Immunomodulatory cytoskeleton of bionic periosteum affects mitochondrial dynamics to promote osteogenesis

Abstract

Purpose

Periosteum is a key regulatory hub in the immune microenvironment of bone repair and the mitochondrial dynamics of bone marrow mesenchymal stem cells (BMSCs) directly determine the osteogenic efficiency; therefore, how to promote bone healing through the immunomodulation of mitochondrial dynamics is a key challenge.

Methods

Inspired by the natural periosteum, we designed a bionic periosteum that integrates topology and controlled Semaphorin 3 A (Sema3A) release. Hyaluronic acid-coated Sema3A granular microsols spontaneously formed a nucleus-sheath structure under a high-voltage electric field to facilitate efficient drug loading.

Results

The periosteum controlled Sema3A release and oriented topography, which drived macrophage M2 polarization via the PI3K/Akt/mTOR pathway, reconstructed the BMSC skeleton via the ROCK2 pathway, and synergistically promoted bone formation by driving mitochondrial fusion through material-appropriate stiffness. In vivo experiments further demonstrated that this biomimetic periosteum efficiently repaired bone defects by activating the immune microenvironment to remodel the BMSC skeleton and modulating mitochondrial dynamics, an endogenous repair cascade.

Conclusion

This immunomodulatory drug delivery strategy utilizing bionic periosteum offers a novel approach for regenerative medicine to regulate immune–osteogenic crosstalk, which is promising for potential applications.

Graphical abstract

This study developed a functionalized material system with biomimetic properties of natural periosteum. This design achieves dual objectives of immune microenvironment remodeling and enhanced osteogenic potency through the integration of oriented topological structures with Sema3A bioactivity. It specifically elucidates the regulatory role of macrophage polarization in modulating the cytoskeletal-mitochondrial kinetic axis during this process. The study systematically deciphers the molecular pathway by which immune signaling influences mitochondrial function and promotes osteogenic differentiation via F-actin cytoskeletal remodeling, thereby clarifying the core regulatory mechanism of the immune -cytoskeleton-mitochondrial kinetic axis in bone regeneration. Employing a progressive experimental design, the study first constructed a biomimetic scaffold with oriented fiber architecture via micro-solution electrospinning, enabling controlled Sema3A release, and systematically characterized its physicochemical properties. Subsequently, a co-culture system optimized bioactive factor concentration gradients to assess material biocompatibility. Conditioned medium simulated the immune microenvironment to decipher how the scaffold promotes bone regeneration by temporally regulating macrophage polarization, inducing BMSC scaffold remodeling, and modulating mitochondrial homeostasis. Finally, the scaffold function was validated in a rat cranial defect model. Early-stage studies examined its immunomodulatory properties and effects on the cytoskeleton-mitochondrial dynamics axis, while late-stage evaluations assessed its bone tissue repair and periosteal regeneration capabilities. This research systematically evaluated the efficacy and mechanisms of AMS@Sema3A in promoting bone regeneration through immune regulation from both macroscopic and microscopic perspectives. Highly consistent in vivo and in vitro results confirm the material's clear clinical application value.

Highlights

• •Bionic periosteum promotes efficient bone tissue regeneration by reshaping the local immune microenvironment and regulating the interaction between the cytoskeleton and mitochondrial dynamics of BMSCs.
• •Bionic periosteum modulates the PI3K/Akt/mTOR pathway through its topological structure in synergy with Sema3A, inducing M2 polarization of macrophages while suppressing M1 polarization. This promotes osteogenesis and alleviates inflammation, thereby sequentially optimizing the immune microenvironment.
• •Biomimetic periosteum activates the RhoA/ROCK2 pathway, promotes cytoskeletal remodeling in bone marrow mesenchymal stem cells (BMSCs), and regulates mitochondrial dynamics during osteogenic differentiation by promoting mitochondrial fusion and inhibiting mitochondrial fission. This enhances bone healing efficiency, endowing the biomimetic periosteum with osteogenic efficacy comparable to that of autologous periosteum.

1. Introduction

The rapid and effective healing of large and difficult-to-heal bone defects, particularly those of critical dimensions, remains a major challenge in clinical medicine [1]. Periosteum can deliver cells and bioactive molecules, and is an indispensable functional unit in bone repair and regeneration [2,3]. Although autologous periosteal grafts are the gold standard for periosteal reconstruction, they are limited by donor-area damage, source scarcity, and remodeling difficulties, while allogeneic/xenografts carry risks of immune rejection and disease transmission. Till date, periosteal tissue engineering has mostly focused on bone repair; however, problems such as induced infections and tissue necrosis have hindered its clinical application [4]. Unlike the complex natural periosteum, artificial periosteum tends to focus on certain functions, such as osteogenesis or vascularization, while neglecting other important functions [5] (see Scheme 1)

Scheme 1.

Preparation and application of biomimetic periosteum. (A) Synthesis of micro-sol gel electrospun fiber membrane (MS) and Sema3A-loaded oriented micro-sol gel electrospun biomimetic periosteum (AMS@Sema3A). (B) The bionic periosteum controls the release of Sema3A to modulate the immune microenvironment in a timely manner, promoting macrophage M2 polarization to obtain an anti-inflammatory microenvironment, rearranging the cytoskeleton, and promoting mitochondrial fusion to enhance osteogenic differentiation and promote bone formation. (C) Schematic diagram of the construction of an SD rat skull defect model, implantation of the bionic periosteum, and subsequent regulation of immunity, cytoskeleton, and promotion of bone regeneration.

During bone regeneration, the periosteum serves as a vital regulator of immunomodulation and osteogenesis [6]. As primary immune response initiators, macrophages play an important role in the specific activation of anti-inflammatory responses and osteogenesis [7]. The timely conversion of pro-inflammatory M1 macrophages to pro-regenerative M2 macrophages in the early stages of injury is essential for effective regeneration [8]. Although biomaterials that promote osteogenesis through immune modulation have been extensively studied, the role of mitochondrial dynamics in immune-promoted osteogenic effects is unclear [[9], [10], [11]].

Mitochondria are not only central organelles for metabolism and energy conversion, but also platforms for cell signaling cascades [12]. Mitochondrial dynamics, which refer to adaptable changes in its morphology and subcellular distribution, encompass mitochondrial fusion and disassembly/disassembly and ultrastructural remodeling of the membrane. These processes can influence and regulate metabolism and participate in complex cellular signaling [13]. Fusion promotes mitochondrial matrix and material exchange to maintain function, while fission mediates quality control by removing damaged mitochondria and triggers selective apoptosis in response to stressful cellular pressures [14]. The dynamic processes of mitochondrial fission and fusion are tightly regulated, which determine mitochondrial shape and influence mitochondrial function [15]. Precise seed cell growth and differentiation regulation is crucial for enhancing the efficacy of bone tissue engineering, and mesenchymal stem cells (MSCs) are the preferred source owing to their osteogenic advantages [16]. Mitochondrial dynamics play a crucial role in BMSC metabolism and differentiation, whereas disturbances in BMSC differentiation process lead to osteoporosis [17]. High-molecular-weight polyacrylic acid (HPAA)-crosslinked collagen membranes promote the osteogenic differentiation of MSCs by promoting mitochondrial fusion and inhibiting mitochondrial fission. Mitochondrial fusion, stimulated by energy demand, induces interconnectional remodeling of the mitochondrial network to differentiate MSCs towards osteoblasts, whereas fission is mediated by, for example, dynamin-related protein 1 (Drp1), which produces smaller, discrete mitochondria [15,18]. Jia Liu et al. have found that mitochondrial fusion occurs early in the osteogenic differentiation of BMSCs [17]. Graphene oxide quantum dots can promote the osteogenic differentiation of human periodontal ligament stem cells (hPDLSCs) by promoting mitochondrial fusion, inhibiting fission, and regulating mitochondrial dynamics, ultimately promoting in vivo bone regeneration [19]. During orthodontic tooth movement (OTM), semaphorin 3 A (Sema3A) promotes the osteogenic differentiation of hPDLSCs through mitochondrial fusion [20]. Till date, current bone engineering research has primarily studied the effect of biomaterials on phenotype and nuclear genotype, while few studies have explored the role of mitochondria in this context [21].

Owing to the clinical needs of bone defect repair and challenges of periosteal tissue engineering, the development of composites with efficient bone regeneration functions has become a core scientific problem. Electrospun silk, which has been inspired by the natural periosteum, has a large specific surface area and high porosity, and is often used as a matrix of bionic periosteum to promote fracture repair [7]. Microsol electrostatic spinning, which spontaneously forms a core-shell structure under a high-voltage electric field, on the contrary, possesses high drug-carrying capacity, slow drug release, and mechanical stability, and oriented spinning has the topographical advantage of regulating macrophage polarization to establish a pro-regenerative immune environment and promote osteogenesis [[22], [23], [24], [25], [26]]. M2 macrophages release anti-inflammatory factors, such as interleukin (IL)-10, and play a key role in various endogenous pathways [25,27]. Elevated extracellular matrix stiffness upregulates Mitofusin1/2 (Mfn1/2) expression and inhibits DRP1 activity, driving the osteogenic differentiation of BMSCs by promoting mitochondrial fusion [28].

As a signaling molecule that inhibits axon extension and growth cone activity, both sensory-neural and exogenous Sema3A promotes osteogenesis and inhibits osteoclastic resorption, exhibiting a dual osteoprotective function [29,30]. In periodontitis, Sema3A modulates the immune response by inducing M1 to M2 macrophage polarization and reducing inflammatory factor expression [31,32]. In addition, Sema3A inhibits mechanical overload-induced F-actin overstretching and maintains mitochondrial fusion through the ROCK2 pathway, thereby reversing the mechanical [20]. In conclusion, Sema3A is a multifunctional osteoprotective bioactive factor involved in immune and mitochondrial regulation.

The bone healing process is complex and sophisticated, and the immune microenvironment, cytoskeleton, and mitochondrial dynamics play important roles in osteogenic differentiation, but are rarely utilized in periosteal tissue engineering presently. Therefore, we constructed an immunomodulatory bionic periosteum that synergizes physical and biological advantages for efficient osteogenesis using Sema3A and topological topography to promote macrophage M2 polarization, rearrange the BMSC cytoskeleton, and modulate mitochondrial dynamics toward fusion, thereby enhancing osteogenic differentiation. Notably, utilizing the immune mechanism, this exquisite design that can crosstalk the F-actin cytoskeleton and mitochondrial dynamics to promote efficient osteogenesis may provide new therapeutic perspectives for clinical bone healing problems, such as massive bone defects and non-unions, chronic osteomyelitis, and osteoporotic fractures.

2. Results

2.1. Material characterization

Rapid advances in tissue engineering have facilitated the development of multifunctional artificial periostea, which has substantially enhanced bone regeneration efficiency [6,7,10]. However, the mechanism of mitochondrial dynamics modulation by artificial periostea, particularly on biomimetic periostea that can simultaneously modulate the cytoskeletal remodeling and mitochondrial dynamics of BMSCs in the immune microenvironment, has not been studied in detail yet. In this study, Sema3A controlled-release fibers with a core-shell structure were prepared by microsol electrostatic spinning. Specifically, hyaluronic acid (HA)-encapsulated Sema3A was emulsified to form microemulsion particles, stably dispersed in dichloromethane solution, and self-assembled into core-shell structured fibers under a high-voltage electric field. A roller receiver collected the resulting randomly drug-loaded fibers (MS@Sema3A), while the unloaded spinning solution yielded randomly oriented (MS) fibers as controls. When the collection device was modified to parallel electrode rods, oriented drug-loaded fibers (AMS@Sema3A) were obtained. The unloaded spinning solution yielded oriented (AMS) fibers as controls.

Microgel electrospinning induces a self-assembled core-shell structure under a high-voltage electric field, which is the physical basis for efficient drug loading and slow-release capability. Transmission electron microscopy (TEM) observations of the internal structure of single fibers confirmed this feature (Fig. 1A). The stability of the bionic periosteum is a key prerequisite for its function. Accordingly, fluorescently labeled bovine serum albumin (BSA) was used to replace Sema3A encapsulated in the HA microsphere and electrostatic spinning was performed in dark. The fluorescence tracing results showed that the labeled protein was uniformly distributed inside the single spun fiber (Fig. 1B), confirming that the microgel electrospinning system possessed stable drug delivery capability.

Fig. 1.

Physicochemical characterization of biomimetic periosteum. (A) TEM image of liquid micro-sol electrospun fibers, scale bar: 500 nm. (B) Distribution of FITC-BSA within micro-sol electrospun fibers, scale bar: 10 μm. (C) SEM image and qualitative analysis of fiber orientation distribution. (D) Fiber orientation distribution maps of different electrospun fibers. (E) Comparison of the percentage of fibers oriented within ±10° across different electrospun samples. (F) Comparison of fiber diameters across different electrospun samples (n = 120 fibers/sample). (G) Tensile mechanical testing of different electrospun fibers. (H) Comparison of Young's modulus across different electrospun fibers. (I) FTIR spectra of different electrospun fibers. (J) Degradation comparison of different electrospun fibers. (K) Water contact angle comparison of different electrospun fibers. (L) Time-course release of Sema3A from MS@Sema3A and AMS@Sema3A fibers. (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test between groups; n = 3; ns: p > 0.05 indicates no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

We characterized the particle size of the HA microsphere loaded with Sema3A using dynamic light scattering (DLS; Fig. S1). The average HA particle size was 0.258 ± 0.003 nm, with electrospun diameters ranging from 100 to 800 nm, and the polymer dispersity index (PDI) was 0.252, indicating that the HA particle size was relatively uniform. The microsol particles were kept at room temperature (20-25°C) for 2 h without significant changes, confirming their excellent stability and suitability for effective embedding of Sema3A.

