Engineered MSC‐sEVs as a Versatile Nanoplatform for Enhanced Osteoarthritis Treatment via Targeted Elimination of Senescent Chondrocytes and Maintenance of Cartilage Matrix Metabolic Homeostasis

First, MSCs were successfully derived from iPSCs using our induction protocol and flow cytometry analysis results showed that the MSCs highly expressed CD29, CD44, CD73, CD90, and CD105, and were negative for CD45, HLA‐DR, and CD34 (Figure S1, Supporting Information). To prepare WPD‐sEVs^siMDM2, MSC‐sEVs were isolated from serum‐free culture medium of MSCs by the ultra‐centrifugation method, and siMDM2 was loaded into sEVs to construct sEVs^siMDM2 by using the electroporation technology. Nanoflow cytometry (NFC) analysis demonstrated efficient siMDM2 loading (25.80% ± 1.16%) into sEVs (Figure S2, Supporting Information). Next, WYRGRL‐PEG_2K‐DSPE (WPD) was synthesized by the amide ligation between WYRGRL and DSPE‐PEG_2K‐NHS, and then incubated with sEVs^siMDM2 to obtain multifunctionally engineered WPD‐sEVs^siMDM2 (Figure2A). MOLDI‐TOF analysis showed the molecular weight of WPD was about 3.5 KDa (Figure S3A, Supporting Information). WPD was cytocompatible as there was no difference between the proliferation of human chondrocytes cultured with medium containing WPD at the concentration range of 0–100 µg ml⁻¹ (Figure S3B,C, Supporting Information), thus, 100 µg ml⁻¹ WPD was used to incubate with sEVs. To verify the successful modification of sEVs, the WPD peptide was labeled with rhodamine B. NFC analysis showed that 100 µg ml⁻¹ cartilage‐targeting peptide WPD was successfully anchored onto the surface of sEVs (Figure 2B). Furthermore, to confirm the possible contaminant of soluble WPD or aggregating WPD, 100 µg ml⁻¹ rhodamine B labeled WPD was treated as the procedure for WPD‐sEVs^siMDM2 isolation. After ultracentrifugation, no fluorescent signal was detected by NFC analysis in the resuspension solution that rinsed the tube bottom, further excluding the possibility of soluble or aggregating WPD contaminant during WPD‐sEVs^siMDM2 isolation. After the successful modification to sEVs, the zeta potential of WPD‐sEVs^siMDM2 was analyzed. The average zeta potential of the unmodified sEVs was ‐17.23 ± 2.42 mV, while WPD‐sEVs^siMDM2 presented an average zeta potential of 9.13 ± 1.31 mV, demonstrating the successful surface charge reverse of sEVs (Figure 2C). Transmission electron microscope (TEM) observation revealed that both sEVs and WPD‐sEVs^siMDM2 were double‐layer vesicles with typical cup shape and intact membrane structure (Figure 2D). Additionally, no signs of sEVs aggregation or WPD aggregating contaminant were observed in the TEM image of WPD‐sEVs^siMDM2. The distribution of particle size was detected by NFC analysis and the average diameter of unmodified sEVs and WPD‐sEVs^siMDM2 were 77.17 ± 4.24 nm and 78.47 ± 3.76 nm, respectively (Figure 2E). Western blot analysis showed that both sEVs and WPD‐sEVs^siMDM2 expressed the specific surface protein makers CD9, CD63, and TSG101, but did not express GM130 (Figure 2F), which well met the identification criteria for sEVs.

Physiological stability of WPD modification is of significant importance in the application of siMDM2 delivery to senescent chondrocytes in joint cavity, thus we evaluated the WPD modification rate, zeta potential, and average size of WPD‐sEVs^siMDM2 in different solutions for a 7‐day period. First, the modification rate, zeta potential, and average diameter of WPD‐sEVs^siMDM2 presented no notable difference after 7 days incubation in phosphate‐buffered saline (PBS) solution (Figure S4, Supporting Information), which indicates WPD‐sEVs^siMDM2 presented an excellent stability in PBS solution. As anionic biomacromolecules hyaluronic acid (HA) and chondroitin sulfate (CS) distribute widely in the joint cavity, and anionic HA or CS can compete with the sEVs to bind WPD by electrostatic interaction, causing the detachment of WPD from sEVs surface and final modification failure. Therefore, the WPD modification stability under the presence of HA or CS was further evaluated by detecting the WPD modification rate. We found only a slightly decrease in modification rate was detected in WPD‐sEVs^siMDM2 incubated with HA and CS solution for 7 days, as about 65% WPD modification rate was kept for WPD‐sEVs^siMDM2 after incubation with HA (65.13% ± 0.57%) or CS (63.83% ± 1.13%) for 7 days (Figure S5, Table S1, Supporting Information). Altogether, these results verified a successful self‐assembly of WPD on sEVs lipid membrane and a relatively stable WPD modification to the lipid membrane of sEVs under mimic physiological environment of the joint.

