Immunomodulatory and anabolic biphasic scaffold with hierarchical biomimetic structure directed osteochondral defect repair

Abstract

Repairing osteochondral tissue is challenging due to its hierarchical structure, mechanical heterogeneity, and the need for spatial control over stem cell differentiation. Advances in tissue engineering have facilitated the development of biphasic cartilage-bone integrated scaffolds for osteochondral repair. The cartilage layer, composed of an IGF-1-loaded polydopamine-ZIF8/HAMA hydrogel, mimicked native tissue and enabled controlled cytokine release. This layer promoted M2 macrophage polarization, enhanced BMSC migration and chondrogenesis, and improved cartilage anabolism. The subchondral bone layer was a nanoclay (XLS)-functionalized 3D bioglass scaffold, which provided superior mechanical strength and supported osteogenic differentiation. These layers were integrated via partial interpenetration of the HAMA hydrogel. Importantly, in vivo studies confirmed that our biphasic scaffold effectively promoted osteochondral defects regeneration in a rat lower femoral osteochondral defect model. Collectively, this biphasic scaffold system presents a promising therapeutic strategy for osteochondral tissue regeneration.

Introduction

Articular cartilage defects caused by trauma, disease, or aging lead to knee joint dysfunction and frequently involve subchondral bone degeneration, ultimately resulting in osteochondral defects¹. The regeneration of osteochondral tissue is challenged due to the distinct chemical compositions, biomechanical properties, and biological functions of cartilage and subchondral bone². Defective cartilage has limited self-healing capacity owing to its avascular and alymphatic nature³. The loss of chondrocytes, disordered ECM production, and excessive inflammatory responses caused by cartilage damage hinder cartilage repair^4,5.

Chondrocyte represent the primary cell type for cartilage repair, but their application is limited by the poor migratory capacity and limited availability⁶. Recently, BMSCs have shown great potential for cartilage repair due to their chondrogenic differentiation capacity^7–9. However, exogenous BMSCs transplantation faces limitations including low cell survival rates, infection risks, and immunocompatibility issues¹⁰. Consequently, recruiting endogenous stem cells to the injury site and inducing their differentiation into chondrocytes has emerged as a promising strategy for cartilage repair. IGF-1 has shown profound effects on chondrocyte biological behavior and fundamentally regulates ECM production during cartilage repair¹¹. Beyond stimulating ECM synthesis, IGF-1 enhances the recruitment, proliferation, and chondrogenic differentiation of BMSCs^11–14. Nevertheless, direct IGF-1 application for cartilage repair is impeded by its short half-life and susceptibility to inflammation-induced degradation¹⁵. Thus, to further improve the efficiency of IGF-1, a protective carrier is needed for the loading and sustained release of IGF-1 to achieve the biological functions of IGF-1. Metal-organic frameworks (MOFs) have recently attracted tremendous interest due to their large surface area and porous structure^16,17. Among these, zeolitic imidazolate nanoparticles (ZIF8 NPs) have been widely employed as drug carriers due to their mild synthesis conditions, pH-responsive properties, and capacity to conjugate with small molecules, peptides, and proteins^18–20. Moreover, the functionalization of polydopamine (PDA) provides nanoparticles adhesive properties and reactive sites and also endows anti-inflammatory capability to the nanoparticles^21,22. These properties enable the preparation of PDA-hybridized ZIF8 (PZIF8) nanoparticles loaded with IGF-1 (PZIF8-IGF-1 NPs).

In this study, we developed a multifunctional bilayer scaffold with distinct composition and mechanical properties to mimic the hierarchical structure of osteochondral tissue. The cartilage layer consists of a PZIF8-IGF-1 functionalized hyaluronic acid methacryloyl hydrogel (HAMA/PZIF8-IGF-1, HPI), which demonstrates endogenous BMSC homing and anti-inflammatory effects. The subchondral bone layer comprises a nanosilicate (XLS, Laponite)-functionalized 45S5 bioactive glass scaffold (BGX). These layers were integrated through semi-infiltration of HAMA into the bone layer. In vitro studies demonstrated that the HAMA/PZIF8-IGF-1 hydrogel promoted SD-BMSC migration, chondrogenic differentiation, and cartilage anabolism while downregulating inflammatory responses in macrophages. The BGX scaffold enhanced osteogenic differentiation, as evidenced by increased ALP activity, upregulation of bone-related genes, and enhanced biomineralization of SD-BMSCs. Furthermore, in vivo results from a lower femoral osteochondral defect model confirmed that the HAMA/PZIF8-IGF-1 hydrogel facilitated cartilage repair while the BGX scaffold promoted new bone formation.

Results

Synthesis and characterization of PZIF8 NPs and HAMA hydrogel

To verify the successful synthesis of polydopamine (PDA)-modified ZIF8 (PZIF8), TEM, DLS, and FTIR tests were carried out. The observed color change following PDA modification served as a preliminary visual indicator of successful polydopamine functionalization on the ZIF-8 nanoparticles (Fig. S1). The TEM images revealed that the morphology of PZIF8 (Fig. 1a, b) was similar to ZIF8 NPs (Fig. S2) and had an average diameter of 350.83 ± 64.40 nm. The DLS data (Fig. 1c) showed that the hydrodynamic diameter of the ZIF8 increased stepwise from 220.57 ± 44.75 nm to PZIF8 (255.07 ± 22.47 nm). The zeta potential data also confirmed the successful functionalization of PDA into ZIF8, since the zeta potential was shifted from ZIF8 (−1.7 mV) to PZIF8 (−13.3 mV) (Fig. 1d). The FTIR spectra (Fig. 1e) showed the same characteristic peaks of PZIF8 and ZIF8 at around 3157 cm⁻¹, 2927 cm⁻¹ and 1579 cm⁻¹, which were attributed to the aromatic and aliphatic C-H and -NH- of imidazole in ZIF8. Moreover, the characteristic peaks of PDA at 1600 cm⁻¹ and 1280 cm⁻¹ were also observed in PZIF8, which belonged to the aromatic nucleus and the vibrational peak of the aromatic O-H, respectively.

Fig. 1. Synthesis and characterization of PZIF8 Nanoparticles and HAMA/PZIF8 hydrogels.

a, b TEM images of PZIF8 NPs; (c) DLS size measurements of ZIF8 and PZIF8; (d) Zeta potential of PDA, ZIF8 and PZIF8 NPs; (e) FTIR spectra of PDA, ZIF8 and PZIF8 NPs; (f) HAMA/PZIF8 hydrogel magnified SEM image at 5000X magnification; (g) swelling behaviors of HAMA, HAMA/PZIF8 hydrogels immersed in DPBS solution for 1 week; (h) degradation behaviors of HAMA, HAMA/PZIF8 hydrogel immersed in DPBS solution for 4 weeks; (i) IGF-1 release profiles from HAMA/IGF-1 and HAMA/PZIF8/IGF-1 hydrogel in DPBS solution. Data are expressed as mean ± SD (n = 3).