The topographical features of biomaterials significantly influence cellular activities and immune responses, and topography plays an important role in immun- –bone crosstalk [33]. Scanning electron microscopy (SEM) characterized the surface topography of different fibers, the degree of orientation was quantified by ImageJ, and pseudo-color distribution maps were generated (Fig. 1C). Microstructural analysis showed that the MS and MS@Sema3A fibers exhibited smooth, randomly aligned, and highly homogeneous spun structures, while the AMS and AMS@Sema3A fibers showed distinct orientation alignment characteristics. Based on color wheel analysis [34], quantitative assessment showed that the AMS@Sema3A fiber had highly uniform orientations, all of which were centrally distributed within ±10°, while the MS and MS@Sema3A fibers showed a trend of orientation concentration (Fig. 1D), with statistically significant (p < 0.05) differences in the degree of orientation between the random and oriented groups (Fig. 1E).

The average diameters of single-spun MS, AMS, MS@Sema3A, and AMS@Sema3A fibers were 0.567 ± 0.071, 0.571 ± 0.072, 0.576 ± 0.08, and 0.573 ± 0.076 μm, respectively, as analyzed using ImageJ (Fig. S2A–D). Quantitative analysis (Fig. 1F) showed that although Sema3A loading increased the diameter of spun fibers, the difference in diameter between samples from the four fibers was not statistically significant (p > 0.05).

The mechanical stability of the bionic periosteum is the basis for its continuous release of bioactive factors [35], and we characterized the mechanical properties and plotted the stress–strain curves by tensile experiments (Fig. 1G). The elastic moduli of MS, AMS, MS@Sema3A, and AMS@Sema3A fibers were 3.0898 ± 1.4594, 3.9013 ± 1.5735, 3.3305 ± 1.5057, and 3.9346 ± 1.4351, respectively. All electrospun fibers exhibited good mechanical properties, with elastic moduli similar to those of native periosteum (3–4 MPa) [36]. AMS@Sema3A achieves the maximum tensile strength (0.375 ± 0.234 MPa), and its Young's modulus and stiffness are higher than those of AMS. Young's modulus and stiffness were not statistically significantly different between AMS and AMS@Sema3A or between MS and MS@Sema3A (p > 0.05). However, the Young's modulus and stiffness of AMS were higher than those of MS (Fig. 1H, Supplementary Fig. 3A). The oriented electrospun fibers significantly enhanced the performance by compensating for the mechanical defects of the core-shell structure, while stiffness optimization further enhanced the osteogenic capacity [28]. This synergistic effect makes oriented microsol electrostatic spinning membranes a key strategy for optimizing the mechanical and osteogenic efficacy of bionic osteochondral membranes.

The molecular structures of different electrospun fibers were characterized using infrared spectroscopy. Analysis of the Fourier transform infrared spectra (Fig. 1I) showed that the MS fiber membranes displayed a characteristic C=O stretching vibration band near 1750 cm−⁻¹ [22] and the infrared spectra of AMS, MS@Sema3A, and AMS@Sema3A fiber membranes exhibited a high degree of similarity. More importantly, the characteristic absorption bands of HA were not detected in all samples, confirming that HA was successfully encapsulated inside the fiber matrix and its signal was completely shielded from detection by IR spectroscopy. This finding validates the effectiveness of the encapsulation process at the molecular level, suggesting that HA is completely encapsulated within the fiber structure without affecting the chemical properties of the fiber body.

Since difficulties in the degradation of tissue-engineered materials may increase the risk of foreign body reaction, we tested the in vitro degradation properties of the four materials [37]. The results (Fig. 1J) revealed that the weights of MS, AMS, MS@Sema3A, and AMS@Sema3A fiber membranes decreased to 90.467 ± 0.802%, 89.567 ± 0.971%, 90.8 ± 0.819%, and 87.467 ± 0.723%, respectively, after the first 10 days. Sema3A loading accelerated MS@Sema3A and AMS@Sema3A fiber membrane degradation rate, which may be because the hydrophilic functional groups present in Sema3A effectively neutralized the inherent hydrophobicity of PLA fibers [24,38]. After 60 days, the weights of MS, AMS, MS@Sema3A, and AMS@Sema3A membranes decreased to 47.933 ± 2.06%, 48.233 ± 2.325%, 42.9 ± 1.136%, and 43.933 ± 0.351%, respectively. The biomimetic bone membrane degradation rate matches the progression of new bone formation in cranial defects [39].

Electrostatically spun scaffolds with appropriate porosity are highly favorable for cell growth [40]. Porosity measurements showed that the porosity of MS, MS@Sema3A, AMS, and AMS@Sema3A fiber membranes was 77.567 ± 1.611%, 76.133 ± 1.805%, 73.533 ± 1.92%, and 71.75 ± 2.613%, respectively (Fig. S3B). The porosity of the oriented fiber structure was slightly lower than that of the non-oriented fiber structure (p < 0.05), but it was conducive to the adhesion and proliferation of osteoblast-like cells.

In tissue engineering, the hydrophilicity of a material is usually related to the fiber diameter, roughness, and porosity of the material; while water contact angle of the hydrophobic material Poly-L-Lactic Acid (PLLA) was greater than 90° [24], the water contact angle in MS, AMS, MS@Sema3A and AMS@Sema3A fibers was 138.017 ± 1.999°, 135.117 ± 2.144°, 137.917 ± 1.763°, and 135.817 ± 1.402°, respectively (Fig. 1K). The water contact angle was similar between MS and MS@Sema3A and not significanty different between AMS and AMS@Sema3A. However, the water contact angles of both AMS and AMS@Sema3A are slightly lower than those of MS and MS@Sema3A. This phenomenon suggests that topology can slightly reduce the water contact angle of the materials and thus enhance their hydrophilicity [24,41,42].

To evaluate the drug retardation performance of the bionic material, this study characterized the in vitro release kinetics of Sema3A protein in microsol electrospun fibres using enzyme-linked immunosorbent assay (ELISA) (Fig. 1L). During the one-month observation period, MS@Sema3A and AMS@Sema3A exhibited similar drug release kinetics. Notably, AMS@Sema3A exhibited significant burst release within the first 4 days, releasing 47.267 ± 2.417% Sema3A protein. In comparison, the release from MS@Sema3A over the same period was 45.133 ± 2.829%. The rapid Sema3A protein release from the bionic periosteum in the initial stage effectively induces macrophage polarization in the local microenvironment at the injury site towards the M2 phenotype. This process significantly alleviates the inflammatory response and achieves a timely transition from the inflammatory phase to the repair phase via a temporal regulatory mechanism. Ultimately, this promotes bone tissue regeneration and repair. Drug release kinetics showed that the cumulative release from AMS@Sema3A and MS@Sema3A on day 14 was 71.367 ± 3.113% and 70.5 ± 3.105%, respectively, indicating a significant slowdown in the release rate. The excellent stability of the microsol supported the sustained release of Sema3A for up to 28 days, with the final cumulative release exceeding 80% initial drug loading. In summary, AMS@Sema3A exhibits mechanical properties close to those of natural periosteum, with degradation behavior compatible with the bone repair process. Its porosity promotes cell ingrowth and Sema3A encapsulation within the spinning fiber core enables effective controlled release, making it an ideal biomimetic periosteum tailored for bone repair.

2.2. In vitro experiments

2.2.1. Biocompatibility of bionic periosteum

The effect of different Sema3A concentrations on the viability of Raw264.7 cells was assessed using the CCK-8 method, and the results showed no significant change in cell viability after treatment with 0–200 ng/mL Sema3A for 6 h. However, treatment with 100 and 200 ng/mL Sema3A for 24 h significantly decreased cell viability (p < 0.05). Based on these results, 50 ng/mL Sema3A was identified as the optimal working concentration for subsequent experiments (Fig. 2A). Based on the drug release kinetics of in vitro micro-soluble electrospun fibers, the actual Sema3A concentration present in the cell culture system is approximately 40 ng/mL. Analysis of the CCK-8 results showed (Fig. 2B) that all four electrospun fiber materials supported the sustained proliferation of BMSCs during the 7-day incubation period. During the first 4 days, cell count did not statistically differ between the MS, AMS, and MS@Sema3A groups. However, day 5 onwards, proliferation significantly enhanced in the AMS@Sema3A group, which peaked by day 7. This result suggests that the AMS@Sema3A composite has the best pro-proliferation effect on BMSCs.

Fig. 2.

In vitro cell compatibility of biomimetic periosteum. (A) At concentrations exceeding 50 ng/mL, Sema3A solution inhibited macrophage growth after 24 h. (B) CCK-8 assay detecting proliferation of bone marrow mesenchymal stem cells (BMSCs) on different group materials at 1, 3, 5, and 7 days. (C) Live/dead (green/red) fluorescence (scale bar = 100 μm) and scanning electron microscopy (SEM) images (scale bar = 3 μm) of BMSCs after 3 days of treatment with different samples. (D) Live/dead (green/red) fluorescence staining and SEM images (scale bar = 50 μm) of RAW264.7 cells after 1 day of treatment with different samples. (E) Immunofluorescence staining of Integrinβ1 protein in different groups (scale bar = 50 μm). (F) Immunofluorescence staining of Vinculin protein in different groups (scale bar = 50 μm). (G) BMSCs scratch assay in different groups (scale bar = 50 μm). (H) Quantitative analysis of migration differences in BMSCs at 24 h and 48 h across different groups. (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test between groups; n = 3; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

Material biocompatibility assessment is crucial in tissue engineering studies [43]. In this study, the compatibility performance of bionic periosteal materials on BMSCs and Raw264.7 cells was systematically evaluated by in vitro experiments using the Live/Dead cell staining method. The experimental design consisted of two sets of parallel tests to examine the interaction of the materials with the two cell types separately, which provided reliable biosafety data support for subsequent studies. SEM observation combined with Live/Dead staining analysis showed that all four electrospun fiber materials supported the normal adherent growth and sustained proliferation of both BMSCs and Raw264.7 cells (Fig. 2C and D). Both cell types showed good survival on the surface of all the materials, and this result confirmed that none of the electrospun fiber materials exhibited cytotoxicity or inhibited proliferation. The experimental data indicated that AMS fibers had higher cellular activity than MS fibers, while MS@Sema3A had more dead cells than AMS@Sema3A. This difference may stem from the fact that the oriented spinning structure is more favorable for cell adhesion and proliferation [24]. Notably, AMS@Sema3A exhibited optimal cellular activity (Fig. S4A), confirming that Sema3A synergies with the topology to promote cell proliferation. The results of the Raw264.7 cell assay (Fig. 2D, Fig. S4B) similarly verified this pattern.

In conclusion, the oriented spinning structure significantly promoted cell proliferation than the non-oriented structure, and the Sema3A factor synergistically enhanced the proliferation efficiency of BMSCs. This dual mechanism of action significantly enhanced the biocompatibility of the material, laying an important foundation for the functional implantation of bionic periosteum.

2.2.2. Bionic periosteum promotes cell adhesion and migration

Cell adhesion ability is a key factor influencing seed cell proliferation and differentiation. In this study, we systematically evaluated the cell adhesion properties of different material surfaces by quantitatively analyzing Integrin β1 and Vinculin expression levels. Integrin β1, as a key transmembrane receptor for promoting cell adhesion, mediates the mechanical connection between the cell and extracellular matrix and bi-directionally modulates cellular microenvironment signaling [44]. Quantitative fluorescence analysis showed that Integrin β1 expression level in oriented microsol electrospun fibers was significantly higher than in the non-oriented fibers (Fig. 2E, Fig. S5A). Sema3A further enhanced Integrin β1 expression intensity on the material surface, with AMS@Sema3A exhibiting the best cell adhesion performance, confirming the effect of Sema3A on the mimetic osteochondral cell adhesion ability of Sema3A.

Vinculin is an actin-binding protein that enhances cell–cell and cell–matrix adhesion [45]. Quantitative immunofluorescence analysis showed that oriented electrospun fibers significantly enhanced Vinculin-mediated F-actin anchoring (Fig. 2F, Fig. S5B). Notably, AMS@Sema3A exhibited the highest Vinculin expression, confirming that the material had excellent cell adhesion properties.

Cell scratch experiments are often used to reveal the effects of different material properties on BMSC migration [46,47]. The results showed that AMS@Sema3A significantly promoted BMSC migration, indicating that the material had an optimal pro-migratory effect. After 24 h, the cell migration rate of the AMS@Sema3A group was significantly higher than that of the AMS (62.963% vs. 15.805%) and MS@Sema3A groups. After 48 h, the migration rate of the AMS@Sema3A group increased to 79.899% (P < 0.0001), while the difference between the MS and AMS groups was not statistically significant (Fig. 2G and H). These data suggest that the AMS@Sema3A bionic periosteum has the greatest pro-BMSC migration effect.

In summary, AMS@Sema3A biomimetic periosteum significantly enhances BMSC adhesion and migration than other materials. Its ability to mimic the extracellular matrix promotes cell proliferation and provides an efficient platform for differentiation.