As senescent chondrocytes accumulation plays essential role in the progression of OA, we sought to determine whether versatile modification could enhance the antiaging effects of sEVs in vitro, and thus a DNA‐damaging chemical agent doxorubicin was applied in human chondrocytes to induce senescent phenotype. First, we determined the optimal concentration of doxorubicin in inducing chondrocyte senescence and we found doxorubicin presented no obvious cytotoxic effects on chondrocytes at the concentration of 0–1 µM (Figure S6, Supporting Information). Thus, doxorubicin at the concentration of 1 µM was used in our following experiments. To evaluate the antiaging effects of WPD‐sEVs^siMDM2 in vitro, doxorubicin‐induced chondrocytes were co‐treated with sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) for 7 days (Figure3A). Then we assessed the cellular uptake of WPD‐sEVs^siMDM2 in human chondrocytes. Our results showed that WPD‐sEVs^siMDM2 displayed a better cellular uptake capacity than unmodified sEVs (Figure S7A, Supporting Information). Flow cytometry analysis further demonstrated a higher mean fluorescence intensity (MFI) of WPD‐sEVs^siMDM2‐treated chondrocytes in comparison with that of unmodified sEVs‐treated chondrocytes (Figure S7B, Supporting Information), suggesting an enhanced chondrocyte uptake ability of WPD‐sEVs^siMDM2. As the improved cellular uptake efficiency might promote the protective effects, then we further evaluated the senescent phenotype and cartilage matrix metabolic homeostasis in doxorubicin‐induced chondrocytes after WPD‐sEVs^siMDM2 treatment. Our results showed the absorbance in WPD‐sEVs^siMDM2‐treated senescent chondrocytes was significantly higher than that of sEVs or sEVs^siMDM2 treatment group (Figure 3B), indicating an enhanced effect of multifunctionally engineered sEVs on chondrocyte proliferation. We next investigated the effect of WPD‐sEVs^siMDM2 treatment on matrix anabolism in senescent chondrocytes, WPD‐sEVs^siMDM2 treatment more effectively promoted the mRNA expressions of Col2a1 (Figure 3C) and Acan (Figure 3D), as compared to the other treatment groups, further validating the enhanced antiaging effects of WPD‐sEVs^siMDM2. Then, the percentage of doxorubicin‐induced senescent chondrocytes after different treatments was detected by senescence‐associated β‐galactosidase (SA‐β‐Gal) staining. Our results showed doxorubicin‐induced senescent chondrocytes (73.78% ± 5.95% senescent cells) were decreased after the treatment of sEVs (63.24% ± 2.99% senescent cells), sEVs^siMDM2 (46.86% ± 2.39% senescent cells), and WPD‐sEVs^siMDM2 (31.13% ± 3.14% senescent cells) (Figure 3E,F). Importantly, WPD‐sEVs^siMDM2 treatment displayed greater scavenging action on senescent cells in contrast to sEVs^siMDM2 treatment group (P = 0.0445). Similarly, immunofluorescence for senescence biomarker P16^INK4a demonstrated WPD‐sEVs^siMDM2 treatment (28.99% ± 4.15%) more effectively reduced the percentage of P16^INK4a‐positive senescent chondrocytes (69.29% ± 3.72%), compared to the sEVs^siMDM2 treatment group (47.55% ± 3.69%, P = 0.0022) (Figure 3G,H). Cdkn1a and Cdkn2a, which are the core senescence‐associated genes were also up‐regulated in doxorubicin‐induced chondrocytes and effectively inhibited after treatment of sEVs, sEVs^siMDM2, or WPD‐sEVs^siMDM2, while WPD‐sEVs^siMDM2 showed greater inhibitory effect compared to the sEVs^siMDM2 treatment group (Figure S8A,B, Supporting Information). Meanwhile, up‐regulation of catabolism markers (Mmp13 and Adamts5) (Figure S8C,D, Supporting Information) and SASP factors (Il6 and Tnfα) (Figure S8E,F, Supporting Information) in doxorubicin‐induced chondrocytes were effectively abolished after the treatment of sEVs, sEVs^siMDM2, or WPD‐sEVs^siMDM2, and WPD‐sEVs^siMDM2 more effectively inhibited the production of SASP factors than sEVs^siMDM2. Enhanced clearance effect of senescent chondrocytes was further confirmed by significantly reduced immunostaining for P21 (Figure S9, Supporting Information) and nuclear γH2AX (Figure 3G) in chondrocytes after WPD‐sEVs^siMDM2 treatment.

To investigate the molecular mechanism of multifunctionally engineered sEVs on attenuating chondrocyte senescence, we further detected MDM2 expression in doxorubicin‐induced chondrocytes after different treatments. Exposure to doxorubicin markedly promoted the expression of MDM2 in chondrocytes, and both sEVs^siMDM2 and WPD‐sEVs^siMDM2 treatment inhibited this effect (Figure S9C,D, Supporting Information). Further, we found WPD‐sEVs^siMDM2 showed a stronger inhibitory effect on MDM2 expression (Figure S9C,D, Supporting Information). Consistently, the results in PCR analysis demonstrated that both sEVs^siMDM2 and WPD‐sEVs^siMDM2 treatment down‐regulated the mRNA expression of Mdm2 in doxorubicin‐induced chondrocytes, while WPD‐sEVs^siMDM2 presented an enhanced inhibitory effect than sEVs^siMDM2 (Figure S9E, Supporting Information). Western blot analysis results also showed WPD‐sEVs^siMDM2 treatment effectively inhibited the protein expressions of P16^INK4a and P21 (Figure 3I). Additionally, we found WPD‐sEVs^siMDM2 treatment promoted MDM2 protein expression and induced P53 phosphorylation in senescent chondrocytes (Figure 3J). To further validate the essential role of MDM2‐P53 pathway in the antiaging function of WPD‐sEVs^siMDM2 on senescent chondrocytes, P53 phosphorylation inhibitor was applied on chondrocytes. We found the inhibitor treatment abrogated the antiaging function of WPD‐sEVs^siMDM2 on chondrocytes as P16^INK4a and P21 expressions were increased (Figure 3K). As expected, P53 phosphorylation was suppressed obviously after inhibitor treatment (Figure 3L). More importantly, we found WPD‐sEVs^siMDM2 promoted the apoptosis of senescent chondrocytes as P16^INK4a+ and TUNEL⁺ double positive cells increased from 2.30% ± 2.37% to 31.03% ± 2.87% after WPD‐sEVs^siMDM2 treatment (Figure S10, Supporting Information). Altogether, these results demonstrated that multifunctional modification significantly improved the antiaging effects of sEVs in vitro.

