam, USA), anti-ACAN (A11691, Abclonal, China), anti-MMP13 (ab39012, Abcam, USA), anti-ADAMTS5 (A2836, Abclonal, China), anti-SEPHS1 (A11435, Abclonal, China), anti-GPX1 (A0968, Abclonal, China), anti-TXNRD1 (A20805, Abclonal, China), anti-ATP5A (A0207, Abclonal, China), anti-TOMM20 (A0208, Abclonal, China), anti-LC3b (A19664, Abclonal, China), anti-MT-ND4 (A11247, Abclonal, China), anti-CTSK (A1782, Abclonal, China), anti-MMP9 (A0289, Abclonal, China), anti-CD40L (A0327, Abclonal, China), anti-CD86 (A1199, Abclonal, China), anti-CD206 (A26948, Abclonal, China), anti-ARG1 (A1847, Abclonal, China), anti-NOS2 (A3774, Abclonal, China), anti-PI3K (A22730, Abclonal, China), anti-AKT (A18675, Abclonal, China), anti-P-PI3K (4228, Cell Signaling Technology, USA), anti-P-AKT (4060, Cell Signaling Technology, USA), anti-P-mTOR (5536, Cell Signaling Technology, USA) and anti-β-Actin (AC038, Abclonal, China). The next day, primary antibodies were removed, and the membrane was incubated with an HRP-conjugated secondary antibody solution (diluted 1:10,000, Abclonal, China) specific to the corresponding species for 1 h at room temperature. The membrane was washed three times for 5 min each with the washing solution. Ultrasensitive Enhanced Chemiluminescent (ECL) substrate (NCM Biological Technology Co. Ltd, China) was prepared and incubated with the membrane for 5 min at room temperature, followed by exposure using a VersaDoc™ imaging system (Bio-Rad, USA). The intensities of the bands were quantified using ImageJ-win64 software for statistical analysis.

Quantitative real-time PCR analysis

Chondrocytes cultured in six-well plates were subjected to intervention treatment, and RNA was extracted by lysing the cells with Trizol reagent (Sigma-Aldrich, USA). The RNA concentration and purity were assessed using a NanoDrop ND-2000 spectrophotometer (Thermo Fisher Scientific, USA). Complementary DNA (cDNA) was then synthesized using a reverse transcription kit (Thermo Fisher Scientific, USA). The synthesized cDNA was mixed with SYBR Green PCR Master Mix (TaKaRa, Kusatsu, Japan), specific primers, and DEPC-treated water to establish the reaction system. Gene expression analysis was performed using the CFX96 real-time PCR system (Bio-Rad, USA), with Gapdh as the reference control. Primer sequences are detailed in Supplementary Tables 2 and 3. The relative expression levels of target genes were calculated using the 2^−ΔΔCt method.

Immunofluorescence staining

Chondrocytes were co-cultured with microsphere groups and treated with IL-1β. After treatment, the samples were fixed with 4% paraformaldehyde (Servicebio, China) for 30 min, followed by a 30-min incubation with an immunostaining blocking solution (Beyotime, China). The primary antibody working solution (1:200) was prepared, and incubation with the primary antibody occurred overnight at 4 °C. The next day, the primary antibody was removed, and the cells were incubated with the Alexa 488 or Alexa 594 fluorescent secondary antibody (1:500, Abcam, USA) at room temperature for 1 h. After washing with PBS, the nuclei were stained with DAPI (Sigma, USA). Imaging was performed using a Zeiss Axiovert 40CFL microscope, and all steps were conducted under light-protected conditions.

RNA sequencing

To evaluate differential gene expression, rat chondrocytes were exposed to IL-1β (10 ng/mL) alone or in combination with HSPHR microspheres. RAW264.7 cells were exposed to RANKL (50 ng/mL) or co-exposed with HSPHR microspheres. Alternatively, RAW264.7 cells were exposed to LPS (100 ng/mL) + IFN-γ (20 ng/mL) or co-exposed with HSPHR microspheres. The resulting samples were submitted to Wekemo Tech Group Co., Ltd. (Shenzhen, China) for comprehensive RNA sequencing analysis using the Affymetrix GeneChip microarray platform (Affymetrix, Santa Clara, CA, USA). RNA expression levels were also assessed with an Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). Differential gene expression was analyzed with Microarray software, and genes with a fold change greater than 2 or less than 0.05 were identified as DEGs. Differential expression analysis between three 30 °C and two 4 °C samples was performed using the R (v.4.0.3) package DESeq2 (v.1.30.0). GSEA was performed with GSEA (v.4.1.0) using the GSEAPreranked tool, whereby genes were preranked on the basis of their P values and fold changes.