2.2.3. Sequential regulation of macrophage polarization alters the local immune microenvironment

The immune signaling cascade coordinates the entire fracture repair process and the cascade reactions it mediates play a decisive role in the initiation of inflammation and subsequent regeneration [48]. Macrophages are highly plastic and can undergo M1/M2 polarization switching in response to microenvironmental signals. In the acute phase, M1-type macrophages secrete pro-inflammatory factors to remove necrotic tissue and recruit stem cells, while in the regenerative phase, they switch to the M2-type macrophages in time to promote tissue repair. This chronological polarization shift is essential for successful tissue regeneration [8]. To elucidate the mechanism by which biomimetic periosteum regulates macrophage polarization, we analyzed the phenotypic changes of mcrophage after co-culture with electrospun fibers with different components.

Firstly, immunofluorescence analysis revealed that the MS group exhibited the strongest INOS fluorescence signal (M1-type marker) 3 and 7 days after co-culture with various components of electrosilk scaffolds (p < 0.05). In contrast, the lowest expression was observed in the AMS@Sema3A group (Fig. 3A), confirming the ability of the scaffold components to temporally regulate macrophage polarization. After 3 days of co-culturing with macrophages, the AMS@Sema3A group exhibited stronger CD206 (M2Mφ marker) fluorescence signals than the other groups. This difference was even more significant after 7 days of co-culture. At the same time, the Raw267.4 cells in the AMS@Sema3A group presented a slender morphology and more pseudopodia. The fluorescence intensity in the MS@Sema3A group was higher than in the AMS group, while the MS group always had the lowest intensity (Fig. 3A). Semi-quantitative analysis confirmed that the fluorescence intensity of INOS and CD206 at 3 and 7 days followed the same trend (Fig. S6A–B). Based on phenotypic analysis, we confirmed that Sema3A released by biomimetic periosteum significantly promoted macrophage polarization to M2 type within the first three days of culture. This finding provides key immunological evidence for its role in promoting bone regeneration.

Fig. 3.

IF and PCR analysis of Mφ polarization. (A) Immunofluorescence staining for INOS and CD206 on days 3 and 7. scale bar = 25 μm. (B) Expression analysis of the inflammatory gene Nos2 at days 3 and 7. (C) Expression analysis of the anti-inflammatory gene Mrc1 at days 3 and 7. (D) Expression analysis of the inflammatory gene Il1b at 3 and 7 days. (E) Expression analysis of the anti-inflammatory gene Il10 at 3 and 7 days. (n = 3 per group, Statistical analysis requires two-way ANOVA, between groups is Tukey's multiple comparison test; ns indicates not statistically significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Next, to systematically verify the molecular mechanism of macrophage phenotypic transformation and functional remodeling, we quantified M1/M2 polarization-related marker genes in macrophages treated with each experimental material using real-time quantitative PCR (qRT-PCR) (Fig. 3B–E). Previous studies have demonstrated that 200 ng/mL lipopolysaccharide (LPS) stimulates macrophages (RAW264.7), effectively mimicking the local inflammatory environment following bone injury [1,49]. Co-culturing the materials with LPS-pretreated Raw267.4 cells for 3 days revealed that the AMS@Sema3A group exhibited higher and lower expression of M2 (Mrc1 and Il10) and M1 (Nos2 and Il1b) polarization markers than the other groups. After 7 days, M1 markers further decreased, while M2 markers increased across all groups, with the most pronounced M2 polarization in the AMS@Sema3A group. Specifically, LPS pretreatment significantly increased Nos2 and Il1b (M1 markers) expression in Raw264.7 cells. In contrast, MS, AMS, and MS@Sema3A groups exhibited reduced Nos2 and Il1b expression, with the lowest levels observed in the AMS@Sema3A group, indicating activated macrophage polarization toward the M2 phenotype. Compared to the LPS and MS groups, the AMS group exhibited higher expression levels of Mrc1 and Il10 (M2 markers). This effect was further enhanced by adding Sema3A, with expression levels exceeding those in the MS@Sema3A group. The experimental data indicated that during the inflammation period (3 days of total culture), the bionic periosteum effectively regulated the immune microenvironment at the injury site through the specific release of Sema3A signaling molecules, which significantly promoted macrophage polarization to the M2-type, and this key immune-regulatory mechanism laid an important immunological foundation for the subsequent bone repair and regeneration.

The results of the macrophage polarization flow analysis for the different materials showed that, after three days of co-culture, the macrophages in each group predominantly expressed the CD11b⁺/CD86⁺ phenotype (Fig. 4A). After seven days, the macrophages predominantly expressed the CD11b⁺/CD206⁺ phenotype (Fig. 4B). These dynamic changes confirmed a temporal shift in macrophage polarization from the M1 to M2 phenotype. Specifically, the AMS@Sema3A group exhibited the highest CD11b⁺/CD206⁺ ratio after 3 days of co-culture, which was significantly higher in the AMS group than in the MS group, but lower than in the MS@Sema3A group. The ratio increased significantly in all groups, particularly in the MS@Sema3A group, after 7 days. Notably, Sema3A treatment increased the ratio in the AMS@Sema3A group by 37.7% than in the AMS group (Fig. 4C), confirming the promoting effect of Sema3A on macrophage polarization. In addition, the CD11b⁺/CD86⁺ ratio was significantly lower in the AMS@Sema3A group than in the other groups after 3 days of co-culture (p < 0.05), declining in all groups after 7 days. The MS group maintained the highest value among them, while the CD11b⁺/CD86⁺ ratio in the MS@Sema3A group was significantly higher than in the AMS group, though still lower than in the AMS@Sema3A group (Fig. 4D). These data confirmed that Sema3A could effectively promote M1 to M2-type polarization and the AMS@Sema3A bionic periosteum showed the strongest pro-M2 polarization ability.

Fig. 4.

Bionic Periosteum Immunomodulatory Effects.(A) Flow cytometry detection of CD86 and CD206 expression in Mφ on day 3. (B) Flow cytometry detection of CD86 and CD206 expression in Mφ on day 7. (C) Quantitative analysis of FCA with CD11b+/CD206+ phenotype. (D) Quantitative analysis of FCA with CD11b+/CD86+ phenotype. (E) Western blot detection of INOS, Arg-1, and IL-1β expression levels. (F-H) Quantitative analysis of Arg-1, INOS, and IL-1β expression levels. (I) Western blot detection of phosphorylation levels of PI3K, AKT, and mTOR. (J-L) Quantitative analysis of phosphorylation levels of PI3K, AKT, and mTOR. (n = 3 per group, Statistical analysis requires one-way ANOVA, between groups is Tukey's multiple comparison test; ns indicates not statistically significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

The PI3K/Akt1 signaling pathway plays a crucial role in regulating macrophage polarization [49]. Finally, to elucidate the molecular mechanism of bionic periosteum for M2 polarization, we examined the key protein expression of Mφ after 7 days of co-culture with each group of electrostatically spun fibres by WB, focusing on the regulatory role of PI3K/Akt/mTOR signalling pathway [31]. Firstly, WB analysis (Fig. 4E–H) showed that Arg-1 expression was significantly higher in the AMS@Sema3A group than in the other groups (p < 0.05), while INOS and IL-1β expression was the lowest, further confirming that the bionic osteochondral membrane had an optimal pro-M2 polarization effect. Next, WB analysis (Fig. 4I–L) showed that the AMS@Sema3A group significantly upregulated the phosphorylation level of the PI3K/Akt/mTOR pathway (p < 0.05), confirming that the Sema3A functionally oriented bionic periosteum induced M2-type macrophage polarization by activating this signalling pathway.

In conclusion, it was shown that the activation of the PI3K/Akt/mTOR signaling pathway in bionic periosteum significantly promoted the polarization of macrophages to M2 type, which accelerated the transition from inflammatory to reparative phase, effectively reduced the inflammatory damage of tissues, and optimized the microenvironment of osteogenesis, and ultimately promoted the regeneration of bone tissue.

2.2.4. Bone formation promotion by bionic periosteum via immune regulation

To elucidate the mechanism by which the bionic periosteum regulates bone formation through M2 macrophage polarization, experiments employed the corresponding conditioned medium (CM, Supplementary Fig. 7) to replace the basal medium. This medium was used to culture BMSCs on different material surfaces, followed by comparative bone formation analysis.

Alkaline phosphatase (ALP) is a core marker of early osteoblast differentiation and bone mineralization, reflecting the mineralization stage of bone formation [50]. To assess the ability of the materials to promote early osteogenic differentiation, BMSCs were cultured in CM from different material groups for 7 days and then subjected to an ALP assay. The results showed that the AMS@Sema3A group had the darkest ALP staining and the best cell morphology (Fig. 5A), and the AMS group stained stronger and weaker than the MS and MS@Sema3A groups, respectively (Fig. 5B). The experimental data indicated that both the oriented material structure and Sema3A protein promote osteogenic differentiation. Therefore, the orientation mimetic periosteum loaded with Sema3A promoted osteogenic differentiation and BMSC mineralization significantly better than a single factor.

Fig. 5.

In vitro osteogenic studies of different biomimetic periosteum models. (A) ALP staining (scale bar: 50 μm for upper image, 500 μm for lower image). (B) Quantitative assessment of ALP activity. (C) Alizarin Red S (ARS) staining (scale bar: 50 μm for upper image, 500 μm for lower image). (D) Semi-quantitative evaluation of ECM mineralization. (E) OCN fluorescence staining (red), F-actin fluorescence staining (green), and nuclear fluorescence staining (blue) in different groups (scale bar = 50 μm). (F) Quantitative analysis of OCN fluorescence. (G) Western blot results showing that both Sema3A and oriented spinning promote the expression of OCN, OPN, and Runx2 proteins. (H) Quantification of Runx2 protein in Western blot. (I) Quantification of OCN protein in Western blot. (J) Quantitative analysis of OPN protein in Western blot. (K) qRT-PCR detection of Col1a1 at day 7 post-BMSCs culture. (L) qRT-PCR detection of Alpl at day 7 post-BMSCs culture. (M) qRT-PCR detection of Bglap at day 10 post-BMSCs culture. (N) qRT-PCR detection of Runx2 at day 10 post-BMSCs culture. (Statistical analysis required one-way/two-way ANOVA; Tukey's multiple comparison test for intergroup comparisons; n = 3; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

Calcium nodule formation detection using alizarin red staining is routinely used to assess osteoblast function and serves as a definitive marker of advanced osteogenic differentiation and bone mineralization [51]. To assess the in vitro osteogenic induction potential of the material, BMSCs were cultured in CM for 21 days and then tested for late osteogenic differentiation using alizarin red staining. Quantitative analysis showed that the AMS group had significantly better calcium nodule formation than the MS group (p < 0.05), but weaker than the MS@Sema3A group; the AMS@Sema3A group exhibited the most significant calcium nodule deposition (Fig. 5C). This result was in perfect agreement with the quantitative data of alizarin red staining (Fig. 5D). AMS@Sema3A bionic periosteum guided the orderly deposition of calcium salts through its topologically oriented structure, while Sema3A signaling molecules continued to promote extracellular matrix mineralization; this physico-biological synergy remarkably enhanced the mineralization efficiency and osteogenic performance at the later stage of bone tissue engineering.

OCN coordinates bone matrix mineralization and systemic metabolism and is an important marker of osteoblast mineralization [52], so we applied immunofluorescence staining to observe OCN protein expression in different groups (Fig. 5E). To assess the in vitro medium-term osteogenic differentiation ability of the materials, we cultured BMSCs in different CM for 14 days and then performed OCN immunofluorescence staining. Confocal fluorescence microscopy results showed that the AMS group had higher and lower OCN expression than the MS and MS@Sema3A groups, respectively; the AMS@Sema3A group had the highest level of OCN protein (Fig. 5F). The results further indicated that both orientation and Sema3A promoted osteogenic mineralization, and their combination resulted in optimal osteogenic mineralization, which is also consistent with the ALP and ARS staining results.

To assess the effect of the immunomodulatory properties of the materials on osteogenic differentiation, BMSCs on different spinning scaffolds were cultured in CM for 7 days, and the expression levels of osteogenesis-related proteins were analyzed using Western blot. Western blot analysis (Fig. 5G) and quantitative results showed that the OCN, OPN, and Runx2 expression levels in the AMS group were higher and lower than those in the MS and MS@Sema3A groups, respectively (Fig. 5H and I). The expression of the three osteogenic markers was significantly better in the AMS@Sema3A group than in other groups, indicating that AMS@Sema3A has the best osteogenic induction ability.

To elucidate the synergistic effect of the material surface morphology features and Sema3A on osteogenesis-related gene expression, the expression levels of key osteogenic genes in BMSCs were detected using qRT-PCR (Fig. 5K–N). qRT-PCR analysis showed that among the different fiber scaffolds cultured in CM , the expression levels of Col1a1, Alpl, Bglap, and Runx2 in BMSCs in the AMS group were between those in the MS group and the MS@Sema3A group, while their expression levels in the AMS@Sema3A group were significantly higher than in other experimental groups. The experimental data confirmed that oriented spun scaffolds significantly enhanced osteogenic gene expression than non-oriented scaffolds, and that AMS@Sema3A bionic periosteum demonstrated optimal osteogenic activity through the slow-release mechanism of Sema3A.

In summary, the biomimetic periosteum constructed by oriented microsol electrostatic spinning has excellent biocompatibility and promotes cellular behaviors, and its oriented structure synergistically induces M2 polarization with the controlled release of Sema3A to optimize the immune microenvironment, which significantly enhances the osteogenic performance of AMS@Sema3A in vitro.