Given the abundant negatively charged biomacromolecules in cartilage matrix, human articular cartilage explants (Φ 5 × 1 mm) were used to examine whether WPD modification could enhance the penetration ability of sEVs. Cartilage explants were stuck in a customized chamber using the one‐way transport penetration mould, and 200 µl DiO labeled unmodified sEVs (1 × 10¹⁰ particles ml⁻¹) or WPD‐sEVs^siMDM2 suspensions (1 × 10¹⁰ particles ml⁻¹) were added to the left side of penetration mould while 200 µl PBS were added to the right side. After penetration ex vivo for different times including 1, 3, and 7 days, the cartilage explants were cryo‐sectioned and observed by laser confocal microscopy. As shown in Figure4A, unmodified sEVs still concentrated in the superficial zone of the cartilage explants after penetration for 3 days, indicating a very limited penetration ability. In contrast, WPD‐sEVs^siMDM2 had diffused into the deep zone of cartilage explants after 1‐day penetration (Figure 4A). Significantly, WPD‐sEVs^siMDM2 had distributed throughout the cartilage explants after penetration for 3 days (Figure 4A). For the uptake evaluation in cartilage explants, DiO labeled unmodified sEVs (1 × 10¹⁰ particles ml⁻¹) or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) were incubated with these cartilage explants (Φ 5 × 1 mm), and the cartilage uptake efficiency was analyzed as the percentage of DiO‐labeled sEVs or WPD‐sEVs^siMDM2 from 200 µl total solution to the cartilage explants. The results in Figure 4B and Table S2 (Supporting Information) demonstrated that WPD‐sEVs^siMDM2 exhibited significantly enhanced cartilage uptake ratio than unmodified sEVs at all time points, which was in consistent with the one‐way penetration assay. In particular, after incubation for 24 h, 47.33% ± 4.50% WPD‐sEVs^siMDM2 were taken up in cartilage explants while it was only 12.33% ± 2.05% unmodified sEVs were taken up. After 7 days penetration, 64.00% ± 2.45% WPD‐sEVs^siMDM2 were taken up and 20.33% ± 1.25% sEVs were taken up in cartilage explants. These results convincingly verified that WPD‐sEVs^siMDM2 are capable of penetrating into the deep zone of the articular cartilage and exhibit improved cartilage penetration ability.

Next, we directly injected unmodified sEVs (1 × 10¹⁰ particles ml⁻¹) or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) labeled with a near‐infrared fluorescence dye DiR into the knee joint to assess the retention capacity. In vivo imaging analysis was used to further evaluate the joint retention time of WPD‐sEVs^siMDM2. For a 35‐day period, DiR fluorescence signal in the mouse knee joint was serially detected and analyzed. Immediately after intra‐articular injection, we detected bright fluorescence signals in both sEVs group and WPD‐sEVs^siMDM2 group (Figure 4C). However, the DiR fluorescence signal was dropped dramatically 1 day after injection in sEVs group, while WPD‐sEVs^siMDM2 group showed significant slower reduction in fluorescence intensity during the same period, indicating that most of the unmodified sEVs were quickly cleared from the joint by the fast clearance system. Notably, the fluorescence signal was still detectable in WPD‐sEVs^siMDM2 group at day 28 after intra‐articular injection, whereas the fluorescence signal was barely detectable in unmodified sEVs group at day14 after injection, suggesting that the retention time of WPD‐sEVs^siMDM2 in knee joint is much longer than that of unmodified sEVs. Applying the fluorescence signal data to a single‐phase exponential decay function, we found that the unmodified sEVs had a joint half‐life of 1.031 days, whereas WPD‐sEVs^siMDM2 significantly prolonged the half‐life to 6.774 days (Figure 4D). Overall, these results demonstrated that multifunctional modification significantly extended the joint retention ability and enhanced the cartilage penetration capacity of sEVs.

We next explored the efficacy of WPD‐sEVs^siMDM2 for treating anterior cruciate ligament transection (ACLT)‐induced PTOA mice. In view of the significantly extended half‐life of WPD‐sEVs^siMDM2 (from 1.031 to 6.774 days), intra‐articular injection of different treatments was performed once a month over two months starting on day 28 after ACLT surgery (Figure5A). Safranin O staining was carried out to evaluate the cartilage destruction. As shown in Figure 5B, obvious cartilage abrasion and loss was presented in PBS‐treated PTOA mice, and these features were also found in sEVs‐treated (1 × 10¹⁰ particles ml⁻¹) and sEVs^siMDM2‐treated (1 × 10¹⁰ particles ml⁻¹) PTOA mice. In contrast, WPD‐sEVs^siMDM2 treatment (1 × 10¹⁰ particles ml⁻¹) effectively slowed the cartilage degeneration and maintained cartilage matrix metabolic homeostasis in PTOA mice. The Osteoarthritis Research Society International (OARSI) scoring system was also applied to quantify the OA severity. Our results showed there was no significant difference between PBS‐treated group (4.00 ± 1.10), sEVs‐treated group (2.80 ± 0.75, P = 0.18), and sEVs^siMDM2‐treated group (3.00 ± 0.63, P = 0.17) (Figure 5C). However, the OARSI score of WPD‐sEVs^siMDM2‐treated group (1.60 ± 0.80) was significantly lower than that of PBS‐treated group (P = 0.02) (Figure 5C), validating the cartilage protective effects of WPD‐sEVs^siMDM2 in PTOA mice. Subsequently, immunohistochemical staining for collagen II, MMP13, and SOX9 was performed to evaluate the regulative effect of WPD‐sEVs^siMDM2 on cartilage anabolism and catabolism. We found the expressions of collagen II and SOX9 was increased and MMP13 expression was decreased in ACLT‐induced PTOA mice after the treatment of WPD‐sEVs^siMDM2, while sEVs and sEVs^siMDM2 treatment showed no significant difference to PBS‐treated group (Figure 5D‐F).