Animals and Ethics

The experimental subjects comprised SPF-grade SD rats and C57BL/6 J mice. All husbandry practices and experimental procedures were conducted in strict accordance with the guidelines for laboratory animal welfare and ethics, receiving approval from the Animal Experiment Management Committee of Soochow University (SUDA20241210A8). The SD rats and C57BL/6 J mice were housed within a SPF barrier environment, maintained at a temperature of 22 ± 2 °C, with 50 ± 10% relative humidity, and subjected to a 12-h light-dark cycle. The animals were accommodated in individually ventilated cages, with 3–5 animals per cage, and were provided with sterile bedding, SPF-grade feed, and acidified sterile water available ad libitum. Health assessments were performed daily to ensure the well-being of the animals.

Gait analysis

Prior to tissue collection, gait analysis was performed on all mice. The forepaws were marked with red pigment, and the hind paws with green pigment. A white recording sheet was placed along a 50 cm long, 10 cm wide, and 10 cm high runway. A black enclosure, with an opening facing the recording sheet, was positioned at the end of the runway to encourage the mice to enter. Each mouse was placed individually at the starting point of the runway and allowed to traverse the path freely, without any external interference. Upon completion of the experiment, consecutive, clearly distinguishable footprints were selected from the recording sheet for subsequent statistical analysis.

Open field test

After the DMM model was successfully developed up to week 8, spontaneous activity and exploratory behavior were evaluated using the VisuTrack system (Shanghai XinRuan Information Technology, China). Mice were placed in a 50 × 3 × 50 cm enclosed arena under dim lighting conditions. A video camera recorded their movement for 3 min, allowing assessment of relative mobility, activity duration, distance traveled, and average speed.

In vivo imaging systems (IVIS) imaging

The degradation of glycopeptide hydrogels within the joint cavity of mice was assessed using the IVIS in vivo imaging system (Spectrum, PerkinElmer). Mice were anesthetized with isopentane before the injection of 100 μL of Cy5.5-labeled microspheres (10 mg/mL) into the joint cavity. Imaging sessions were conducted on days 1, 3, 5, 7, and 14, with anesthesia maintained using the same protocol. The rats were carefully positioned on the imaging chamber stage, and the testing temperature was kept at 37 °C during each exposure. Exposure duration and imaging parameters were optimized for visualization and applied consistently across all sessions. Kinetic imaging curves were generated after all imaging sessions. The region of interest (ROI) was delineated as the joint cavity area in each mouse. The IVIS imaging system (Xenogen) was used to integrate bioluminescence signals, and quantification of these signals determined the average radiance. The resulting ROI photon flux was used as an indicator of hydrogel retention within the joint cavity.

Micro-CT

At eight weeks postoperatively, the rats were euthanized, and knee specimens were collected for further analysis. These specimens were fixed in a 4% paraformaldehyde solution (Servicebio, China) for 48 h. Following fixation, the specimens were scanned and analyzed using a high-resolution Micro-CT system (SkyScan 1176, Aartselaar, Belgium). Initial scans were performed at 60 kV, 170 mA, and 0.7° rotation with a resolution of 18 μm, maintaining consistent scanning parameters across all samples. The scanned data were reconstructed and subjected to three-dimensional quantitative analysis using NRecon, Data Viewer, CTAn, and Mimics software. Key metrics, including bone volume to tissue volume ratio (BV/TV), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th), were evaluated to assess the extent of OA.

Histological staining

After fixation, specimens were decalcified using 10% EDTA (pH 7.4, Sigma-Aldrich, USA) over approximately two months. Post-decalcification, the samples underwent dehydration and were embedded in paraffin (Leica, Germany). Tissue blocks were sectioned into 6 μm slices. These sections were deparaffinized in xylene, dehydrated through a graded ethanol series, and stained with hematoxylin and eosin, Safranin O/fast green, and toluidine blue. Microscopic images were captured using a light microscope (CX23). Cartilage degeneration was evaluated using the OARSI scoring system and the ratio of hyaline cartilage (HC) to calcified cartilage (CC). Cartilage defect repair was assessed using the International Cartilage Repair Society (ICRS) and the Modified O’Driscoll histological (MODS) scoring systems. Additionally, sections of embedded rat visceral specimens were prepared to evaluate the in vivo toxicity of the materials, with hematoxylin and eosin staining used for assessment.

Immunohistological staining and immunohistofluorescence

Sections underwent deparaffinization in xylene and dehydration through a graded ethanol series. They were then incubated with 3% hydrogen peroxide for 15 min and subjected to antigen retrieval using 2 mg/mL hyaluronidase (Sigma-Aldrich, USA) at 37 °C. The sections were blocked with 1.5% goat serum and incubated overnight with the primary antibody at a dilution of 1:200 at 4 °C. The next day, the primary antibodies were removed, and the sections were incubated with HRP-conjugated secondary antibodies at a dilution of 1:500 for 1 h at room temperature. The sections were then stained using the 3,3′-diaminobenzidine (DAB) working solution (Vector Laboratories, Burlingame, CA, USA) and counterstained with hematoxylin. For immunofluorescence staining, the sections were incubated with Alexa 488 or Alexa 594 fluorescent secondary antibodies at a dilution of 1:500 for 1 h at room temperature, protected from light, and counterstained with DAPI. Imaging was performed using a Zeiss Axiovert 40CFL microscope. The percentage of positive cells was quantified using ImageJ-win64 software.