2.2.5. Sema3A promotes osteogenesis by regulating the actin cytoskeleton through the RhoA/ROCK2 pathway

Immunofluorescence showed that cytoskeletal differences in the four groups of BMSCs may affect osteogenic differentiation. Accordingly, the BMSCs were cultured in different scaffolds and treated with CM for 3 days. Then, comparative analyses were performed with Runx2 and ROCK2 immunofluorescence staining. Runx2 expression analysis showed that the expression in the AMS group was between those in the MS and MS@Sema3A groups, and the highest expression was observed in the AMS@Sema3A group (Fig. 6A and B). Immunofluorescence images revealed that the cells in the AMS and MS@Sema3A groups appeared more elongated than in the MS group, while cells in the AMS@Sema3A group exhibit the most elongated morphology. Fluorescence images of ROCK2 expression across the four groups revealed the strongest ROCK2 expression and most elongated cells in the AMS@Sema3A group, followed by the MS@Sema3A group with moderate expression and relatively rounder cell morphology. The AMS group exhibited higher ROCK2 expression than the MS group with comparatively elongated cells (Fig. 6C and D). In summary, oriented topography guided cell elongation, promoting ROCK2 and Runx2 expression. Sema3A application exhibited a similar trend, elongating BMSCs, enhancing ROCK2 protein expression, and increasing Runx2 expression levels.

Fig. 6.

Biomimetic periosteum modulates cytoskeletal organization to promote osteogenesis. (A) Runx2 immunofluorescence staining reveals that both Sema3A and orientation enhance Runx2 expression, with AMS@Sema3A exhibiting the brightest Runx2 fluorescence. Scale bar: 50 μm. (B) Quantitative analysis of Runx2 protein expression in BMSCs across groups in Figure A (n = 5). (C) Immunofluorescence staining shows Sema3A increases ROCK2 expression, with the brightest ROCK2 fluorescence observed in AMS@Sema3A. Scale bar: 50 μm. (D) Quantitative comparison of ROCK2 protein expression among BMSCs groups in Figure C (n = 5). (E) Western blot results show that spinning-oriented stretching and Sema3A promote ROCK2 and Runx2 protein expression, with AMS@Sema3A exhibiting the highest ROCK2 and Runx2 expression. (F) Quantitative comparison of ROCK2 protein levels among groups in Figure E (n = 3). (G) Quantitative comparison of Runx2 protein levels among groups in Figure E (n = 3). (H) Immunofluorescence shows Sema3A reverses Y27632-induced decreases in F-actin spreading and ROCK2 expression. Scale bar: 50 μm. (I) Quantitative comparison of ROCK2 expression among groups in Figure H (n = 5). (J) Immunofluorescence showing Y27632-induced reduction in Runx2 expression in AMS@Sema3A. Scale bar: 50 μm. (K) Quantitative comparison of Runx2 expression among groups in Figure J (n = 5). (L) Western blot results show Sema3A promotes ROCK2 and Runx2 protein expression in BMSCs, while ROCK2 inhibitor Y27632 suppresses ROCK2 and Runx2 protein expression. (M) Quantification of ROCK2 protein expression across groups in Figure L (n = 3). (N) Quantification of Runx2 protein expression across groups in Figure L (n = 3). (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test for intergroup comparisons; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

Then, the differences in ROCK2 and Runx2 expression between the different groups were further analyzed by Western blot (WB). As demonstrated in WB (Fig. 6E) and quantitative analysis of the data (Fig. 6F and G), the increase in the ROCK2 and Runx2 expression levels in the AMS group was lower than those in the MS@Sema3A group compared to those in the MS group. However, the increase in the ROCK2 and Runx2 expression levels in the AMS@Sema3A group were significantly higher than that of all other groups. Immunofluorescence and WB results showed consistent trends of ROCK2 and Runx2 expression in BMSCs in all groups. Material orientation promotes ROCK2 expression and osteogenesis and Sema3A enhances osteogenesis by modulating the cytoskeleton, giving the AMS@Sema3A scaffold an optimal osteogenic effect.

Owing to the previously observed differential changes in the cytoskeleton and osteogenesis between different groups, whether Sema3A regulates the actin cytoskeleton via the RhoA/ROCK2 pathway and thus affects the osteogenic function [53] was investigated using 10 μM RhoA/ROCK2 pathway-specific inhibitor Y27632 (HY-10071, MedChemexpress, USA) for intervention validation.

Immunofluorescence results showed that Sema3A promoted F-actin polymerization through ROCK2 activation (Fig. 6H and I), which was completely blocked by Y27632, and the Y27632-induced skeletal disorders in BMSCs and Runx2 expression inhibition were effectively reversed by Sema3A (Fig. 6J and K), confirming a key role of the ROCK2 pathway in Sema3A-mediated osteogenic differentiation. Next, protein blotting and quantitative analysis showed that the AMS@Y27632 and AMS@Sema3A groups exhibited the lowest and highest ROCK2 and Runx2 expression levels, respectively (Fig. 6E–G). Notably, the ROCK2 expression levels in the AMS@Sema3A@Y27632 group were comparable to those in the AMS group, and Y27632 completely antagonized Runx2 upregulation by Sema3A. The results confirmed that Sema3A regulates the actin cytoskeleton and thus promotes osteogenic differentiation through the RhoA/ROCK2 pathway.

2.2.6. Sema3A regulates the cytoskeleton and promotes mitochondrial fusion

Mitochondria are multifunctional organelles closely associated with the cytoskeleton [[54], [55], [56]]. Changes in the cytoskeletal structure of BMSCs are associated with changes in mitochondrial morphology. Therefore, to investigate their relationship, BMSCs were seeded on different scaffolds and cultured in corresponding CM for 3 days, followed by immunofluorescence staining analysis using Tom20 and phalloidin.

Tom20 immunofluorescence showed that the mitochondrial morphological characteristics were significantly different in each group, with short punctate mitochondria accompanied by reticular F-actin in the MS group, short and thin tubular mitochondria in the AMS group, more elongated mitochondria in the MS@Sema3A group, and the highest degree of mitochondrial fusion in the AMS@Sema3A group. Notably, mitochondrial distribution in the AMS group was more fused than in the MS group and Sema3A significantly promoted the formation of networked fusion structures in mitochondria (Fig. 7A).

Fig. 7.

Study on the Regulation of BMSCs Mitochondria by a Biomimetic Periosteum. A) Tom20 and phalloidin staining reveal short dot-like mitochondria and a reticular distribution of F-actin in the MS group. Oriented stretching induces elongated tubular mitochondria and F-actin polymerization. Sema3A promotes mitochondrial fusion accompanied by F-actin polymerization. The area within the dotted box is a close-up view. Scale bar: 50 μm. (B) Aspect ratio of BMSCs in each group in Figure A (n = 5). (C) Comparison of roundness among BMSCs in each group in Figure A (n = 5). (D) Immunofluorescence reveals that ROCK2 inhibitor Y27632 disrupts the cytoskeleton, causing disintegration of the mitochondrial network in BMSCs. The area within the dotted box is a close-up view. Scale bar: 50 μm. (E) Western blot results show that oriented stretching promotes Mfn2 protein expression while suppressing Drp1 protein expression. Sema3A inhibits Drp1 protein expression and promotes Mfn2 protein expression. (F) Quantitative comparison of Drp1 protein levels between groups in Figure E (n = 3). (G) Quantitative comparison of Mfn2 protein levels between groups in Figure E (n = 3). (H) Immunofluorescence revealed that orientation reduced Drp1 expression relative to non-orientation, while Sema3A further promoted this decrease. Scale bar: 50 μm. (I) Quantitative comparison of Drp1 expression among different groups in Figure H (n = 5). (J) Immunofluorescence shows that orientation increases Mfn2 expression compared to non-orientation, and Sema3A enhances Mfn2 expression. Scale bar: 50 μm. (K) Quantitative comparison of Mfn2 expression among different groups in Figure J (n = 5). (L) Western blot results show Sema3A promotes Mfn2 protein expression and inhibits Drp1 protein expression in BMSCs, while ROCK2 inhibitor Y27632 suppresses this effect. (M) Quantification of Mfn2 protein expression among groups in Figure L (n = 3). (N) Quantification of Drp1 protein expression among groups in Figure L (n = 3). (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

Sema3A regulates axon contraction through the actin cytoskeleton [57], a process that involves ROCK1/2 kinases and regulates F-actin function through the RhoA kinase pathway [58,59]. Immunofluorescence analysis demonstrated that orientation altered the F-actin distribution in BMSCs, as evidenced by decreased cell roundness and increased aspect ratio. After Sema3A treatment, BMSCs exhibited a similar trend, significantly increasing the cell aspect ratio and promoting cell elongation (Fig. 7B and C). Notably, the ROCK inhibitor Y27632 blocked the Sema3A-induced F-actin remodeling effect and led to mitochondrial fragmentation (Fig. 7D, Fig. S8A–B).

The experimental results showed that Sema3A mediated F-actin rearrangement and promoted mitochondrial fusion in BMSCs on oriented spins by regulating the RhoA/ROCK2 pathway.

2.2.7. The decisive role of mitochondrial dynamics in BMSC cytoskeletal rearrangement-promoted bone formation

Mitochondrial dynamics maintain their morphological distribution and functional integrity through a coordinated fission and fusion cycle, thereby safeguarding cellular homeostasis [60,61], where the Mfn2 protein regulates mitochondrial fusion process and the Drp1 protein dominates mitochondrial fission events [62]. Based on the close association between mitochondria and the cytoskeleton, to investigate the role of mitochondrial dynamics in immune-induced bone formation, BMSCs were inoculated into different scaffolds and cultured in corresponding CM for 3 days and the differences in mitochondrial dynamics between the groups were analyzed by protein blotting and immunofluorescence.

Protein blotting (Fig. 7E) and quantitative analysis (Fig. 7F and G) showed that Mfn2 expression was upregulated and Drp1 expression was downregulated in the AMS group than in the MS group, and the AMS@Sema3A group showed the most significant overexpression of Mfn2 and downregulaion of Drp1, while the expression level in the MS@Sema3A group was in between these two groups. Orientation spinning is more likely to cause mitochondrial fusion than random spinning, which may be related to its higher stiffness [28]. The results of Sema3A promoting mitochondrial fusion are consistent with the literature [20]. Immunofluorescence showed that both Sema3A and orientation decreased Drp1 expression (Fig. 7H and I) and increased Mfn2 expression (Fig. 7J and K), with Mfn2 expression being the highest in the AMS@Sema3A group, which was consistent with the protein blotting results.

Next, to elucidate the regulatory role of the RhoA/ROCK2 pathway on cytoskeleton remodeling and mitochondrial dynamics [20], the ROCK inhibitor Y27632 was used to carry out mechanistic studies. First, Western blot (WB) assay showed that Mfn2 expression was significantly upregulated while Drp1 expression was downregulated in the AMS@Sema3A group (Fig. 7L–N), and the effect was reversed by Y27632 treatment, confirming that Sema3A regulates mitochondrial fusion through the RhoA/ROCK2 pathway.

Immunofluorescence analyses further validated the role of the RhoA/ROCK2 signaling pathway in coordinating cytoskeletal remodeling with the mitochondrial dynamic homeostasis. Immunofluorescence staining confirmed that Sema3A on orientation spins upregulated Mfn2 and repressed Drp1 (Fig. 8A and B), and Y27632 reversed this regulation, inhibiting Mfn2 while promoting Drp1 expression (Fig. 8C and D), ultimately maintaining the dominant mode of mitochondrial division (Fig. 8A–D). Meanwhile, Sema3A treatment changed the morphology of AMS-cultured BMSCs to more elongated form, an effect that was inhibited by Y27632, confirming that it regulates the cytoskeleton and mitochondrial fusion status through the RhoA/ROCK2 pathway.

Fig. 8.

Rescue experiment demonstrating bionic periosteum regulation of BMSCs mitochondria. (A) Immunofluorescence reveals that Sema3A promotes Mfn2 expression, while the ROCK2 inhibitor Y27632 suppresses this effect. Scale bar: 50 μm. (B) Immunofluorescence shows that Sema3A inhibits Drp1 expression, and Y27632 reverses this suppression. Scale bar: 50 μm. (C) Quantitative analysis of Mfn2 protein expression between groups in Figure A (n = 5).(D) Quantitative analysis of Drp1 protein expression between groups in Figure B (n = 5).(E) Immunofluorescence shows that the mitochondrial fusion inhibitor MFI8 induces mitochondrial fission in BMSCs. Scale bar: 50 μm. (F) Western blot results show MFI8 inhibits Runx2 expression in BMSCs with minimal effect on ROCK2 expression. (G) Quantification of Runx2 protein expression across groups in Figure F. (H) Quantification of ROCK2 protein expression across groups in Figure F. (I) Immunofluorescence shows Sema3A promotes Runx2 expression, which is blocked by MFI8. Scale bar: 50 μm. (J) Immunofluorescence shows Sema3A promotes ROCK2 expression, while MFI8 has little effect on ROCK2 expression. Scale bar: 50 μm. (K) Quantification of Runx2 protein expression between groups in Fig. I(n = 5).(L) Quantification of ROCK2 protein expression between groups in Figure J (n = 5).(M) Western blot results show that the mitochondrial fission promoter CCCP has little effect on ROCK2 expression but inhibits Runx2 expression in BMSCs. (N) Quantification of ROCK2 protein between groups in Figure M. (O) Quantification of Runx2 protein between groups in Figure M. (P) Immunofluorescence shows that the mitochondrial fission promoter CCCP induces mitochondrial fission in BMSCs. Scale bar: 50 μm (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test; n = 3; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001).