Moreover, the antiaging effects of WPD‐sEVs^siMDM2 in articular cartilage of ACLT‐induced PTOA mice were evaluated. Consistent with our in vitro results, immunohistochemical staining for P16^INK4a showed that the percentage of senescent cells was increased after ACLT surgery (from 6.62% ± 2.53% to 61.76% ± 5.81%), while only WPD‐sEVs^siMDM2 treatment effectively decreased the number of senescent cells (31.12% ± 7.92%) (Figure 5G,H). Next, enhanced clearance function of senescent cells was further confirmed by the reduced expressions of P21 (Figure 5G,I), SASP factors (IL6 and TNFα) (Figure 5G,J, Figure S11A,B, Supporting Information), and HMGB1 (Figure S11A,C, Supporting Information) in WPD‐sEVs^siMDM2‐treated PTOA mice, compared to the sEVs^siMDM2 treatment group. To confirm the role of MDM2‐P53 pathway in the antiaging function of WPD‐sEVs^siMDM2 in OA cartilage in vivo. The expression of MDM2 in articular cartilage was further evaluated in PTOA mice after different treatments. immunohistochemical staining results showed MDM2‐positive chondrocytes was increased in PBS‐treated PTOA mice (from 2.55% ± 1.60% to 57.56% ± 8.09%), while only WPD‐sEVs^siMDM2 treatment effectively down‐regulated the percentage of MDM2‐positive chondrocytes (19.64% ± 3.18%) (Figure S11D,E, Supporting Information). Additionally, we found only WPD‐sEVs^siMDM2 treatment effectively inhibited the expression of P53 and promoted the expression of p‐P53 in mice cartilage (Figure S11F,G, Supporting Information).

Interestingly, our results also demonstrated WPD‐sEVs^siMDM2 treatment markedly decreased the number of inflammatory cells in the synovium (Figure S12A, Supporting Information) and synovial inflammation score (from 1.80 ± 0.40 to 0.60 ± 0.49, P = 0.03) (Figure S12B Supporting Information) of ACLT‐induced PTOA mice, suggesting therapeutic effects on articular cartilage mediated by WPD‐sEVs^siMDM2 could ameliorate synovitis in the joint. In addition, our results showed WPD‐sEVs^siMDM2 treatment decreased subchondral bone plate score (P = 0.0476) and bone volume score (P = 0.0397) in PTOA mice knee joint (Figure S12C‐E, Supporting Information). Together, these in vivo results suggested that WPD‐sEVs^siMDM2 treatment can alleviate senescent phenotype and maintain matrix metabolic homeostasis in ACLT‐induced PTOA mice.

Aging in mice causes the development of spontaneous OA, similar to what occurs in human patients. We therefore investigated the antiaging effects of WPD‐sEVs^siMDM2 treatment in naturally aged mice. Starting at the age of 12‐month‐old, WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) was intra‐articularly injected into the joint cavity of mice once a month for 8 months (Figure6A). Safranin O staining results showed obvious cartilage destruction was observed in PBS‐treated 20‐month‐old naturally aged mice, and only WPD‐sEVs^siMDM2 treatment significantly reduced cartilage loss and maintained the intact structure of articular cartilage in aged mice (Figure 6B). In aged mice, WPD‐sEVs^siMDM2 treatment obviously reduced OARSI score (from 3.00 ± 0.63 to 1.80 ± 0.75), but it showed no significant difference (P = 0.1032) (Figure 6C). Additionally, the expression of anabolism‐related marker collagen II and SOX9 was significantly increased and catabolism‐related marker MMP13 was reduced in aged mice after WPD‐sEVs^siMDM2 treatment (Figure 6D‐F), suggesting a certain degree of enhanced cartilage protective effects of multifunctionally engineered sEVs.

Similar to the antiaging effects of WPD‐sEVs^siMDM2 in ACLT‐induced PTOA mice, only WPD‐sEVs^siMDM2 treatment effectively eliminated P16^INK4a‐positive senescent cells (from 55.04% ± 4.31% to 27.28% ± 5.30%) in articular cartilage of naturally aged mice (Figure 6G,H). Furthermore, the expressions of P21 (Figure 6G,I), SASP factors (IL6 and TNFα) (Figure 6G,J, Figure S13A,B, Supporting Information), and HMGB1 (Figure S13A,C, Supporting Information) were notably increased in PBS‐treated aged mice, while only WPD‐sEVs^siMDM2 treatment significantly reduced the expressions of P21, HMGB1, and SASP factors. To further verify the role of MDM2‐P53 pathway in naturally aged mice, the expression of MDM2 in articular cartilage of aged mice was analyzed by immunohistochemical staining after different treatments. In particular, the percentage of MDM2‐positive chondrocytes in articular cartilage was significantly up‐regulated in PBS‐treated aged mice (65.71% ± 5.61%), compared to the young mice (4.84% ± 1.36%). Consistent with our results in ACLT‐induced PTOA mice, only WPD‐sEVs^siMDM2 treatment effectively decreased the percentage of MDM2‐positive chondrocytes in aged mice (from 65.71% ± 5.61% to 33.52% ± 5.45%) (Figure S13D,E, Supporting Information). Additionally, our results showed WPD‐sEVs^siMDM2 treatment effectively inhibited the expression of P53 and up‐regulated the expression of p‐P53 in mice cartilage (Figure S13F,G, Supporting Information).

