Materials

Iron acetylacetonate, N,N-carbonyl diimidazole, oleylamine, oleic acid, calcium hydroxide, polyethyleneimine (Mw ~ 1800), hydroxy polyethylene glycol (Mw ~ 5000), sodium hyaluronate, and 3-aminobenzeneboronic acid were purchased from Aladdin Bioscience Co. (Shanghai, China). Fetal bovine serum, DMEM high glucose medium and other biologically related reagents or kits were purchased from KeyGen Biotechnology Co. (Nanjing, China).

Preparation of Fe3O4 NPs

To synthesize iron nanoparticles, initially, 1.4 g of iron acetylacetonate was weighed and mixed with 8 mL of oleic acid and 12 mL of oleylamine. The reaction commenced by purging with N2 at room temperature for 15 min, followed by vacuuming using an oil pump for 15 min while maintaining the nitrogen purge. The reaction mixture was heated to 120 °C and held for 2 h. Subsequently, the temperature was raised to 220 °C and maintained for 30 min, then increased at a rate of 2 °C/min to 300 °C and held for 30 min. The solution was cooled to 120 °C and kept for an additional 30 min. After cooling to 120 °C, nitrogen flow was stopped, and air was introduced to continue the reaction for 90 min. Upon completion, the heating was turned off, and the liquid was transferred to a 50 mL centrifuge tube. Acetone (50 mL) was added, and centrifugation at 8000 rpm for 5 min separated the supernatant, which was discarded. The precipitate was dissolved in 5–10 mL of hexane, followed by addition of 40 mL of ethanol and repeated centrifugation. This washing process was repeated twice. Finally, the nanoparticles were dissolved in 5 mL of hexane, sealed, and stored at room temperature.

Preparation of PEI-PEG

Anhydrous and alcohol-free chloroform solvent was prepared by initially mixing chloroform with deionized water, subjecting it to triple extraction, and treating it with calcium hydride under reflux at 65 ℃ for 4 h to obtain redistilled chloroform. In the subsequent step, HO-PEG2000-OH was activated by combining 2 g of HO-PEG2000-OH with 0.821 g of CDI in 20 mL of the prepared chloroform solvent, stirring overnight under a nitrogen atmosphere, and precipitating the resulting PEG-CDI with cold ether followed by vacuum drying. Further, 1.4 g of PEG-CDI and 1.9 g of PEI (Mw ≈ 1800 Da) were dissolved in 7.5 mL of chloroform, mixed, and reacted for 12 h under nitrogen. The product, PEI-PEG, was obtained after precipitation with cold ether, washing, and vacuum drying, and stored at −20 ℃. Structural confirmation was performed by dissolving 5 mg of PEI-PEG in 500 μL of deuterated chloroform (CDCl3), transferring it to an NMR tube, and analyzing its structure using 1H NMR spectroscopy.

Preparation of Fe3O4-PEI-PEG NPs

Fe3O4 nanoparticles (20–30 μL), dispersed in hexane, were transferred to an EP tube and ethanol was added to induce precipitation. After centrifugation, the supernatant was discarded, and the precipitate was purged with nitrogen, dried, and weighed to determine the nanoparticle mass. The nanoparticles were then redispersed in chloroform and mixed with a PEI-PEG solution, followed by sonication for 30 min to 1 h and overnight shaking. The resulting mixture was dried using rotary evaporation and further subjected to vacuum drying for 24 h. The product was subsequently dispersed in water via sonication and subjected to dialysis for three days with three changes of dialysate. Finally, the product was stored under sealed conditions.

Determination of total iron concentration in Fe3O4-PEI-PEG NPs

A small volume (approximately 10 μL) of nanoparticles was treated with nitric acid and heated to 200 °C, with continuous shaking to prevent evaporation. Upon evaporation of most of the nitric acid, the residue was diluted with water, and the dilution factor was recorded. Phenanthroline and hydroxylamine hydrochloride were dissolved in acetic acid-ammonium acetate solution and 2M hydrochloric acid, and added to an EP tube. After incubating for 2 h in the dark, the solution was transferred to a 96-well plate and the absorbance was measured. The iron concentration of the sample was determined from a standard curve, and multiplied by the dilution factor to obtain the total iron concentration of the original sample.