MFI8, a small molecule mitochondrial fusion inhibitor, effectively promotes mitochondrial fission by blocking the mitochondrial fusion process and disrupting MFN1/2 complex assembly [63]. In this study, 10 μmol/L MFI8 (Med Chem Express, HY-150031) was used to induce mitochondrial fracture in BMSCs (Fig. 8E). To verify whether Sema3A regulates cytoskeletal rearrangement to promote osteogenesis through mitochondrial fusion, MFI8 was used to block mitochondrial dynamics and its effect on skeletal remodeling and osteogenesis was assessed. First, protein blotting showed that MFI8 inhibited Runx2 expression (Fig. 8F). Notably, Sema3A could not completely reverse the inhibitory effect of MFI8, although it promoted Runx2 expression (Fig. 8G). In addition, MFI8 had no significant effect on ROCK2 expression and cytoskeleton (Fig. 8H), consistent with the Runx2 inhibition results. Next, immunofluorescence results showed that Sema3A promoted ROCK2 and Runx2 expression (Fig. 8I and J), whereas MFI8 only inhibited Runx2 without affecting ROCK2 (Fig. 8K and L), suggesting that Sema3A regulates osteogenic differentiation through mitochondrial fusion.

To rule out the off-target effects of MFI8 on BMSCs, we conducted supplementary experiments. Mitochondrial fragmentation occurs during bioenergetic stress responses, a process mediated by Drp1 recruitment to mitochondria, while the mitochondrial stress agent carbonylcyanide-3-chlorophenylhydrazone (CCCP) promotes mitochondrial fragmentation [64]. First, we further validated the effects of Sema3A on the cytoskeleton and osteogenesis by adding CCCP in Western blot experiments (Fig. 8M − O). Briefly, 10 μmol/L CCCP (Sigma-Aldrich, C2759) induced mitochondrial fragmentation in BMSCs (Fig. 8P). Western blot analysis revealed that the mitochondrial fragmentation inducer CCCP had minimal effect on ROCK2 expression, but suppressed Runx2 expression in BMSCs (Fig. 8M). Quantitative analysis confirmed the finding that CCCP exerted no significant impact on the cytoskeleton and did not affect ROCK2 expression (Fig. 8N). Sema3A promoted Runx2 expression, partially reversing the inhibitory effect of CCCP (Fig. 8O). Furthermore, immunofluorescence results showed that CCCP suppressed the pro-expressive effect of Sema3A on Runx2 (Fig. 9A and B) but did not affect the Sema3A-mediated promotion of ROCK2 expression (Fig. 9C and D), further indicating that Sema3A regulates osteogenic differentiation through mitochondrial fusion.

Fig. 9.

Experimental verification of biomimetic periosteum regulation of mitochondrial biogenesis in BMSCs. (A) Immunofluorescence reveals that CCCP can block Sema3A's promotion of Runx2 expression. Scale bar: 50 μm. (B) Quantitative analysis of Runx2 protein expression between groups in Figure A. (C) Immunofluorescence shows that CCCP has little effect on ROCK2 expression, whereas Sema3A promotes ROCK2 expression. Scale bar: 50 μm. (D) Quantitative analysis of ROCK2 protein expression between groups in Figure C. (Statistical analysis required one-way ANOVA with Tukey's multiple comparison test; n = 3; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001).

The results showed that Sema3A in the oriented bionic periosteum remodeled the cytoskeleton by upregulating ROCK2 and enhanced the osteogenic differentiation efficacy of BMSCs via a mitochondrial fusion mechanism in the local immune microenvironment at the early stage of osteogenic differentiation.

2.3. In vivo experiments

In this study, we established a critical-sized cranial defect model (5 mm in diameter) in SD rats, and the regulatory effects of different electrospun fiber membranes as bionic periosteum on the bone regeneration process in vivo were systematically evaluated. The experimental design adopted a dual time-node sampling strategy, focusing on resolving the dynamic changes of the immune microenvironment and its regulatory effects on the skeleton remodeling and mitochondrial behavior of BMSCs in the early post-operative period (3 and 7 days), and systematically assessing the regenerative effects of bone tissue in the later period (4 and 8 weeks).

2.3.1. Macrophage regulation in bionic periosteum

Macrophages dynamically regulate the immune microenvironment through phenotypic polarization, a key mechanism throughout tissue regeneration [65]. During bone repair, M1 macrophages promote MSC recruitment and neovascularization by maintaining a moderate inflammatory response, while M2 macrophages dominate the subsequent osteogenic differentiation process [22]. In vitro experiments showed that the controlled release of Sema3A from bionic periosteum drove macrophage M2 polarization and accelerated bone regeneration (Fig. 10A), which was verified by in vivo immunofluorescence analysis.

Fig. 10.

Bionic periosteum modulates the osteogenic immune microenvironment in vivo. (A) Schematic diagram of bionic periosteum-induced polarization of M1 to M2 macrophages. (B) INOS tissue immunofluorescence staining at 3 and 7 days. Overall image scale bar: 500 μm. Magnified section scale bar: 50 μm. (C) Arg-1 tissue immunofluorescence staining at 3 and 7 days. Overall image scale bar: 500 μm. Lower image scale bar: 50 μm. (D) Comparison of Arg-1 tissue immunofluorescence intensity among groups at 3 and 7 days. (E) H&E staining of cranial defects at 4 and 8 weeks. Overall image scale bar: 500 μm. Magnified section scale bar: 50 μm. (F) Masson's trichrome staining of cranial defects at 4 and 8 weeks. Overall image scale bar: 500 μm. Magnified section scale bar: 50 μm. (G) Quantitative analysis of Masson's staining (red-stained areas) at 4 and 8 weeks. (Statistical analysis required two-way ANOVA with Tukey's multiple comparison test between groups; n = 5; ns: p > 0.05, no statistical significance; ∗: p < 0.05; ∗∗: p < 0.01; ∗∗∗: p < 0.001; ∗∗∗∗: p < 0.0001.)

Immunofluorescence analysis revealed the strongest INOS expression in the MS group at 3 days post-surgery. The AMS@Sema3A and Control (autologous periosteal transplantation group) groups exhibited similar minimal INOS expression, while the MS@Sema3A group showed intermediate expression levels between AMS and AMS@Sema3A groups (Fig. 10B, Supplementary Fig. 9). In contrast, M2 marker Arg-1 expression was similar between the AMS@Sema3A and Control groups, with the remaining groups exhibiting a distribution pattern opposite to that of INOS (Fig. 10C and D). At 7 days post-surgery, Arg-1 expression in the AMS@Sema3A group resembled that of the Control group but was significantly higher than in other groups. The AMS group exhibited expression levels intermediate between the MS and MS@Sema3A groups (Fig. 10C and D). At this time point, INOS expression patterns showed a negative correlation with Arg-1 (Fig. 10B, Supplementary Fig. 9).

In summary, at the early stage of bone defects, orientation and Sema3A synergistically promoted M2 polarization and inhibited M1 expression, so that AMS@Sema3A demonstrated an optimal pro-M2 polarization effect.

2.4. Bone repair and anti-inflammatory evaluation

Histological analyses of rat skull specimens were used to evaluate the repair effect of bionic periosteum. H&E staining analysis revealed that the biomimetic periosteal material in the AMS@Sema3A group demonstrated optimal bone repair outcomes at 4- and 8-weeks post-surgery. The quantity, volume, and continuity of new bone formation in this group significantly outperformed those in other experimental groups and approached the levels observed in the control group. Quantitative analysis results further validated this finding (Fig. 10E, Supplementary Fig. 10).

Similarly, Masson staining confirmed that the AMS@Sema3A and Control groups exhibited the largest and most continuous red-stained areas of new bone tissue. Quantitative analysis revealed that the repair efficacy of the AMS@Sema3A group was approximately equivalent to that of the Control group and significantly superior to that of other groups (Fig. 10F and G). This indicates that the biomimetic fiber scaffold, similar to autologous periosteal transplantation, possesses optimal bone repair performance.

Macrophage polarization significantly regulates anti-inflammatory and osteogenic processes [66]. TNF-α, primarily secreted by M1 macrophages, is associated with persistent excess inflammatory responses and impedes the repair process [1]. Compared to the MS and AMS groups, the control and AMS@Sema3A groups exhibited the lowest TNF-α expression, followed by the MS@Sema3A group (Fig. 11A and B). This demonstrates that AMS@Sema3A suppresses local inflammatory factors during bone defect repair. Its potent anti-inflammatory effect creates a local microenvironment similar to that of autologous periosteal transplantation.

Fig. 11.

Evaluation of anti-inflammatory and osteogenic effects of biomimetic periosteum in vivo. (A) TNF-α immunofluorescence staining at 3 and 7 days. Overall image scale bar: 500 μm. Magnified section scale bar: 50 μm. (B) Comparison of TNF-α immunofluorescence intensity among groups at 3 and 7 days. (C) IL-10 immunofluorescence staining at 3 and 7 days. Scale bar for whole image: 500 μm. Scale bar for magnified section: 50 μm. (D) Comparison of IL-10 immunofluorescence intensity among groups at 3 and 7 days. (E) TGF-β immunofluorescence staining at 3 and 7 days. Overall image scale bar: 500 μm. Magnified image scale bar: 50 μm. (F) Comparison of TGF-β immunofluorescence intensity among groups at 3 and 7 days. (G) OCN immunohistochemical staining at 4 and 8 weeks. (H) 3D micro-CT reconstruction images of cranial defect regions across different groups. the image on the right shows a lateral view of the rat skull, with the skull defect location indicated by the dashed box. (I) Intergroup comparison of BMD at 4 and 8 weeks (n = 6). (J) Bone volume to total volume ratio (BV/TV) at 4 and 8 weeks (n = 6). (Statistical analysis required one-way/two-way ANOVA followed by Tukey's multiple comparison test; n = 5.)

M2 macrophages promote the osteogenic differentiation of MSCs through the secretion of the key anti-inflammatory factor IL-10 [[67], [68], [69]]. In vitro experiments have shown that M2 polarization promotes increased IL-10 expression, a phenomenon validated in vivo by tissue immunofluorescence. On day 3 post-surgery, IL-10 expression levels were low across all groups. The fluorescence intensity was higher in the Control and AMS@Sema3A groups than in the MS@Sema3A group, while the MS group exhibited the lowest fluorescence intensity, which was also lower than that in the AMS group. At 7 days post-surgery, IL-10 expression significantly increased in the AMS@Sema3A group, surpassing that in all other groups and approaching that in the Control group (Fig. 11C). Semi-quantitative analysis revealed consistent trends in fluorescence intensity changes at both 3- and 7-days post-surgery (Fig. 11D).

Transforming growth factor beta (TGF-β) is a key anti-inflammatory factor that modulates immune responses and cell proliferation and differentiation while promoting bone healing [[70], [71], [72]]. In vivo tissue immunofluorescence revealed low TGF-β expression levels across all groups at 3 days post-surgery. The AMS@Sema3A group exhibited the highest fluorescence intensity, comparable to the control group, followed by the MS@Sema3A group, while the MS group showed lower fluorescence intensity than the AMS group. By day 7 post-surgery, TGF-β expression in the AMS@Sema3A group approached that in the Control group, significantly increasing from that in other groups (Fig. 11E). Semi-quantitative analysis revealed consistent trends in fluorescence intensity changes between 3- and 7-days post-surgery (Fig. 11F). These results indicate that, similar to autologous periosteal transplantation, the AMS@Sema3A membrane not only effectively controls local inflammatory responses, but also holds potential for establishing a microenvironment conducive to bone repair. In summary, these findings strongly support that the AMS@Sema3A biomimetic periosteum exhibits properties similar to autologous periosteal transplantation in vivo. It effectively suppresses inflammatory factors, upregulates M2 macrophages and anti-inflammatory factors, promotes timely transition from inflammation to proliferative repair, and creates a favorable environment for subsequent bone repair and remodeling.

Osteocalcin OCN, as a marker of mid-to late-stage osteogenic differentiation, is not only the most abundant non-collagenous protein component in bone tissue, but also plays a key role in regulating bone matrix mineralization and systemic metabolic homeostasis [52]. Immunohistochemical results showed that the OCN expression levels in the AMS group at 4- and 8-weeks post-surgery were between those in the MS and MS@Sema3A groups, while OCN expression in the AMS@Sema3A group was similar to that in the Control group and significantly better than that in the other experimental groups (Fig. 11G, Supplementary Fig. 11). These data suggest that both material orientation and Sema3A can promote bone repair. Their synergistic effect renders the AMS@Sema3A biomimetic periosteum similar to autologous periosteum, thereby exhibiting the most outstanding bone repair performance.

Micro CT scan reconstruction analyses of rat skull specimens at 4- and 8-weeks post-surgery were used to assess the effects of bone repair in each group. CT reconstruction images revealed that bone mineral density (BMD) and bone volume-to-total volume ratio (BV/TV) in the AMS group were between those in the MS and MS@Sema3A groups at both 4- and 8-weeks post-surgery (Fig. 11H), while the AMS@Sema3A group exhibited values similar to those in the Control group (Fig. 11I and J). Topography orientation may enhance the repair effect of the AMS group by promoting BMSC adhesion and mineralization [24]. Meanwhile, the osteoprotective effect of Sema3A further improves the repair advantage of the AMS@Sema3A group.

Periostin proteins specifically expressed by the periosteum play a key regulatory role in bone tissue remodeling and damage repair [73]. Immunohistochemical analysis confirmed that the AMS@Sema3A group exhibited the most pronounced enhancement and linear arrangement of periosteal protein staining similar to the Control group (Fig. 12A and B). Its expression intensity progressively exceeded that of the MS@Sema3A, AMS, and MS groups. This gradient pattern clearly demonstrates that this biomimetic periosteum possesses optimal performance in promoting periosteal repair, comparable to autologous periosteal transplantation.