In consistent with the results in ACLT‐induced PTOA mice, Hematoxylin & Eosin (H&E) staining of synovium results demonstrated only WPD‐sEVs^siMDM2 treatment decreased the synovial inflammation score of aged mice (from 1.60 ± 0.49 to 0.60 ± 0.49, P = 0.0794) (Figure S14A,B, Supporting Information). Moreover, we found WPD‐sEVs^siMDM2 treatment decreased subchondral bone plate score and bone volume score in naturally aged mice, but it also presented no obvious difference compared with the PBS‐treated mice (Figure S14C–E, Supporting Information). Altogether, our data indicated that WPD‐sEVs^siMDM2 treatment effectively alleviated senescent phenotype and inhibited cartilage matrix catabolism in naturally aged mice.

Last, to validate the feasibility of the versatile engineered sEVs strategy in clinical OA patients, we evaluated the effects of WPD‐sEVs^siMDM2 in a three‐dimensional environment by using ex vivo cultured OA articular cartilage explants from patients undergoing total knee arthroplasty, and OA‐affected cartilage explants were divided into four groups that received PBS, sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), and WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) treatments, respectively. Non‐OA affected normal cartilage was used as the control group. As shown in Figure7A, significant proteoglycan loss and cartilage degradation was observed in OA‐affected cartilage in contrast to the non‐OA affected cartilage, while multifunctionally modification of sEVs effectively promoted the production of cartilage matrix and ameliorated cartilage destruction (Figure 7A), further confirming the enhanced protective effects of WPD‐sEVs^siMDM2 treatment. Consistently, the mRNA expressions of Acan and Col2a1 were markedly down‐regulated in OA‐affected cartilage, while only WPD‐sEVs^siMDM2 treatment effectively increased the mRNA expressions of Col2a1 (Figure 7B) and Acan (Figure 7C). Consistent with our in vivo and in vitro results, immunohistochemical staining for P16^INK4a and IL6 showed WPD‐sEVs^siMDM2 treatment effectively decreased the expressions of P16^INK4a and IL6 in OA cartilage (Figure 7D). Enzyme‐Linked Immunosorbent Assay (ELISA) results further confirmed that the protein expressions of SASP factors (IL6 and TNFα) were significantly down‐regulated after the treatment of WPD‐sEVs^siMDM2 (Figure 7E,F). Next, senescent phenotype was also confirmed in OA‐affected cartilage by the enhanced gene expressions of catabolism‐related genes (Adamts5 and Mmp13) (Figure 7G,H), senescence marker (Cdkn1a and Cdkn2a) (Figure 7I,J), and pro‐inflammatory factor Il6 (Figure 7K). In contrast to sEVs‐treated group and sEVs^siMDM2‐treated group, WPD‐sEVs^siMDM2 treatment more effectively suppressed the mRNA expressions of these senescence‐related genes in OA‐affected cartilage, corroborating our results from cultured human chondrocytes, PTOA mice, and naturally aged mice. Together, these data provide ex vivo evidence that WPD‐sEVs^siMDM2 treatment has significant antiaging effects and protective action on senescent chondrocytes and OA cartilage.

WPD and rhodamine B labeled WPD were purchased from ChinaPeptides (Suzhou, China). siMDM2 was designed and synthesized by Hanbio (Shanghai, China). Doxorubicin, bovine serum albumin (BSA), and dimethyl sulfoxide (DMSO) were obtained from Sigma‐Aldrich (MO, USA). The serum‐free ncMission hMSC Medium was obtained from Shownin Biotechnologies Co., Ltd. (RP02010, China). Dulbecco's modified Eagle's medium (DMEM)‐F12 and PBS were obtained from HyClone (UT, USA). Fetal bovine serum (FBS), collagenase II, and Penicillin–Streptomycin were purchased from Gibco (CA, USA). Lipid membrane fluorescent dyes DiO and DiR were purchased from Thermo Fisher Scientific (MA, USA). 4′,6‐Diamidino‐2‐phenylindole (DAPI) was purchased from Beyotime Biotechnology (Jiangsu, China). P53 phosphorylation inhibitor 17‐DMAG was obtained from MedChemEexpress (HY‐10389, China). The information of primary and secondary antibodies for Western blot analysis, immunofluorescent staining, and immunohistochemical staining was listed in Table(Supporting Information). S3

sEVs derived from induced pluripotent stem cells derived MSCs were used in this study. The iPSC cell line (iPS‐S‐01) was provided by the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences. Briefly, the medium was replaced with a serum‐free MSC culture medium containing basal medium (Nuwacel Nova Missoin Basal Medium, Nuwacell Biotechnology, RP02010‐01) and supplement (Nuwacell Nova Missoin Supplement, Nuwacell Biotechnology, RP02010‐02) after 14 days of culturing in iPSCs culture medium (Nuwacell Biotechnology, RP01001). Then, the surface antigens of MSCs were analyzed by flow cytometry. The cells were continually passaged after reaching 80% confluence, and cells from passages five to ten were used in the subsequent experiments. After reaching 80% confluency, MSC‐sEVs were obtained by the serial ultracentrifugation method.^[15, 25^] Briefly, the culture medium was centrifuged at 300×g for 10 min and 2000×g for 30 min under 4 °C. Then large EVs were removed by centrifugated at 10 000×g for 60 min and filtration through a 0.22 µm filter (Millipore, USA). Next, the supernatant was ultracentrifuged at 100 000×g for 70 min under 4 °C (ultracentrifuge Optima XPN with a SW32 Ti rotor, Beckmann, USA). After removing the supernatant, sEVs pellets were resuspended in PBS and further ultracentrifuged at 100 000×g for 70 min under 4 °C to harvest the pelleted sEVs.