Fe3O4-PEI-PEG NPs loaded siRNA method

A solution of 1 OD260 of siRNA (approximately 33 μg) (Shanghai GenePharma Co., Ltd.) was dissolved in 40 μL DEPC water. The concentration of siRNA was precisely measured using a μDrop Plate (Thermo Scientific™, USA) to determine the nucleic acid concentration. Fe3O4-PEI-PEG NPs were diluted accordingly, and a proportional volume was aspirated using a pipette. The NP solution was then added dropwise to the siRNA solution while gently mixing with a pipette gun. The mixture was thoroughly blended and allowed to stand at room temperature for 30 min to form complexes of siRNA and Fe3O4-PEI-PEG NPs (si-Fe NPs).

Evaluation of the ability of Fe3O4-PEI-PEG NPs to protect siRNAs

Free siRNA and si-Fe NPs were mixed with various concentrations of RNA degrading enzyme (RNase) in 0.2 mL EP tubes, sealed with film, and incubated for 1 h in a 37 °C water bath. Subsequently, gel electrophoresis was conducted.

Evaluation of cellular uptake capacity of si-Fe NPs

The C28/I2 cell line was divided into two groups in a 24-well plate: the naked siRNA group and the siRNA-Fe NPs group. For the siRNA-Fe NPs group (siRNA: 80 pmol/mL), FAM-siNC-Fe NPs were prepared, while equal amounts of free FAM-siNC were used for the free siRNA group. Cells were incubated in 24-well plates for 12 h, followed by medium aspiration and two PBS washes to remove extracellular free siRNA or nanoparticles. Next, 200 μl of 4% paraformaldehyde was added for fixation. After fixation, the cells were washed twice with PBS and stained with 200 μl of DAPI solution. Following 2–3 PBS washes, fluorescence microscopy was used for observation.

Preparation of HA-PBA polymers

HA (2 g) was dissolved in 200 mL of deionized water and stirred overnight until fully dissolved, yielding a clear and viscous liquid solution. Subsequently, PBA (940 mg) was dissolved in 20 mL of deionized water under sonication until complete dissolution. This solution was then combined with the HA solution. The pH of the resulting solution was adjusted to 6.5~7 using 1 M NaOH and 1 M hydrochloric acid solutions. DMTMM (1.6 g) was then dissolved in 10 mL of deionized water by ultrasonication, followed by mixing with the HA + PBA solution. The reaction mixture was stirred at room temperature under a nitrogen atmosphere for 72 h.

Preparation of HPP hydrogels in different ratios

Firstly, PVA solutions of different concentrations were prepared. PVA (9 g) was added to 100 mL of deionized water and heated to 90 ℃ to dissolve. After cooling to room temperature, the solution was transferred to a sample bottle, resulting in a 9% (w/v) PVA solution. Portions of this solution were diluted to obtain 3 and 6% PVA solutions with deionized water. Next, HA-PBA solutions of varying concentrations were prepared. HA-PBA samples (2, 4, and 6 mg) were individually weighed into 1.5 mL EP tubes. To dissolve the samples, 200 μL of PBS was added to each tube, followed by continuous sonication. This process yielded 1, 2, and 3% (w/v) HA-PBA solutions in total. Subsequently, 66.7 μL of 3, 6, and 9% PVA solution was added to corresponding portions of the HA-PBA solutions. A pipette gun was used to mix the PVA solution into the HA-PBA solution while stirring, ensuring thorough mixing before continuing stirring to maintain uniformity. After stirring for 5–10 s, the liquid in the EP tube condensed into a viscous solid. The mixture was then centrifuged at 10,000 rpm for 3 min to remove air bubbles. Successful formation of the HPP hydrogel was confirmed by inverting the EP tube; no liquid flow indicated successful gelation.

Hydrogel rheology testing

The HPP and si-Fe-HPP hydrogel were molded into cylinders with a diameter of 8 mm and height of 2 mm using a PTFE mold. These cylinders were carefully removed and positioned on the base plate of a rheometer. Subsequently, a series of rheological tests were conducted using the rheometer. A strain sweep was performed to evaluate the response over a strain range of 0.1–1000% at a fixed angular velocity frequency of 1 Hz. An angular velocity frequency sweep followed, maintaining a strain of 1% while scanning angular velocities from 0.1 to 20 Hz. A time sweep was conducted at 1% strain and 1 Hz frequency, spanning 120 s with time as the x-axis. Lastly, a step strain test involved varying strains (1–1000%) at 1 Hz frequency, with each strain step lasting 2 min. These tests aimed to characterize the mechanical properties and viscoelastic behavior of the hydrogel formulations under different conditions, providing insights into their potential applications in biomedical settings.

In vitro degradation of hydrogels with different ratios of HPP

Nine different hydrogel formulations were prepared with varying ratios of HA-PBA and PVA: 1% HA-PBA + 3% PVA (denoted as 1% + 3%, and so forth), up to 3% HA-PBA + 9% PVA. Each formulation was duplicated, resulting in a total of 18 hydrogel samples. Initially, each hydrogel sample was immersed in 2 mL of PBS to reach swelling equilibrium, and the initial weight was recorded. Subsequently, pairs of hydrogels with identical compositions were separately incubated in 2 mL of PBS and 2 mL of 100 μM H2O2 solution at 37 °C. Every 48 h, the incubation medium was replaced, and the hydrogels were gently centrifuged at low speed. After removing the liquid, the surface of each hydrogel was dried with filter paper, reweighed, and the weight recorded. This process was repeated until complete degradation of the hydrogel occurred in the PBS medium. These experiments were conducted to assess the degradation behavior of the HPP hydrogels under oxidative conditions, mimicking environments relevant to biomedical applications.

Investigation of ROS-responsive degradation behavior of HPP hydrogels

Four identical HPP hydrogel samples (2% + 9%) were prepared and individually placed into separate 2 mL EP tubes. Each hydrogel sample was initially immersed in 2 mL of PBS until swelling equilibrium was reached. The initial weight of each hydrogel was recorded as 100%. Subsequently, the PBS was replaced with solutions containing 50 μM H2O2, 100 μM H2O2, and 500 μM H2O2 in separate EP tubes containing one hydrogel sample each. The samples were then incubated at 37 °C, and every 48 h, they were gently centrifuged at low speed. After removing the supernatant, the surface of each HPP hydrogel was carefully dried with filter paper, reweighed, and the weight recorded. The incubation medium (PBS or H2O2 solution) was replenished after each measurement, and the process was repeated until complete degradation of the HPP hydrogel was observed under the respective conditions. This experimental setup aimed to evaluate the degradation kinetics of the HPP hydrogel under oxidative stress conditions induced by varying concentrations of H2O2, simulating environments relevant to biomedical applications.

Investigation of H2O2 scavenging capacity of HPP hydrogel in vitro

Four different concentrations of H2O2 solution (50, 100, 500 μM, 1 mM) were prepared in duplicate in 1.5 mL EP tubes, totaling eight samples, each with a volume of 500 μL. One hundred microliters (100 μL) of HPP hydrogel was immersed in each H2O2 solution and incubated at room temperature for 2 h in darkness. Subsequently, the hydrogel samples were centrifuged to collect 100 μL of the supernatant, which was then transferred to a 96-well plate. To this, an equal volume of 1 M aqueous sodium iodide solution was added and allowed to react for five minutes in the absence of light. UV absorption spectra were then measured using a spectrophotometer within the wavelength range of 300~500 nm to analyze the reaction products formed under oxidative conditions induced by different concentrations of H2O2. This methodological approach aimed to assess the oxidative degradation behavior of the HPP hydrogel and characterize its response to varying levels of oxidative stress.

Cell compatibility

C28/I2 cells were seeded into 96-well plates. Si-Fe nanoparticles (NPs) were prepared at various concentrations based on iron mass, dispersed in complete medium, and added to the 96-well plates. Final concentrations of iron (Fe) in the wells were adjusted to 0, 0.049, 0.088, 0.195, 0.390, and 0.781 μg/mL, with each concentration tested in triplicate wells. The same experimental setup was applied to HPP and si-Fe-HPP hydrogels. This approach ensured consistent testing across different materials (Si-Fe NPs, HPP, and si-Fe-HPP hydrogels).

RT-qPCR

Cellular mRNA was extracted from chondrocytes following the manufacturer’s instructions using the RNA-quick Purification Kit (#RN001, ES Science, Shanghai, China). The HiScript-TS 5′/3′ RACE Kit (RA101, Vazyme, China) was utilized for the detection of MMP-13, IL-1β, ACAN, and Col II expression levels. RT-qPCR analysis was conducted on a LightCycler 480 PCR system (Roche, Switzerland) using ChamQ Universal SYBR qPCR Master Mix (Q711, Vazyme, China). Primer sequences for the assays are provided in Table S2.

Animal experiments

All animal experiments were conducted with the approval of the Ethics Committee of Drum Tower Hospital affiliated with Nanjing University. The experiments adhered strictly to the ARRIVE guidelines and were carried out in compliance with the U.K. Animals (Scientific Procedures) Act of 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments, and the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 8023, revised 1978). Male 12-week-old C57BL/6 mice were procured from the Animal Model Research Center of Nanjing University and were housed under specific pathogen-free conditions with ad libitum access to food and water. The OA mouse model was induced by surgical destabilization of the medial meniscus under isoflurane anesthesia. Twelve-week-old male C57BL/6 mice were prepped by sterilizing the skin with iodophor. During surgery, the joint cavity of the right knee was opened, and the medial meniscotibial ligament was severed to destabilize the medial meniscus from the tibial plateau under a stereomicroscope. Loosening was confirmed using microscopic forceps, after which the incision was sutured layer by layer and sterilized again. The same surgical procedure was performed on the Sham group without ligament resection. Following Sham or DMM surgery on the right knee joint, mice were randomly allocated into five groups: (1) Sham group; (2) DMM group; (3) DMM with HPP injection group; (4) DMM with si-Fe NPs injection group; and (5) DMM with si-Fe-HPP injection group. One week post-surgery, the si-Fe-HPP group received intra-articular injections of 10 μL si-Fe-HPP (siMMP-13: 1.5 nmol/10 μL) once per month for a period of 12 weeks. The HPP and si-Fe NPs groups were administered equivalent amounts of their respective components found in si-Fe-HPP, while the Sham and DMM control groups received an equal volume of PBS. After 12 weeks of treatment, mice were euthanized, and joint tissues, major organs, and serum samples were collected for efficacy and safety evaluation. Motor function was assessed using the open field test, gait analysis, and Von Frey test before the animals were sacrificed for sampling (Fig. 5a).

Histological analysis

A 10% ethylenediaminetetraacetic acid (EDTA) solution (#1340, Biofroxx, Germany) was utilized for decalcifying mouse knee joints, followed by embedding in paraffin blocks. The knee joints were sectioned coronally into continuous 5 μm slices using a microtome (Thermo, Germany). Sections were subjected to H&E staining (#C0105S, Beyotime) and S.O. staining (#G1371, Solarbio) to evaluate synovitis and cartilage lesions, respectively. Synovitis severity was assessed using a scoring system ranging from 0 to 3, and cartilage degeneration was evaluated using the Osteoarthritis Research Society International (OARSI) grading system (0–6) by two independent blinded observers. The maximum scores for synovitis and OARSI were recorded, and their average values were calculated. Furthermore, to assess the in vivo biocompatibility of si-Fe-HPP, sections of major organs (heart, liver, spleen, lungs, and kidneys) were stained with H&E (#C0105S, Beyotime). Serum biochemical indicators were also analyzed to evaluate safety profiles.

Immunohistochemical staining

Endogenous peroxidase activity was quenched using 3% (v/v) H2O2. Horseradish peroxidase-conjugated secondary antibodies, specifically goat anti-rabbit or anti-mouse immunoglobulin G (IgG) (Biosharp, Shanghai, China), were used. Immunohistochemical staining was visualized using the Ultrasensitive DAB kit (#1205250, Typing, Nanjing, China). Non-immune IgG served as a negative control during the immunohistochemical staining process.

Micro-CT analysis

Micro-CT scans were conducted using a VivaCT 80 scanner (Scanco Medical AG, Switzerland) equipped with a 70 kVp light source. 3D reconstructions of mouse knee joints were generated using Scanco Medical software (Scanco, Switzerland) with a threshold set at 220 to evaluate bone morphology. Quantitative analysis of osteophyte formation was performed based on the 3D reconstructed images.

Behavior tests

The open field test was conducted in a 50 cm × 50 cm square arena with 25 cm high walls under quiet, dimly-lit conditions. Mice movement trajectories were recorded using a tracking system (Zhenghua Technology). For gait analysis, mice freely walked on a 70 cm × 20 cm white runway. Prior to testing, the forepaws were marked with red ink and hindpaws with blue ink to record footprints. Measurements were analyzed by two independent observers blinded to the experimental conditions. Pain sensitivity was assessed using an electronic von Frey anesthesiometer (IITC, Woodland Hills, USA), recording the mechanical paw withdrawal threshold.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism software (version 8.0) and SPSS software (version 25.0). Quantitative results represent findings from at least three independent experiments. Graphical Analysis Using GraphPad Prism 8.0 and Origin 2022. No samples or animals were excluded from the analysis. To assess data variance equality and normal distribution, Levene’s test and Shapiro–Wilk’s test were applied, respectively. For comparisons among multiple groups, either one-way or two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was performed. Data are presented as mean ± standard deviation, and statistical significance was set at P < 0.05.