Fig. 12.

In vivo study of bionic periosteum regulating cytoskeletal organization to promote mitochondrial homeostasis. (A) Immunohistochemical staining of periostin tissue at 4 and 8 weeks. Overall image scale bar: 500 μm. Magnified section scale bar: 50 μm. (B) Statistical analysis of periostin tissue immunohistochemical staining at 4 and 8 weeks. (C) Immunofluorescence colocalization of Runx2 and Mfn2 at 3 and 7 days across groups. Overall image scale bar: 500 μm. Magnified image scale bar: 50 μm. (D) Comparison of Runx2 tissue immunofluorescence intensity at 3 and 7 days. (E) Comparison of Mfn2 tissue immunofluorescence intensity at 3 and 7 days. (F) Immunofluorescence staining for Drp1 at 3 and 7 days. Scale bar for whole image: 500 μm. Scale bar for magnified image: 50 μm. (G) Immunofluorescence intensity of Drp1 at 3 and 7 days. (H) Immunofluorescence intensity of ROCK2 at 3 and 7 days. (I) ROCK2 immunohistochemical staining at 3 days and 7 days. Scale bar for whole image: 500 μm. Scale bar for magnified section: 50 μm. (J) Macroscopic photograph of rat cranial defect implanted with biomimetic periosteum. (K) Macroscopic photograph showing biomimetic periosteum promoting rat cranial defect repair. (n = 5 per group; statistical analysis performed using two-way ANOVA with Tukey's multiple comparison test; ns indicates no statistical significance; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

In summary, similar to autologous periosteal transplantation, AMS@Sema3A biomimetic periosteum exhibits optimal osteogenic, anti-inflammatory, and periosteal regeneration properties, making it an ideal periosteal repair scaffold.

2.4.1. Bionic periosteum promotes mitochondrial fusion and enhances osteogenesis by modulating the cytoskeleton

In vitro experiments confirmed that the bionic periosteum retains sustained osteogenic induction activity; however, the consistency of its mitochondrial dynamics-mediated osteogenic effect both in vivo and ex vivo still needs to be further verified. Accordingly, we first examined Mfn2 expression in vivo by tissue immunofluorescence to verify the in vivo efficacy of this osteochondral membrane in regulating mitochondrial fusion.

Rat skull specimens were collected on 3- and 7-days post-surgery for the co-localization of Runx2 and Mfn2 by immunofluorescence (Fig. 12C). Quantitative analysis showed that, compared with that in the MS group, Runx2 expression in the remaining three groups increased on day 3 and reached a significant difference on day 7 (Fig. 12D). Specifically, on day 3, Runx2 expression in the AMS group was between that in the MS and MS@Sema3A groups, with the highest expression levels observed in the AMS@Sema3A and Control groups. By day 7, expression intensity maintained the same gradient distribution across groups, with the AMS@Sema3A and Control groups sustaining the highest expression levels. The AMS group had significantly higher and lower expression than the MS and MS@Sema3A groups, respectively. Furthermore, quantitative analysis of Mfn2 expression across all four groups revealed low expression levels at 3 days post-surgery. The MS group exhibited the lowest fluorescence intensity, while the MS@Sema3A group showed levels between those in the AMS and AMS@Sema3A groups. The AMS@Sema3A and Control groups demonstrated the highest fluorescence intensities. At 7 days post-surgery, Mfn2 expression in the AMS@Sema3A and Control groups was significantly superior to that in other groups. Semi-quantitative analysis confirmed consistent trends in Mfn2 fluorescence intensity changes at both 3- and 7-days post-surgery (Fig. 12E).

Immunofluorescence detection of the mitochondrial fission marker Drp1 (Fig. 12F) revealed that the strongest expression in the MS group on postoperative day 3, with the MS@Sema3A group exhibiting expression levels between those in the AMS and AMS@Sema3A groups. The AMS@Sema3A and Control groups showed the lowest fluorescence intensity. The expression gradient remained stable at day 7 post-surgery, with the AMS@Sema3A and Control groups maintaining the lowest expression levels. Semi-quantitative analysis confirmed consistency in this expression pattern at both 3- and 7-days post-surgery (Fig. 12G). Thus, AMS@Sema3A promotes osteogenic differentiation by regulating mitochondrial homeostasis, upregulating the osteogenic marker Runx2 and the mitochondrial fusion protein Mfn2, while downregulating the mitotic marker Drp1, demonstrating a mechanism analogous to autologous periosteal transplantation.

To verify the synergistic changes of F-actin cytoskeleton remodeling and mitochondrial fusion, tissue specimens were subjected to ROCK2 immunohistochemical assay at 3- and 7-days post-surgery (Fig. 12H). Immunohistochemical results at 3-days post-surgery showed that ROCK2 expression in the AMS@Sema3A and Control groups were significantly higher than that in other groups and the lowest in the MS group, while its expression in the AMS group was between that in the MS and MS@Sema3A groups. At 7 days post-surgery, ROCK2 expression level increased in all groups but maintained the same expression gradient, and the highest expression was maintained in the AMS@Sema3A and Control groups (Fig. 12I), indicating that this treatment had the most significant effect on F-actin cytoskeletal rearrangement and was consistent with the in vitro results.

In vitro experiments demonstrated that the AMS@Sema3A biomimetic periosteum significantly promoted the osteogenic differentiation of BMSCs compared to other periostea. To investigate whether this biomaterial exhibits similar osteogenic efficacy in vivo, we applied AMS@Sema3A to critical-sized defects in rat skulls. Results revealed effective bone repair at the defect site within 8 weeks (Fig. 12J and K). Compared to direct administration of Sema3A at the defect site, the AMS@Sema3A biomimetic periosteum induced timely M1 to M2 polarization of macrophages around the cranial defect through controlled release of Sema3A. This sequential regulation of the immune microenvironment suppressed inflammation, promoted osteogenesis, and aligned with the progressive bone repair course. Furthermore, Sema3A effectively promotes immune-mediated osteogenesis and the oriented topographical advantage enhances BMSCs adhesion, migration, and immune regulation to facilitate bone formation. The synergistic effect of both components enables the AMS@Sema3A biomimetic periosteum to achieve osteogenic promotion through immune regulation comparable to that of autologous periosteum. It also avoids the pain and secondary injury associated with autologous harvesting, demonstrating promising clinical application prospects. In summary, AMS@Sema3A bionic periosteum mimics the characteristics of natural periosteum through its oriented topology, and has dual functions of immunomodulation, osteogenesis, and anti-inflammation. The mechanism involves the ‌synergistic action‌ of controlled Sema3A release and topographical guidance to induce M2 polarization of the immune microenvironment, while simultaneously promoting mitochondrial fusion through F-actin cytoskeleton reconstruction in BMSCs, thereby driving osteogenic differentiation and ultimately achieving efficient bone regeneration.

2.4.2. In vivo biocompatibility of the bionic periosteum

To evaluate the in vivo biocompatibility of the materials, they were implanted subcutaneously in the backs of SD rats for up to 8 weeks. During the observation period, all rats remained healthy with no surgical-related infections or other complications. H&E staining (Fig. 13A) revealed inflammatory cell infiltration around the implanted materials at 4 weeks. The MS group exhibited the highest number of inflammatory cells, while the MS@Sema3A group showed an intermediate level between the AMS and AMS@Sema3A groups. H&E staining of soft tissue surrounding implants harvested at 8 weeks post-surgery revealed the lowest inflammatory cell count in the AMS@Sema3A group, followed by that in the MS@Sema3A group, with the AMS group exhibiting an intermediate inflammatory cell count between the MS (highest count) and MS@Sema3A groups (Fig. 13B). This indicates that both orientation and Sema3A possess anti-inflammatory properties, resulting in the lowest inflammatory cell count in the AMS@Sema3A group. The anti-inflammatory effects of both factors are likely related to their ability to promote macrophage M2 polarization and inhibit inflammatory cytokines, further confirming the in vivo anti-inflammatory efficacy of this biomimetic periosteum. Furthermore, similar to the control group, histological staining of major organs at 8 weeks post-implantation showed no significant pathological or toxic changes across all experimental groups, further confirming the in vivo safety of the biomimetic periosteal scaffold (Fig. 13C).

Fig. 13.

In vivo biocompatibility study of the biomimetic periosteum. (A) H&E staining of subcutaneous tissue containing the material at 4 and 8 weeks in rats. Scale bar: 50 μm. (B) Comparison of inflammatory cells surrounding each material group in the subcutaneous implant model at 4 and 8 weeks in rats. (C) H&E staining of major organs in rats at 8 weeks. (n = 5 per group; statistical analysis performed using two-way ANOVA with Tukey's multiple comparison test; ns indicates no statistical significance, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Although current artificial periostea used for tissue engineering are structurally diversified, their functions still have limitations. Despite the increasing number of studies promoting osteogenesis through immunomodulation, these systems cannot actively regulate cytoskeletal and mitochondrial dynamics. Unlike other bone tissue engineering scaffolds, this biomimetic periosteum reconfigures the local immune microenvironment through biological and physical factors, modulates the BMSC cytoskeleton to enhance mitochondrial fusion, thereby significantly boosting osteogenic differentiation efficiency. This cascade regulation achieves highly efficient bone regeneration comparable to that of native periosteum. Oriented microsol electrostatic spinning was used to stabilize the loading and sustain slow release of Sema3A with appropriate stiffness support and topological advantages. Microsol electrostatic spinning-loaded Sema3A induces macrophage M2 polarization promptly through early burst release, which establishes an anti-inflammatory-promoting regenerative microenvironment and inhibits the inflammatory damage associated with M1 polarization, and its reconstituted immune homeostasis drives osteogenic differentiation. In addition, the released Sema3A could rearrange the BMSC cytoskeleton and maintain mitochondrial fusion to promote osteogenesis. Therefore, bionic periosteum can precisely and chronologically regulate the local immune microenvironment, rearrange the cytoskeleton, and induce osteogenic differentiation through the mitochondrial kinetic cascade of BMSCs for efficient bone regeneration using the synergistic axis of immune–cytoskeletal–mitochondrial kinetic, which should solve the regeneration bottleneck of bone defects in the clinic.

3. Discussion

Bone regeneration is a complex process involving endocrine, immune, neural, cellular, and cytokine factors, and neural and immune crosstalk cannot explain the whole osteogenesis process [74]. Shifting from the traditional single osteogenic/vascularization strategy, this study focuses on the crosstalk between cytoskeleton and mitochondrial dynamics in immune-induced osteogenesis and develops a novel biomimetic periosteum to coordinate the local immune microenvironment and BMSC cytoskeleton, modulate mitochondrial dynamics, and drive efficient bone regeneration.

Through in vitro experiments, this study demonstrated that Sema3A-loaded oriented bionic periosteum promotes macrophage M2 polarization through the PI3K/Akt/mTOR pathway. In addition, in the local immune microenvironment, Sema3A co-oriented spinning topology activated the RhoA/ROCK2 pathway, drove F-actin rearrangement, and promoted mitochondrial fusion in BMSCs, which synergistically achieved efficient osteogenesis. Mechanistic studies have shown that biomimetic periosteum significantly enhances osteogenic potency by regulating mitochondrial dynamics, namely promoting mitochondrial fusion and inhibiting mitochondrial fission. Within the body, this bionic periosteum not only exhibits excellent biocompatibility and safety while promoting periosteal regeneration, but also possesses cascade-like regulatory capabilities similar to those in vitro to enhance osteogenesis. The fibrous scaffold effectively facilitates the timely M1 to M2 polarization of macrophages, suppresses the inflammatory cytokine TNF-α, and promotes the expression of anti-inflammatory factors IL-10 and TGF-β, thereby creating a highly efficient anti-inflammatory and repair-promoting local immune microenvironment. Furthermore, the biomimetic periosteum enhances the expression of periosteal proteins, influences the cytoskeleton, and increases ROCK2 protein expression. By upregulating mitochondrial fusion protein Mfn2 expression and downregulating drp1 levels, it maintains mitochondrial dynamic homeostasis to elevate the osteogenic marker Runx2 and promote osteocalcin (OCN) expression, thereby enhancing bone mineral density and bone mass. In summary, through the synergistic action of bioactive factor Sema3A and oriented architecture, the biomimetic periosteum scaffold effectively suppresses inflammation, modulates the local immune microenvironment at the defect site, influences the cytoskeleton, maintains mitochondrial dynamic homeostasis, and promotes periosteal regeneration. This achieves an efficient osteogenesis comparable to autologous periosteal transplantation.

In addition, the periosteum achieves functional innovation by synergistically integrating the oriented topology with osteoprotective bioactive factors. First, the biomimetic scaffold is constructed using widely used and well-biocompatible PLLA, with carbon dioxide and water as end-products, avoiding toxicity caused by scaffold raw materials [23]. Much like the ECM, electrostatic spinning has high porosity and specific surface area and is now widely used in osteochondral tissue engineering [7,10,75]. Microsol electrostatic spinning has become a preferred drug delivery platform for spinal cord repair and periosteal tissue engineering owing to its device simplicity, ultra-high drug loading rate, and long-lasting release properties [22,23,[76], [77], [78], [79]]. Previous studies have demonstrated that oriented nanofibers guide the morphological remodeling of BMSCs, enhance migration, and efficiently induce osteogenesis [24]. Oriented spun scaffolds have relatively low porosity, high stiffness, and low water contact angle than random fibers [24], and we have also found that the topology is anti-inflammatory and promotes osteogenesis by promoting macrophage polarization towards the M2 phenotype, which further corroborates the results of previous studies [25,26].