To construct multifunctionally engineered MSC‐sEVs, siMDM2 (Table S4, Supporting Information) was loaded into MSC‐sEVs by the classical electroporation method.^[10, 26^] Briefly, 10 µl MSC‐sEVs (1 × 10¹² particles ml⁻¹) and 100 µg siMDM2 were lightly mixed in 400 µl of electroporation buffer (1.15 mM potassium phosphate, pH = 7.2, 21% Optiprep, 25 mM potassium chloride) and electroporated in a 4 mm cuvette by the electroporation apparatus with a 0.35‐s pulse 20 times at 700 V as previously described.^[23b^] Then, the mixture was incubated for 30 min at 37 °C to recover the membrane structure of MSC‐sEVs. Finally, free siMDM2 was removed from the mixture by ultracentrifugation for 70 min at 100 000×g under 4 °C, and the sEVs^siMDM2 pellets were resuspended in PBS.

To achieve cartilage‐targeting ability and surface charge reverse, sEVs^siMDM2 (1 × 10¹¹ particles ml⁻¹) were then incubated with the PBS solution containing various concentration of WPD (or rhodamine B labeled WPD) for 60 min at room temperature.^[15, 27^] After incubation, 1 ml mixed solution were diluted with 100 ml sterile PBS to eliminate WPD aggregation. Then, the diluted solutions were purified by ultracentrifugation at 100 000×g for 70 min to obtain WPD‐sEVs^siMDM2. To confirm whether WPD could be removed from WPD‐sEVs^siMDM2 by ultracentrifugation, 1 ml rhodamine B labeled WPD in PBS solution at the concentration of 100 µg ml⁻¹ was diluted in 100 ml PBS solution and then purified by ultracentrifugation. Next, the ultracentrifuge tube bottom was washed with PBS and detected by a multifunctional microplate reader (Varioskan LUX, Thermo Fisher Scientific, USA).

To further validate whether WPD modification could alter the integrity and contents of sEVs, a series of experiments were performed. First, NFC analysis (N30 NanoFCM, China) was used to detect the modification rate and size distribution of WPD‐sEVs^siMDM2. Surface charge of unmodified sEVs and WPD‐sEVs^siMDM2 was detected by a nanoparticle analyzer (DelsaMax Pro, Beckman Coulter, USA). TEM (HitachiH‐7650, Japan) was used for morphology observation of unmodified sEVs and WPD‐sEVs^siMDM2. To detect the expressions of surface markers in unmodified sEVs and WPD‐sEVs^siMDM2, the protein samples were collected from MSCs, unmodified sEVs, and WPD‐sEVs^siMDM2 using the RIPA lysis buffer (Beyotime, China). Western blot analysis were performed as described previously.^[15^] Briefly, after being incubated with 5% BSA for 60 min at room temperature for blocking the nonspecific binding, the polyvinylidene difluoride membranes were incubated with various primary antibodies including CD9 (1:1000), CD63 (1:1000), TSG101 (1:1000), and GM130 (1:1000) at 4 °C overnight. Next, the membranes were incubated with horseradish peroxidase (HRP)‐labeled secondary antibody for 60 min at room temperature, and chemiluminescent signal of each band was visualized by using the ECL detection kit and Bio‐RAD imaging system.

To evaluate the stability of WPD modification, the modification rate and average particle diameter and surface charge of WPD‐sEVs^siMDM2 were detected after incubation for different time points by NFC analysis and the nanoparticle analyzer. WPD‐sEVs^siMDM2 or rhodamine B labeled WPD‐sEVs^siMDM2 were placed in PBS for a 7‐day period (1, 2, 4, and 7 days) at room temperature. Additionally, to explore the time‐dependent stability of WPD‐sEVs^siMDM2 under physiologically mimic joint microenvironment, WPD‐sEVs^siMDM2 were incubated with 10 mg ml⁻¹ HA or CS at 37 °C for 1, 2, 4, and 7 days. After incubation, WPD‐sEVs^siMDM2 were isolated by 0.22 µm filtration and serial ultracentrifugation, then the modification rate was detected by NFC.

All procedures in this study were approved by the Independent Ethics Committee of Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (Approval Number: 2022‐0611, October 24, 2022, Shanghai, China). Written informed consent was provided from the patients. Human articular chondrocytes were isolated from knee joint cartilage of OA patients. As previously described,^[15^] the harvested articular cartilage was cut into 1 mm³ and incubated with 0.25% collagenase II in DMEM‐F12 medium supplemented with 1% Penicillin–Streptomycin for 4 h at 37 °C. After digestion, human chondrocytes were resuspended and filtered through a 40 µm cell filter (Corning, USA) before seeding at a density of 1 × 10⁵ cells ml⁻¹ in DMEM‐F12 medium supplemented with 1% Penicillin–Streptomycin and 10% FBS at 37 °C under 5% CO₂. Chondrocytes at passage 2 were used in the following experiments. To mimic the senescent phenotype of chondrocytes in OA cartilage, a DNA‐damaging chemical agent doxorubicin at the concentration of 1 µM was used to treat the human chondrocytes once every 2 days for 14 days,^[19b^] and then treated with sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) for 7 days.

Lipid dye DiO was used to label sEVs according to previously published article.^[15^] Briefly, unmodified sEVs (1 × 10¹⁰ particles ml⁻¹) and WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) were incubated with 10 µM DiO for 30 min at 37 °C in the dark. Subsequently, the suspension was filtered by 0.22 µm membrane and centrifuged at 100,000×g for 140 min to remove the free DiO. Next, the purified DiO‐labeled sEVs and WPD‐sEVs^siMDM2 were incubated with chondrocytes for 12 h at 37 °C under 5% CO₂. After incubation, the chondrocytes were fixed with 4% PFA for 30 min and then stained with DAPI for 5 min. Images in each group were observed by the confocal microscopy (Leica Microsystems, Germany). In addition, the DiO staining of PBS only group were treated with the same isolation procedure to exclude the possibility of the nonspecific labelling. The results were also analyzed by the CytoFLEX flow cytometry (Beckman Coulter, USA) and the mean fluorescence intensity of DiO‐positive unmodified sEVs and WPD‐sEVs^siMDM2 were evaluated.

A Cell Counting kit‐8 (CCK8) kit (Dojindo, Japan) was applied to evaluate the chondrocyte proliferation according to manufacturers’ instruction. As described previously,^[15, 28^] human chondrocytes were seeded into 96‐well plate at 3000 cells per well. After being incubated with WPD, unmodified sEVs, or WPD‐sEVs^siMDM2 for different time points, the absorbance in each group was detected at 450 nm by a multifunctional microplate reader (Varioskan LUX, Thermo Fisher Scientific, USA).

Total RNA was isolated from human chondrocytes by using the RNeasy kit following the manufacturer's instruction (Qiagen, Germany). Total RNA from human cartilage explants was isolated by using the TRIzol reagent (0.1 ml TRIzol g⁻¹ tissue, Invitrogen, USA), followed by the chloroform and phenol acid extraction. The concentration of total RNA was measured at OD260. The complementary DNA (cDNA) was obtained by extracted RNA by reverse transcription using the PrimeScriptTM RT reagent Kit (RR047A, TaKaRa, Japan). SYBR Green (RR420A, TaKaRa, Japan) was applied to proceed with PCR reactions following the manufacturer's recommendation and human gene β‐actin was set as the housekeeping gene. The relative gene expressions of the mRNAs were analyzed by using the 2^−ΔΔCT method.^[29^] All primers used for PCR analysis were purchased from Sangon Biotech (Shanghai, China) and are listed in Table S5 (Supporting Information).

The protein samples were collected from chondrocyte lysates using RIPA solution. The Western blot analysis was performed as described previously,^[30^] after blocked with 5% BSA for 60 min at room temperature to block the unspecific absorption and then the PVDF membranes were incubated with various primary antibodies overnight at 4 °C. Subsequently, the membranes were incubated with horseradish peroxidase (HRP)‐labeled secondary antibody at room temperature for 60 min. The chemiluminescent signals were visualized by the ECL Western Blot detection kit and Bio‐RAD imaging system.

Immunofluorescence staining was performed to detect the expressions of P16^INK4a, γH2AX, and MDM2 in chondrocytes. The chondrocytes were fixed in 4% PFA and immersed in 5% BSA for 1 h at 37 °C. Then, the cells were incubated with different primary antibodies including P16^INK4a (1:100), γH2AX (1:100), P21 (1:100), and MDM2 (1:100) at 4 °C overnight. Next, the secondary antibody (1:200) with fluorescence was added to the cells for 1 h at room temperature in the dark. After incubation, cells were stained with DAPI for 5 min and then observed using a confocal microscope (Leica Microsystems, Germany).

After different treatments, the apoptosis rates of chondrocytes in all groups were detected using a TUNEL staining kit (Beyotime, China) according to the protocol.

SA‐β‐gal staining was performed as previously described.^[31^] The cell senescence β‐galactosidase staining kit (Beyotime, China) was applied according to the manufacturer's instruction. Senescent chondrocytes were identified as blue‐stained cells under light microscopy (Leica Microsystems, Germany). Total cells and SA‐β‐Gal positive cells were counted in three random fields per culture dish to analyzed the percentage of SA‐β‐gal positive cells.

Two types of human articular cartilage specimens were collected in this study. OA‐affected articular cartilage (OA cartilage) and non‐OA affected articular cartilage (normal cartilage) were both collected from six OA patients (mean ± S.D. age, 75.17 ± 6.62 years old; range, 67–85 years old) undergoing total knee joint replacement. For ex vivo culture of OA cartilage explants, explants were harvested from OA‐affected area or non‐OA affected area of articular cartilage using biopsy punch (5 mm in diameter and 1 mm in thickness) and washed with PBS supplemented with 1% Penicillin–Streptomycin for three times. Next, the cartilage explants were cultured in DMEM‐F12 medium supplemented with 10% FBS and 1% Penicillin–Streptomycin in a 48‐well‐plate. For the treatments on OA cartilage explants, sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) were applied for 7 days.

