Ultra-small iron-based nanoparticles Mitigate rheumatoid arthritis inflammation via macrophage Repolarization and SLC7A11/GPX4-mediated ferroptosis inhibition

Abstract

The characteristic pathological manifestations of rheumatoid arthritis (RA) include inflammatory cell infiltration, abnormal synoviocyte proliferation, and progressive bone and cartilage destruction. An excessive buildup of reactive oxygen species (ROS) within the joints is a critical factor promoting RA pathological progression. In this study, we innovatively employed a hard template-restricted controlled sintering carbonization strategy to fabricate ultra-small Fe₃O₄@C nanoparticles with hierarchical structures. The ultra-small Fe₃O₄@C nanoparticles exhibit multiple natural enzyme-mimic catalytic activity, effectively diminishing intracellular ROS levels in macrophages, while also facilitating the polarization toward the M2 phenotype, and significantly suppresses the production of pro-inflammatory cytokines. Mechanistic investigations reveal that Fe₃O₄@C significantly suppresses ferroptosis in synoviocytes and chondrocytes through regulation of the SLC7A11/GPX4 signalling pathway, thereby alleviating synovial tissue erosion and promoting type II collagen synthesis. In the collagen-induced arthritis mouse model, Fe₃O₄@C exhibited remarkable anti-inflammatory and chondroprotective effects, providing an innovative nanozyme therapeutic approach for RA treatment.

Graphical abstract

1. Introduction

Rheumatoid arthritis (RA) is classified as a systemic chronic autoimmune disorder, primarily distinguished by symmetric polyarticular synovitis. Pathological features of RA include progressive synovial hyperplasia, inflammatory cell infiltration, and, in the end, the degradation of articular cartilage along with bone erosion. Recent studies have demonstrated that oxidative stress plays a pivotal role in the pathogenesis of RA [1]. The overproduction of reactive oxygen species (ROS) leads to oxidative injury in joint tissues and promotes the polarization of macrophages towards the pro-inflammatory M1 phenotype, exacerbating synovitis and arthritis progression [2,3]. Notably, M1 macrophages further promote excessive release of pro-inflammatory cytokines through activation of NF-κB signaling pathways [4]. Given the pivotal function of ROS in various cell types within the synovial microenvironment of RA, modulating ROS emerges as an innovative approach to disrupt the vicious "ROS-inflammation" cycle and achieve effective RA treatment [5]. Current clinical approach to managing RA primarily depends on medications including non-steroidal anti-inflammatory drugs, disease-modifying anti-rheumatic drugs, or traditional Chinese medicine [6,7]. However, despite achieving certain efficacy in symptom control, existing therapeutic regimens still face numerous challenges: on one hand, they fail to achieve sustained remission of synovial inflammation; on the other hand, their protective effects on articular cartilage remain limited. Therefore, developing novel therapeutic strategies that combine efficient ROS scavenging capacity, significant anti-inflammatory effects, and chondroprotective properties has become a critical breakthrough in current RA treatment research.

Nanozymes represent a category of nanomaterials which possess catalytic properties analogous to those of natural enzymes [[8], [9], [10], [11], [12], [13], [14]]. Among them, nanomaterials capable of simulating two or more natural enzyme activities are termed multi-enzyme nanozymes, with metal-based nanozymes particularly attracting attention due to their multifunctional catalytic characteristics [[14], [15], [16], [17]]. Among various nanozymes, iron-based nanozymes have emerged as research hotspots due to their superior catalytic efficiency, affordability, and excellent stability. Studies have demonstrated that precise modulation of iron oxide nanoparticle size can effectively regulate their enzymatic activity [[17], [18], [19], [20], [21]]. With their larger specific surface area, ultrasmall iron oxide nanoparticles (USIO NPs) can expose more active sites, thereby significantly enhancing catalytic efficiency [[21], [22], [23]]. Meanwhile, carbon-based nanozymes, as an emerging class of nanomaterials, also exhibit multi-enzyme mimetic properties, capable of simultaneously mimicking both catalase (CAT) and superoxide dismutase (SOD) enzymatic activities [[24], [25], [26]]. For instance, research has revealed that carbon dots (C-dots) SOD nanozymes not only demonstrate specific targeting capability toward oxidatively damaged cells but also effectively scavenge superoxide anion radicals, significantly reducing intracellular ROS levels [27]. This unique multi-enzyme activity enables iron oxide nanozymes and carbon-based nanozymes to exhibit exceptional performance in ROS scavenging [28]. Through rational design of cascade catalytic reactions, these nanozymes can more efficiently modulate ROS levels, thereby achieving precise intervention in the inflammatory microenvironment of RA. Specifically, ultrasmall iron oxide nanozymes and carbon-based nanozymes can reconstruct the inflammatory microenvironment by inhibiting pro-inflammatory cytokine release on the one hand and promoting anti-inflammatory cell conversion on the other hand, ultimately enhancing joint mobility and improving the overall quality of life for patients [29]. Despite the tremendous potential of ultrasmall iron oxide and carbon-based nanozymes in modulating oxidative stress and inflammatory responses, current research on their functional mechanisms in cartilage protection remains notably insufficient. In-depth exploration of this field will contribute to elucidating the biological effects of nanozymes and provide novel therapeutic strategies and theoretical foundations for RA treatment.

Combining nanozymes with other enzymatically active nanomaterials can achieve multi-enzyme synergistic catalysis, enhancing catalytic efficiency and specificity. Metals doped in carbon-based nanozymes can improve catalytic reaction rates [24]. In this study, ultrasmall Fe₃O₄@C nanoparticles with hierarchical structures were successfully designed and synthesized through a hard template-confined controlled sintering and carbonization strategy involving hydrothermal processing followed by high-temperature calcination. The resulting materials exhibit excellent anti-inflammatory and antioxidant properties, demonstrating significant potential for RA treatment (Scheme 1). Through systematic characterization analysis, we confirmed that the synthesized materials possess multi-layered structural characteristics, with an inner core composed of ultrasmall Fe₃O₄ nanoparticles (3.09 ± 0.18 nm) and an outer layer encapsulated by amorphous carbon. Performance experiments demonstrated that this material exhibits cascade catalytic activities of SOD and CAT. Furthermore, in vitro experiments revealed that Fe₃O₄@C nanozymes displayed exceptional ROS scavenging capacity, effectively regulating macrophage polarization through modulation of redox homeostasis, thereby disrupting the vicious "ROS-inflammation" cycle. Fe₃O₄@C significantly decreased the concentrations of the pro-inflammatory cytokines IL-1β and TNF-α, concurrently enhancing the release of the anti-inflammatory mediators IL-10 and TGF-β. Additionally, Fe₃O₄@C promoted the upregulation of Solute Carrier Family 7 Member 11 (SLC7A11) and Glutathione Peroxidase 4 (GPX4), as well as facilitated the production of reduced glutathione (GSH), reducing lipid peroxidation levels in synoviocytes and chondrocytes. Moreover, Fe₃O₄@C decreased Matrix Metallopeptidase 3 (MMP3) levels in synoviocytes and promoted type II collagen synthesis in chondrocytes, maintaining articular synovium and cartilage homeostasis. In the CIA mouse model, Fe₃O₄@C nanozyme treatment significantly alleviated joint swelling and synovial inflammation, while histological analysis confirmed its ability to reduce inflammation levels and promote cartilage repair. These results demonstrate that Fe₃O₄@C nanozymes exhibit promising therapeutic potential in RA treatment through their unique multi-synergistic effects of antioxidation, anti-inflammation, and chondroprotection, providing important evidence for developing novel RA therapeutic strategies.