Sema3A is a multifunctional neural axon rejection protein that is a key regulator of morphogenesis and homeostasis in various organ systems and maintains bone homeostasis by reducing bone resorption and enhancing bone formation [80,81]. Notably, neural-derived Sema3A has been used in bone tissue engineering as it promotes osteogenesis [[82], [83], [84], [85]], while recombinant Sema3A promotes osteogenesis and analgesia [32,86]. Previous studies have demonstrated that Sema3A has significant osteogenic capacity and promotes macrophage polarization towards the M2 phenotype to exert anti-inflammatory and pro-regenerative effects [31,32]. Therefore, we further explored the mechanism by which Sema3A promotes osteogenesis through immunity. Sema3A promotes alternative macrophage activation in inflammation through PI3K/AKT/mTOR signaling to alleviate periodontitis [31], and we found that Sema3A promotes bone regeneration by facilitating macrophage M2 polarization through the PI3K/Akt/mTOR pathway. Therefore, we used oriented electrostatic spinning as a substrate for the bionic periosteum for the temporal modulation of macrophage M2 polarization to promote osteogenesis by micro-solvent-controlled release of Sema3A.

This study also explored the topological topography and downstream mechanisms by which Sema3A modulates the immunopromotion of osteogenesis. Previous studies have found that changes in fiber patterning induced changes in cell morphology and resulted in different stresses on the cytoskeleton, while actin filaments in the cytoplasm were aligned along the stress direction [87,88]. Orientation spinning promotes cell adhesion, cell elongation, and contact guidance. Actin filaments follow contact guidance along the fiber direction, while cytoskeletal alignment contributes to differentiation onset [89]. Topography-induced MSC differentiation and morphological changes are dependent on cytoskeletal tension [90]. Furthermore, under mechanical loading, Sema3A regulates the cytoskeleton of human periodontal ligament cells (hPDLCs) [20]. Our results reveal that oriented scaffolds induce spindle-shaped morphology and parallel actin arrangement in BMSCs, while Sema3A increases the aspect ratio and promotes cell elongation. Moreover, oriented scaffolds enhance bone regeneration by upregulating ROCK2 expression in BMSCs, and Sema3A synergistically amplifies this pathway to boost osteogenic potential.

Mitochondria are highly dynamic organelles that constantly undergo fission and fusion, and disruption of mitochondrial dynamics affects their function [61]. F-actin is widely distributed throughout the cell, forming a series of cytoskeletal structures with varying functions, whereas the cytoskeleton and associated proteins play a crucial role in regulating mitochondrial dynamics, organization, and function [91]. Mitochondrial fission begins with actin/non-actinomyosin II-driven pre-contraction, where the F-actin cloud co-localizes with the contractile site and induces random deformation; terminal cleavage is mediated and executed by Drp1 [92,93]. Actin polymerization, dependent on the actin nucleating protein complex Arp2/3, plays a crucial role in regulating mitochondrial fusion [94]. Sema3A induces growth cone collapse by causing F-actin reorganization in neurons through the RhoA/ROCK2 pathway [95]. In addition, Sema3A affects CD8⁺ T cell F-actin, inhibiting immune synapse formation and motility [96]. Sema3A promotes the osteogenic differentiation of hPDLCs by upregulating ROCK2 in a stress microenvironment, thereby stabilizing the cytoskeleton and maintaining mitochondrial fusion [20]. We found that orientation topography induced BMSC skeleton elongation and mitochondrial fusion to promote osteogenesis, which was possibly related to the relative increase in material stiffness [28]. Sema3A upregulates ROCK2 protein in BMSCs, increases the aspect ratio of BMSCs, and promotes mitochondrial fusion to enhance osteogenesis. In summary, we demonstrate that in an immune environment, both Sema3A and orientation can modulate the cytoskeleton, promote mitochondrial fusion, and enhance osteogenic potential.

4. Conclusions

In this study, we constructed a cascade-regulated bionic periosteum that reprogrammed the local early immune microenvironment of injury through the integration of topographical dominance and controlled release of Sema3A, which in turn regulated the cytoskeleton and promoted mitochondrial fusion of BMSCs to enhance osteogenesis. The biomimetic periosteum promoted M2 macrophage polarization through its topographical features, guided directional alignment of BMSCs, and enhanced mitochondrial fusion to facilitate bone formation. Surprisingly, Sema3A enhanced orientation-induced cytoskeletal elongation and promoted mitochondrial fusion in BMSCs to enhance osteogenesis. The microsol electrospun fibers loaded with Sema3A mimiced the effects of natural periosteum on cell adhesion, proliferation, and differentiation through the orientation topology; induced osteogenic differentiation of BMSCs; and enhanced bone regeneration by releasing Sema3A to modulate the RhoA/ROCK2 pathway, mediating F-actin rearrangement of BMSCs on the oriented spun filaments, and regulating mitochondrial dynamics. In vivo experiments confirmed that the biomimetic periosteum exhibits excellent biocompatibility and safety as well as promotes periosteal regeneration. More importantly, in vivo validation confirmed the highly efficient immune-mediated osteogenic mitochondrial kinetic homeostasis regulation pathway exhibited by this biomimetic periosteum in vitro. In conclusion, the composite fibrous scaffolds constructed in this study induced chronotropic immune reprogramming, which subtly cross-talked the immune microenvironment and cytoskeleton as well as mitochondrial dynamics to promote osteogenesis through the combination of topographical structures and bioactive factors, thus providing a biomaterial therapeutic strategy that effectively promotes bone repair.

5. Experimental part

Preparation of electrostatically spun fibres: At room temperature, we added 0.1 g of hyaluronic acid (HA, Yuancheng, China) to 9.9 g of deionised water and stirred it well to obtain a 1 wt% aqueous solution of HA. 10 μl of Sema3A (25 ng/ul) (R&D, USA) was mixed thoroughly with 50 μl of 1 wt% HA solution to obtain a homogeneous 1% HA-Sema3A aqueous solution. Then, 0.01 g Span 80 (Sigma, USA) and 4 g dichloromethane (DCM, Aladdin, China) were mixed and stirred well at room temperature, and 0.5 g Poly-L-Lactic Acid (PLLA, Mv = 100,000 Da,Idle Gang, China) and 2 g N, N-dimethylformamide (DMF, Qiangshun, China) were mixed and stirred well to obtain the microsolvent electrostatic spinning solution. The preparation of different electrostatic spinning membranes was carried out as in the previous literature [22,23]. Briefly, the prepared microsol electrostatic spinning solution was first aspirated with a 10 ml syringe, a flat-tipped steel needle with an inner diameter of 0.9 mm was mounted and its end was securely clamped with a metal clip from a DC high-voltage power supply (Dongwen, China), whereas the spinning receiver device was placed at a distance of 15-20 cm from the tip of the needle for the collection of electrospun fibres. The metal clip at the other end of the high-voltage power supply was connected to a grounded spinning receiver device, such as a cylindrical drum receiver wrapped with aluminium foil (rotating at 120 rpm) for collecting random microsol electrostatic spinning fibres (MS) or between electrode rods placed parallel to each other for collecting orientation microsol electrostatic spinning fibres (AMS), and in the same way, Sema3A-carrying orientation microsol electrostatic spinning fibres (AMS@Sema3A) were prepared and Sema3A-loaded random microsol electrostatic spinning (MS@Sema3A). A precision pump (LSP02-1B, Lange Precision Pump, China) was used to control the injection rate of 60 μ L/min and the voltage was 15 kV. The ambient environment required that the ambient temperature should be controlled at 20-25 °C, the humidity should be stabilised at 30-50%, and the fluctuation of temperature and humidity should be <±2 °C/±5%. The prepared electrostatically spun fibres were dried in a vacuum oven for 72 h, and the residual organic solvent was removed and prepared for use.

Material characterization: A Dynamic light scattering particle size analyzer (Malvern, UK) was used to detect the diameter distribution of hyaluronic acid (HA) microspheres encapsulating Sema3A.Different groups of electrospun fibers were fixed on copper plates using conductive glue and examined using SEM (Hitachi, Japan). Under an acceleration voltage of 10 kV, the morphological characteristics, fiber diameter, and orientation differences of different materials were observed. In HSB mode, colour saturation and luminance indicate coherence and source image, respectively, while different shades of the image correspond to different degrees of orientation of the fibres [33]. We used the OrientationJ Distribution tool to quantify the percentage of different groups of electrospun fibres in each angular orientation. The degree of orientation of fibres within ±θ angle is equivalent to the area of the curve peaks within the corresponding angle range. In this study, we mainly counted the degree of fibre orientation within ±10°, which was calculated as follows:

$$y_1{\int }_{-10}^{10}\beta d\theta$$ (1)

Among them, β is the fiber orientation ratio in the θ direction. TEM (Hitachi, Japan) was used to evaluate the core-shell structure formed inside micro-sol gel electrospun fibers by micro-sol gel particles stretched under a high-voltage electric field. BSA labeled with fluorescein isothiocyanate (FITC) (Solarbio, China) was used to replace Sema3A to observe the distribution of cytokines encapsulated in microcapsules during spinning. FTIR (ThermoScientific, USA) was used to evaluate the chemical composition of the spinning fibers in each group. A water contact angle meter (KRUSS, Germany) was used to compare the hydrophilicity of different fiber scaffolds. Fibers from different groups were cut into fixed sizes (15 × 3 × 0.1 mm) and placed in a mechanical testing machine (Hengyi, China) to test the tensile mechanical properties of the samples. The results were plotted as stress-strain curves using Origin software, and then the tensile stress-strain curves of the fibres were obtained. The Young's modulus (E) of different samples was calculated from the tensile stress-strain curve, and the stiffness (k) of different samples was calculated using formula (2). Finally, the differences in Young's modulus and stiffness of different groups of materials were compared.

$$\mathrm{k}=\frac{E\times w\times t}{L}$$ (2)

Among them, E, w, t, and L represent the Young's modulus, width, thickness, and length of the fiber, respectively. The porosity and degradation rate of the materials were used to compare the basic properties of different groups of materials as periosteal tissue engineering scaffolds. ELISA was used to evaluate the release pattern and temporal efficacy of Sema3A in drug-loaded micro-soluble electrospun fibers.

Preparation of in vitro materials and cell culture: The groups of materials used for cell culture were first prepared as follows: circular cell crawlers with a diameter of 14 mm and a thickness of 100 μm were used to collect different groups of electrospun fibres. The cell crawler sheets with collected electrospun fibres were dried in a vacuum oven for 72 h and then sterilized in a 75% ethanol solution for half an hour. They were rinsed three times with an appropriate amount of PBS to remove the residual ethanol, placed in 24-well plates, and sterilized by UV irradiation for half an hour, then soaked in culture medium for 1 h and kept as a reserve. Rat BMSCs were obtained by extracting from the bone marrow cavity of rat femur and tibia, isolated and cultured, and Raw267.4 cells were purchased from Wuhan Pusai Life Science and Technology Co. The average concentration of cells was obtained by digesting the cells in a Petri dish and counting them under a light microscope three times. In different experiments, appropriate amounts of cells were inoculated on different groups of electrospun fibres, and then appropriate amounts of the corresponding medium were added and placed in a cell culture incubator for cultivation, in which the medium was changed every 2- 3 days. The conditions of the cell culture incubator were set at a temperature of 37°C, relative humidity of 95% and CO2 content of 5%.

Cell proliferation: RAW264.7 cells were inoculated in 96-well plates at a concentration of 2 × 10³ per well, and stimulated with different concentrations of Sema3A (0, 25, 50, 100, and 200 ng/mL) for 0 h, 6 h, 24 h, and 48 h, and 10 μL of Cell Counting Kit 8 Reagent (CCK-8, Dojindo, Japan) were added to each well. In the dark after incubation for 2 h, the 96-well plates were placed in an enzyme labeller (BioTek, UK) and the absorbance was measured at 450 nm. Place cell monolayers containing different groups of electrospun fibers into a 24-well plate. The control group consists of cell monolayers without electrospun fibers. BMSCs with a concentration of 1 × 10⁴ were seeded in each well of the culture plate. On days 1, 3, 5, and 7 of culture, the culture medium was replaced with a mixture of 100 μL of CCK8 reagent (Dojindo, Japan) and 900 μL of culture medium. After 3 h of incubation in a cell incubator, 100 μL of the mixture was extracted and added to a 96-well plate. Then the absorbance at 450 nm was measured by an enzyme marker. The cell proliferation curves were plotted separately. In this experiment, 3 replicates were used for each group.

Live/dead staining: In a 24-well plate, cells were seeded onto cell crawls containing electrospun fibers of different groups at appropriate concentrations of BMSCs or Raw267.4 cells and incubated with the corresponding culture medium for 3 days. After removing the medium, the cells were washed three times with pre-warmed PBS at 37°C. After preparing the cell live/dead staining working solution according to the instructions of the live/dead staining kit (Solarbio, China), 300 μL of the live/dead staining working solution was added to each well, and the cells were incubated at room temperature and protected from light for 30 min. After removing the working solution, the cell crawls were placed under an inverted fluorescence microscope (Zeiss, Germany) for observation. Live cells were labeled with green fluorescence and dead cells with red fluorescence, which were analyzed for live cell fluorescence intensity using ImageJ software.