To investigate the synthesis of HAMA, ¹H NMR was performed. Compared the ¹H NMR spectrum of HA (Fig. S3b), new peaks (Fig. S3c) at 5.67 ppm and 6.08 ppm attributed to the -C=CH2 in methacrylic anhydride (MA) and the characteristic absorption peak at 1.88 ppm belonging to the -CH3 in MA were observed in HAMA. The degree of substitution (DS) of methacryloyl groups calculated from the ¹H NMR data was about 67.5%. Fig. S3d showed the SEM images of different concentrations of HAMA hydrogels after freeze-drying. The pore distribution in 1% HAMA hydrogel was not uniform, while no interconnected pores were observed in 2% HAMA hydrogel. The 1.5% HAMA hydrogel had an interconnected uniform pore size distribution, with a pore size around 120–200 μm, which was suitable for cell adhesion and growth²³. Therefore, 1.5% HAMA was used in the subsequent experiments. Fig. 1(f) showed that in PZIF8 loaded HAMA hydrogel, PZIF8 NPs uniformly dispersed in the HAMA hydrogel. Time-sweep rheological analysis of the HAMA hydrogel revealed a sol-gel transition at 43 seconds, identified by the crossover point where G′=G′′. The storage modulus G′ subsequently reached a plateau after approximately 150 s, indicating complete hydrogel network formation (Figure S4). After 150 s, G’ reaches a plateau, indicating that the hydrogel crosslinking is completed. Swelling behaviors of HAMA and HAMA/PZIF8 hydrogels were studied in DPBS solution at 37 °C (Fig. 1g). Similar swelling profiles, including the swelling equilibrium time and equilibrium swelling ratio, were observed between HAMA and HAMA/PZIF8 hydrogels in a period of 7 days. These results indicated that the hydrogels will be structurally stable after implantation into the articular cartilage of joint injury and are beneficial for the repair of articular cartilage defects. We also measured the degradation behaviors of HAMA and HAMA/PZIF8 hydrogels (Fig. 1h). The degradation curves of the two hydrogels were similar during the 28-day degradation period, and the degradation rates of HAMA/PZIF8 hydrogels were slightly lower than those of HAMA hydrogels. The results indicated that the modification of PZIF8 did not affect the degradation behavior of HAMA hydrogels. After 28 days (i.e., 672 hours) of degradation, the residual weight ratios for HAMA and HAMA/PZIF8 hydrogels were around 70.19% and 65.75%, respectively.

The in vitro IGF-1 release profiles from HAMA/IGF-1 and HAMA/PZIF8-IGF-1 hydrogels were shown in Fig. 1i. An initial burst release was observed in both hydrogels in the first 24 h, with a cumulative release about 330 ng. The HAMA/IGF-1 hydrogel reached a release plateau at 7 days, with a cumulative IGF-1 release amount around 490 ng. On the other hand, a sustained release profile was observed in the HAMA/PZIF8/IGF-1 hydrogel over the next 13 days, indicating that the mesoporous structure of PZIF8 prolonged the release duration of IGF-1 and prevented the its degradation. Additionally, a higher amount of IGF-1 was released from the HAMA/PZIF8-IGF-1 hydrogel compared to the HAMA/IGF-1 group (600 vs 490 ng). This finding is consistent with our expectation that an efficient initial release of IGF-1 in the early stage after implantation could promote the recruitment of BMSCs. Subsequently, a sustained release of IGF-1 would induce the differentiation of BMSCs towards chondrocytes.

Synthesis and characterization of BG-based scaffolds

BG and BGX scaffolds were obtained using the foam replica and sintering techniques. The SEM showed that both BG and BGX scaffolds exhibited a highly interconnected and uniform porous structure with a pore size between 300–500 μm, and the surface morphology was relative rough Fig. 2(a, c). After HAMA modification, the open pore structure of BG and BGX scaffolds was maintained, and HAMA was covered on the struts of BG and BGX scaffolds with a relative smooth morphology Fig. 2(b, d). The compressive strength of BG scaffolds was significantly enhanced after XLS incorporation, from 0.11 MPa to 0.25 MPa (Fig. 2i). The compressive properties of BG and BGX scaffolds were also markedly increased after HAMA coating, from 0.11 MPa to 0.80 MPa and 0.25 MPa to 1.03 MPa, respectively. Figure 2(j) showed the cross-section microstructure of the bilayered scaffold, where the HPI hydrogel was used for the cartilage layer and the HBGX was used as the subchondral bone layer, and both layers were connected by the HPI infiltration into the HBGX to form a hybrid layer. These data indicated that a gradient layer existed between the cartilage layer and the bone layer, which provides the integrity of the osteochondral structure.

Fig. 2. Synthesis and characterization of BG-based scaffolds.

SEM images of (a, e) BG, (b, f) HBG, (c, g) BGX and (d, h) HBGX scaffolds; (i) Compressive modulus of BG, HBG, BGX and HBGX scaffolds; (j) SEM image of the cross-section of HPI hydrogel and HBGX scaffolds. Data are expressed as mean ± SD (n = 3). **p < 0.01, ***p < 0.001.

Chondrogenic properties of IGF-1

The combination of transforming growth factor-β (TGF-β) and IGF-1 is nowadays a common protocol for inducing chondrogenic differentiation of stem cells. However, it has also been reported that IGF-1 alone can induce stem cells differentiation into cartilage cells²⁴. Therefore, in this study, we investigated the effects of IGF-1 on the proliferation, migration and pro-chondrogenic differentiation of SD-BMSCs. As shown in Figs. S5, S6, both the CCK8 assay and live/dead cell staining results demonstrated that IGF-1, at concentrations between 10 and 500 ng/mL, significantly promoted the proliferation of SD-BMSCs, especially at the concentration of 100 ng/mL. Additionally, the wound-healing data revealed that IGF-1 promoted the migration of SD-BMSCs in a dose-dependent manner, but there was no significant difference between the 100 ng/mL and 500 ng/mL concentrations of IGF-1 (Fig. 3a–c). Since cartilage tissue is primarily composed of collagenous extracellular matrix and a small number of chondrocytes, we also investigated the role of IGF-1 in the invasion of SD-BMSCs using a transwell assay. As shown in Fig. 3d, e, the highest number of SD-BMSCs penetrated into the lower chamber in the presence of 100 ng/mL of IGF-1. It is worth noting that when the concentration reached 500 ng/mL, IGF-1 inhibited the invasion of SD-BMSCs compared to the 100 ng/mL group.

Fig. 3. Effects of IGF-1 on BMSC migration, invasion, and chondrogenic gene expression.

a Wound-healing assay images after removing the ibidi culture-insert for 0, 8 and 24 h treatment with different concentration of IGF-1; (b, c) Quantification of wound healing data for 8 and 24 h, respectively; (d) Crystalline violet staining images of invasive cells treatment with different concentrations of IGF-1 (d1-d4 corresponeds to IGF-1concentration of 0, 10, 100 and 500 ng/mL) and (e) Quantitative analysis of invasive cell numbers; The relative gene expression levels of cartilage anabolic related genes markers (Aggrecan, Col-2α1, Sox-9) and catabolic genes markers (MMP13, Adamts4, Adamts5) of SD-BMSC cells cultured on different concentrations of IGF-1 after (f) 7 days and (h) 14 days; Western blot analysis of Sox-9 and Col-2α1 in SD-BMSC cells after cultured on different concentrations of IGF-1 after (g) 7 days and (i) 14 days. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