Scheme 1.

Schematic illustration of Fe₃O₄@C nanoparticles for therapy in rheumatoid arthritis. (A) Schematic illustration of the synthesis process for Fe₃O₄@C nanoparticles. (B) Therapeutic mechanism of Fe₃O₄@C nanoparticles in RA by regulating macrophage polarization and inhibiting ferroptosis in synovial and chondrocytes.

2. Materials and methods

2.1. Chemical materials and reagents

Tetraethyl orthosilicate (TEOS), hexadecyl trimethyl ammonium bromide (CTAB), triethanolamine, and iron acetylacetonate (Fe(acac)₃) were obtained from Tokyo Chemical Industry (TCI, Shanghai, China). Sodium hydroxide (NaOH) and pyrogallol were sourced from Aladdin (Shanghai, China).

2.2. Synthesis of mesoporous silica (mSiO₂) nanoparticles

The preparation of the solution involved the dissolution of CTAB (1.822 g, 5 mmol) and triethylamine (0.202 g, 2 mmol) in 200 mL of deionized water, with continuous magnetic stirring at a temperature of 80 °C for a duration of 4 h. Following this, TEOS (14.58 g, 70 mmol) was introduced to the mixture in a dropwise manner, and the reaction was kept at 80 °C for an additional 2 h. The resulting products were isolated via centrifugation at 10,000 rpm for 15 min and were subsequently washed three times with deionized water to remove any unreacted materials or by-products. Following the drying process, the products were finely pulverized and subjected to calcination at 600 °C for a period of 6 h to eliminate the soft template CTAB.

2.3. Synthesis of Fe₃O₄@C nanoparticles

Fe(acac)₃, with a mass of 0.211 g corresponding to 6 mmol, was solubilized in 20 mL of anhydrous ethanol under magnetic stirring (500 rpm) at 25 °C for 10 min until complete dissolution. Concurrently, A total of 0.5 g of mSiO₂ nanoparticles was suspended in 20 mL of deionized water and subjected to magnetic stirring at a speed of 500 rpm for a duration of 10 min at a temperature of 25 °C, ensuring thorough dissolution. Then, the two precursor solutions were combined under magnetic stirring at a speed of 500 rpm for a duration of 1 h at a temperature of 25 °C to achieve a homogeneous mixture. The suspension was hydrothermally treated at 180 °C for 24 h. The solid products were obtained through centrifugation at 10,000 rpm for a duration of 15 min, followed by three separate washings with deionized water. After the washing process, the products were dried and subsequently ground into a fine powder, which was then subjected to calcination at a temperature of 800 °C for 3 h. Following this, an etching solution was prepared by dissolving 10 g of NaOH and 1 g of 3,4-dihydroxybenzoic acid in 100 mL of deionized water. The calcined powder was then immersed in 50 mL of this etching solution and stirred at a speed of 500 rpm for 24 h to facilitate the removal of the mSiO₂ hard template. The final products were subsequently collected using dialysis and subjected to freeze-drying.

2.4. Characterization

The morphology and structure of Fe₃O₄@C were analyzed by transmission electron microscopy (TEM, FEI Tecnai G2 F30). X-ray diffraction (XRD) patterns were acquired employing a PANalytical Empyrean powder diffractometer with Cu Kα radiation (λ = 0.1541 nm). The chemical composition was conducted through X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha). Additionally, hydrodynamic diameter and zeta potential measurements were performed with a Malvern Zetasizer Nano ZS90 system. Ultraviolet–visible (UV–Vis) absorption spectra were recorded by Persee TU1810 and SpectraMax 190 spectrophotometers. Changes of dissolved oxygen concentration in the reaction system were recorded through a dissolved oxygen analyzer (HQ40d).

2.5. SOD-mimic activity assessment

Xanthine/Xanthine Oxidase Method: the specific SOD-mimic activity of Fe₃O₄@C nanoparticles was quantified using a xanthine/xanthine oxidase superoxide generation system. Superoxide anion (O₂·-) scavenging capacity was determined by measuring the inhibition rate with a commercial SOD assay kit (Dojindo, Japan), adhering strictly to the manufacturer's instructions.

Pyrogallol Autoxidation Method: the pyrogallol autoxidation method also measured the O₂·- inhibition rate. Reaction mixtures contained 2.95 mL of Fe₃O₄@C nanoparticle suspensions (6.25–200 μg/mL in 50 mmol/L Tris-HCl buffer, pH 8.2) and 50 μL of freshly prepared pyrogallol solution (60 mmol/l). After rapid mixing in quartz cuvettes (final volume 3 mL), the kinetics of absorbance at a wavelength of 325 nm were measured at 30-s intervals over a total duration of 300 s using a Persee TU1810 UV–Vis spectrophotometer maintained at 25 °C. The O₂·- inhibition rate was calculated as: Inhibition (%) = [(A₁ – A₂)/A₁] × 100, where A₁ represents the autoxidation rate of pyrogallol without nanoparticles, and A₂ denotes the rate in nanoparticle-containing systems. The more detailed calculation method for the SOD mimetic activity (U/mg) is provided in the Supplementary Information.

2.6. CAT-mimic activity assessment

CAT-mimic activity was quantified by monitoring the dissolved oxygen evolution in a H₂O₂ decomposition reaction with a calibrated dissolved oxygen meter. All measurements were performed at 25 °C within a 0.1 M PBS solution at a pH 7.4. a. To investigate the correlation between the concentration of nanoparticles and the activity of the CAT-mimic reaction, 125, 250, 375, and 500 μL Fe₃O₄@C nanoparticles suspensions (1.0 mg/mL) were subjected to serial dilution in a PBS solution, resulting in a final volume of 4.95 mL. Reactions were initiated by rapidly adding 50 μL of 30 % H₂O₂. Real-time dissolved oxygen concentration was recorded every 30 s for 600 s. b. To determine the substrate-concentration dependence of the CAT-like reaction, 50 μg/mL of the Fe₃O₄@C was dispersed in PBS buffer, achieving a final volume of 5.0 mL. The reaction commenced with the swift adding different volumes (ranging from 25 to 100 μL) of 30 % H₂O₂, yielding final H₂O₂ concentrations ranging from 50 to 200 mM. Real-time dissolved oxygen concentration was recorded every 30 s for 600 s.

2.7. Cell culture

The RAW264.7 macrophage cell line and mouse synovial mesenchymal stem cells (SMSCs) were obtained commercially from Procell (Wuhan, China). The immortalized human chondrocyte and synoviocyte cell lines were purchased from iCell Bioscience (Shanghai, China). All cell cultures were maintained in DMEM enriched with 10 % fetal bovine serum and 1 % penicillin-streptomycin, under controlled conditions at 37 °C within a humidified environment containing 5 % CO₂. All cells were passaged when confluence reached approximately 80 %. RAW264.7 cells, SMSCs, chondrocytes, and synoviocytes were seeded in multi-well plates at 1 × 10⁵ cells/mL.

2.8. Cell viability assessment

2.8.1. The cell counting kit-8 (CCK-8) assay

RAW264.7 cells were plated at a density of 10,000 cells per well in 96-well plates for the CCK-8 assay. Then, RAW264.7 cells were treated with Fe₃O₄@C nanoparticles at various concentrations or PBS as a control. CCK-8 was purchased from Beyotime (Shanghai, China). 10 μL of CCK-8 reagent was added to each well (100 μL medium/well) and incubated for 30 min at 37 °C in a humidified 5 % CO₂ atmosphere. Absorbance measurements were performed at 450 nm.