Cell Scratching Assay: Cultures from different groups of materials and Mφ after 3 days of incubation were collected, centrifuged, and purified to prepare CM (Fig. S5).In 6-well plates, each well was inoculated with BMSCs at a cell density of 1 × 10⁶ and first incubated with basal medium for 48 h. The bottom of each well was scratched with a 200 μL pipette tip. Each well was washed 3 times with PBS and incubated for 48 h with the addition of CM as described above. The cells were observed with an inverted microscope (Zeiss, Germany), and images were taken at the desired time points (0, 24, 48 h) to monitor the migration of the cells towards the scratched area.

Cell immunofluorescence staining: Immunofluorescence staining was used to detect the expression of proteins related to bone marrow mesenchymal stem cells and macrophages. Co-culture each group of materials with Raw267.4 cells and incubate them for 3 days and 7 days, respectively, to study the temporal changes in macrophage polarization status. When studying the role of immune factors, use CM from Raw267.4 cells and different groups of materials co-cultured for 3 days instead of BMSCs' basal medium. BMSCs of appropriate concentration were seeded on electrospun fibers in different groups and cultured in CM for 24 h to study cell adhesion, for 3 days to study BMSCs mitochondrial dynamics and cytoskeleton, and for 7 days to study osteogenic differentiation. Collect cells and fix them with 4% paraformaldehyde at room temperature for 30 min. Then permeate the cells with 0.5% Triton X-100, incubate at room temperature for 25 min, and block with 5% BSA at 4°C overnight. Remove BSA, then add the following primary antibodies: INOS (ab178945, Abcam, USA), CD206 (ab64693, Abcam, USA), Integrinβ1 (A21234, Abclonal, USA), Vinculin (ab129002, Abcam, USA), Tom20 (11802-1-AP, Proteintech, USA), Rock2 (ER1706-48, Huabio, China), Runx2 (ab76956, Abcam, USA), Drp1 (ab184247, Abcam, USA), Mfn2 (ab124773, Abcam, USA), OCN (A6205, Abclonal, USA)), followed by incubation at 4°C overnight. Then, add the appropriate secondary antibody (ab150084, Abcam, USA), FITC Phalloidin (40735ES75, Yeasen, China), and DAPI (40728ES03, Yeasen, China), and incubate at 37°C for 2 h. Cells were washed three times with PBS before proceeding to the next step, with each wash lasting 15 min. Cell immunofluorescence was observed using a confocal fluorescence microscope (Leica, Germany), and images were captured. Semi-quantitative analysis was performed using ImageJ.

Flow analysis of Mφ polarization: After the appropriate number of Raw267.4 cells were co-cultured with different groups of electrospun scaffolds for 3 and 7 days, respectively, the cells were collected and examined by flow cytometry (BD, Canto II, USA). Cells were collected, and it was ensured that the remaining cells on the scaffolds were less than 20% in 5 random fields of view. In summary, the collected cells were transferred to 1.5 ml Eppendorf (EP) tubes and centrifuged at 4°C, 2000 rpm for 5 min before sequentially adding anti-CD11b-pecy7 (561098, BD Pharmingen, USA), anti-CD86-apc (561964, BD Pharmingen, USA), and anti-CD206-pe (565250, BD Pharmingen, USA). Labeled cells were incubated at 4°C for 30 min, and the number of scaffold cells in each group was detected by flow cytometry to be greater than 1 × 10⁴. Cells were analyzed using FlowJo V10.8.1 software, and the cells were first surrounded by CD11b, followed by CD86 and CD206 in that order. Statistical analyses of the cell proportions were carried out using Graphpad Prism 9.0, and statistical plots were generated.

qRT-PCR: After co-culturing Raw267.4 cells with different groups of materials for 3 and 7 days, the expression levels of macrophage polarization-related genes Mrc1, Il10, Nos2, and Il1b were detected by qRT-PCR. For the osteogenic differentiation experiment, 2 × 10⁴ BMSCs were seeded onto the surface of the materials in a 24-well plate and cultured in CM for 7 days before detecting the expression of Col1a1 and Alpl genes. After 10 days of culture, the expression of Runx2 and Bglap genes was detected. Total RNA was extracted from the cells using an RNA extraction kit (Vazyme, China), and after concentration and purity testing, the concentration of all samples was standardized using EDTA. Reverse transcription was performed at 70°C in a metal bath for 5 min. After the reaction was terminated, 100 μL of nuclease-free water was added to dilute the cDNA. A qPCR reaction system was prepared for target gene expression analysis. Gapdh and Actb were used as internal reference genes for macrophage experiments and osteogenic experiments, respectively. All primer sequences were designed using PubMed and synthesized by Shanghai Generay Company (Tables S1–2).

Western Blot: After co-culturing with different groups of materials for 7 days, Raw267.4 cells were collected for WB detection of Mφ polarization and pathway-related protein expression. Similarly, BMSCs at an appropriate concentration were seeded on different groups of materials and cultured in CM for 14 days, after which the expression of osteogenesis-related proteins was detected by WB. In brief, total protein was extracted using RIPA lysis buffer (Beyotime, China), and the lysis products were quantified using a bicinchoninic acid (BCA) kit (Solarbio, China). Subsequently, 20 μg of protein was loaded onto a 10% SDS-PAGE gel (Vazyme, China), followed by electrophoresis, membrane transfer, and blocking. The membranes were then incubated overnight with the following primary antibodies: p-PI3K (17366, CST, USA), PI3K (4249, CST, USA), p-AKT (4060, CST, USA), AKT (9272, CST, USA), mTOR (21214, SAB, USA), p-mTOR (11221, SAB, USA), OPN (ab63856, Abcam, USA), OCN (ab93876, Abcam, USA), RUNX2 (ab76956, Abcam, USA), and GAPDH (41549, SAB, USA). The membrane was then incubated with a secondary antibody solution labeled with horseradish peroxidase (HRP), and protein bands were detected using an enhanced chemiluminescence (ECL) kit (Vazyme, China), followed by scanning with a gel imaging system (shstbiosystems, China).

Osteogenic differentiation: Materials from different groups were inoculated with appropriate concentrations of BMSCs and cultured in CM for 7 days before ALP staining and quantitative analysis using Alkaline Phosphatase Colouring Solution Kit (Beyotime, China) and Alkaline Phosphatase Activity Quantification Kit (Built, China). Next, after 21 days of culture in CM, ARS staining was performed with Alizarin Red Staining Kit (Cyagen, China). Calcium nodules were observed under a light microscope, and images were captured, and then their absorbance was measured at 562 nm on an enzyme marker after dissolving the calcium nodules with perchloric acid.

Observation of cell morphology: We observed the cell morphology on the fibrous membrane by SEM(Hitachi, Japan). 24-well plates were inoculated with appropriate amounts of BMSCs or Raw267.4 cells on different groups of materials. After 2 days of culture, the culture medium was aspirated, washed three times with PBS, and then 4% paraformaldehyde was added. Then the plates were wrapped with aluminium foil and fixed in a refrigerator at 4 °C for 2.5 h. Next, the paraformaldehyde was aspirated and washed three times with PBS, and then the cells were dehydrated with a gradient of ethanol (30%, 50%, 75%, 80%, 95%, and 100%) twice, and the solution was changed every 30 min. The dehydration-treated cells were freeze-dried and processed for SEM observation at an accelerating voltage of 15 kV.

Preparation of experimental animals: SD rats were purchased from Zhaoyan (Suzhou) New Drug Research Centre Co. All rats were males, with an average weight of 200-220 g. The rats were used for the experiments. All surgical operations and experimental treatments involving experimental animals were approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (Luncheon No. SUDA20241113A0) during this experiment.

Construction of rat skull defect model and implantation of bionic periosteum: Firstly, SD rats were anaesthetised with 2% pentobarbital 2.5 mL/kg, and after the rats were completely anaesthetised, the cranial parietal hair of the rats was scraped off, and the skin incision was sterilized. A median incision of approximately 2.5 cm in length was made along the longitudinal axis of the skull, and the skin and fascia were incised layer by layer, and the periosteum was peeled off until the skull was exposed. A circular critical-size bone defect was created on the right side of the skull using a 5-mm-diameter dental drill. After cleaning and hemostasis, sterilized materials from different groups were placed over the defect area. The incision was sutured layer by layer and disinfected with povidone-iodine. The control group underwent autologous periosteal grafting, which involved elevating the periosteum in the area to one side prior to drilling and repositioning it over the defect site after drilling and removal of the bone fragment. To prevent the occurrence of infectious complications, 400,000 units of penicillin were injected intramuscularly per rat per day for 3 consecutive days from the day of surgery.

Collection of animal specimens: SD rats were euthanized at 3 days, 7 days, 4 weeks, and 8 weeks after surgery. Cranial specimens from each group were collected and fixed in 10% formalin for 24 h. Specimens from 3 to 7 days postoperatively were used to analyse the fluorescence expression of Mφ polarization indexes INOS, Arg-1, inflammation and anti-inflammation-related index TNF-α, IL-10, TGF-β, cytoskeletal ROCK2, mitochondrial dynamics Drp1 immunofluorescence, and immunofluorescence co-localization of Runx2 and Mfn2; use samples taken at four and eight weeks post-surgery to evaluate the expression of OCN and Periostin proteins and the repair of bone defects.

Micro-CT analysis: The cranial samples of rats in each group at 4 and 8 weeks after surgery were collected, scanned, and processed using Micro-CT (SkyScan1176, Belgium), with the scanning parameters set to 65 kV voltage, 385 mA current, and 7 μm resolution. NRecon, Dataview, and CTAn software were used to analyse the obtained scan data to obtain BV/TV and BMD. Mimics Research 20.0 software was used to carry out surface reconstruction of the scanned cranial sample data to observe the bone repair.

Histological analysis: The collected cranial bone specimens from different groups of rats were fixed in 10% formalin for 24 h and then decalcified in ethylenediaminetetraacetic acid (EDTA) solution for 2 weeks, followed by gradient dehydration with different concentrations of ethanol solutions. The dehydrated samples were immersed in pure xylene solution and then embedded in paraffin. Histological analyses were divided into two parts; specimens from 3 to 7 days after surgery were decalcified, embedded and sectioned as described above, and then analyzed with INOS antibody (GB11119-100, Servicebio, China), Arg-1 antibody (GB115724-100, Servicebio, China), IL-10 antibody (GB11534-100, Servicebio, China), TNF-α antibody (GB11188-100, Servicebio, China), TGF-β antibody (GB115739-100, Servicebio, China), Mfn2 antibody (ab124773, Abcam, USA) and Runx2 antibody (ab192256, Abcam, USA), OCN (GB11233-100, Servicebio, China), Drp1 antibody (ab184247, Abcam, USA) were used for immunofluorescence staining, ROCK2 antibody (ER1706-48, huabio, China) for immunohistochemical staining to detect Mφ phenotypic changes, cytoskeletal and mitochondrial kinetic index parameters in vivo. The specimens at 4 and 8 weeks postoperatively were stained with hematoxylin and eosin (H&E) staining and Masson staining to analyse the repair, and immunohistochemical staining with OCN (ab93876, Abcam, USA), and Periostin (ab215199, Abcam, USA) to assess bone tissue and periosteum regeneration, respectively.

In Vivo Biocompatibility of Membranes: To evaluate the in vivo biocompatibility of MS, AMS, MS@Sema3A, and AMS@Sema3A membranes, each material group was implanted subcutaneously in the back of male Sprague-Dawley (SD) rats for 8 weeks. Subcutaneous tissue inflammatory responses and visceral pathological changes in the rats were observed. Briefly, SD rats were anaesthetised with 2.5 mL/kg of 2% pentobarbital. A longitudinal incision was made on the rat's back, and membranes from each group were implanted subcutaneously. At 4 and 8 weeks post-implantation, rats were euthanized with an overdose of pentobarbital. Subcutaneous tissue containing the implanted materials was collected for H&E staining. At 8 weeks, major organs including heart, liver, spleen, lungs, and kidneys were collected for hematoxylin and eosin (H&E) staining.

Statistical analysis: All experimental data were expressed as mean ± standard deviation and were statistically analyzed and plotted using Graph Prism 9.0 software or OriginPro 2024 10.1 software. An unpaired t-test is used for statistical comparisons between two samples, which requires verification of normal distribution and variance homogeneity. Comparisons between multiple samples use one-way or two-way analysis of variance, and differences between groups are analyzed using Tukey's multiple comparison test. ns indicate not statistically significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

6. Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Received: ((will be filled in by the editorial staff))Revised: ((will be filled in by the editorial staff)) Published online: ((will be filled in by the editorial staff))

Ethics approval and consent to participate

All surgical operations and experimental treatments involving experimental animals were approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (Luncheon No. SUDA20241113A0) during this experiment.

CRediT authorship contribution statement

Tianyu Zhou: Data curation, Formal analysis, Project administration, Validation, Writing – original draft. Zhengxia Ni: Data curation, Methodology, Software. Yiyang Huang: Formal analysis, Methodology, Validation. Ziyan Huang: Data curation, Formal analysis. Qiangqiang Guo: Software, Validation. Jie Wu: Investigation, Methodology. Hongyi Zhu: Formal analysis, Visualization. Xinzhao Jiang: Formal analysis. Liang Zhou: Methodology. Wei Wang: Investigation. Kun Xi: Methodology, Writing – review & editing. Yong Gu: Funding acquisition, Writing – review & editing. Wenguo Cui: Conceptualization, Project administration, Writing – review & editing. Liang Chen: Conceptualization, Funding acquisition, Writing – review & editing.

Declarations of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article: