Spatiotemporally programming the immune-osteogenic cascade with a dual-immunomodulatory scaffold for functional bone regeneration

Abstract

Functional repair of critical-sized bone defects is hindered by uncontrolled inflammatory microenvironments that disrupt endogenous regeneration. Conventional osteoimmunomodulatory biomaterials mainly emphasize direct regulating macrophage phenotype transformation, while neglecting the pivotal role of activating the immunomodulatory function of homing mesenchymal stem cells (MSCs) in immune-osteogenic cascade regulation. Herein, we present a dual-immunomodulatory bone scaffold (DIBS) to sequentially regulate the immune functions of macrophages and MSCs for repairing critical-sized bone defects. This scaffold integrates pH-responsive hydrogels containing oxidized xyloglucan (OXG) and metal polyphenol (Sr-PA) nanoparticles into a 3D-printed hydroxyapatite framework, enabling controlled release of immunomodulators and mineralization ions. The OXG first directs early M1-to-M2 macrophage transition to mitigate inflammation; then acidic-triggered release of Sr²⁺ and protocatechualdehyde from nanoparticles enhances the immunoregulatory function of homing MSCs. Additionally, hydroxyapatite framework provides essential mechanical stability and ion sources for late-stage osteogenic mineralization. Single-cell RNA sequencing and validation demonstrate that DIBS effectively induces the generation of immunoregulatory MSC subpopulations, which is associated with M2 macrophage activation through CCL2/CCR2 signaling axis, promoting angiogenesis and osteogenic differentiation. In vivo, DIBS effectively remodels the osteoimmune microenvironment, induces organized collagen arrangement and H-type vascularization, facilitating functional bone tissue repair. This study provides a new multicellular immunomodulatory strategy for endogenous bone repair.

Graphical abstract

Highlights

• •The scaffold with dual immunomodulatory properties mimics the hierarchical structure and function of bone, enabling effective immune-osteogenic regulation.
• •The scaffold can first modulate macrophage phenotype and subsequently enhance the immunoregulatory function of homing MSCs.
• •The scaffold-educated immunoregulatory BMSCs subsets promote immune-osteogenic cascade via the CCL2/CCR2 signaling axis.
• •The scaffold guides organized collagen deposition and H-type vessel formation, achieving functional bone regeneration.

1. Introduction

Using bioactive materials to induce endogenous bone regeneration has emerged as an increasingly prospective strategy for reconstructing critical-sized bone defects [1]. This approach relies on activating endogenous cells migration into the damaged site to induce regeneration through modulation of the extracellular microenvironment or driving cellular reprogramming [2]. Although current active materials can recruit sufficient endogenous functional stem/progenitor cells to the injury site, it is struggle to fully restore normal tissue function. The mechanism of failure of endogenous repair is due to the uncontrolled inflammatory microenvironment that impairs differentiation of functional stem/progenitor cells, ultimately leading to failure of bone tissue repair [3]. Recent advances in osteoimmunity have profoundly revealed the pivotal role of immune regulation in the bone healing process, wherein the adequate immune-inflammatory regulation serves as a central player in establishing an effective link between bone grafts and the stem cell recruitment, angiogenesis and new bone formation [4,5]. Hence, designing a biomaterial that can actively regulate the inflammatory microenvironment is of paramount importance for the advancement of endogenous bone regeneration research.

Among various osteoimmunomodulatory strategies, the timely transition of macrophages from the pro-inflammatory (M1) to the reparative (M2) phenotype is widely recognized as the pivotal link between material implantation and successful regeneration [6,7]. During the initial healing phase at injury sites, pro-inflammatory M1 macrophages rapidly infiltrate to clear necrotic tissue and initiate inflammatory responses. Subsequently, these M1 phenotypes polarize to anti-inflammatory M2 phenotypes, resolving inflammation and promoting bone repair by secreting anti-inflammatory and osteoinductive factors [8,9]. Accordingly, strategically engineering the physical and biochemical properties of biomaterials to induce an M2-dominant immune microenvironment has already become a mainstream strategy, resulting in significant advances in promoting vascularization and new bone formation [10,11]. However, the ideal bone healing process requires not only the precise sequential transition from early pro-inflammatory (M1) to later anti-inflammatory/reparative (M2) states, but also the subsequent self-maintenance of a regenerative microenvironment primarily mediated by crosstalk between M2 macrophages and endogenous mesenchymal stem cells (ESCs) [12,13]. However, relying solely on the initial regulation of macrophage phenotype transformation appears insufficient to recapitulate the full range of immune regulatory mechanisms required throughout the repairing process [14], in particular, this approach overlooks the essential role of the sustained immunomodulatory of mesenchymal stem cells (MSCs) during the post-inflammatory, osteogenic differentiation phase [15]. The MSCs not only possess multipotent differentiation potential but also exhibit significant immunomodulatory capabilities, profoundly influencing the bone repair process. The MSCs can sense their microenvironment and adjust their immune phenotype accordingly. By secreting factors such as TGF-β, HGF, and PGE2, MSCs can directly suppress excessive inflammation that benefit bone regeneration in inflammation. More importantly, MSCs still exert immunomodulatory effects during osteogenic differentiation, which can regulate macrophage chemotaxis and target macrophages and their activation through selective repolarization [16,17]. Potential approaches to facilitating bone regeneration using immunomodulation are the preconditioning of MSCs, multistage loading and delivering anti-inflammatory cytokines, or other substances to empower their immunomodulatory properties [18]. Nonetheless, the clinical use of these strategies is restricted because of the low survival rate of pretreated MSCs after transplantation, potential off-target effects and uncontrollable long-term biosafety issues [17,19]. Therefore, integrating specific functional molecules and constructing a biomimetic microenvironment to engineer biomaterials with “dual immunomodulatory” properties that can initially direct macrophage polarization and synchronously enhance the immunoregulatory function of MSCs presents an exceptional opportunity to enhance functional regeneration in critical-sized bone defects, which remains largely unexplored.