2.8.2. Live/dead staining

Target cells (RAW264.7 cells, SMSCs, chondrocytes) were cultured in 96-well plates and subjected to treatment with Fe₃O₄@C (200 μg/mL) for 24 h, while control groups were maintained without any treatment. Cell viability was assessed through live/dead staining with Calcein-AM and propidium iodide (PI). The assay kit for Calcein-AM/PI Cell Viability and Cytotoxicity was obtained from Beyotime (Shanghai, China) and was employed under the instructions. Target cells were incubated with the staining solution for 10 min in a cell culture incubator and observed under an inverted fluorescence microscope. Live cells were stained green (calcein-AM) while dead cells showed red fluorescence (PI).

2.9. Co-culture system for RA microenvironment simulation

To simulate the RA inflammatory microenvironment, a co-culture system was established using RAW264.7 macrophages and target cells (chondrocytes, synoviocytes, or SMSCs) as shown in Fig. 4A. RAW264.7 cells were first polarized to a pro-inflammatory (M1) phenotype by 24-h pretreatment with 100 ng/mL lipopolysaccharide (LPS, Sigma, Germany). The activated RAW264.7 cells were then co-cultured alongside target cells with the upper compartment containing Fe₃O₄@C nanoparticles (200 μg/mL).

Fig. 4.

Anti-ferroptotic effect of Fe₃O₄@C nanoparticles via the GPX4/SLC7A11 axis. (A) Schematic illustration of the co-culture system for assessing Fe₃O₄@C protective effects. (B) Volcano plot analysis of DEGs in SMSCs between the inflammatory and Fe₃O₄@C treatment groups. (C–D) qPCR validation of ferroptosis and inflammatory markers in synoviocytes and chondrocytes under different treatment conditions. DEGs, differentially expressed genes (n = 3). (E–F) Ratio of reduced to oxidized fluorescence intensity of BODIPY 581/591C11 in synoviocytes (E) and chondrocytes (F), demonstrating lipid peroxidation levels (n = 3). (G–H) Total GSH content in synoviocytes (G) and chondrocytes (H), reflecting cellular antioxidant capacity (n = 3).∗(P < 0.05), ∗∗ (P < 0.01), ∗∗∗(P < 0.001), ∗∗∗∗(P < 0.0001).

2.10. Cellular uptake of Fe₃O₄@C nanoparticles

Fluorescein isothiocyanate (FITC) was purchased from Bioss (Beijing, China). A stock solution of FITC (5 mg/ml) was formulated in dimethyl sulfoxide (DMSO) as the solvent and protected from light. For fluorescent labelling, the well-dispersed nanoparticle solution was incubated with FITC solution under constant stirring at 4 °C in complete darkness for 12 h. RAW264.7 cells were incubated with FITC-labelled Fe₃O₄@C (200 μg/ml) for durations of 6, 12, and 24 h, respectively. The control group did not receive any treatment. After PBS washing, cells were stained with Hoechst 33342 for 10 min, after which further washes with PBS were performed. The fluorescence emitted by the cells was subsequently examined under an inverted fluorescence microscope to evaluate Fe₃O₄@C nanoparticle uptake at different time points.

2.11. ROS scavenging activity assessment

The ability of Fe₃O₄@C to scavenge ROS was assessed utilizing the DCFH-DA fluorescent probe (MedChemExpress, USA). RAW264.7 cells were categorized into two experimental groups: (1) Fe₃O₄@C-treated group, which underwent a pretreatment with 200 μg/mL Fe₃O₄@C nanoparticles for 24 h, and (2) untreated control group. Then, both groups were exposed to 100 ng/mL LPS for an additional period of 24 h to induce oxidative stress before ROS measurement. Following this, the cells were treated with DCFH-DA at a concentration of 10 μM for a duration of 30 min at 37 °C, ensuring that the process was conducted in a dark environment. The levels of intracellular ROS were then visualized using an inverted fluorescence microscope and quantified by measuring the fluorescence intensity.

2.12. High-throughput RNA sequencing and bioinformatics interpretation

RNA sequencing, along with bioinformatics analysis, was conducted on total RNA extracted from SMSCs after co-culture treatment. RNA libraries were prepared and sequenced on a high-throughput sequencing platform. Differential gene (DEG) expression analysis was performed to identify significantly regulated genes between different groups. DEGs were identified and visualized using volcano plots. Functional enrichment analysis was performed to analyze potential biological mechanisms.

2.13. Quantitative real-time PCR (qRT-PCR) analysis

The mRNA expression levels of macrophage polarization markers and cellular protection indicators were systematically evaluated through qRT-PCR analysis. Total RNA was extracted from treated cells by RNA Extraction Kit (Vazyme, Nanjing, China) according to the instructions provided by the manufacturer. CDNA synthesis was performed using the Reverse Transcriptase Kit (Vazyme, Nanjing, China). qRT-PCR was performed using SYBR Green Kit (Yeasen, Shanghai, China).

RAW264.7 macrophages were cultured and divided into two experimental groups: Fe₃O₄@C-treated with 200 μg/mL Fe₃O₄@C nanoparticles for a duration of 24 h, and an untreated control group maintained in complete medium. After pretreatment, both groups underwent stimulation with 100 ng/mL LPS for 24 h. Subsequently, macrophages were collected to facilitate RNA extraction and qRT-PCR analysis of polarization markers. We measured the levels of expression of M1 polarization indicators, specifically IL-1β and IL-6, alongside the M2 marker, CD206. Synoviocytes and chondrocytes cultured in the lower chamber were then processed separately for RNA isolation and qRT-PCR analysis. For synoviocyte protection assessment, we analyzed the synovial invasion marker MMP3 and ferroptosis-related markers (GPX4 and SLC7A11). Chondroprotective effects were evaluated by examining the cartilage anabolic marker COL2A1 in conjunction with the same ferroptosis-related markers (GPX4 and SLC7A11). Gene expression was normalized to β-actin and calculated using the 2^−ΔΔCt method.

2.14. ELISA analysis

For cytokine profiling, RAW264.7 macrophages were subjected to three experimental conditions in a standardized ELISA protocol: (1) LPS + Fe₃O₄@C group, pretreatment with 200 μg/mL Fe₃O₄@C nanoparticles for 24 h, subsequently followed by stimulation with 100 ng/mL of LPS for 24 h; (2) LPS control, which involved treatment with LPS alone for 24 h; and (3) untreated normal control (NC). Following treatments, conditioned media were collected and centrifuged at 3000×g for 20 min at a temperature of 4 °C to eliminate cellular debris. The resulting clarified supernatants were then aliquoted into sterile microtubes and preserved at −80 °C in preparation for ELISA analysis. The concentrations of inflammatory cytokines were quantified, focusing on M1 markers such as IL-1β and TNF-α, alongside anti-inflammatory cytokines represented by M2 markers like TGF-β and IL-10. This quantification was carried out using commercial ELISA kits sourced from Bioswamp, Wuhan, China. Absorbance readings were taken at a wavelength of 450 nm.

2.15. Lipid peroxidation analysis

Synoviocytes and chondrocytes were co-cultured using a Transwell system, with experimental and control groups treated as previously described. Lipid peroxidation was evaluated using the BODIPY 581/591C11 (Beyotime, Shanghai, China). To prepare a working solution, the probe was diluted at a ratio of 1:1000 in PBS. Subsequently, the target cells were incubated with this working solution for a duration of 30 min, after which they were visualized using an inverted fluorescence microscope. The quantification of fluorescence intensity was performed utilizing ImageJ software. The degree of lipid peroxidation was determined by calculating the red to green fluorescence intensity, where decreased ratios indicated increased lipid peroxidation levels.