To further explore the chondrogenic effect of IGF-1 on SD-BMSCs, RT-PCR and western-blot experiments were conducted. At day 7, IGF-1 increased the expression of cartilage anabolic-related gene markers (Aggrecan, Sox-9) and inhibited the expression of catabolic-related markers (MMP13, Adamts4, Adamts5) at concentrations of 100 and 500 ng/mL (Fig. 3f). Furthermore, the groups treated with IGF-1 (100 and 500 ng/mL) showed increased expression of anabolic-related genes (Aggrecan, Col-2α1, Sox-9) and decreased expression of catabolic-related genes (MMP13, Adamts4) after 14 days of treatment (Fig. 3h). Our western blot data (Fig. 3h, i, S7, S8) showed that at day 7, Sox-9 expression increased in an IGF-1 dose-dependent manner. While all tested IGF-1 concentrations (10, 100, and 500 ng/mL) enhanced Col-2α1 expression, the highest level was observed at 100 ng/mL. By day 14, all IGF-1 concentrations continued to significantly upregulate Sox-9, with the most pronounced effect at 10 ng/mL. In contrast, Col-2α1 expression remained maximally elevated at 100 ng/mL. The presence of glycosaminoglycan on SD-BMSCs was determined by toluidine blue and safranin O staining after culturing with IGF-1 for 10 and 21 days, and our data revealed that IGF-1 promoted chondrogenic differentiation in a time and dose-dependent manner (10–100 ng/mL) (Figure S9, S10). Based on the above data, IGF-1 at a concentration of 100 ng/mL was used in the subsequent experiments.

Anti-inflammatory Effects of PZIF8 NPs

The capacity of cartilage tissues to repair itself is limited due to a lack of blood supply, nerves, and lymphatic vessels²⁵. The inflammatory microenvironment created after cartilage injury leads to chondrocyte death and hypertrophy, further undergo the abnormal differentiation of stem cells and inhibition of extracellular matrix synthesis, ultimately impeding cartilage repair²⁶. Therefore, it is crucial to modulate the immune microenvironment during the process of cartilage repair. In this study, a polydopamine (PDA) functionalized ZIF8 (PZIF8) nanoparticle was developed to modulate the inflammatory activity of macrophages. Lipopolysaccharide (LPS), the most potent inducer of M1-type polarization in macrophages, was used to induce a pro-inflammatory response. Both live/dead cell staining and CCK8 assay results showed that PZIF8 nanoparticles had good cell compatibility within the concentration range from 0 to 10 μg/mL (Fig. 4a, b). As seen in Fig. 4(c), LPS significantly promoted the expression of pro-inflammatory gene markers (IL-1β, IL-6, and TNF-α) after 1 and 3 days of incubation. After 1 day of co-culture with Raw264.7 cells, the PZIF8 NPs markedly decreased M1-related gene expressions (IL-1β, IL-6) (Fig. 4c) and increased M2-related gene expressions (IL-4, IL-10 and Arg-1) (Fig. 4d). In contrast, the ZIF8 NPs had pro-inflammatory effects on Raw264.7 cells (Fig. 4e, f). Similar results were also observed with PZIF8 and ZIF8 NPs after 3 days of co-culture with Raw 264.7 cells (Fig. 4g–l). These data demonstrate that ZIF8 NPs alone had no anti-inflammatory capabilities, but the functionalization of PDA in ZIF8 NPs provides an anti-inflammatory effect on macrophages. In cartilage tissue, M1-type macrophages cause degradation of the cartilage matrix, inhibit the proliferation and viability of BMSCs, impair their ability to immunosuppress the environment, and further prevent cartilage repair^27,28. In contrast, M2-type macrophages play a role in maintaining the phenotype of chondrocytes, as IL-4, IL-10 and Arg-1 secreted by M2-type macrophages have an effect on maintaining the phenotype of chondrocytes and promoting wound healing. Therefore, the developed PZIF8 NPs could promote cartilage tissue repair by regulating the homeostasis of pro-inflammatory and anti-inflammatory factors in the inflammatory microenvironment.

Fig. 4. Morphological polarization of Raw 264.7 cells toward the M2 phenotype stimulated by PZIF8 NPs.

a Live/dead staining of Raw264.7 cells cultured on (a1, a3) control and (a2, a4) 5 μg/mL PZIF8 NPs for 1 day and 3 days, respectively; (b) Cytotoxicity of different concentration of PZIF8 NPs on Raw264.7 cells cultured for 1 and 3 days; The relative gene expression levels of IL-1β, IL-6, TNF-α of Raw 264.7 cells cultured on LPS and PZIF8 NPs+LPS after (c) 1 day and (g) 3 days; The relative gene expression levels of IL-4, IL-10, Arg-1 of Raw 264.7 cells cultured on LPS and PZIF8 NPs+LPS after (d) 1 day and (h) 3 days; The relative gene expression levels of IL-1β, IL-6, TNF-α of Raw 264.7 cells cultured on LPS and ZIF8 NPs+LPS after (e) 1 day and (i) 3 days; The relative gene expression levels of IL-4, IL-10 and Arg-1 of Raw 264.7 cells cultured on LPS and ZIF8 NPs + LPS after (f) 1 day and (j) 3 days. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Biological functions of HPI hydrogels in vitro

The cytotoxicity of hydrogels was assessed in SD-BMSCs usign the CCK8 assay, and HAMA (H) hydrogel was selected as the blank control (Fig. 5a). Significantly higher cell viability was observed on HAMA/PZIF8-IGF-1 (HPI) hydrogel, and no toxicity was detected in the HAMA/PZIF8 (HP) and HAMA-IGF-1 (HI) groups compared to the HAMA (H) group after 1 and 4 days of culture. On day 7, all HP, HI and HPI hydrogels showed higher cell viability than the H hydrogel, although the HPI hydrogel had the highest cell viability. Consistent with the CCK8 data, similar results were observed in the live-dead staining images of SD-BMSCs cultured on the four groups of hydrogels after 4 and 7 days of culture (Fig. 5b). Moreover, 3D reconstruction of live and dead cell images also showed that SD-BMSCs spread well and grew faster on HPI hydrogels than on the other hydrogel groups (Figs. S11–13). Similarly, no toxic effects were observed in Raw 264.7 cells after culture on all four groups of hydrogels for 1 and 4 days (Fig. S14). It is worth noting that significantly higher cell viability was observed in IGF-1-containing hydrogels (HI, HPI) compared to the H hydrogel, indicating that IGF-1 could promote the proliferation of Raw 264.7 cells.

Fig. 5. Biological functions of HPI hydrogels on different cells.

a Cytotoxicity of SD-BMSCs on different group of hydrogels after culture for 1, 4 and 7 days; (b) Live/dead staining of SD-BMSCs cultured on different group of hydrogels for 4 and 7 days; (c) Influence of different hydrogels extracts on SD-BMSCs migration after removing the ibidi culture-insert for 0, 8 and 24 h; (d, e) The quantification results of cell migration rates for 8 and 24 h, respectively; (f) The relative gene expression levels of Aggrecan, Col-2α1, Sox-9, Adamts4, Adamts5 of SD-BMSC cells cultured on different group hydrogels for 7 days; (g) Western Blot analysis of the protein expression level of Col-2α1, MMP13, Adamts5 in SD-BMSC cells after cultured on different group of hydrogels for 7 days; (h–j) Quantification results of the Col-2α1, MMP13, Adamts5 protein levels determined by western blot; (k, l) The relative gene expression levels of IL-1β, IL-6, TNF-α, IL-4, IL-10, Arg-1 of Raw 264.7 cells cultured on different group of hydrogels for 1 day. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

To investigate the effect of hydrogels on the migration of SD-BMSCs, a scratch migration assay was performed. As shown in Fig. 5c-e, both HI and HPI hydrogels exhibited significantly higher migratory activity of SD-BMSCs than the H hydrogel after 8 and 24 hours of treatment, especially the HPI hydrogel group. Moreover, no obvious wound healing capability was observed in the HP hydrogel compared to the H hydrogel. These results suggest that the biological functions of IGF-1 were maintained after being physically mixed into the HAMA hydrogel or encapsulated into the PZIF8 NPs, which were further mixed with the HAMA hydrogel, and the sustained release of IGF-1 from PZIF8 NPs had positive effects on the wound healing of BMSCs.