Herein, we report a dual-immunomodulatory bone scaffold (DIBS) designed to regulate the immune-osteogenic microenvironment throughout the entire healing phase for critical-sized bone defects regeneration. The DIBS is formed through Schiff base and polyphenol interfacial modification reactions between a dynamic functional hydrogel module and a 3D-printed directional porous bioceramic framework module. It multidimensionally recapitulates the hierarchical structure and bioactivity of natural bone tissue to support cell recruitment, adhesion, proliferation, and differentiation. The functional hydrogel module utilizes xyloglucan (XG) and metal-polyphenol (Sr-PA) nanoparticles to achieve dual immunomodulation of macrophages and MSCs. Specifically, XG is a naturally occurring polysaccharide with significant anti-inflammatory properties [20,21]. Following DIBS implantation, XG first transforms infiltrating macrophages from the M1 phenotypes to the M2 phenotypes, alleviating the early inflammatory response and establishing a favorable regenerative microenvironment. Secondly, protocatechualdehyde (PA), as a polyphenolic compound, exhibits potent anti-inflammatory and antioxidant properties [22]. Recent studies suggest that PA may also have immunomodulatory properties [23,24]. Meanwhile, the Sr²⁺ renowned for its pro-osteogenic and pro-angiogenic effects, has recently been reported to induce MSCs to secrete immunomodulatory factors such as transforming growth factor-β (TGF-β) and prostaglandin E2 (PGE2) [25,26]. The formation of metal-polyphenol complexes provides an ideal platform for the combined delivery of PA and Sr²⁺ [27], which could enable the synergistic programming of BMSCs' immunoregulatory capabilities. Consequently, Sr-PA nanoparticles act as functional agents by participating in hydrogel network entanglement via the Schiff base chemistry. In acidic inflammatory microenvironments, the bonds of the Schiff base and metal coordination break down rapidly [28,29], leading to the release of PA and strontium ions and enhancing the immunomodulatory function of MSCs. Finally, the directionally porous hydroxyapatite (HA) ceramic modules can mimic the mechanical strength and inorganic composition of bone, providing the necessary mechanical stability and ion sources for osteogenic mineralization in bone repair. Notably, the macropores of bioceramic framework are organically integrated with the micropores of the dynamic hydrogel, forming a layered interpenetrating porous structure in DIBS. This structure can recruit endogenous cells and guide their residency and functional differentiation [30,31]. Consequently, DIBS achieves immune-osteogenic cascade regulation throughout the entire bone healing process by utilizing XG to modulate inflammation in the early stage, employing Sr-PA nanoparticles to enhance the immunomodulatory function of BMSCs in the midstages, and providing essential calcium and phosphate for biomineralization via the HA ceramic framework in the later osteogenic stage (Scheme 1).

Scheme 1.

Design and function of DIBS. Schematic illustration of the preparation of the DIBS. The DIBS achieved immune-osteogenic cascade regulation throughout the entire bone healing process by utilizing XG to modulate inflammation in the early stage, employing Sr-PA nanoparticles to enhance the crosstalk between BMSCs and macrophages in the midstages, and providing essential calcium and phosphate for biomineralization via the HA ceramic framework in the later osteogenic stage.

Single-cell sequencing combined with in vitro and in vivo validation demonstrate that DIBS not only directly induces macrophage polarization toward the M2 anti-inflammatory phenotype but also simultaneously enhances the immunomodulatory capacity of BMSCs. The DIBS induced BMSCs to secrete C-C motif chemokine ligand 2 (CCL2), which could bind to the macrophage surface target C-C motif chemokine receptor 2 (CCR2), triggering M2-type polarization and release of regenerative factors for accelerating angiogenesis and osteogenic differentiation. This work not only underscores the importance of incorporating mesenchymal stem cells as active immunomodulatory units into the design of osteoimmunity materials, but also provides a novel approach for developing highly functional biomaterials that enhance bone regeneration through multicellular synergistic immunoregulatory networks.

2. Results and discussion

2.1. Preparation and characterization of the DIBS

To achieve in situ sequential regulation of the immune-osteogenic microenvironment, the DIBS was engineered using a dual-module strategy featuring a rigid inner layer and a soft outer layer to enable controlled release of immunomodulators and mineralization ions. The outer layer of DIBS comprises a pH-responsive immunomodulatory hydrogel module. Oxidized xyloglucan (OXG) was synthesized by partially oxidizing vicinal diols to aldehyde moieties [32]. A distinctive peak at 1725 cm⁻¹, corresponding to the C=O stretching vibration, signifies the formation of aldehyde groups on the OXG structure (Fig. S1, Supporting Information), allowing Schiff base cross-linking with the amino groups of carboxymethyl chitosan (CMCS). This reaction formed a stable viscoelastic hydrogel after 145s (Fig. S2A and B, Supporting Information). XG as a natural polysaccharide, possesses significant anti-inflammatory properties and excellent biocompatibility. It can directly modulate the phenotype and function of infiltrating macrophages during the early inflammatory process, conferring the initial immunomodulatory capacity to DIBS. To further program the immunoregulatory capacity of MSCs recruited to defect sites, metal polyphenol (Sr-PA) nanoparticles were synthesized via a one-pot method employing a one-step ion/molecule assembly process. The Sr-PA nanoparticles respond to acidic inflammatory microenvironments via Schiff base imine bond and metal coordination bond, releasing Sr²⁺ and PA in situ. This enables programmable immunomodulatory regulation of MSCs.

As shown in Fig. 1A, the heating conditions within the reactor accelerated the coordination process, promoting the formation of Sr-PA nanoparticles. Scanning electron microscopy (SEM) revealed that Sr-PA nanoparticles exhibited a needle-like nanoflower morphology, with dimensions spanning from 500 nm to 800 nm (Fig. 1C). Fourier transform infrared spectroscopy (FT-IR) further confirmed the successful synthesis of Sr-PA nanoparticles. Compared to SrCl₂, the characteristic peaks at 1591 cm⁻¹,1455 cm⁻¹ and 1643 cm⁻¹observed in Sr-PA suggested the absorbance associated with the aromatic rings and C=O in the PA molecules [33] (Fig. 1B). Thermogravimetric analysis (TGA) revealed that Sr-PA sustained approximately 40% mass loss when subjected to temperatures of up to 1000 °C under controlled heating conditions (Fig. 1E). These results confirm the successful incorporation of the organic component, with an organic-to-inorganic mass ratio of approximately 2:3. Furthermore, the energy-dispersive X-ray spectroscopy (EDS) of Sr-PA nanoparticles further confirm the successful inclusion of PA in Sr-PA and demonstrated the uniform distribution of Sr²⁺ within the Sr-PA nanoparticles (Fig. 1D).

Fig. 1.

Characterization of dynamic functional hydrogel module. A) Schematic synthesis of Sr-PA by hydrothermal method. B) FT-IR spectra of SrCl₂, PA, and Sr-PA. C) SEM images of Sr-PA. D) Elemental mapping of Sr-PA. E) TGA curves of Sr-PA. F) Schematic synthesis of dynamic functional hydrogel based on Schiff base cross-linking. G) Time-modulus of the dynamic functional hydrogel with different concentrations of Sr-PA. H) Stress relaxation curves and half-time of relaxation for dynamic functional hydrogels. I) Damping factor (tanδ) of dynamic functional hydrogels. J) Wide-scan XPS spectra of OXG, CMCS, and Sr-PA@OC hydrogel. K) C 1s deconvoluted XPS curves of Sr-PA@OC hydrogel. L) N 1s deconvoluted XPS curves of Sr-PA@OC hydrogel. M) Storage modulus (G′) and loss modulus (G″) of hydrogels during oscillatory strain amplitude sweep. N) Self-healing analysis of the Sr-PA@OC hydrogel with alternate strain switched from 1% to 250% for four cycles. O) Schematic diagram of PA-modified HA ceramic framework and its modified morphology. P) FT-IR spectra of PA-modified HA framework. Q) C 1s deconvoluted XPS curves of PA-modified HA framework.

Furthermore, the mechanical properties of pH-responsive immunomodulatory hydrogel (Fig. 1F) could be regulated by varying the additive ratios of Sr-PA. The initial storage modulus increased from 100 to 400 Pa as the Sr-PA concentration rose from 50 to 200 μg/mL at 37 °C (Fig. 1G). Viscoelasticity was evaluated by measuring the half-time of relaxation, which represents the time required to relax 50% of the stress in a stress relaxation test [34]. Stress relaxation tests revealed that Sr-PA notably affected the viscoelasticity of the dynamic hydrogels, with relaxation half times significantly larger with increasing additions. The 50Sr-PA group had a speedy stress relaxation rate, with a half-stress relaxation time of ≈41 s. Consistent with these results, rheological analysis revealed that 50Sr-PA hydrogels have a significantly higher loss tangent as compared to other hydrogels (Fig. 1H and I). These results suggested that proper addition of Sr-PA nanoparticle favors highly dynamic network structures, in which stem cells can timely acquire mechanical signaling to regulate their expansion, migration, and gene expression, and to direct cell fate decisions [35]. Considering the effect of the addition amount on bioactivity (Fig. S3, Supporting Information) and mechanical properties, we chose the 50-100 Sr-PA dynamic hydrogel (Sr-PA@OC) for subsequent experiments.

The network structure of pH-responsive immunomodulatory hydrogel was characterized using XPS. Fig. 1J presented the complete XPS profiles of the lyophilized OXG, CMCS, and Sr-PA@OC hydrogels. The spectra of CMCS and Sr-PA@OC exhibited C 1s and N 1s peaks, while the OXG spectrum showed no N 1s peaks due to the absence of nitrogen in the XG macromolecular chains. Comparison of the high-resolution C 1s spectrum of the Sr-PA@OC hydrogel (Fig. 1K) with those of OXG (Fig. S4A, Supporting Information) and CMCS (Fig. S4B, Supporting Information) revealed a new peak at 288.8 eV in the Sr-PA@OC hydrogel, attributed to the C=N bond in the hydrogel [36]. Similarly, a comparison of the high-resolution N 1s spectra of the Sr-PA@OC hydrogel (Fig. 1L) with that of CMCS (Fig. S4C, Supporting Information) revealed a new peak at 397.3 eV, also attributed to the C=N bond in the hydrogel. These high-resolution C 1s and N 1s spectra confirmed the formation of Schiff base bonds in the hydrogel. The strain sweep indicated that when the strain exceeded 500%, the G′ value of the dynamic hydrogels notably decreased, highlighting the strong resistance of the hydrogel network to deformation (Fig. 1M). The alternate-step strain-amplitude sweep experiments demonstrate Sr-PA@OC hydrogels possessed shear-thinning and self-healing properties (Fig. 1N). Evidence of macroscopic self-healing demonstrated that hydrogels could rapidly form the corresponding shape after passing through a 27G needle, and the severed hydrogel could reform into a complete hydrogel within 30 min (Fig. S5, Supporting Information). These observations collectively demonstrate the excellent dynamic characteristics and injectability of Sr-PA@OC hydrogel. However, providing stable mechanical support and essential calcium and phosphate is of great importance for the mineralization and functional reconstruction of the bone matrix [37]. Therefore, an essential step involved incorporating the 3D-printed hydroxyapatite (HA) bioceramic framework with macropores into the instructive niche, thereby supporting cascade regulation of the regenerative microenvironment. The HA bioceramic framework used for bionic niche fabrication was made by DLP-3D printing and degreasing sintering post-treatment process, and its filaments (≈500 μm) were arranged in a 90°crisscross pattern at a 30% filling rate, accurately forming the 3D-connected structure with a pore size of 350 μm. As shown in Fig. S6A, SEM images revealed that the 3D-printed HA framework retained its complete shape and formed a rough surface morphology. EDS characterization of Ca, P, and O elements confirmed the uniform distribution of HA framework. Of note, the PA can not only act as an active agent for regulating immunity, but also as a multifunctional crosslinking agent, mediating multiple chemical bonds between organic and inorganic substances [23]. Hence, it was employed as an organic and inorganic interface linker to improve the bonding between the hydrogel and HA ceramic interfaces. SEM images showed that aggregated PA nanoparticles form micro- and nano-morphologies on the ceramic surface (Fig. 1O). The FT-IR revealed that peaks at 1260 cm⁻¹ and 1690 cm⁻¹ corresponding to C-O and C=O, respectively (Fig. 1P), proved that PA was successfully coated onto the surface of the HA framework (Fig. S6B, Supporting Information). Additionally, visible signals for C-O at around 286.3 eV and C=O at around 288.3 eV suggested PA on the HA framework [23,38] (Fig. 1Q). Similarly, the results regarding interfacial bonding strength indicate that the PA-modified ceramic bonds more firmly to the hydrogel (Fig. S7, Supporting Information). Therefore, the unique physical morphology would synergize with the aldehyde groups of PA to promote reliable integration between the dynamic hydrogel and the HA ceramic interface.

For comparison, an OC hydrogel (OXG and CMCS) without Sr-PA nanoparticles was also prepared and integrated with the PA-modified HA ceramic framework, named OCHA. Fig. 2B–D demonstrated gross images of HA, OCHA and DIBS, each with a unique morphology. The dynamic hydrogel network, embedded within the macropores of the 3D-printed HA framework, formed a unique interconnected microporous structure. The average pore diameters of the micropores of the OCHA and DIBS were 47.42 ± 0.77 μm (Fig. 2E) and 41.20 ± 1.15 μm (Fig. 2F), respectively. Notably, Sr-PA nanoparticles addition had little effect on the pore size of the micropores but led to sharper edges, attributed to Sr-PA acting as active inorganic connectors that increase the modulus of the dynamic hydrogel network.

Fig. 2.