2.16. Total glutathione analysis

Synoviocytes and chondrocytes were co-cultured using a Transwell system and treated as previously described. Target cell pellets were lysed for sample preparation using protein removal reagent, followed by two freeze-thaw cycles in liquid nitrogen to ensure complete cell disruption. The quantification of total glutathione levels was determined utilizing a commercial assay kit (Beyotime, Shanghai, China). Briefly, 10 μl of cell lysate or glutathione standard was dispensed into individual wells of a 96-well plate, followed by the addition of 150 μL of a freshly prepared assay working solution. After ensuring thorough mixing and allowing the mixture to incubate at room temperature for 5 min, shielded from light, 50 μL of NADPH solution was introduced into each well. The reaction mixture was then further incubated for an additional 30 min at 37 °C, after which the absorbance was recorded at 412 nm using a microplate reader. A standard curve was established with reduced GSH standard, and sample GSH concentrations were calculated accordingly.

2.17. Experimental animals

A total of thirty male DBA/1 mice, each seven weeks of age, were acquired from Jinan Pengyue Laboratory Animal Breeding Co., Ltd. These mice were maintained in specific pathogen-free (SPF) barrier facilities and subjected to stable environmental conditions. The mice had unrestricted access to a standard rodent diet and sterilized drinking water. The protocol for animal experimentation received approval from the Laboratory Animal Ethics Committee of the Affiliated Hospital of Qingdao University (AHQU-MAL20240724YMY).

2.18. CIA model induction and treatment

All DBA/1 mice underwent a 1-week acclimatization period under SPF conditions before experimental initiation. To establish the CIA model, mice received subcutaneous tail injection of 100 μL emulsion on day 0. This emulsion was composed of bovine type II collagen at a concentration of 2 mg/mL (Chondrex, USA), which was mixed with an equivalent volume of complete Freund's adjuvant at a concentration of 1 mg/mL (Chondrex, USA). On day 21, a booster immunization was administered using 100 μL of an equal mixture of bovine type II collagen (at a concentration of 2 mg/mL) emulsified in incomplete Freund's adjuvant (Chondrex, USA). The mice were randomly divided into three groups, each comprising ten individuals: (1) NC group (non-induced healthy mice); (2) CIA model group (PBS); (3) Fe₃O₄@C treatment group. All interventional procedures were performed under 2.5 % isoflurane anesthesia via intra-articular injection into the knee joint. The experimental group received 10 μL of Fe₃O₄@C nanoparticles (200 μg/mL) on days 28, 31, 34, and 36, while the control group was administered equal volumes of PBS buffer at the same timepoints. All mice were sacrificed at the experimental endpoint (day 45).

2.19. Arthritis clinical scoring

This study employed a standardized clinical scoring system to systematically evaluate disease progression in the CIA model. Beginning from the primary immunization (day 0), arthritis symptoms in mice were scored weekly; after booster immunization (day 21), scoring frequency was increased to every 3 days until the experimental endpoint. The scoring criteria employed a 5-point scale: 0 = the absence of any signs of erythema or swelling in any limb joints; 1 = the presence of mild erythema and swelling specifically in the wrist or ankle joints; 2 = moderate erythema accompanied by mild swelling in either the ankle or wrist joints; 3 = significant erythema and swelling in paws and digits; 4 = severe inflammatory reactions in multiple joints of all limbs. Finally, arthritis severity in each mouse was quantitatively assessed by summing the independent scores of all four limbs (maximum score of 16).

2.20. Micro-CT analysis

In order to quantitatively evaluate the degree of bone damage within knee joints, this study employed high-resolution micro-CT (Hiscan XM Micro CT, Suzhou Hiscan Information Technology Co., Ltd) for three-dimensional imaging analysis on knee joint specimens from sacrificed mice. The X-ray tube parameters were configured at 60 kV and 134 μA, facilitating image acquisition at an isotropic resolution of 10 μm. A rotational increment of 0.5° was implemented over a full 360° angular span, with each step featuring a 500 ms exposure duration. The resultant images were reconstructed utilizing Hiscan Reconstruct software (Version 3.0, Suzhou Hiscan Information Technology Co., Ltd) and subsequently subjected to quantitative analysis of bone microstructural variables through Hiscan Analyzer software (Version 3.0, Suzhou Hiscan Information Technology Co., Ltd). This analysis yielded metrics, including bone volume fraction (BV/TV), trabecular thickness (Tb.Th), and trabecular separation.

2.21. Histological analysis and immunohistochemical staining

To systematically evaluate histopathological changes in knee joints, collected mouse knee joint specimens were placed in 10 % EDTA decalcification solution and subjected to decalcification treatment at 4 °C for 30 days. Following graded ethanol dehydration, 5 μm thick coronal serial sections were prepared using the paraffin embedding. Hematoxylin and eosin (HE) staining was employed to assess overall morphological changes in joint tissue. At the same time, Safranin O-Fast Green (SO/FG) staining was implemented to visualize structural characteristics of cartilage tissue and the distribution of proteoglycans. To further analyze cartilage matrix metabolic characteristics, immunohistochemical (IHC) methods were employed to detect key marker expression: primary antibodies included COL2 (Affinity Biosciences, Changzhou, China) and ADAMTS5 (Affinity Biosciences, Changzhou, China). Cartilage tissue sections subjected to IHC staining were evaluated in a blinded manner by experienced pathologists using light microscopy. A widely recognized German semi-quantitative scoring system was employed to assess both staining intensity and the extent of stained areas. Staining intensity was graded as follows: 0 (no staining), 1 (light yellow), 2 (brownish-yellow), and 3 (brown). The percentage of stained area was scored as: 0 (<5 %), 1 (5 %–25 %), 2 (26 %–50 %), 3 (51 %–75 %), and 4 (>75 %). The final IHC score was calculated by multiplying the scores for staining intensity and stained area.

2.22. Biosafety analysis

To systematically evaluate the biosafety of Fe₃O₄@C nanomaterials, whole blood specimens were collected from mice via cardiac puncture at the experimental endpoint (day 45). An automated biochemical analyzer detected key hepatic and renal function indicators, including serum ALT, AST, ALP, UREA, and CRE. In contrast, a hematology analyzer was employed to measure hematological parameters, including WBC, RBC, HGB, and NEUT. Concurrently, vital organ tissues were harvested and subsequently preserved in a 4 % paraformaldehyde solution for 48 h, followed by routine paraffin embedding to prepare 4 μm serial sections for HE staining observation.