DMM model

Eight-week-old male SD rats were used as the experimental model. All microsurgical instruments were rigorously sterilized prior to the start of the experiment. Anesthesia was induced via intraperitoneal injection of 1% sodium pentobarbital at a dose of 400 μL per 100 g of body weight. Following anesthesia, the patellar region was shaved and disinfected, and a sterile surgical drape was applied. A longitudinal incision was made along the medial side of the patella, followed by sequential dissection of the skin, subcutaneous tissue, and joint capsule. The patella was then displaced laterally to expose the medial condyles of the femur and tibia, as well as the meniscal structures. The adipose tissue beneath the patella was carefully blunt-dissected, and the medial anterior tibial ligament was transected cautiously to avoid damaging the cartilage. Subsequently, the meniscus was dislocated from its original stable position, and the surrounding tissues were sutured at multiple layers. In the control group, the joint capsule was incised and then closed without further manipulation. Postoperatively, the surgical site was thoroughly cleaned and disinfected, followed by intramuscular administration of penicillin to prevent infection. One week after arthroplasty, the animals were randomly assigned to experimental groups and received intra-articular injections according to the treatment protocol. At 8 weeks post-surgery, the rats were euthanized, and their knee joints were harvested for subsequent analysis.

Cartilage defect model

Eight-week-old male SD rats were used for the experimental procedures. All microsurgical instruments were strictly sterilized prior to surgery. Anesthesia was induced via intraperitoneal injection of 1% sodium pentobarbital at a dose of 0.4 mL per 100 g of body weight. After confirming anesthesia and ensuring stabilization, the rats were placed in a supine position. The surgical site was shaved, with particular attention to the knee joint, followed by local disinfection and sterile draping. A longitudinal incision was made along the medial side of the patella using a sterile scalpel, and sequential incisions were performed through the subcutaneous tissue and joint capsule. The patella was laterally dislocated using hemostatic forceps to expose the femoral condyle and trochanter. Thereafter, a full-thickness cartilage defect measuring 1.5 mm in diameter and 1.5 mm in depth was created on the femoral trochlear groove using an electric drill. The surgical site was irrigated with saline to remove debris, after which the patella was repositioned, and the joint cavity was closed by suturing the tissue layers sequentially. In the sham-operated group, only exposure and closure of the joint cavity were performed without creating a cartilage defect. After completing the surgical procedures, local disinfection was carried out, and penicillin (200,000 U/kg) was administered intramuscularly to prevent infection. Once the rats had recovered from anesthesia, they were returned to a controlled environment for postoperative care. One week after surgery, rats with full-thickness articular cartilage defects were randomly assigned to one of four treatment groups (excluding the sham group) and received intra-articular injections of 200 μL PBS, HAMA (2 mg/mL), or HSPHR (2 mg/mL). At 8 weeks post-surgery, the rats were euthanized, and their knee joints were harvested for further analysis.

In vivo toxicity

To assess the safety profile of HAMA and HSR microspheres, rat serum samples were collected following intra-articular injection. Total protein (TP), albumin (ALB), aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN), and creatinine (Cre) concentrations were measured using standard assay kits according to the manufacturers’ instructions. Histological examination of cardiac, hepatic, splenic, pulmonary, and renal tissues was then conducted using H&E staining.

Statistical analysis

Data are presented as mean ± standard deviation (SD). To evaluate differences between groups, Student’s t test was used for comparisons between two groups, and one-way ANOVA was used for multiple group comparisons. A p value < 0.05 was considered statistically significant. All data analysis was performed using GraphPad Prism 10.1.2 software.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

X.S.Z. and Y.X. designed the study. Y. L., Y.J.Z., C.Q.Y., X.W.X., K.K., Y.B.W., Y.G.D., J.F.Y., and M.Z.H. conducted the study and collected the data. Y.L., Y.J.Z., C.Q.Y. and X.W.X. analyzed the data. Y. L.and Y.J.Z. interpreted the data and the experiments. Z.W.L., H.Y., Y.X., and X.Z. drafted the manuscript. All authors contributed to the manuscript and approved the final version.

Peer review

Peer review information

Nature Communications thanks Yiting Lei and the other anonymous reviewers for their contribution to the peer review of this work. [A peer review file is available].

Data availability

All supporting data for these study findings described in this paper are included in the main text and supplementary information, with the source data provided alongside this paper. Source data for Figs. 1–6 and Supplementary Figs. 1–18 are available in the associated source data file. The RNA-seq data generated in this study have been deposited in the Sequence Read Archive of NCBI database under accession code PRJNA1392454. The mass spectrometry proteomics data generated in this study have been deposited into the ProteomeXchange Consortium via the PRIDE partner repository with the dataset under accession code PXD072313. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-68817-2.