Cartilage is comprised of 95% ECM and 3–5% chondrocytes⁴, and the dynamic balance of cartilage matrix metabolism plays a crucial role in the process of cartilage repair, since the ECM can maintain cartilage structure integrity and regulate chondrocyte functions⁷. To evaluate the anabolism and catabolism of SD-BMSCs, RT-PCR and western blot were carried out to assess the influence of hydrogels on the metabolism of SD-BMSCs. As shown in Fig. 5f, we found that the cartilage anabolic-related genes markers aggrecan, Col-2α1, and Sox-9 were significantly increased in HPI hydrogel compared to H hydrogel, while the catabolism-related gene expressions (MMP13, Adamts4 and Adamts5) were markedly decreased. Interestingly, although HI hydrogel increased aggrecan expression, no significant increase was observed in Col-2α1 and Sox-9 compared to H hydrogel after 7 days of treatment. To further investigate this finding, our western blot analysis revealed that HPI hydrogels increased the Col-2α1 expression, downregulating the MMP13 and Adamts5 protein expressions compared to H hydrogel (Fig. 5g-j). Additionally, HP and HI hydrogel could also upregulate Col-2α1 and downregulate Adamts5 expressions compared to the H group. All these data indicate that the sustained release of IGF-1 combined with the incorporation of PZIF8 NPs synergistically facilitate the cartilage matrix metabolism.

To further investigate the anti-inflammatory performance of hydrogels on Raw264.7 cells, RT-PCR was performed. In the inflammatory microenvironment generated by LPS, the HP and HPI hydrogels had a significant effect on inhibiting M1 macrophage-related gene expression and promoting M2 macrophage-related gene expression, with the effect of HPI hydrogels being the most obvious (Fig. 5k–l). Interestingly, the H hydrogel also has the ability to remarkably reduce M1 marker gene expressions (IL-6 and TNF-α) and promote M2 marker gene (Arg-1) expression, while HI could decrease IL-1β and TNF-α expressions and increase IL-4 expression. Miraculously, after incubating with all groups of hydrogels for 1 day, both M1 phenotype genes (TNF-α, IL-1β, IL-6) and M2 phenotype genes (IL-4, IL-10, Arg-1) were dramatically increased (Figs. S15, S16). In summary, we developed a hydrogel that exhibited good biocompatibility, enhanced SD-BMSCs migration and proliferation abilities, increased anabolism, and reduced catabolism of cartilage matrix, and regulated macrophage immune response. Therefore, the HPI hydrogel is a promising alternative for cartilage tissue regeneration.

Biological functions of BG-based scaffold in vitro

The cytotoxicity of SD-BMSCs on BG, HBG, BGX, HBGX scaffolds were measured by the CCK8 assay after 1 and 4 days of culture (Fig. 6a). HBG, BGX, HBGX scaffolds exhibited significantly higher cell viability compared to the BG scaffold at both 1 day and 4 days. To further investigate the osteogenic differentiation abilities of BG-based scaffolds, alkaline phosphatase (ALP) assay, osteogenic-related gene expression (RT-PCR) and calcium nodule formation tests were performed. As shown in Fig. 6b, Fig. S17, the HBG, BGX, HBGX scaffolds exhibited significantly higher ALP expression compared to the BG scaffolds after 7 days of culture. The BGX scaffold promotes ALP activity, mainly due to the degradation products of XLS, such as SiO₄^4-, Li⁺ and Mg²⁺, which can induce osteogenic differentiation of stem cells^29,30. It is worth noting that HAMA modified BG scaffolds also showed the ability to enhance the ALP activity of SD-BMSCs, and no synergistically effects between HAMA and XLS on ALP activity were detected. COL-1α1, the early osteogenic marker, and BSP and OCN the late osteogenic marker, were studied after 7 and 21 days of culture. On day 7, HBGX scaffolds showed highest BSP expression, and HBG and HBGX groups showed higher COL-1α1 expression relative to BG group. Intriguingly, the HBG scaffold also exhibited a significant promotion of osteogenic gene expression, BSP, either at 7 or at 21 days, which may be due to the HAMA modification providing a microenvironment similar to endochondral osteogenesis for the BG scaffold. Since Hyaluronic acid (HA), a non-sulfated glycosaminoglycan, had positive effects on COL-1α1 expression (Fig. 6c, d)³¹. After 21 days of co-culture, both the BGX and HBGX scaffold showed a significant advantage of BSP and OCN gene expressions over the BG scaffold. Additionally, the biomineralization of SD-BMSCs on scaffold extracts was studied through analyzing the calcium nodules after 21 days of culture. As shown in Fig. 6g, a higher amount of calcium nodules was observed in the HBGX group compared to the BG group, which is beneficial for the subchondral bone regeneration.

Fig. 6. BG-based scaffolds promote osteogenic differentiation of SD-BMSCs.

a Cytotoxicity of BG, BGX, HBG and HBGX scaffolds on SD-BMSCs; (b) ALP activities of SD-BMSCs cultured on different BG-based scaffolds extracts for 7 d; The relative gene expression levels of (c) BSP and (d) Col-1α1 of SD-BMSCs cultured on BG-based scaffolds for 7 days; The relative gene expression levels of (e) BSP and (f) OCN of SD-BMSCs cultured on BG-based scaffolds for 21 days; (g) Effects of BG, HBG, BGX and HBGX scaffolds extracts on calcium nodule formation of SD-BMSCs after 21 d of culture. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Osteochondral regeneration in vivo

To further study the effect of the biphasic scaffolds on osteochondral regeneration in vivo, the biphasic scaffolds were implanted into the lower femoral osteochondral defect of SD rats for up to 10 weeks. The HAMA + TGF-β/HBGX + BMP2 (HT/HBGXB) group was used as the positive control (Fig. 7a). Micro-CT and 3D reconstructed data were used to evaluate the role of the biphasic scaffolds in promoting bone regeneration. As shown in Fig. 7b, the micro-CT reconstruction images showed that new bone was formed in all groups implanted with biphasic scaffolds, while there was still a large cavity in the defect region of the blank control group. And the new bone tissue tightly connected to the periphery and showed a tendency to grow from the edge to the center, indicating that new bone could grow into the interconnected porous scaffolds. Furthermore, quantitative data showed that similar bone volume fraction (BV/TV), bone mineral density (BMD), bone trabecular number (Tb.N) was observed in all groups although lowest bone trabecular separation (Tb.Sp) was detected in HP/HBGX group (Fig. 7c–f). In contrast to the in vitro cell results, HBG scaffolds played the main role in promoting subchondral bone regeneration, while the addition of XLS into BG scaffolds did not provide significant advantages in new bone formation, which may be due to the fact that the content of XLS in BG scaffold is only 3 wt%. Additionally, the composition of the upper layer cartilage also showed no significant effects on subchondral bone repair.