2.23. Statistical analysis

The data were expressed as the mean accompanied by the standard error of the mean. Statistical evaluations were conducted via one-way ANOVA, supplemented by Tukey's post-hoc test to facilitate multiple comparisons. For intergroup analyses, independent samples t-tests were utilized. A P value of less than 0.05 was deemed statistically significant, with significance levels indicated as follows: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

3. Results and discussion

3.1. Synthesis and structure characterization

The ultrasmall Fe₃O₄@C nanoparticles were successfully synthesized through a hard-template-assisted controlled sintering strategy [30,31]. mSiO₂ templates with well-defined porosity were first synthesized via a CTAB-assisted sol-gel process using TEOS as the silica source. The resulting mSiO₂ particles exhibited excellent dispersibility and a uniformly porous architecture, which functioned effectively as nanoreactors for confining metal precursors during subsequent steps. Structural and morphological analysis by TEM (Fig. 1A) indicated a narrow size distribution with a statistical average diameter of 50.3 ± 4.99 nm, σ = 8.93 % (Fig. S1). Subsequent hydrothermal treatment and calcination with iron precursors were precisely confined within the mesoporous structure of mSiO₂, effectively preventing particle overgrowth and agglomeration. This nanoconfinement strategy proved essential for attaining ultrasmall dimensions (<5 nm), as validated by TEM characterization of the final product, revealing near-spherical nanoparticle morphology with exceptional monodispersity (3.09 ± 0.18 nm, σ = 5.83 %, Fig. 1B&S2). As shown in Fig. 1C, the analysis conducted using high-resolution TEM substantiated high crystallinity with lattice fringes matching the d-spacing of magnetite (0.253 nm). High-resolution structural characterization revealed a crystalline-amorphous hetero-architecture. Further, the lattice distortion detected in the core region was characterized through aberration-corrected atomic-resolution high-angle annular dark-field scanning TEM. This region was enveloped by 2–3 nm amorphous carbon shells, evidenced by gradient contrast (Fig. S3). XRD analysis (Fig. 1D) confirmed face-centered cubic Fe₃O₄ (PDF #88–0315) with characteristic peaks (2θ) at 30.2° (220), 35.5° (311), 43.1° (400), 53.4° (422), and 57.0° (511), while the amorphous carbon layer exhibited featureless diffraction patterns. XPS depth profiling afforded atomic-level insights into the surface composition and chemical states of Fe₃O₄@C nanoparticles, revealing distinct photoelectron peaks for Fe 2p, O 1s, and C 1s levels (Fig. S4). The high-resolution Fe 2p spectrum displayed characteristic doublets at 710.7 eV (Fe 2p3/2) and 724.3 eV (Fe 2p1/2) with satellite features, while deconvolution confirmed the coexistence of Fe²⁺ (710.5 eV) and Fe³⁺ (713.6 eV) states (Fig. 1E) [32,33]. The O 1s spectrum revealed three chemical environments corresponding to lattice oxygen (O_lattice, 529.9 eV), oxygen vacancies (O_vacancy, 531.3 eV), and surface-adsorbed oxygen species (O_surface, 532.0 eV) (Fig. 1F). Carbon speciation analysis identified four distinct bonding configurations corresponding to C-Fe (284.4 eV), C-C (284.8 eV), C-O (286.1 eV), and C=O (288.3 eV), suggesting covalent interactions at the core-shell interface (Fig. 1G).

Fig. 1.

Characterization of Fe₃O₄@C nanoparticles. (A) TEM image depicting mSiO₂ nanoparticles. (B) TEM image illustrating the morphology of Fe₃O₄@C nanoparticles. (C) High-resolution TEM image of Fe₃O₄@C nanoparticles (inset: local enlarged image). (D) XRD pattern representing the crystalline structure of Fe₃O₄@C nanoparticles. (E) High-resolution Fe 2p XPS spectrum of Fe₃O₄@C nanoparticles. (F) High-resolution O 1s XPS spectrum of Fe₃O₄@C nanoparticles. (G) High-resolution C 1s XPS spectrum of Fe₃O₄@C nanoparticles. (H) Hydrodynamic size distribution of mSiO₂, mSiO₂@Fe₃O₄, and Fe₃O₄@C. (I) Surface zeta potential of mSiO₂, mSiO₂@Fe₃O₄, and Fe₃O₄@C.

Dynamic light scattering systematically monitored the colloidal evolution during synthesis, revealing three distinct hydrodynamic diameter regimes corresponding to each fabrication stage (Fig. 1H). The pristine mSiO₂ templates exhibited monodisperse distribution (average diameter 58.6 nm), while the mSiO₂@Fe₃O₄ intermediate exhibited a bimodal distribution, characterized by a prominent peak observed at 122.5 nm, alongside a lesser peak located at 1128 nm. This size inflation originated from the partial collapse of mesoporous frameworks during synthesis. This significant enlargement and secondary peak emergence originated from partial hard template structure collapse, which compromised the nanoconfinement effect and permitted uncontrolled particle growth. To address this, gradient centrifugation selectively removed oversized aggregates, yielding purified Fe₃O₄@C nanoparticles with a monomodal average hydrodynamic diameter of 12.5 nm. Zeta potential measurements (Fig. 1I) demonstrated charge evolution from −45.2 ± 2.76 mV (mSiO₂) to −22.8 ± 4.55 mV (mSiO₂@Fe₃O₄) and finally −31.8 ± 1.92 mV (Fe₃O₄@C), where the restored negative potential suggests carboxyl group incorporation during carbonization. This significant negative surface charge promotes strong electrostatic repulsion between particles, effectively suppressing the aggregation of nanoparticles. In addition, size distribution evaluation conducted within 7 days showed excellent colloidal stability (Fig. S5).

3.2. Cascade enzyme-mimic property of Fe₃O₄@C nanoparticles

We systematically evaluated their antioxidant enzyme-mimetic activities by leveraging the ultrasmall dimensions and crystalline-amorphous heterointerface of Fe₃O₄@C nanoparticles. Quantitative analysis via the cytochrome c/xanthine oxidase assay revealed superior SOD-mimic activity with a specific activity of 103.6 U/mg, exceeding the commonly used antioxidant ascorbic acid (AA, 45.1 U/mg) by 2.3-fold (Fig. 2A). This enhancement is attributed to facilitated electron transfer at the Fe₃O₄-carbon interface. It is worth noting that in nanozyme-catalyzed reactions, especially for iron-based nanomaterials, the surface active sites may become depleted as the reaction proceeds, and efficient internal electron transfer is essential to replenish these active sites and sustain the catalytic cycle [20,[34], [35], [36]]. Oxygen vacancies might serve as active sites for O₂·- disproportionation in the interface,and we propose that the formation of oxygen vacancies introduces localized electron-deficient regions on the material surface, which may facilitate the adsorption and catalytic conversion of the electron-rich species (O₂·-) at these sites [[37], [38], [39], [40]]. Further validation through pyrogallol autoxidation kinetics demonstrated higher O₂·- scavenging efficiency versus AA at equivalent mass concentrations (Fig. 2B). CAT-mimic activity was systematically verified by monitoring the temporal evolution of dissolved oxygen under two orthogonal conditions: (i) variable Fe₃O₄@C nanoparticles concentrations (25–100 μg/ml, Fig. 2C) and (ii) variable H₂O₂ substrate concentrations (50–200 μM, Fig. 2D). In all cases, the dissolved oxygen signal increased monotonically with reaction time. Leveraging the dual SOD and CAT activities of Fe₃O₄@C nanoparticles, we further engineered a one-component SOD-CAT cascade capable of neutralizing the primary inflammatory initiator O₂·- (Fig. 2E). In this cascade reaction, Fe₃O₄@C nanoparticles first convert O₂·- into H₂O₂ and O₂ via SOD-mimic activity, and subsequently decompose the resultant H₂O₂ into non-toxic H₂O and O₂ through CAT-mimic activity. The cascade enzyme-mimic catalytic activity of Fe₃O₄@C nanoparticles ensures complete detoxification of ROS without accumulation of harmful intermediates, providing a robust therapeutic strategy for alleviating ROS-mediated inflammatory responses. Besides, we evaluated the peroxidase (POD)-like activity of Fe₃O₄@C using the TMB colorimetric assay to assess its potential for ROS generation. The specific activity was calculated by fitting the kinetic data obtained at multiple concentrations, yielding a value of approximately 0.0563 U/mg (Fig. S6). Our results clearly demonstrate that the POD-like activity of our material is significantly lower than those reported in relevant literature for typical iron-based nanozymes. Therefore, it can be concluded that Fe₃O₄@C contributes negligibly to ROS generation and functions predominantly as an ROS-scavenging agent.