Articular cartilage explants (Φ 5 mm, 1 mm thickness) were harvested from the knee joints of OA patients by using an electric drill. After extraction, the cartilage explants were washed with PBS containing 1% Penicillin–Streptomycin. In order to evaluate the cartilage uptake and penetration capacity of unmodified sEVs and WPD‐sEVs^siMDM2. Lipophilic fluorescence dye DiO was used to label unmodified sEVs and WPD‐sEVs^siMDM2. The cartilage uptake efficiency was measured as the percentage of sEVs from the bath solution to the cartilage explants. The cartilage explants were immersed in 200 µl PBS containing DiO labeled unmodified sEVs and WPD‐sEVs^siMDM2 in a 96‐well plate at 37 °C for 1, 3, and 7 days. After that, the cartilage explants were removed, and the DiO fluorescent intensity was recorded by the multifunctional microplate reader (Varioskan LUX, Thermo Fisher Scientific, USA).

Next, a self‐designed one‐way transport teflon mould was applied to assess the cartilage penetration ability of unmodified sEVs and WPD‐sEVs^siMDM2 as described previously.^[15^] Briefly, the freshly harvested cartilage explants were stuck into the hole in the removable baffle which divided the mould to two chambers. 200 µl DiO labeled unmodified sEVs or WPD‐sEVs^siMDM2 were added to one side chamber of the mould, while 200 µl sterilized PBS was added to the other chamber of the mould. After penetration for different time points, the cartilage explants were immersed with OCT glue and sectioned to 5 µm slides by a vibrating microtome (Leica, Germany) followed by an immediate microscopy observation (Leica Microsystems, Germany).

All animal experimental procedures were approved by the Animal Research Committee of Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (Approval Number: 2022‐0611, October 24, 2022, Shanghai, China). The joint retention capacity of unmodified sEVs or WPD‐sEVs^siMDM2 in mouse was detected over a period of 35 days by using an IVIS Spectrum imaging system (PerkinElmer, USA). For in vivo imaging study, a near‐infrared fluorescence lipid dye DiR was used to label unmodified sEVs and WPD‐sEVs^siMDM2. Briefly, 10 µl DiR‐labeled unmodified sEVs or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) was injected to the mouse knee joint cavity. After intra‐articular injection, images of each mouse joint were recorded by the IVIS Spectrum imaging system at different time points (0, 1, 3, 7, 14, 28, and 35 days). The data of total radiant efficiency within the region of each knee joint was calculated and presented along with time‐based on previous studies.^[15, 32^]

To establish PTOA model, 8‐week‐old male mice were subjected to sham surgery or ACLT surgery as described previously.^[19, 20^] Briefly, the right joint capsule was opened and the anterior cruciate ligament was transected to destabilize the right knee joint in ACLT group mice, whereas the joint capsule was opened without any other operation in sham group mice. For the administration of different treatments in ACLT‐induced PTOA mice, 10 µl PBS, sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) was intra‐articularly injected into the right knee joint cavity of mice once every four weeks by using a 10 µl micro syringe (Hamilton, USA) four weeks after surgery. All mice were euthanized eight weeks after different treatments. For the administration of different treatments in naturally aged OA mice, 10 µl PBS, sEVs (1 × 10¹⁰ particles ml⁻¹), sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹), or WPD‐sEVs^siMDM2 (1 × 10¹⁰ particles ml⁻¹) was intra‐articularly injected into the right knee joint cavity of 12‐month‐old mice once a month. Mice were euthanized after 8 months treatments. 8‐month‐old mice were used as the young control group.

The harvested knee joints in each group were fixed immediately in 4% PFA for 24 h, dehydrated using the gradient alcohol, vitrified with dimethyl benzene, and embedded in the paraffin. The paraffin sections (thickness: 5 µm) were deparaffinized in xylene, hydrated with gradient ethanol, and stained with Safranin O and H&E solutions. Additionally, the score of cartilage loss and destruction in each group was evaluated by using the mouse OARSI scoring system,^[33^] and the synovial inflammation was also assessed according to previously published article.^[19b^] In addition, subchondral bone was evaluated by a method which was published previously.^[34^]

For evaluation of the immunoreactivity to collagen II, SOX9, MMP13, P16^INK4a, P21, IL6, HMGB1, TNFα, MDM2, P53, and p‐P53, histological sections were deparaffinized and then incubated with 0.3% hydrogen peroxide at room temperature for 30 min. Primary antibodies used in this study were listed in Table S3 (Supporting Information). Subsequently, the enzyme‐activated antigen retrieval (0.1% trypsin) for each section was performed for 10 min at room temperature. Then, all sections were blocked with 5% BSA for 1 h at room temperature and incubated with primary antibodies overnight at 4 °C. Next, appropriate HRP‐labeled secondary antibodies (1:200) were applied to incubation with the sections for 1 h at room temperature. After that, all sections were observed and analyzed under a light microscopy (Leica Microsystems, Germany).

The cartilage tissues were also harvested and homogenized in 1 ml PBS.^[35^] After centrifugation at 10 000×g for 15 min at 4 °C, the supernatants in different groups were used for examining the expressions of inflammatory mediators IL6 and TNFα by the corresponding ELISA kits. The microplate reader was applied to measure the absorbance at 450 nm, and the final presented data for each group was normalized with the control group.

Statistical analysis was performed by using GraphPad Prism software 8. All the experimental results were expressed as the mean ± standard deviation (SD). Differences between groups were evaluated by the two‐sided Student's t‐test (for comparison of two groups) or the one‐way analysis of variance (ANOVA) with Tukey's test (for comparison of more than two groups). Data quantified based on ordinal grading systems including the OARSI grade, synovial inflammation score, and the subchondral bone score, whose data points are not continuous and do not follow a normal distribution, were analyzed using nonparametric statistical method based on Mann–Whitney U test was used. Statistical significance level was defined as P < 0.05 for hypothesis testing.