Fabrication and characterization of DIBS. Created with BioRender.com. A) Schematic illustration of the fabrication process of DIBS. B-D) Optical images, low magnification SEM and high magnification SEM images showing the top view of HA, OCHA and DIBS (the SEM images show the morphology after freeze-drying). E, F) Microporous diameter distribution of OCHA and DIBS. G, H) SEM images showing the cross-section and interface integration of OCHA and DIBS (Yellow: interface integration). I) SEM and EDS elemental mapping images of DIBS. J) In vitro degradation testing of DIBS under different pH conditions. The release characteristics of PA (K) and Sr²⁺(L) from DIBS under different pH conditions. M) Compression strength, N) toughness, and O) elastic modulus of different scaffolds. P) Specific surface area of different scaffolds. Data are represented as means ± SD, ∗p < 0.05. ns, not significant.

SEM images demonstrated that the interior of the scaffold exhibited a hierarchically interpenetrating morphology between the microporous hydrogel network and the HA framework (Fig. 2G and H). Gel filaments radiated from the center of the ceramic macropores, forming distinct transitional layers at the ceramic-gel interfaces. This was a critical outcome of the PA modification process. EDS mapping confirmed the uniform spatial distribution of the characteristic bioactive elements Sr, Ca, and P from the DIBS group, in line with the positioning of the materials (Fig. 2I). Meanwhile, the dynamic hydrogel network formed a 3D-like spiderweb structure in the macropores of the ceramics, which would provide a larger surface area for cell 3D adhesion and growth (Fig. 2P), and further serve as a ladder for cells to climb, ascend, and extend inward.

In addition, we systematically investigated the degradation and release characteristics of DIBS. Under acidic conditions (pH 6.0), DIBS exhibited a rapid degradation rate, with approximately 25-30% mass loss after 20 days, whereas under neutral conditions (pH 7.4), the mass loss was only about 15% (Fig. 2J). Importantly, Sr-PA nanoparticles encapsulated within the dynamic hydrogel modules also exhibited significant pH-responsive release behavior. The Sr²⁺ and PA were rapidly released under acidic conditions, with total release amounts within 20 days reaching 1.8 times and 1.25 times those under neutral conditions, respectively (Fig. 2K and L).

As shown in Fig. S8 (Supporting Information), in the acidic condition, the zeta potential of DIBS to become increasingly negative over the first 4-16 days (from approximately −3.4 mV to −7.2 mV), reflecting the progressive exposure of PA-rich fragments. In the neutral condition (pH 7.4), the slower degradation rate would result in a less negative Zeta potential (−2.0 mV to −2.4 mV) with smaller fluctuations. This pH-responsive degradation is designed to align with the inflammatory phase (acidic) and then slow down as the microenvironment normalizes (neutral), preventing premature scaffold collapse [39,40]. The acidic inflammatory microenvironment triggers Schiff base bond and metal-polyphenol coordination bond cleavage through ligand protonation and competitive coordination, leading to hydrogel network and Sr-PA nanoparticle degradation and sustained release of strontium ions and PA [28,41]. It is noteworthy that the recruitment of stem cells and crosstalk with macrophages during bone regeneration occur within 14 days post-injury [4,13]. This timeframe aligns closely with the peak release period of Sr-PA in this study, which spans approximately days 4-16. This temporal concordance is crucial for enhancing the immunoregulatory activity of MSCs, thereby promoting in vivo osteogenic and angiogenic processes.

Possessing good mechanical properties to withstand forces from physiologic systems and external factors is a crucial requirement for scaffolds for repairing bone defects [42]. The results of compression tests demonstrate that dynamic hydrogel embedded in HA framework significantly improved the compressive strength of OCHA and DIBS, resulting in a compressive modulus of 6.02 ± 0.37 MPa and 8.56 ± 1.19 MPa for the bionic niche, respectively (Fig. 2M). The higher brittle stiffness of bioceramics needs to be considered for improvement when used in BTE scaffolds. Our results showed that the flexible dynamic hydrogel significantly improved the toughness of the ceramic scaffolds, in which the toughness of DIBS was 58,405 J/m², which was about 6.5 times higher than that of HA scaffolds (Fig. 2N). The enhanced toughness would improve the overall structural stability and prevent the scaffold from premature fracture failure in vivo, safeguarding the success of the repair process. Similarly, the dynamic hydrogel decreased the elastic modulus of pure HA ceramic scaffolds. The elastic modulus of the DIBS was reduced to 841.9 MPa (Fig. 2O), approaching the level of natural cancellous bone (0.1-2 GPa) [38]. This reduced the stress shielding effect, enabling newly formed bone tissue to withstand more reasonable mechanical stimulation and promoting functional bone regeneration and remodeling.