Fig. 2.

Enzyme-mimic activity of Fe₃O₄@C nanoparticles. (A) SOD specific activity of Fe₃O₄@C and AA. (B) O₂·- inhibition of Fe₃O₄@C and AA. (C) Dissolved oxygen profiles under varied Fe₃O₄@C nanoparticles concentrations. (D) Dissolved oxygen profiles under varied substrate H₂O₂ concentrations. (E) Schematic illustration of the SOD-CAT cascade reaction for ROS elimination driven by Fe₃O₄@C enzyme-mimic activity.

It is worth noting that in nanozyme-catalyzed reactions, especially for iron-based nanomaterials, the surface active sites may become depleted as the reaction proceeds. Efficient internal electron transfer is essential to replenish these active sites and sustain the catalytic cycle [20,34]. Therefore, the enzyme-mimic activity is closely related to the intrinsic electron transfer capability of the material. This interpretation is supported by previous studies on redox-active nanozymes. The pyrogallol autoxidation method was employed to evaluate the SOD-mimic activity of our material, following well-established protocols reported in the literature [41,42]. The principle of this assay is as follows: under alkaline conditions, pyrogallol undergoes autoxidation, generating superoxide anion radicals (O₂·-), and the reaction kinetics can be monitored by measuring the increase in absorbance. SOD-mimetic nanozymes catalyze the dismutation of O₂·-, thereby inhibiting the autoxidation rate of pyrogallol. The extent of inhibition can be quantitatively correlated with the SOD-mimic activity. This method has been widely used for the evaluation of both natural SOD enzymes and SOD-mimetic materials.

Both size and composition are considered crucial factors for enzyme-mimetic activities in the design of our material. In previous studies, iron-based nanomaterials gained wide recognition due to their POD-mimic activity, which typically promotes the generation of hydroxyl radicals (•OH) and increases the levels of ROS [8]. However, subsequent studies have revealed that iron-based nanozymes can exhibit multiple enzyme-mimicking activities, most commonly the four redox-related activities: POD, oxidase, SOD, and CAT. Importantly, recent research has indicated that as the size of iron oxide nanozymes decreases, their predominant POD-mimic activity can be significantly suppressed [[43], [44], [45]]. As for composition, carbon-based nanomaterials have been widely reported to exhibit multiple enzyme-mimicking activities [[24], [25], [26],46]. Our design strategy intentionally integrates the potential advantages of both controlled size and hybrid composition to achieve the desired enzymatic profile.

3.3. In vitro biocompatibility and cellular uptake assessment of Fe₃O₄@C nanoparticles

To evaluate Fe₃O₄@C nanozymes' biosafety and determine the appropriate working concentration for subsequent functional studies, we first assessed the cytotoxicity on RAW264.7 macrophages. The RAW264.7 cells represent a well-established and widely utilized in vitro model system for investigating macrophage biology and function [47]. The impact of varying concentrations of Fe₃O₄@C nanozymes on cell viability was assessed utilizing the CCK-8 assay. As shown in Fig. 3A, cell viability exhibited a concentration-dependent response, following treatment with Fe₃O₄@C at concentrations of 100 μg/ml, 200 μg/ml, and 300 μg/ml. Notably, the 200 μg/ml Fe₃O₄@C treatment group demonstrated excellent cellular compatibility with maintained high cell viability, while the 300 μg/ml concentration began to show a slight reduction in cell viability. Following the analysis of these findings, a working concentration of 200 μg/ml was determined for use in the subsequent experimental procedures.

Fig. 3.

ROS-scavenging and anti-inflammatory effects of Fe₃O₄@C. (A) Cytotoxicity assessment of Fe₃O₄@C by CCK-8 assay (n = 3). (B–C) Live/dead staining assessment and quantitative analysis of RAW264.7 macrophages treated with Fe₃O₄@C; scale bar, 200 μm. (D–E) DCFH-DA fluorescence (green) demonstrated antioxidant capacity and quantitative analysis of ROS scavenging efficiency (n = 3); scale bar, 200 μm. (F) qRT-PCR analysis of macrophage polarization markers (n = 3). (G) Secretion profiles of inflammatory cytokines measured by ELISA (n = 3). ∗(P < 0.05), ∗∗ (P < 0.01), ∗∗∗(P < 0.001). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To systematically evaluate the biocompatibility profile of Fe₃O₄@C nanozymes beyond RAW264.7 macrophages, we performed dual-fluorescence live/dead staining using Calcein-AM and PI on two additional clinically relevant cell types: SMSCs and chondrocytes. SMSCs represent the primary cellular constituents of synovial tissue and are crucial for maintaining joint homeostasis and contributing to pathological conditions via their paracrine signalling activities [48,49]. In the pathological microenvironment of RA, activated SMSCs undergo phenotypic transformation into aggressive fibroblast-like synoviocytes, contributing to synovitis progression and cartilage destruction [50]. As demonstrated in Fig. 3B, under inverted fluorescence microscopy observation, Fe₃O₄@C-treated groups showed minimal differences in live/dead staining compared to untreated control groups. The majority of the cells exhibited green fluorescence, signifying their viability, whereas the presence of red fluorescence-labelled dead cells was minimal, confirming the excellent biocompatibility of Fe₃O₄@C. Fe₃O₄@C nanozymes also exhibited excellent biocompatibility toward SMSCs and chondrocytes after treatment (Figs. S7–8). The live/dead staining results showed minimal differences in PI fluorescence intensity between the control group and Fe₃O₄@C-treated group, indicating negligible cytotoxic effects on target cells. These findings suggested that Fe₃O₄@C nanozymes possessed favorable biocompatibility profiles, making them suitable candidates for intra-articular therapeutic applications without causing substantial cellular damage to joint tissues.

The cellular uptake capacity of Fe₃O₄@C nanozymes by macrophages was investigated using the FITC labelling technique. Time-dependent uptake experiments revealed that Fe₃O₄@C nanozymes could effectively penetrate cells. As illustrated in Fig. S9, with incubation time extending from 6 to 24 h, intracellular FITC fluorescence intensity gradually increased, indicating that Fe₃O₄@C nanozymes cellular uptake exhibited time-dependent enhancement. After 24 h of incubation, distinct green fluorescence signals were observed within cells, confirming that Fe₃O₄@C nanozymes possess excellent cellular permeability and uptake efficiency, establishing the foundation for their intracellular biological functions.

3.4. ROS scavenging capacity of Fe₃O₄@C nanoparticles

Activated macrophages generate excessive ROS during inflammation, with pathological accumulation driving chronic inflammation and tissue damage [51]. In RA, the joint microenvironment exhibits elevated ROS levels that promote: (1) synovial inflammation and hyperplasia, (2) extracellular matrix (ECM) degradation via activated metalloproteinases, and (3) progressive cartilage destruction [5,52]. While endogenous antioxidant enzymes provide natural ROS defense, their capacity is overwhelmed in RA. This redox imbalance highlights the therapeutic potential of Fe₃O₄@C that mimics antioxidant enzyme activities, offering a novel approach for RA treatment by restoring oxidative homeostasis.