Fig. 7. The repair efficacy of osteochondral defects by HPI/HBGX biphasic scaffolds.

a Photographs of surgical process of HPI/HBGX biphasic scaffolds implanted into the lower femur of SD rats; (b) Representative micro-CT images of subchondral bone repair after 10 weeks of treatment; (c–f) Quantification results of micro-CT images. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001; scale bar = 1 mm.

The upper layer cartilage repair was evaluated by hematoxylin and eosin (H&E) and Safranin-O (Saf-O) staining (Fig. 8a, b). Smaller cartilage lacuna was observed in the HPI group compared to other groups. Notably, new bone was found in the deep layer of the defect region in the H/HBG and HP/HBGX groups, while less subchondral bone was observed in the H/HBGX and HPI/HBGX groups. According to Fig. 8c, the HPI/HBGX group had the best cartilage repair capacities, even better than the HT/HBGXB group (positive control), and the HP/HBGX group showed similar cartilage repair ability compared with the HT/HBGXB group. To study the mechanism of HP hydrogel on cartilage repair, here, Raw 264.7 cells were co-cultured with ZIF8 and PZIF8 NPs in an LPS-induced inflammatory microenvironment, and our data revealed that PZIF8 significantly promoted TGF-β3 expression after 1 day of treatment, while ZIF8 NPs had no such effects (Fig. 8d and S18). These results were also confirmed by the western blot analysis data (Fig. 8f, h). Similarly, under the LPS-induced condition, the HP and HPI groups showed higher TGF-β3 promoting effects than H and HI hydrogels due to the presence of PDA in PZIF8 NPs (Fig. 8e, g, h). These data indicated that PDA could induce the expression of TGF-β3, which was consistent with the previous studies^32,33. TGF-β3 is an important chondrogenic growth factor that can activate TGF-β type I receptor (TβR-I), further promote Smad 2 and Smad 3 phosphorylation and transport to the nucleus, finally facilitating the expression of Sox-9 and Col-2α^34,35. Besides the Smad 2/3 pathway, MAPK pathway is also involved in the TGF-β3-induced Sox-9 and Col-2α expressions, and these two pathways together contributes to the upregulation of aggrecan^36,37. In conclusion, our results demonstrated that TGF-β3 secreted by M2 macrophages induced by PZIF8 NPs together with IGF-1 play synergistic effects on the chondrocyte differentiation (Sox-9) of BMSCs and cartilage extracellular matrix formation (Col-2α, aggrecan), finally efficiently accelerating the regeneration of cartilage.

Fig. 8. The repair efficacy of osteochondral defects by HPI/HBGX biphasic scaffolds.

a H&E tissue staining of biphasic scaffolds in each group after 10 weeks of treatment; (b) Safranine O tissue staining of biphasic scaffolds in each group after 10 weeks of treatment; (c) Quantification results of Safranine O tissue staining images; (d) The relative gene expression levels of TGF-β3 of Raw 264.7 cells cultured with PZIF8 NPs after 1 day; (e) The relative gene expression levels of TGF-β3 of Raw 264.7 cells cultured on different group hydrogel after 1 day; (f) Western blot analysis of the protein expression level of TGF-β3 in Raw264.7 cells after cultured with ZIF8 and PZIF8 NPs for 1 day; (g) Western blot analysis of the protein expression level of TGF-β3 in Raw264.7 cells after cultured on different group of hydrogels for 1 day; (h-i) Quantification results of the TGF-β3 protein levels determined by western blot; (j) Schematic illustration of PZIF8-IGF-1 NPs effects on the BMSCs chondrogenesis and ECM formation. Data are expressed as mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001; scale bar = 1 mm.

Discussion

The regeneration of osteochondral tissue remains a formidable challenge in orthopedics due to its complex hierarchical structure, pronounced mechanical heterogeneity, and the distinct biological microenvironments required for cartilage and bone repair. This study successfully designed and evaluated a biomimetic biphasic scaffold that addresses these multifaceted challenges through a combination of immunomodulatory cues, sustained growth factor release, and enhanced mechanical properties, demonstrating significant efficacy in promoting osteochondral regeneration both in vitro and in vivo.

A central finding of this work is the critical role of immunomodulation in creating a pro-regenerative microenvironment. The inflammatory response following injury is a major impediment to cartilage repair, often leading to chondrocyte apoptosis, matrix degradation, and failed integration of repair tissue^4,26,38,39. Our results demonstrate that the polydopamine-functionalized ZIF8 nanoparticles (PZIF8 NPs) within the cartilage layer effectively polarized macrophages from a pro-inflammatory (M1) to an anti-inflammatory (M2) phenotype. This shift was evidenced by the downregulation of M1 markers (IL-1β, IL-6, TNF-α) and the upregulation of M2 markers (IL-4, IL-10, Arg-1) in Raw 264.7 cells. This is a crucial advantage, as M2 macrophages not only mitigate destructive inflammation but also contribute to tissue repair by secreting beneficial cytokines^40,41. Importantly, we identified a novel pro-regenerative pathway by which PZIF8 NPs, specifically the polydopamine component, stimulated the secretion of TGF-β3 from macrophages. TGF-β3 is a potent chondrogenic inducer that activates the Smad and MAPK pathways, leading to the upregulation of key chondrogenic markers like Sox-9 and Col-2α1^34–37. This creates a powerful positive feedback loop where the scaffold’s immunomodulatory action directly enhances the chondrogenic differentiation of recruited BMSCs, synergizing with the sustained release of IGF-1.

The sustained and controlled delivery of IGF-1 from the PZIF8/HAMA (HPI) hydrogel was another cornerstone of our strategy. IGF-1 is a master regulator of cartilage anabolism, promoting BMSC migration, proliferation, and chondrogenesis^11–14. However, its clinical application is limited by a short half-life and rapid degradation in inflamed joints¹⁵. Our data confirm that the mesoporous structure of PZIF8 NPs effectively protected IGF-1 and facilitated a sustained release profile, overcoming the initial burst release observed with IGF-1 directly loaded into HAMA hydrogel. This controlled release was functionally critical, enabling potent BMSC chemotaxis to the defect site and providing sustained chondrogenic differentiation signals. The in vitro results clearly showed that the HPI hydrogel significantly enhanced SD-BMSC migration, upregulated anabolic genes (Aggrecan, Col-2α1, Sox-9), and suppressed catabolic genes (MMP13, Adamts4/5), confirming a shift in metabolic balance towards matrix synthesis and deposition.

For the subchondral bone layer, the incorporation of XLS into the bioglass scaffold (BGX) yielded two significant benefits. First, it provided a dramatic improvement in mechanical strength, with the compressive strength increasing from 0.11 MPa (BG) to 1.03 MPa after XLS incorporation and HAMA coating (HBGX). This enhanced mechanical property is essential for withstanding physiological loads in the subchondral bone region and providing stable mechanical support for the overlying cartilage. Second, the degradation products of XLS (SiO₄⁴⁻, Li⁺, Mg²⁺) exerted potent osteoinductive effects, as demonstrated by increased ALP activity, upregulation of osteogenic genes (COL-1α1, BSP, OCN), and enhanced biomineralization of SD-BMSCs. The HAMA coating on the bone scaffold (HBG, HBGX) further improved cell viability and provided a more favorable microenvironment for osteogenesis, potentially mimicking the endochondral ossification process.