2.2. Promotion of cell recruitment, adhesion, and migration by DIBS

As an immunomodulatory niche tailored for the microenvironment, superior biocompatibility is paramount for fostering the adhesion and proliferation of endogenous stem/progenitor cells [43]. We first evaluated the effect of the scaffolds on cell viability and proliferative capacity. Cytotoxicity results indicated cell survival rates exceeding 100%, confirming the absence of toxicity for all scaffolds (Fig. S9B, Supporting Information). Furthermore, whether culturing BMSCs with extracts or co-culturing BMSCs with scaffolds, live/dead staining revealed a higher density of viable cells (green) within the DIBS (Fig. S10, Supporting Information). This suggested that both the bioactive agent and the hierarchically interconnected architecture of DIBS were conducive to enhanced cell activity. Then, the ability of different scaffolds to promote BMSCs adhesion and migration into the hierarchical microporous structure was further comprehensively assessed over varying co-culture durations. After 1 day of seeding, the expression of vinculin, a protein associated with cell-matrix interactions, was evaluated [44]. Results demonstrated that BMSCs on DIBS exhibited more centrally located and elongated vinculin-rich focal adhesions (Fig. S11, Supporting Information). At days 3 and 5 post-seeding, 2D-view images of BMSCs stained for cytoskeleton and nucleus showed their remarkable growth on all scaffolds. BMSCs had already developed spindle-shaped morphologies by day 3 (Fig. 3A), with the DIBS group exhibiting a significantly higher cell number compared to the OCHA and HA groups (Fig. 3C). After 5 days, BMSCs exhibited more extensive spreading and a higher cell aspect ratio. Compared to HA scaffolds, both OCHA and DIBS supported the most extensive BMSC coverage area and viability rate (Fig. 3B–F). The distribution and growth behavior of cells within the pore structures of different scaffolds were further explored. 3D-view images of BMSCs revealed that within the macropores of the HA scaffold, cells could only form a monolayer along the ceramic walls, migrating inward via intercellular connections (Fig. 3D). In contrast, within the DIBS, the hierarchical hydrogel network provided a 3D space for cell adhesion and infiltration. Cells utilized hydrogel fibers to achieve suspended 3D growth within the center of ceramic macroporous structures (Fig. 3E).

Fig. 3.

Biocompatibility and cellular response of DIBS. A, B) 2D-CLSM images showed BMSCs cultured on different scaffolds for 3 d and 5 d in vitro (F-actin, green; DAPI, blue). C) Quantification of cell number of BMSCs on different scaffolds after 3 d and 5 d of culture in vitro. D, E) 3D-view images of BMSCs stained for cytoskeleton and nucleus showed their migration and distribution in the interiors and pores of different scaffolds. F) CCK-8 assay of the cellular viability of BMSCs co-cultured with different scaffolds for 1, 3, and 5 days. G) Immunofluorescence staining and H) quantitative analysis of CD86 and CD206 after one week of implantation. I) Immunofluorescence staining and J) quantitative analysis of CD44 and CD105 after two weeks of implantation. K-N) Molecular docking results of OXG and CMCS with the FNH domain of FN. Data are represented as means ± SD, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To further evaluate the biocompatibility of DIBS and its capacity to recruit endogenous cells in vivo, cell-free scaffolds were implanted into 8 mm rat calvarial defects. Following the initial introduction of the biomaterial, an immune response ensues, with macrophages acting as central regulators in this inflammatory cascade. An ideal bone tissue engineering scaffold (BTE) should possess the ability to modulate macrophage phenotype from the pro-inflammatory M1 state towards the anti-inflammatory M2 state, thereby fostering optimal bone regeneration [45]. Immunofluorescence (IF) staining indicated that both OCHA and DIBS promoted the polarization and infiltration of M2 macrophages (CD206⁺ cells) compared to the HA scaffold after one week of implantation (Fig. 3G and H), demonstrating superior immunomodulatory capacity.

Furthermore, the recruitment efficacy of endogenous stem cells (ESCs, CD44⁺CD105⁺) was also assessed two weeks post-implantation. IF staining revealed significant recruitment of CD44⁺CD105⁺ ESCs within the DIBS. Although lacking the bioactive PA and Sr²⁺, the hierarchical physical structure of the OCHA also recruited a notable number of ESCs, while minimal ESCs migrated into the HA scaffold (Fig. 3I and J). This pronounced recruitment effect could be attributed to the synergistic combination of structural advantages and Sr²⁺ induction, as reported [46]. Interestingly, results from both in vitro and in vivo experiments consistently demonstrated that the dynamic polysaccharide hydrogel network within the ceramic macropores significantly enhanced MSCs adhesion. Beyond the physical anchoring effect conferred by the hierarchical topography, the specific adsorption of adhesion proteins on the material surface may serve as a crucial chemical signal driving cell adhesion. To validate this mechanism, molecular docking was employed to investigate the interactions between the dynamic hydrogel components and the core adhesion protein, fibronectin (FN). Docking results revealed favorable alignment of OXG and CMCS with the FNH domain of FN [47], indicating attractive interactions between them. The binding forces between OXG/CMCS and the amino acid residues of the FNH binding site were primarily contributed by van der Waals, carbon hydrogen bonds, and conventional hydrogen bonds. Their calculated binding energies were −6.6 kcal mol⁻¹ and -5.4 kcal mol⁻¹, respectively, signifying favorable binding interactions (Fig. 3K–N). Additionally, the affinities of OXG and CMCS for the RGD-binding domain of FN were compared, yielding binding energies of −4.2 kcal mol⁻¹ and -3.5 kcal mol⁻¹, respectively. These values indicated lower binding energies compared to their interactions with the FNH domain. Therefore, the functional groups of the dynamic polysaccharide hydrogel (e.g., aldehyde, carboxyl) may preferentially bind to non-cell-binding domains of FN (e.g., regions distinct from the RGD sequence), achieving chemical anchoring of FN. Then, in turn, enhanced MSCs adhesion via the RGD-integrin binding site. In summary, the synergistic interplay between the hierarchical topography and the material-specific protein adsorption enables the rapid recruitment, adhesion, and retention of endogenous functional cells.