To evaluate the antioxidant capacity of Fe₃O₄@C nanozymes, we evaluated the levels of intracellular ROS utilizing the DCFH-DA fluorescent probe. Fig. 3C–D shows that LPS stimulation significantly induced massive ROS production in RAW264.7 cells, manifested as intense green fluorescence signals. Remarkably, ROS levels in the Fe₃O₄@C treatment group exhibited a considerable decrease compared to those in the LPS group. with markedly diminished fluorescence intensity, Fe₃O₄@C nanozymes possess superior ROS scavenging capability. These results indicate that Fe₃O₄@C nanozymes can effectively eliminate excess ROS generated during inflammatory processes, exerting antioxidant effects.

3.5. Molecular mechanisms of Fe₃O₄@C mediated M1-to-M2 macrophage polarization

The regulation of macrophage polarization states represents a crucial mechanism for inflammation resolution and tissue repair in RA [53]. Macrophages can differentiate into two distinct functional phenotypes: M1 and M2 macrophages. Given that oxidative stress significantly influences macrophage polarization in RA [54], we investigated the immunomodulatory effects of Fe₃O₄@C nanozymes by co-culturing them with RAW264.7 macrophages to assess their impact on polarization state regulation.

In order to examine the impact of Fe₃O₄@C nanozymes on the polarization of macrophages, characteristic markers of M1 and M2 macrophages were examined using qPCR technology. In comparison to the LPS group, the treatment with Fe₃O₄@C resulted in a notable decrease in the mRNA expression levels of M1-type markers, specifically IL-1β and IL-6, while concurrently enhancing the expression of the M2-type marker CD206 (Fig. 3E). The findings suggest that Fe₃O₄@C nanozymes possess the capability to significantly inhibit pro-inflammatory reactions in macrophages while facilitating their transition to an anti-inflammatory phenotype. To further validate this phenomenon at the protein level, we employed ELISA technology to detect secretion levels of related cytokines in cell culture supernatants. As illustrated in Fig. 3F, Fe₃O₄@C treatment significantly reduced secretion levels of pro-inflammatory factors IL-1β and TNF-α, while simultaneously enhancing secretion of anti-inflammatory factors TGF-β and IL-10. The findings strongly aligned with the mRNA expression data, confirming that Fe₃O₄@C nanozymes can regulate macrophage polarization at both transcriptional and translational levels, promoting conversion from pro-inflammatory M1-type to anti-inflammatory M2-type. This demonstrates that Fe₃O₄@C nanozymes, through their excellent ROS scavenging capacity, effectively modulate the oxidative stress status of macrophages, subsequently influencing their polarization process and ultimately achieving beneficial conversion from pro-inflammatory to anti-inflammatory states. This dual-action mechanism provides a solid theoretical foundation for applying Fe₃O₄@C nanozymes in treating inflammatory diseases like RA.

3.6. Transcriptomic analysis and validation of Fe₃O₄@C-mediated synovial and cartilage protection

To comprehensively understand the molecular mechanisms underlying Fe₃O₄@C-mediated protection of RA, RNA sequencing analysis was performed on SMSCs under inflammatory conditions. The experimental design involved co-culturing SMSCs with LPS-activated RAW264.7 macrophages to simulate the inflammatory microenvironment characteristic of RA joints, followed by Fe₃O₄@C treatment. Volcano plot analysis revealed significant DEG expression patterns when comparing the SMSCs inflammatory group, co-cultured with LPS-induced macrophages, with those treated with Fe₃O₄@C nanozymes (Fig. 4B). The transcriptomic profiling identified GPX4 and SLC7A11 were markedly up-regulated and MMP3 was down-regulated upon Fe₃O₄@C intervention, suggesting dual mechanisms whereby Fe₃O₄@C nanozymes may attenuate RA progression by: (1) suppressing ferroptotic cell death through GPX4 activation and (2) inhibiting ECM degradation via MMP3 reduction. The Gene Ontology enrichment analysis demonstrated that Fe₃O₄@C primarily targeted pathways associated with ion channel activity and calcium ion binding (Fig. S10). Since the modulation of calcium channels/transport can restore glutathione peroxidase activity and GSH levels to inhibit ferroptosis [55], the observed disturbances in calcium homeostasis in our system provide compelling corroborating evidence for the involvement of Fe₃O₄@C nanozyme in regulating ferroptosis.

To substantiate the transcriptomic findings and elucidate the cytoprotective mechanisms of Fe₃O₄@C nanozymes, qPCR analysis was performed to examine key molecular markers associated with ferroptosis regulation, synovial invasion, and cartilage degradation. Synoviocytes were subjected to inflammatory stimulation by co-culturing with LPS-activated RAW264.7 macrophages, followed by Fe₃O₄@C treatment. As shown in Fig. 4C, Fe₃O₄@C treatment significantly modulated the expression of crucial biomarkers involved in synovial pathology and ferroptosis. MMP3, a key enzyme responsible for synovial invasion and cartilage degradation, was markedly down-regulated in the Fe₃O₄@C treatment group. Simultaneously, ferroptosis-protective genes, including GPX4 and SLC7A11, were significantly up-regulated upon Fe₃O₄@C treatment. These results indicated that Fe₃O₄@C nanozymes exerted synovial protection through dual mechanisms: suppressing inflammatory-mediated tissue invasion via MMP3 down-regulation and enhancing cellular resistance to ferroptosis through GPX4/SLC7A11 pathway activation. Furthermore, the chondroprotective effects of Fe₃O₄@C were evaluated using human chondrocytes under similar inflammatory conditions. According to Fig. 4D, Fe₃O₄@C treatment remarkably enhanced the expression of COL2A1, a critical marker for cartilage matrix synthesis and chondrocyte functionality. Concurrently, ferroptosis-protective genes GPX4 and SLC7A11 expression levels were significantly elevated in Fe₃O₄@C-treated chondrocytes, demonstrating consistent anti-ferroptosis activity across different joint cell types.

To systematically evaluate the regulatory effects of Fe₃O₄@C nanozymes on ferroptosis in synoviocytes and chondrocytes, we measured lipid peroxidation levels and GSH content. The results demonstrated that Fe₃O₄@C nanozyme group significantly increased intracellular GSH levels and markedly reduced lipid peroxidation compared to untreated controls (Fig. 4E–H, Figs. S11-12). These findings indicate that Fe₃O₄@C nanozymes effectively suppress ferroptosis under inflammatory conditions by enhancing GSH biosynthesis and alleviating oxidative lipid damage.

These molecular findings collectively demonstrated that Fe₃O₄@C nanozymes provided comprehensive joint protection through orchestrated regulation of inflammatory responses, ferroptosis inhibition, and promotion of tissue-specific regenerative processes via a multi-target network involving "ion channel-GPX4/SLC7A11-MMP3" regulation. It is reported that targeting the GPX4-SLC7A11 axis significantly ameliorates inflammatory conditions and RA progression [56,57]. The dual targeting of synovial invasion and cartilage degradation and enhanced cellular antioxidant capacity positioned Fe₃O₄@C nanozymes as promising therapeutic agents for RA intervention with superior biocompatibility and multi-faceted protective mechanisms.

3.7. Therapeutic efficacy in collagen-induced arthritis mouse model

This study evaluated the therapeutic efficacy of Fe₃O₄@C nanomaterials against RA by establishing a CIA mouse model, with the experimental workflow designed as shown in Fig. 5A. The CIA model has been extensively validated to mimic significant pathological characteristics of human RA, which encompasses progressive development of chronic synovitis as well as progressive bone destruction. There is substantial evidence that estrogen signaling may promote the initiation or exacerbation of autoimmune diseases, including RA [58]. Therefore, male DBA/1 mice were employed for the CIA modeling to exclude the influence of estrogen on RA disease pathogenesis.

Fig. 5.

In vivo Therapeutic Efficacy of Fe₃O₄@C Nanoparticles after Intra-articular Injection. (A) Experimental timeline illustrating CIA induction. (B) Representative micro-CT images and three-dimensional reconstructions at day 45 across three groups. (C) Quantitative analysis of bone microstructure (n = 5). CIA, collagen-induced arthritis; BMD, Bone Mineral Density; BV/TV, bone volume to tissue volume; Tb.Th, trabecular thickness; SMI, structure model index. ∗(P < 0.05), ∗∗ (P < 0.01), ∗∗∗(P < 0.001), ∗∗∗∗(P < 0.0001).

Systematic clinical arthritis scoring revealed that the Fe₃O₄@C treatment group demonstrated a significant reduction in arthritis scores as early as 5 days post-intervention, with this improvement sustained until the experimental endpoint (maximum score <8), indicating that Fe₃O₄@C effectively controlled arthritis progression (Fig. S13).

To further elucidate the therapeutic mechanisms of Fe₃O₄@C, we conducted multimodal assessments targeting RA's three core pathological features: bone damage, cartilage destruction, and synovial infiltration. The analysis conducted via Micro-CT indicated that the model group displayed typical bone surface erosion and structural destruction. In contrast, the Fe₃O₄@C treatment group demonstrated minimal bone destruction, indicating substantial osteoprotective effects (Fig. 5B, Fig. S14). Key parameters, including bone mineral density (BMD), BV/TV, structure model index (SMI), and Tb.Th in the Fe₃O₄@C treatment group showed statistically significant differences from the model group (Fig. 5C).

Histopathological analysis further confirmed the therapeutic efficacy of Fe₃O₄@C. The results of the HE staining indicated that the experimental group exhibited a significant decrease in inflammatory infiltration within the synovial tissues (Fig. 6A). SO/FG staining assessment of cartilage integrity revealed that the model group exhibited typical full-thickness cartilage destruction. In contrast, the experimental group maintained intact cartilage structure with histological features essentially consistent with the healthy control group (Fig. 6B). IHC analysis revealed that Fe₃O₄@C treatment up-regulated the expression of COL2 in the cartilage matrix while effectively inhibiting the ADAMTS5 production (Fig. 6C–F). The findings suggest that Fe₃O₄@C may facilitate the protective effects on cartilage through promoting the synthesis of COL2 and inhibiting ADAMTS5-mediated matrix degradation, thereby achieving effective intervention in the pathological progression of RA. Combined with subsequent Micro-CT analysis results, these data demonstrate that Fe₃O₄@C nanomaterials can alleviate bone damage in RA, inhibit cartilage destruction and synovial infiltration, and exhibit significant therapeutic efficacy against RA.

Fig. 6.

Histopathological Evaluation Following Intra-articular Injection of Fe₃O₄@C Nanoparticles. Comparative histological analysis of articular tissues demonstrating: (A) HE staining, scale bar, 20 μm; (B) SO/FG staining for proteoglycan content, scale bar, 20 μm; and (C–F) Immunohistochemical staining patterns in three groups and quantitative statistical analysis. (n = 3). Scale bar, 50 μm. HE, Hematoxylin and eosin; SO/FG, Safranin O-Fast Green. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.8. Biosafety evaluation

This study evaluated the biosafety of Fe₃O₄@C nanomaterials through complete blood count, blood biochemistry analysis, and tissue HE staining. Hematological analysis results demonstrated that major hematological parameters, including WBC, RBC, HGB, and NEUT in the experimental group, did not exhibit any statistically significant differences compared to those in the NC group (Fig. S15). Blood biochemical analysis indicated that key indicators reflecting hepatic and renal function (ALT, AST, ALP, UREA, CRE) all fluctuated within normal physiological ranges and no statistically notable differences across the groups (Fig. S16). Histopathological examination further confirmed that vital organs in the experimental group showed no significant pathological changes, with intact tissue structures and absence of toxic reactions such as inflammatory infiltration, cellular degeneration, or necrosis (Fig. S17). These data demonstrate that Fe₃O₄@C nanomaterials possess excellent biocompatibility, providing important safety evidence for their further clinical translation applications.

Although this study has comprehensively demonstrated the excellent therapeutic efficacy of ultrasmall Fe₃O₄@C nanoparticles in RA through their multiple enzyme-mimetic activities at both cellular and animal levels, and preliminary validation has confirmed their biosafety, key scientific questions regarding their in vivo biodistribution and metabolism must be addressed to facilitate clinical translation. On one hand, the ultrasmall size of the nanoparticles confers a high surface-area-to-volume ratio, which may promote uptake by local tissues and subsequent entry into systemic circulation via lymphatic drainage and blood flow. Once in the bloodstream, nanoparticles are prone to capture by the mononuclear phagocyte system, with the liver and spleen likely serving as the primary clearance routes for iron oxide nanoparticles [59]. On the other hand, the ultrasmall size may facilitate accumulation at inflammatory sites through the enhanced permeability and retention effect. Although we have not yet conducted in-depth in vivo studies on the pharmacokinetics, biodistribution, and clearance of these nanoparticles, their application in RA treatment has been increasingly recognized [60,61].

3.9. Limitations and future perspectives

While this study successfully demonstrates the potent multi-functional capabilities of the Fe₃O₄@C nanozyme in vitro, we acknowledge its limitations. The functional validation was conducted using a single, optimized concentration to provide a clear proof-of-concept. A comprehensive investigation into the precise dose-response relationships for each individual protective function (antioxidation, anti-inflammation, chondroprotection) was beyond the scope of this initial report but represents a critical direction for future work. Furthermore, the in vivo biodistribution, metabolic fate, and long-term biosafety of the nanozymes remain essential subjects for subsequent research prior to any clinical translation.

Our future studies will, therefore, be focused on: (1) Systematically elucidating the dose-dependent effects of the nanozyme in more complex 2D and 3D culture models; (2) Employing small animal models to rigorously evaluate its pharmacokinetics, retention time, and safety profile in a physiological environment. These investigations will build directly upon the foundational findings presented here and are essential for advancing this nanozyme towards therapeutic applications.

4. Conclusion

This study demonstrates that Fe₃O₄@C nanozymes exert therapeutic effects against rheumatoid arthritis through three synergistic mechanisms: (1) Fe₃O₄@C nanoparticles exhibit dual enzyme-mimetic activities, functioning as both SOD and CAT mimics, which effectively scavenge ROS, thereby facilitating the phenotypic switch of macrophages from inflammatory M1 to anti-inflammatory M2; (2) Activation of the GPX4-SLC7A11 axis inhibits ferroptosis in synoviocytes and chondrocytes, maintaining redox homeostasis; and (3) They markedly suppress synovial invasion while enhancing cartilage matrix synthesis, thereby preserving joint integrity. These findings establish Fe₃O₄@C nanozymes as a novel therapeutic strategy for RA through their combined antioxidant, anti-inflammatory, and chondroprotective actions.

CRediT authorship contribution statement

Mingyue Yan: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization. Kehao Hou: Writing – original draft, Methodology, Investigation. Jinpeng Zhao: Writing – original draft, Methodology, Investigation. Shuangshan Li: Writing – review & editing, Methodology. Xiaolin Wu: Supervision, Funding acquisition. Shichao Bi: Supervision, Funding acquisition, Conceptualization. Jing Yu: Supervision, Conceptualization. Tianrui Wang: Supervision, Funding acquisition, Conceptualization. Yingze Zhang: Project administration, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.