The integration of the two layers via semi-infiltration of the HAMA hydrogel created a stable, hybrid interface, which is vital for preventing delamination under load and facilitating the formation of a continuous osteochondral unit. The in vivo results in a rat femoral defect model validated the overall efficacy of this biphasic design. While the BGX scaffold showed robust osteogenic potential in vitro, the in vivo bone regeneration was primarily driven by the HAMA-modified scaffolds (HBG, HBGX), suggesting that the in vivo microenvironment and improved cell-scaffold interactions offered by HAMA are critical. Most importantly, the HPI/HBGX group exhibited the most superior cartilage repair, with hyaline-like cartilage, abundant glycosaminoglycan deposition, and well-integrated subchondral bone. Its performance even surpassed the positive control (TGF-β3/BMP-2 loaded scaffold), highlighting the advantage of our strategy that leverages endogenous cell recruitment and immunomodulation over the administration of high doses of exogenous growth factors.

In conclusion, this study presents a comprehensive and innovative strategy for osteochondral regeneration. By engineering a biphasic scaffold that mimics the native tissue structure and function, we successfully overcame the key challenges of immunomodulation, sustained biochemical cue presentation, and mechanical support. The PZIF8-based cartilage layer not only delivers IGF-1 in a sustained manner but also actively remodels the injury microenvironment towards an anti-inflammatory and pro-chondrogenic state, partly through the induction of TGF-β3. The XLS-functionalized bone layer provides strong mechanical support and actively promotes osteogenesis. The synergistic effect of these two layers leads to the simultaneous and integrated regeneration of both cartilage and bone, making this biphasic scaffold a highly promising candidate for clinical translation in the treatment of osteochondral defects and osteoarthritis. Building on these promising results, the study also defines a clear pathway for further development. While the robust structural and functional validation provided here is a critical foundation, certain limitations point directly toward essential future research. A more detailed biochemical and molecular characterization—including immunohistochemistry for collagen types II and X, along with transcriptomic analysis—would precisely define the phenotype, maturation, and hypertrophy risk of the regenerated cartilage. To address these points, future work will (1) extend in vivo observation to assess long-term repair stability; (2) evaluate the scaffold in a large animal model (e.g., porcine) to better gauge translational potential; and (3) employ transcriptomic profiling to elucidate the molecular mechanisms governing the osteochondral regeneration observed.

Methods

Materials

Zinc acetate dihydrate (C₄H₆O₄Zn·2H₂O), 2-methylimidazole (HMIM), methacrylic anhydride (MA), methanol and dimethyl carbonate (DMC) were purchased from Macklin Biochemical Co. (Shanghai, China). BSA was obtained from Beyotime Biotechnology (Shanghai, China). 45S5 Bioactive glass was purchased from Dingan Tec (Suzhou, China). XLS NPs were purchased from Nanocor (Beijing, China). 2-Hydroxy-4’-(2-hydroxyethoxy)-2-methylpropiophenone was purchased from YANGFAN NEW MATERIALS (Zhejiang, China).

PZIF8 NPs synthesis

The preparation of PZIF8 was as previous reported with some modifications²¹. Briefly, 0.1 M C₄H₆O₄Zn·2H₂O and 1.6 M HMIM were dissolved in 100 mL of Tris-HCl buffer with pH = 8.5, and 50 mg of 3-Hydroxytyramine hydrochloride was added after stirring for 1 min. After stirring for another 2 h, the mixture was collected, washed 3 times, and the pellet was collected. After freeze-drying to obtain PDA@ZIF8 (PZIF8) NPs.

Preparation of IGF-1 loaded PZIF8 (PZIF8@IGF-1) NPs

In the preparation of PZIF8 NPs, after adding 3-Hydroxytyramine hydrochloride for 1 min, IGF-1 solution (8.25 μg/mL) was slowly added. After stirring for 2 h, the upper solution was collected by centrifugation and measured the volume; followed by the same procedure as that used for the PZIF8 NPs. The loading efficiency was calculated by the Eq. (1).

$$\mathit{Loading}\mathit{Efficiency}=\frac{\mathit{Weight}\mathit{of}\mathit{loaded}\mathit{protien}}{\mathit{Weight}\mathit{of}\mathit{initially}\mathit{added}\mathit{protien}}\times 100\%$$ 1

Preparation of HAMA/PZIF8@IGF-1 hydrogel

1 g of hyaluronic acid (HA) was dissolved in 100 mL of ultrapure water, 5 mL of methacrylic anhydride was added and the pH was adjusted to 8, and stirred at 0 °C for 24 h. After that, ice ethanol was used to pellet the mixture and centrifuge to collect the pellet. This pellet was dialyzed in DI water for 3 d. Finally, the resultant solution was frozen at −80 °C overnight, followed by freeze drying for 3 d to obtain Hyaluronic acid methacryloyl (HAMA) powder.

To prepare HAMA hydrogel, HAMA was dissolved in DI water to make a 1.5% (w/v) solution. Then photoinitiator 2959 solution (0.5% (w/v)) was added and stirred for 2 h at 50 °C. The mixed solution was transferred to the mold and cross-linked under ultraviolet (UV) light of 365 nm for 5 min to obtain the HAMA hydrogel.

To prepare HAMA/PZIF8@IGF-1 hydrogel, PZIF8@IGF-1 NPs were added in the above prepared HAMA solution, and stirred for 2 h, followed by the same photo-crosslinking procedure as that used for the HAMA hydrogel.

Preparation of 45S5 BG and BG-XLS (BGX) scaffolds

The 45S5 BG scaffolds were prepared as previous reported²⁹. Briefly, 0.4 g of polycaprolactone (PCL) was dissolved in 10 mL of dimethyl carbonate (DMC), and then 4 g of 45S5 BG powder was added and stirred for uniform dispersion, the polyurethane film (PU) was soaked in the above 45S5 ink, and then the excess ink was extruded and dried overnight in an oven at 60°C. Repeating the above procedure twice. Finally, the 45S5 BG scaffold was sintered by gradient heating to 1100 °C. The preparation process of BGX scaffold was the same as that of BG scaffold, except that 3 wt% XLS was added to the 45S5 BG powder to prepare the BGX scaffolds.

Preparation of HAMA/BG (HBG) and HAMA/BGX (HBGX) scaffolds

The BG and BGX scaffolds were merged in 1% HAMA solution and cross-linked under UV light at 365 nm for 5 min. The cross-linked scaffolds were dried overnight at 60 °C to obtain HBG and HBGX scaffolds.

Physicochemical characterization

The morphologies of PZIF8 NPs were observed on a transmission electron microscope JEM-1200 (JOEL, Japan) at 120 kV acceleration voltage. The particle sizes and zeta potentials of the prepared PZIF8 NPs were obtained by a Malvern Zetasizer (NanoZS90, Westborough, MA). FTIR spectra of PZIF8 NPs were performed on a Nicolet iS-50 FTIR Spectrometer (Thermo Scientific, USA) and collected in a mid-IR range (4000–400 cm⁻¹). To detect the degree of substitution of HAMA, proton nuclear magnetic resonance (¹H NMR, Qone AS400.) measurements were performed. The morphologies of HAMA-based hydrogels and BG-based scaffolds were observed by a scanning electron microscope S-3000N/H (Hitachi, Japan).

In vitro swelling experiments

Hydrogels were prepared in a circular mold with a diameter of 10 mm and a height of 2 mm. Weigh the hydrogels under dry conditions, recorded as W_d. The hydrogels were then merged in DPBS, and replaced the DPBS every 24 h. At each time point (2 h, 4 h, 6 h, 8 h, 24 h, 2 d, 3 d, 5 d, 7 d), the hydrogel weight was weighed and recorded as Ws after removing the excess solution. The swelling rate of the hydrogel was calculated according to the Eqs. (2), and 3 samples are measured in each group.

$$\mathit{ESRs}=\frac{\mathit{Ws}-\mathit{Wd}}{\mathit{Wd}}\times 100\%$$ 2

In vitro degradation experiments

Hydrogels were prepared in a circular mold with a diameter of 10 mm and a height of 2 mm. The prepared hydrogels were merged in DPBS, and DPBS was replaced every 24 h. The weight of the hydrogel was recorded as W₀ under dry condition, and the weight was weighed at 0 h, 2 h, 4 h, 6 h, 8 h, 24 h, 2 d, 3 d, 5 d, 7 d, 14 d, 21 d, 28 d, and recorded as W. The degradation rate of the hydrogel was calculated according to the Eqs. (3), and 3 samples are measured in each group as averaged.

$$Ds=\left(1-\frac{W}{W0}\right)\times 100\%$$ 3

Mechanical properties of BG-based scaffolds

The BG-based scaffolds were prepared according to method in 2.5, with the size of 0.7 × 0.7 × 0.7 cm³. The scaffolds were tested using a dynamic thermo mechanical analyzer (TA DMA Q850) in the compression mode, and 10 scaffolds were tested for each group.

In vitro IGF-1 release

PZIF8@IGF-1 and IGF-1 were added into HAMA hydrogel, and the input amount of IGF-1 was determined according to the loaded content of IGF-1 in the PZIF8@IGF-1 NPs. HAMA/PZIF8@IGF-1 and HAMA/IGF-1 hydrogels were prepared in circular molds with a diameter of 10 mm and a height of 2 mm, and 3 samples were prepared per group; The hydrogel was merged in 2 mL DPBS, shaker at 37 °C, 90 rpm. The samples were taken at 1 h, 4 h, 8 h, 1 d, 3 d, 7 d, and 14 d, respectively. The absorbance of the collected sample was measured by the IGF-1 kit to calculate the release of IGF-1 and obtain the in vitro release curve of IGF-1.

Preparation of scaffolds extracts solution

BG, HBG, BGX and HBGX scaffolds were immersed in SD-BMSCs complete medium at a ratio of 1 mL of medium per 10 mg of scaffolds, and the mixture was placed on a constant temperature shaker at 37 °C for 90 rpm. The same method was used to place the medium without the immersed scaffolds on the shaker. The culture media were collected after 24 h of immersion and used for cell culture. (Scheme 1).

Scheme 1.

Biphasic multifunctional scaffold promotes osteochondral defects repair by releasing IGF-1 and regulating the immune microenvironment.

In vitro cell culture

Mouse macrophages cell line (Raw 264.7) was purchased from Procell Life Science & Technology Company (Wuhan, China). SD-BMSC was purchased from OriCell (Shanghai, China). Raw 246.7 cells were cultured with DMEM high-glucose. SD-BMSCs were cultured with OriCell^® Complete culture medium of SD-BMSC. All medium supplemented with 10% FBS and 1% penicillin/streptomycin (P/S, Beyotime). Moreover, all cells were cultured in a humid incubator at 37 °C with CO2 level of 5%.

Cell cytotoxicity

The cytotoxicity of the PZIF8 and PZIF8@IGF-1 NPs were measured by the CCK8 assay (Dojindo Molecular Technologies, Inc., Japan). Firstly, 1 × 10⁴ SD-BMSCs were seeded into 24-well plates, and after 1 d of culture, PZIF8 and PZIF8@IGF-1 NPs were added. After 1 and 3 d of culture, washed 2 times with DPBS to remove the residual NPs. CCK8 containing culture medium was added and cultured for 30 min at 37 °C, the supernatant of plate was transferred to a 96-well plate and measured the absorbance value at 450 nm. Raw 264.7 cells were evaluated using the same methodology. 200 μL/well of hydrogels were pre-laid in the 24-well plates, and 2 × 10⁴ cells/well was seeded onto the hydrogels, followed by the same treated procedure as NPs.

The cytotoxicity of the scaffolds was studied by scaffold extracts, followed by the same cell culture procedure as that used for the NPs.

Live/dead cell staining

To visualize the proliferative effect of IGF-1 on SD-BMSC and the toxicities of NPs and hydrogels, cell live/dead staining was used for evaluation. 5 × 10³ SD-BMSCs were seeded into 24-well plate and incubated overnight. After that IGF-1 was added. After cultured for 1 and 3 days, the supernatant was removed. Calcein AM and PI were mixed with serum-free culture medium in a certain proportion was added. The plates were placed at 37 °C in darkness for 15 min, cells morphologies were observed by DMi8 fluorescence microscopy. The same procedure of SD-BMSCs on PZIF8, PZIF8@IGF-1 NPs were conducted as the IGF-1. 1 × 10⁵ cells were seeded onto the pre-laid hydrogels in 24-well plates, Calcein AM/PI mixed solution were then added and culture for 15 min at 37 °C.

In vitro cell migration

The ibidi scratch was inserted into the 24-well plate, 5 × 10⁴ SD-BMSCs were seeded into the insert, then cultured for 1 day. To eliminate the effect of cell proliferation on migration, the cells were starved for 2 h using serum free medium before seeding. Then, the insert was removed. Observe cell migration under DMi8 fluorescence microscopy and take pictures at 0 h, 8 h and 24 h. Finally, the cell migration capacity was analyzed by Image J software (Image-win64, USA).

Transwell invasion assay

Store the matrix glue, chamber, orifice plate and tips in a 4 °C freezer overnight. Cold serum-free medium was mixed with Matrigel matrix (Corning, USA), and dropwise added to the upper side of the chamber, placed in the cell culture incubator for 2 h, and waited for the Matrigel to condense. Gently rinsed the Matrigel using the serum-free medium, carefully removed unbounded Matrigel. Using serum-free medium hydrated matrix gel containing 10 mg/mL BSA in a 37 °C incubator for 30 min, washed the collected SD-BMSCs using DPBS and resuspend with serum-free medium containing 10 mg/mL BSA for later use. 5 × 10⁴ SD-BMSCs are seeded in the chamber, while IGF-1 containing culture medium was added to the lower chamber and placed in the 37 °C incubator for 1 day. The uninvaded cells in the upper chamber were wiped off, the cells on the lower side of the small chamber were fixed using ethanol, washed with DI water, and air-dried completely. The invaded cells were stained using crystalline violet staining solution, washed with DI water, placed under a microscope for observation, and randomly photographed and recorded. The invasion ability of cells was analyzed by ImageJ 1.54p java 1.8.0_172(64 bit) software.