2.3. Initial immunomodulation by directly regulating macrophage polarization from DIBS

Inflammation regulation plays the crucial role in the initial stages of DIBS-driven bone regeneration, whereby a timely transition from the pro-inflammatory M1 phenotype to the anti-inflammatory M2 phenotype of macrophages is essential for creating a reparative immune microenvironment conducive to tissue regeneration [48]. In this study, the anti-inflammatory properties of the scaffolds were assessed during the early stages of inflammation (Fig. 4A). First, lipopolysaccharide (LPS) was used to induce macrophage polarization toward the M1 phenotype, simulating an initial inflammatory microenvironment. IF staining (Fig. S12) showed that DIBS and OCHA significantly reduce the percentage of M1 macrophages (CD86⁺) as early as day 3 and increase M2 macrophages (CD206⁺), compared to HA. By day 5 and day 7 (Fig. 4B), the M2-dominant phenotype is well established. After 7 days of co-culture, a comprehensive analysis employing IF (Fig. 4B), Western blotting (Fig. S13, Supporting Information), and flow cytometry (Fig. 4D) revealed that DIBS significantly induced polarization of macrophages towards the M2 phenotype. In contrast, the majority of macrophages of HA group underwent predominant polarization towards the M1 phenotype. In the HA group, macrophages exhibit a rounded or amoeboid shape characteristic of M1 phenotype, whereas in DIBS and OCHA groups, cells display an elongated, spindle-like morphology typical of M2 macrophages (Fig. 4B). Furthermore, quantitative real-time polymerase chain reaction (qRT-PCR) results confirmed that, compared to the HA group, the OCHA and DIBS significantly upregulated the expression of the Arginase 1 (Arg-1) and CD206 genes, while downregulating the expression of the nitric oxide synthase (iNOS) and TNF-α genes (Fig. 4C). Enzyme-linked immunosorbent assay (ELISA) data further corroborated this finding (Fig. 4E), revealing that both the OCHA and DIBS reduced the secretion of pro-inflammatory cytokines (TNF-α and IL-6) while increasing the secretion of the anti-inflammatory cytokine (IL-10) and the pro-osteogenic/pro-angiogenic growth factor (TGF-β). Similarly, these were consistent with in vivo findings, where OCHA and DIBS promoted macrophage polarization toward the M2 phenotype in vivo (Fig. 3G and H). All these results can be attributed to the scaffold's outer hydrogel module preferentially regulating the inflammatory microenvironment.

Fig. 4.

Macrophage polarization regulation by DIBS. A) Schematic illustration of DIBS promoting M2 macrophage polarization and anti-inflammation. Created with BioRender.com. B) Immunofluorescence and morphological images of macrophages treated with different scaffolds. C) Related gene expression of M1 and M2 polarization markers of macrophages co-cultured with different scaffolds. D) Flow cytometry graph of macrophage polarization. E) ELISA analysis of pro-inflammatory factors and anti-inflammatory factors secreted by macrophages at 7 days. Data are represented as means ± SD, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

XG is a natural polysaccharide that is often directly oxidized (OXG) to form aldehyde cross-linking sites which bind to substances like gelatin and carboxymethyl chitosan to enhance biological activity. Owing to its characteristic galactose units and a backbone composed of β-(1  →  4)-D-xylopyranosyl residues, XG/OXG exhibits unique immunomodulatory functions, playing an active role in osteoimmunity for regeneration [20]. Chen et al. developed a multifunctional bioink based on XG as the matrix, where XG enhanced the scaffold's bioactivity while also alleviating the inflammatory response post-implantation, thereby accelerating bone tissue repair [21]. Our previous study developed an injectable composite hydrogel containing OXG, which demonstrated significant anti-inflammatory properties and sustained recruitment of anti-inflammatory M2 macrophages for wound healing [18,49]. In this study, OXG-modified hydrogel layers were designed, wherein the rapid degradation of the outer hydrogel layer during the early phase directly modulate macrophage polarization (Fig. 2J), thereby rapidly establishing an initial reparative immune microenvironment.

2.4. Subsequent immunomodulation by regulating BMSCs reprogramming towards immunomodulatory phenotype from DIBS

Temporal regulation of the immune microenvironment is crucial for functional regeneration in critical-sized bone defects. Following initial inflammatory modulation, endogenous MSCs are recruited to the defect site to mediate subsequent regenerative repair [43,50]. Given the pivotal role of the immunomodulatory function of MSCs in regulating macrophage polarization and maintaining immune homeostasis during the subsequent regenerative phase, the effects of DIBS on the immunomodulatory function of BMSCs were further investigated.

