The synthesis of mPEG-PCL was performed using ring-opening polymerization, using poly (ethylene glycol) (mPEG) as the initiator and stannous octoate as the catalyst. Caprolactone and mPEG were mixed in a polymerization reaction vessel, with the addition of stannous octoate (0.3%, w/w). The synthesis reaction was carried out at 150°C under nitrogen protection for 6 h. The obtained mPEG-PCL was purified through re-precipitation in petroleum ether and dried in a vacuum oven at 40°C. The purified mPEG-PCL was analyzed using nuclear magnetic resonance (NMR) spectroscopy in deuterated chloroform. 1H NMR (400 MHz, CDCl₃) δ, ppm: 1.36 (s, 2H), 1.64 (s, 4H), 2.31 (s, 2H), 3.38 (s, 3H), 3.64 (s, 4H), 4.04 (s, 2H). The spectrum of 1H NMR was presented in Supplementary Figure S4.

Co-loaded polymersomes were prepared using an emulsion-solvent evaporation method, which facilitated the encapsulation of both hydrophilic cordycepin and hydrophobic PBA. This method included two phases: emulsification and dilution. During the emulsification phase, mPEG-PCL and PBA were dissolved in 4 mL of chloroform to create the oil phase, while 0.8 mL of an aqueous cordycepin phosphate-buffered saline (PBS) solution (pH 7.4) was added into the mixture. The primary emulsion was formed via ultrasonication at 150 W for 5 min in an ice-water bath. This primary emulsion was then added dropwise to 10 mL of 0.5% polyvinyl alcohol (PVA) solution, followed by ultrasonication at 300 W for 3 additional min. During the dilution phase, the secondary emulsion was gradually added to 40 mL of 0.1% PVA solution under vigorous stirring at 600 rpm for 30 min. Finally, chloroform and excess water were removed using a rotary evaporator to obtain the required polymersomes.

The purification of co-loaded polymersomes were carried out using a combination of membrane filtration and dialysis. Briefly, an appropriate volume of polymersome solution was first passed through a 0.22 μM membrane to remove large aggregates. The resulting filtrate was transferred into the dialysis tubing (MWCO: 3 kDa) and immersed in 250 mL of PBS, stirred gently for 2 h at room temperature to remove free cordycepin. After dialysis, the sample was centrifuged at 10,000 rpm for 10 min to remove any precipitated PBA or particulate residues. The obtained polymersomes were freeze-dried for future use.

To determine encapsulation efficiency (EE%) and drug loading efficiency (DL%), 5 mg of lyophilized drug-loaded polymersomes were dispersed in 5 mL DI water. The solution was mixed with methanol and sonicated to ensure complete release of encapsulated compounds. The concentrations of cordycepin and PBA were then determined using the HPLC method described above. EE% and DL% were calculated as follows: Encapsulation efficiency  EE % = W encapsulated   compund W feeding   compound × 100 % Drug loading efficiency  DL % = W e n c a p s u l a t e d   c o m p o u n d W t o t a l   p o l y m e r s o m e × 100 % where W _{encapsulated compound} represented the weight of the compound encapsulated in the loaded polymersomes, W _{feeding compound} represented the total weight of the feeding compound, and W _{total polymersomes} represented the total weight of the compound-loaded polymersomes.

OFAT experiments were carried out to investigate the effects of various variables on encapsulation efficiency, including ultrasonic power for primary emulsification, ultrasonic power for secondary emulsification, PVA concentration in the aqueous phase, organic-to-aqueous phase ratio, and emulsification stirring speed. During each OFAT experiment, all variables were held constant except for the parameter under investigation. The ultrasonic power for primary emulsification was varied at 100 W, 150 W, 200 W, while the secondary emulsification power was varied from 150 W to 300 W in increments of 50 W. PVA concentration in the aqueous phase was varied from 0.5% to 2%, the organic-to-aqueous phase ratio varied between 1:4 and 1:7, and the emulsification stirring speed ranged from 200 rpm to 800 rpm.

Based on the significant parameters identified from the OFAT experiments, a Taguchi orthogonal experimental design was applied to further optimize the formulation. Selected factors were evaluated using a L₉ (3⁴) orthogonal array, which involved 9 experiments with three levels across four factors. Encapsulation efficacy was selected as the response variable, and the results were analyzed statistically using range analysis. Subsequently, optimal polymersomes were prepared under the best conditions to investigate the maximum encapsulation efficiency.

The particle size and polydispersity index (PDI) were determined using dynamic light scattering (Nano-ZS 90, Malvern). Measurements were carried out in triplicate at 25°C after equilibration for 2 min.

The stability of the polymersomes was evaluated by monitoring changes in particle size distribution and PDI under 4°C over a 1-month period. Observations were recorded at intervals of 0, 10, 20, and 30 days.

For morphological analysis, co-loaded polymersomes were placed onto a specialized copper mesh, negatively stained with 2% phosphotungstic acid, and then dried at room temperature. The morphology of the co-loaded polymersomes was observed under a transmission electron microscopy (TEM).

XRD analysis was carried out to determine the crystallographic structure of the co-loaded polymersomes. The patterns of free compound combination, blank polymersomes, a physical mixture of free compound combination and blank polymersomes, and co-loaded polymersomes were analyzed using an X-ray diffractometer with Cu Ká radiation. Measurements were performed at a voltage of 40 kV and a tubing current of 30 mA. The scanned angle range was set at 2θ from 10° to 80° with the scanning speed of 0.02°/min.

The in vitro release profile of co-loaded polymersomes was determined using the dialysis method. Briefly, co-loaded polymersomes were placed in dialysis bags (MWCO = 3 kDa) and incubated in 50 mL of PBS buffer containing varying concentrations of H₂O₂ (0, 50, and 100 µM) to mimic the osteoarthritic environment. The incubation was carried out at 37°C with gentle shaking. Aliquots were taken at predetermined time points, replaced with fresh media, and stored at −20°C for further analysis. After dialysis, the concentrations of cordycepin and PBA were determined using the HPLC method as described above.

The ROS scavenging ability of co-loaded polymersomes was assessed using luminol-based chemiluminescence (Khan et al., 2014). Briefly, a series of co-loaded polymersome solutions with varying concentrations of PBA (0, 10, 100, 250, 500, and 750 μg/mL) were prepared. In each well of a white 96-well microplate, 100 μL of H₂O₂ solution (100 μM) was mixed with 100 μL of each polymersome solution. Then, 10 μL of CuCl₂ solution (100 μM) and 50 μL of luminol solution (50 μM) were added in sequence. After thorough mixing, the chemiluminescence intensity was measured using a microplate reader. The ROS scavenging efficiency of the co-loaded polymersomes at a concentration of x μg/mL was calculated using the following equation: S c a v e n i n g   e f f i c i e n c y x   μ g / m L % = I n t e n s i t y 0   μ g / m L − I n t e n s i t y x   μ g / m L I n t e n s i t y 0   μ g / m L × 100

Murine RAW264.7 macrophage cell line was purchased from the Cell Bank of Type Culture Collection, Chinese Academy of Science (Shanghai, China). Cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin under the condition of 37°C and 5% CO₂.

The cytotoxicity of free compounds, their combination, and co-loaded polymersomes against RAW264.7 cells were assessed using the CCK-8 assay. Briefly, cells were seeded into 96-well plates at a density of 2 × 10⁴ cells/well and incubated overnight. Subsequently, cells were treated with varying sample concentrations for 48 h. For the combination group and co-loaded polymersomes, cordycepin and PBA were mixed in equal or near-equal molar ratios, with each compound contributing approximately half of the total concentration. Subsequently, 10 μL of CCK-8 solution was added into each well and incubated for another 2 h. Finally, absorbance was measured at 490 nm using a microplate reader (BioTek XS2, United States).

RAW264.7 cells were seeded in 24-well plate at a density of 1× 10⁵ cells/well. After overnight culture, cells were pre-treated with lipopolysaccharide (LPS, 1 μg/mL) for 12 h to induce an inflammatory response. Subsequently, cells were exposed to free cordycepin, free PBA, their free compound combination, or co-loaded polymersomes for 24 h. Cytokine contents, including IL-1β and TNF-α, in the supernatants were determined using commercially available ELISA kits (LianKe Bio Co., Ltd., Hangzhou, China) according to the manufacturer’s instructions. Absorbance was then measured at 520 nm using a microplate reader.

ROS production in RAW264.77 cells were measured using the Amplex Red hydrogen peroxide assay (Mishin et al., 2010; Karakuzu et al., 2019). This assay employs horseradish peroxidase (HRP) to catalyze the oxidation of 10-acetyl-3,7- dihydroxypenoxazine in the presence of H₂O₂, resulting in the red fluorescent product resorufin. In brief, cells were seeded in 96-well plate with phenol-free complete DMEM medium, treated with LPS, and subsequent treated with various samples. Then, 50 μL of culture supernatants were mixed with 50 μL of Amplex Red working buffer containing 50 μM Amplex Red and 10 U/mL HRP. The mixture was incubated in the dark at room temperature for 30 min. The fluorescence was detected with an excitation wavelength of 530 nm and an emission wavelength of 590 nm in a microplate reader.

Specific pathogen-free (SPF) grade Sprague Dawley (SD) rats (weight 200 ± 20 g, 50 females) were purchased from Jiangsu University. All rats were maintained in a temperature-controlled facility with a 12-h light/dark cycle at 24°C and provided free access to standard diet and tap water. All experimental procedures were ethically approved by the Animal Use and Care Committee of Jiangsu University (Approval number: 11,657), and all procedures were conducted in accordance with the NIH guidelines for the Care and Use of Laboratory Animals.

The OA rat model was established according to the reference (Wang et al., 2014; Cheng et al., 2019; Liu et al., 2021). A solution of 5% papain and 0.03 M L-cysteine was mixed at a volume ratio of 1:2, where L-cysteine served to preserve the activity of papain. The obtained mixture (10 μL) was injected into the left knee joint of the rats on days 0, 4 and 7 at the same dose. Rats in the control group were injected with equal volumes of sterile saline in the left knee joint. Two weeks post-injection, the OA rat model was successfully established. Prior to treatments, the diameters of right and left knee were determined every other day in the medial-lateral direction, with the caliper positioned perpendicular to the leg axis. The joint swelling was expressed as the ratio of the knee diameter to the baseline mean knee diameter.

To investigate the biodistribution, co-loaded polymersomes were labeled with the water-soluble near-infrared dye sulfo-Cyanine7 (sulfo-Cy7). In brief, 1 μmol of sulfo-Cy7 NHS ester was dissolved in 1 mL of anhydrous DMF, and 1.1 equivalents of mPEG-PCL-NH₂ and DIPEA were added. The mixture was stirred in the dark for 1 h at room temperature, followed by dialysis in the dark to remove the organic solvent. Finally, the Cy7-labeled polymers were incorporated into polymersomes at the percentage of 0.5% during preparation.

Cy7-labeled co-loaded polymersomes were then injected intraperitoneally into OA rats. The rats were anaesthetized with 2% isoflurance and whole-body images were captured after administration using the Xtreme In vivo Imaging System at 0, 12 and 24 h. For imaging Cy7, an excitation filter of 750 nm and an emission filter of 830 nm were used. Additionally, images of various organs were also captured after sacrifice of the OA rat 24 h post-administration.

OA rats were randomly divided into six groups (n = 7) and treated intraperitoneally (IP) every other day for 4 weeks with: (1) saline, (2) free cordycepin (15 mg/kg), (3) free PBA (7.5 mg/kg), (4) free combination of cordycepin and PBA (15 mg/kg cordycepin and 7.5 mg/kg PBA), (5) co-loaded polymersomes (15 mg/kg cordycepin and 7.5 mg/kg PBA), and (6) blank polymersomes. Healthy rats were used as control. The IP injection volume was kept at 0.75 mL. Periodic measurements of body weights and joint swelling of rats were performed throughout the study.

After 4 weeks of treatment, rats were euthanized, and sacrificed for further analysis. Serum levels of IL-1β and TNF-α were detected using ELISA kits (Lianke Bio, Hangzhou, China).

All collected knee joints were fixed in 10% formalin for 3 days, decalcified for 5 days, embedded in paraffin, and sectioned into 5-μm thick slices with 200-μm intervals. The paraffin sections were dewaxed, hydrated in graded alcohol, and stained using hematoxylin and eosin (H&E) solution. The pathological changes in the articular cartilage were analyzed using Image-Pro image analysis software.

Due to the distinct physicochemical properties, cordycepin and PBA exhibited different encapsulation trends. The encapsulation efficiency peaked at an optimal sonication power of 150 W for primary emulsification (Figure 1A). In detail, at lower sonication powers, phase separation within the vesicle material was observed, resulting in reduced encapsulation efficiency. Beyond the optimal level, further increases in sonication power led to a decline in encapsulation efficiency, which may be associated with disruption of the vesicle structure due to excessive ultrasonic energy (Yamaguchi et al., 2009; Yang et al., 2021). However, secondary emulsification power had a minimal effect on encapsulation efficiency (Figure 1B). Stirring speed during emulsification significantly impacted encapsulation; increasing speed improved encapsulation until a threshold was reached, beyond which encapsulation efficiency declined (Figure 1C). Stirring speed had a stronger influence on the encapsulation efficiency of cordycepin than PBA, likely due to differences in their solubility and molecular interactions with the polymer membranes.

Formulation factors were also investigated. As shown in Figure 1D, cordycepin exhibited a marked decrease in encapsulation efficiency when the PVA concentration increased from 1.5% to 2%, suggesting that higher PVA content may interfere with effective entrapment. In contrast, PBA showed a slight increase in encapsulation efficiency between 1% and 1.5% PVA, followed by a minor decline at 2%, indicating that intermediate PVA concentrations are more favorable for PBA loading. Additionally, increasing the organic-to-aqueous phase ratio slightly decreased cordycepin encapsulation efficiency while significantly reducing PBA encapsulation efficiency, dropping from 86% to approximately 50% (Figure 1E). This decline was attributed to the instability of the primary emulsion. When the volume of the aqueous phase increased as the organic-to-aqueous phase ratio reduced from 1/4 to 1/7, the size of the resulting primary emulsion significantly increased, compromising its structural integrity. This loss of stability eventually led to the breakdown of the emulsion and the subsequent release of PBA from the organic layer. Overall, PBA exhibited consistently lower encapsulation efficiency than cordycepin during the OFAT optimization.

The results of OFAT experiments indicated that secondary emulsification power had minimal influence on encapsulation efficiency, so four other factors were selected for further optimization using a Taguchi orthogonal design. Based on the OFAT results, the ultrasonic power for primary emulsification was evaluated between 100 W and 200 W, while PVA concentration was examined between 0.5% and 1.5%. The organic/aqueous phase ratio was tested at 1:4, 1:5, and 1:6. Lastly, stirring speed was set at 200, 400, and 600 rpm.

An L ₉ (3⁴) Taguchi orthogonal design was applied to identify the optimal preparation conditions for maximizing the encapsulation efficiency of cordycepin and PBA. After 9 experimental runs, the results were presented in Table 1, 2. We analyzed the effects of each factor level on the encapsulation efficiency for cordycepin and PBA using range analysis. For cordycepin, the key factors influencing encapsulation efficiency were ranked as follows: PVA concentration in the aqueous phase (B₁, 0.5%) > ultrasonic power (A₁, 100 W) > stirring speed (C₃, 800 rpm) > organic/aqueous phase ratio (D₁, 1:4). Similarly, the factors influencing PBA encapsulation were ranked, with a slight variation: ultrasonic power (A₁, 100 W) > PVA concentration (B₁, 0.5%) > organic/aqueous phase ratio (D₁, 1:4) > stirring speed (C₃, 800 rpm).

The observed differences in factor rankings for the two compounds required a strategic compromise to optimize both encapsulation efficiencies simultaneously. Consequently, we prioritized maximizing cordycepin encapsulation while ensuring high PBA encapsulation. Based on the Taguchi experimental design, the optimal formulation conditions were determined to be a PVA concentration in aqueous phase of 0.5%, ultrasonic power of 100 W, stirring speed of 800 rpm, and an organic/aqueous phase ratio of 1:4.

The confirmation experiments validated the optimal formulation conditions. As shown in Table 3, the experiment achieved encapsulation efficiencies of 97.61% ± 0.40% for cordycepin and 86.12% ± 0.95% for PBA. In terms of drug loading efficiency, the optimized formulation showed a DL% of 4.47 ± 0.02 for cordycepin and 3.95 ± 0.04 for PBA. The result demonstrated the reproducibility of the optimized formulation, further confirming its suitability for co-loaded application.

As shown in Figure 2A, DLS measurements revealed that the optimal co-loaded polymersomes had an average size of 101.03 ± 0.42 nm with a polydispersity index (PDI) of 0.248 ± 0.014, indicating a relatively narrow size distribution. TEM images further confirmed that co-loaded polymersomes were relatively spherical in shape, with sizes consistent with the DLS data (Figure 2B).

Additionally, the co-loaded polymersomes exhibited a relatively high colloidal stability and could maintain their size and size distribution in PBS at 4°C for over 1 month, as shown in Figure 2C. At the end of the 30-day stability study, the encapsulation efficiencies decreased to 80.53% ± 3.63% for cordycepin and 65.43% ± 8.72% for PBA, indicating moderate drug retention under 4°C.

The crystalline form of the encapsulated compounds could affect their therapeutic efficacy. XRD analysis was performed to evaluate the polycrystalline structure of the co-loaded polymersomes. As shown in Figure 2D, the combination of free cordycepin and PBA exhibited multiple crystal diffraction peaks, indicating their crystalline states.

In contrast, blank polymersomes had no obvious characteristic peaks, consistent with an amorphous structure. Notably, the XRD pattern of the co-loaded polymersomes demonstrated that the characteristic diffraction peaks of cordycepin and PBA were nearly eliminated, suggesting that both compounds existed in an amorphous form within the polymersomes (Thakral et al., 2016). This transition to an amorphous state likely enhance their encapsulation and stability, demonstrating the effectiveness of the polymersomes in preventing crystallization.

The in vitro release profiles of cordycepin and PBA from the polymersomes were investigated under varying H₂O₂ concentrations. The H₂O₂ concentrations selected for the release study (0, 50, and 100 µM) were used to simulate both physiological (baseline) and pathological (inflamed) conditions of the OA microenvironment, indicating that synovial fluid in osteoarthritic joints exhibits elevated oxidative stress (Zhang et al., 2021; Jiang et al., 2022; Wang et al., 2023). As shown in Figure 2E, approximately 35% of cordycepin was released after 72 h under 0 μM or 50 μM H₂O₂. In contrast, at 100 μM H₂O₂, the cumulative release rose significantly to 59%. Similarly, PBA exhibited a slower release profile, with less than 20% released after 72 h in PBS alone (Figure 2F). Under oxidative conditions of 50 μM and 100 μM H₂O₂ for 72 h, the cumulative release of PBA increased to 48% and 55%, respectively. These results indicated that the presence of H₂O₂ could significantly accelerate the release of both compounds from the polymersomes in a concentration-dependent manner.

The ROS scavenging ability of the co-loaded polymersomes was evaluated using a chemiluminescence assay in a high ROS environment. In this study, H₂O₂ was used to mimic a high-ROS environment, with the H₂O₂-Cu²⁺ reaction generating hydroxyl radicals (·OH). Subsequently, luminol-based chemiluminescence was employed to measure the scavenging effect of the compounds on hydrogen peroxide.

As shown in Figure 2G, the in vitro ROS scavenging capability increased with compound concentration. Free PBA exhibited greater ROS scavenging activity compared to free cordycepin, and their combination demonstrated enhanced activity. Notably, the co-loaded polymersomes exhibited the highest scavenging efficacy, reaching up to 97% at a concentration of 750 μM, indicating a strong dose-dependent response. In contrast, blank polymersomes showed minimal ROS scavenging ability across the tested concentration range. These results confirmed that the co-loaded polymersomes exhibited ROS scavenging efficacy comparable to that of the free compound combination, effectively reducing ROS levels at the site of OA.

The cytotoxicity of different groups on RAW264.7 cells were evaluated using the CCK-8 assay. As shown in Figure 3A, blank polymersomes had negligible effects on cell viability, confirming their biocompatibility. Among the treatments, free cordycepin exhibited the highest cytotoxicity, likely due to its ability to suppress NF-κB signaling pathway (Huang et al., 2021), with an IC₅₀ value of 101.4 ± 18.96 μM. In contrast, free PBA exhibited relatively low cytotoxicity, indicating a favorable safety profile. The free compound combination revealed intermediate cytotoxicity between free cordycepin and free PBA. Co-loaded polymersomes exhibited similar cytotoxicity to the free compound combination in RAW 264.7 cells across all concentrations tested (Figure 3B). These results suggest that the encapsulated formulation maintains the biocompatibility profile of the free combination and may reduce the direct exposure of cells to free cordycepin at high local concentrations.

To investigate the anti-inflammatory effects of various samples, RAW264.7 cells were pre-treated with LPS to simulate an inflammatory environment of OA. As shown in Figures 3C,E, the exposure of LPS for 24 h resulted in a significant increase in the secretion of pro-inflammatory cytokines, including IL-1β and TNF-α, from RAW264.7 cells. Treatment with various groups prior to LPS activation could attenuate the release of these pro-inflammatory cytokines in a dose-dependent manner (Figures 3D,F). The free compound combination at 50 μM exhibited greater inhibition of cytokines production compared to either compound alone. Furthermore, at the same concentration (50 μM), the co-loaded polymersomes exhibited inhibitory effects on cytokine secretion that were statistically comparable to those of the free compound combination. This result may be attributed to the co-delivery and protective properties of the polymersomes, which help maintain the stability and localized availability of both compounds.

Unlike the luminol assay, the Amplex Red method cannot penetrate the cell membrane. Therefore, this method was used to determine the concentration of H₂O₂ in the cell supernatant following various treatments to assess the antioxidant effects of these treatments.

As shown in Figure 3G, blank polymersomes exhibited negligible antioxidant activity at the test concentrations. In contrast, free compounds and co-loaded polymersomes demonstrated dose-dependent antioxidant activity. Free PBA showed superior efficacy in reducing H₂O₂ production in RAW264.7 cells compared to free cordycepin. The combination of both compounds exhibited significantly higher antioxidant activity than either free compound alone. Notably, co-loaded polymersomes demonstrated the greatest inhibition of H₂O₂ production at higher concentration compared to the free compound combination. The improved inhibitory effect of co-loaded polymersomes is likely due to their sustained release properties, where provided prolonged antioxidant effects.

The in vitro activity of the co-loaded polymersomes allowed us to investigate their therapeutic effects in an OA rat model. We first assessed the targeting potential of the co-loaded polymersomes using Cy7-labeled polymersomes. These polymersomes were administered via intraperitoneal (IP) injected in OA rats, with induced OA in the left knee joint and an unaffected right knee joint serving as the healthy control. As shown in Figure 4A, Cy7 fluorescence signals were detected in the inflamed left knee joint at 12 h post-injection, with signals persisting for up to 24 h. In contrast, minimal fluorescence was observed in the right, healthy knee joint at both time points. This result clearly demonstrated the ability of the co-loaded polymersomes to target the OA-affected site. The rats were subsequently sacrificed for ex vivo imaging of the knee joint and major organs. As shown in Figure 4B, the left knee joint exhibited strong Cy7 fluorescence signals at 12 h post-injection, which weakened at 24 h. To further investigate the distribution of co-loaded polymersomes in major organs, these tissues were harvested at 12 h and 24 h post-injections (Figure 4C). The Cy7 signals were detected in the liver and kidney at 12 h post-injection, but the signal intensity diminished at 24 h post-injection. These results indicated the co-loaded polymersomes efficiently targeted the OA site and accumulated in the inflamed joint.

To evaluate the therapeutic effects of the treatment on joint swelling (edema), changes in joint width and body weight were measured.

As shown in Figure 5A, OA was induced via intra-articular injection of papain and L-cysteine on days −14 and −10, followed by intraperitoneal administration of different treatment formulations from day 0 to day 28. Key experimental time points for measurements are highlighted, including joint swelling and body weight assessments at days 14 and 28.