Immunomodulatory effects of PZIF8 NPs

The anti-inflammation activity of PZIF8 NPs were evaluated through RT-PCR assay as previous reported⁴². Lipopolysaccharide (LPS) acts as the strongest inducer of M1-type polarization in macrophages. Raw 264.7 cells were treated with LPS and then co-cultured with PZIF8 for 1 day. Total RNA was extracted from macrophages using the GeneJET™ RNA Purifcation Kit and reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcript kit according to the manufacturer’s instructions. Expression of anti-inflammatory related genes marker (IL-4, IL-10, Arg-1, TGF-β3) and pro-inflammatory related genes marker (IL-1β, IL-6, TNF-α) was quantified using RT-PCR. β-actin was the housekeeper gene. The relevant primer information is shown in Table S1.

Chondrogenic differentiation of SD-BMSCs on hydrogels

To investigate the chondrogenic differentiation of SD-BMSCs on hydrogels, toluidine blue and Safranin-O staining were used. SD-BMSCs (5 × 10⁴ cells/well) were seeded into 24-well plates and treated with hydrogel extracts for 1 d. Culture media were removed after 7 or 14 days, washed with DPBS, and then fixed with 4% paraformaldehyde for 30 min, and 0.5% toluidine blue staining solution or saffron O staining solution was added and stained for another 20 min. After washed with DPBS several times, cells were observed under the microscope.

To investigate the chondrogenic differentiation effect of prepared hydrogels on SD-BMSCs, the chondrogenic expression-related proteins was also performed. Cells were lysed by RIPA lysis solution and the cell suspensions were treated with protein loading buffer. Protein separation was performed in 10% SDS-PAGE, followed by electrophoretic separation of polyacrylamide gels electro-transferred to polyvinylidene fluoride membranes. The antibodies were blocked for 2 h at room temperature, using 5% milk incubated overnight at 4 °C, followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibody and finally visualized by fully automated chemiluminescence imager (Fluachem E) for presentation. The relevant antibodies are as follows: Rabbit Anti-Collagen Ⅱ Polyclonal (Bioss, China), MMP13 Rabbit Polyclonal antibody (Proteintech, USA), Anti-ADAMTS5 antibody ab41037 (Abcam, USA), Anti-SOX9 antibody [EPR14335-78] ab185966 (Abcam, USA), Anti-TGF beta 3 antibody [EPR27093-72] (ab308300).

To further investigate the effect of the prepared hydrogels on SD-BMSCs to promote cartilage differentiation, the chondrogenic synthesis-related genes marker (Sox-9, Aggrecan, Col-2α1), and chondrolysis-related genes marker (MMP13, Adamts4, Adamts5) were also investigated. The relevant gene primers are shown in Table S2.

Osteogenic differentiation of SD-BMSCs on scaffolds

To study the osteogenic differentiation of SD-BMSCs on scaffolds, we evaluated the relative activity of ALP. SD-BMSCs (1 × 10⁴ cells/well) were seeded in 24-well plates and cultured for 7 d using scaffold extract solution. ALP assay was assayed using the standard instructions of the ALP assay kit (Anorun, Beijing, China), and total protein content was calculated using the BCA protein kit (Thermo Fisher, USA). To visualize the expression of ALP, we used ALP staining kit (Sigma, USA) and performed according to the manufacturer’s instructions.

The biomineralization effect of scaffold extracts on SD-BMSCs was also evaluated by alizarin red staining. SD-BMSCs (1 × 10⁴ cells/well) were seeded in 24-well plates and cultured using scaffold extract for 21 d. Cells were fixed with 4% paraformaldehyde for 30 min and stained with alizarin red staining solution for 15 min, washed with DI water to remove unstained alizarin red, observed under microscope and photographed.

To further investigate the effect of the prepared hydrogels on SD-BMSCs to promote osteogenic differentiation, osteogenic-related genes markers (BSP, Col-1α1) were also investigated. The same procedure was conducted as those mentioned in the promotion of chondrogenic differentiation. The relevant primer stock numbers are as follows: GAPDH (Mm99999915_g1, Thermo Fisher), BSP (Mm00436767_m1, Thermo Fisher), OCN (Mm03413826_mH, Thermo Fisher), Col-1α1 (Mm00801666_g1, Thermo Fisher).

In vivo osteochondral regeneration

Care and use of the laboratory animals followed the protocol approved by the Institutional Animal Care and Use Committee of Wenzhou Medical University (China, xmsq2023-0174). Male Sprague-Dawley (SD) rats (6 weeks old) were housed in the Animal Experiment Center of Ophthalmology& Optometry, eye Hospital affiliated to Wenzhou medical University. After one week of adaptive culture, SD rat anesthetized with isoflurane gas, the 2.5 mm diameter osteochondral defect was created in the lower right femur of SD rat by dental instruments. Following in vitro preparation of the scaffolds (Methods 2.3–2.6), we fabricated biphasic scaffolds matching the rat wound dimensions using a 2.5 mm diameter hole punch. After complete hemostasis, different biphasic scaffolds were implanted into the joint defect, the wound was closed carefully, the muscle tissue was closed with 5-0 absorbable sutures, and the skin was closed with 5-0 mousse sutures. A total of 12 rats were used in our animal experiments, and they were divided into 4 groups (n = 3). Rats were sacrificed after 10 weeks, anesthesia with isoflurane gas, euthanized by removing their necks and fresh right knee joint tissues were collected. Intact joints isolated from each rat were fixed with 4% paraformaldehyde for subsequent analysis.

Micro-CT analysis

The rat paws scanned by a micro-CT scanner (#Skyscan1175, Aartselaar, Belgium) using an X-ray energy of 380 μA/65 kV at a resolution of 18 μm. The scanned original images were reconstructed and converted into REC files and processed by CT-vox software to obtain 3D images and analysis results.

Histological testing and analysis

Rat posterior joints were fixed with 4% paraformaldehyde for 2 days, fully decalcified in EDTA, embedded in paraffin, and cut into 4 μm thick sections. For the observation of tissue morphology, randomly selected sections were stained with H&E and Safranin O/Fast Green for structural analysis.

Data Analysis

Before statistical comparison, the normality of data distribution was assessed using the Shapiro-Wilk test. The One-way ANOVA with Tukey’s correction (compare among groups) and Student’s t tests (between two groups) were used to calculate statistical significance, statistical significance was denoted as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. All the data are presented as the mean ± standard deviation (SD).

Author contributions

The following are the individual contributions of each author: Hao Yu, Wei Wang and Hongning Wang contribute to the Methodology, Software, Formal analysis and Writing - Original Draft; Wenyi Zhang and Yan Zheng contribute to the Investigation and Data Curation; Luya Chen contributes to the Validation and Formal analysis; Shenbin Huang contributes to the Project administration, Funding acquisition and Writing-Review & Editing; Wentao Yan contributes to the Supervision and Writing-Review & Editing; Qingqing Yao contributes to the Conceptualization, Supervision and Writing-Review & Editing.

Data availability

All data are available in the main text or the supplementary materials. The raw datasets used and/or analyzed during the current study are available from the corresponding author.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41536-026-00463-0.