As a result, Figure 5B indicated that a significant increase in joint width was observed in OA rats compared to normal rats prior to treatment, indicating that the injection of papain and L-cysteine successfully induced joint inflammation in the rats. Compared with untreated OA rats, various treatments reduced the joint width in OA rats over the 4-week study period. At 2 weeks post-administration, rats treated with co-loaded polymersomes showed an 11.45% ± 4.31% reduction in joint width, compared to a 7.76% ± 2.59% reduction in rats treated with the free compound combination. At 4 weeks post-administration, the reduction in joint width was 19.17% ± 3.58% in the co-loaded polymersome group, compared to 9.31% ± 2.00% for the free compound combination group. These results indicated that joint width for co-loaded polymersomes treated rats were significantly lower than those of other groups, showing the efficacy of co-loaded polymersomes in reducing OA rats joint swelling.

The body weights of the rats were also monitored alongside joint width to evaluate the systemic safety and therapeutic effects of co-loaded polymersomes. As shown in Figure 5C, the average body weight of normal rats increased by 1.46-fold over 4 weeks, reaching 403 ± 5.80 g. In contrast, untreated OA rats exhibited a 1.48-fold increase in body weight, from 263.6 ± 9.80 g to 390.4 ± 10.20 g, although their final body weight remained significantly lower than that of normal rats.

Among the treated groups, OA rats treated with co-loaded polymersomes demonstrated the most significant recovery in body weight, which approached that of normal rats by the end of treatment. However, rats treated with free cordycepin or free PBA showed relatively lower body weights compared to untreated OA rats. The superior therapeutic effect of co-loaded polymersomes could be attributed to the enhanced effects of the combined compounds and the sustained release profile of the polymersomes.

To evaluate the anti-inflammatory effects of co-loaded polymersomes, serum levels of pro-inflammatory cytokines, including IL-1β and TNF-α, were detected using ELISA. As shown in Figures 6A,B, untreated OA rats had significantly elevated levels of IL-1β and TNF-α, contributing to abnormal cartilage matrix metabolism and increased cell apoptosis.

After 4 weeks of treatment, the levels of IL-1β and TNF-α were significantly reduced in all treated groups compared to untreated OA rats. Among them, the administration of co-loaded polymersomes demonstrated the lowest levels of both cytokines, which were similar to those in the normal rat group. It was noteworthy that the serum level of TNF-α in co-loaded polymersomes group was even lower than those in normal rat group, although no statistically significant difference was found between these two groups. These results indicated that co-loaded polymersomes effectively reduced systemic inflammation in OA rats.

The cartilage morphology of knee joints in OA rats was observed using hematoxylin and eosin (HE) staining to further evaluate the therapeutic effects of various treatments. As shown in Supplementary Figure S5, the surface of articular cartilage in the normal rat group was flat and smooth, with neatly arranged chondrocytes. In contrast, untreated OA rats exhibited severe cartilage damage, characterized by numerous defects, disorganized chondrocyte layers, and reduced cell density. The structural abnormalities confirmed the successful establishment of the OA model.

After 4 weeks of treatment, OA rats treated with co-loaded polymersome showed significant improvement in cartilage morphology. The cartilage surface was smooth, and the chondrocytes were arranged in an orderly manner, resembling the structure observed in normal rats (Supplementary Figure S5). In contrast, OA rats treated with blank polymersomes showed negligible effects on cartilage repair, with a large number of defects on the cartilage surface and disorganized chondrocyte layers. Uneven cell distribution and hemorrhage were also detected, similar to that in untreated OA rat model. The groups treated with free cordycepin or free PBA showed moderate improvements in cartilage surface smoothness and chondrocyte organization. However, irregular cell distribution and macrophage infiltration were still observed. Compared with either free compound alone, the free compound combination group demonstrated a smoother cartilage surface and more regular chondrocyte arrangement, but small focal hemorrhages were still present. Taken together, co-loaded polymersomes were the most effective treatment in reducing pathological cartilage damage and promoting cartilage regeneration in OA rats. The combination of cordycepin and PBA within the polymersome delivery system contributed to this superior therapeutic outcome, likely due to their enhanced effect and sustained release properties.