To gain an unbiased, high-resolution view of the cellular heterogeneity and transcriptional reprogramming of BMSCs within the different scaffold microenvironments, after 14 days of co-culture, BMSCs were dissociated from the scaffolds for single-cell RNA sequencing (scRNA-seq) (Fig. 5A). Subsequent to standard data processing and cell quality control, a total of 13,758 cells from the HA group, 11,259 cells from the OCHA group, and 12,413 cells from the DIBS group were obtained for further analysis. The dataset underwent log-normalization and scaling via the Seurat pipeline, with cell heterogeneity visualized using t-distributed Stochastic Neighbor Embedding (t-SNE). Unsupervised graph clustering partitioned the cells into eleven distinct clusters (Fig. S14, Supporting Information). These clusters were classified into specific cell types based on marker gene expression and function (Fig. 5B–D): 1) Immunoregulatory BMSCs with high expression of CCL2, CCL7, CXCL1 and CXCL3; 2) Proliferative BMSCs with high expression of cenpf, Mki67, and Top2a; 3) Osteo BMSCs with high expression of Col1a1, Col12a1, Acta2 and Sparc; 4) Osteo progenitors with high expression of Mgp, Fos, Plac8, Spp1 and Stc1; 5) Early preosteoblasts with high expression of Fosb and Spp1; 6) Preosteoblasts with high expression of Fosb, VEGFA and Spp1. These cell types are consistent with previously reported findings [[51], [52], [53], [54]]. The preosteoblast cluster from the OCHA and DIBS groups comprised a significantly higher proportion of cells at the transcriptome level compared to the HA group (Fig. 5C). This suggested that BMSCs cultured on composite scaffolds exhibited gene expression patterns indicative of higher osteogenic differentiation potential. Notably, a cluster of BMSCs with an immunoregulatory gene expression profile was predominantly identified in the OCHA and DIBS groups. This cluster was characterized by high expression of immunomodulatory genes (Cxcl3, Cxcl6, CCL2, CCL20) (Fig. S15, Supporting Information). Gene Ontology (GO) functional enrichment analysis indicated that the functions of this cluster were mainly related to mesenchymal cell differentiation, stem cell proliferation, and bone development; cellular response to chemokine, chemokine-mediated signaling pathway; and angiogenesis (Fig. S16, Supporting Information). Furthermore, single-cell trajectory analysis further revealed that the immunoregulatory BMSCs were located near the developmental beginning of the cellular population, suggesting they may contribute to establishing the immunoregenerative microenvironment (Fig. 5E–G). These results suggest that, under the regulation of the instructive niche, BMSCs exhibited multipotency and heterogeneity, and the immunoregulatory BMSCs had stronger plasticity in immunomodulation and stemness. We further analyzed the functional profile of the most abundant subset in DIBS, the Osteo progenitors. GO analysis highlighted significant enrichment in biological processes such as positive regulation of inflammatory response, regulation of tumor necrosis factor production, transition metal ion transport, and positive regulation of mitochondrion organization (Fig. S17, Supporting Information). These processes were intimately linked to immunomodulation, cellular metabolism, and ion transport, further emphasizing that the DIBS could enhance the immunoregulatory capacity and metabolic activity of stem cell subsets. This is attributed to the dynamic hydrogel modules based on Schiff base bonds responding to the acidic inflammatory microenvironment, continuously releasing PA and Sr²⁺ to enhance the immunomodulatory capacity of BMSCs during the osteogenic differentiation process. Next, we analyzed the pseudotime kinetics of differentially expressed genes during the development of Immunoregulatory BMSC subsets toward Osteo progenitors (Fig. 5H–K) or Preosteoblasts (Fig. S18, Supporting Information) subsets within DIBS. The results demonstrated that, compared to HA scaffolds, most immunomodulatory-related genes (e.g., the chemokine CCL2, Stat3) exhibit higher expression levels during progression to Osteo progenitors subsets in DIBS (Fig. 5I–L). This enhanced expression would be essential for the recruitment of functional repair cells and the establishment of an immunomodulatory regenerative process [8]. Conversely, during differentiation toward Preosteoblasts subsets, immunomodulatory gene expression declined (Fig. S19, Supporting Information), with a concomitant surge in pro-angiogenic gene (VEGF) expression (Fig. 5J–M). This dynamic transition between immunomodulatory and pro-angiogenic gene programs may play a critical role in restoring the regenerative potential of the cascade for critical-sized bone defects.

Fig. 5.

Overview of the single-cell transcriptomic sequencing results for BMSCs in different microenvironments. Created with BioRender.com. A) Schematic illustration showing 3D co-culture and cell isolation for scRNA sequencing based on 10X Genomics. B) Dot plots showing the expression of feature genes in six subsets. The dot size represents the proportion of cells expressing specific genes, and the dot color represents the expression level of genes. C) Stacked bar graph showing the proportion of each subpopulation in different scaffolds. D) t-SNE visualization displaying clusters and annotated cell types. E) CytoTRACE trajectory analysis of six cell subpopulations. F) Pseudotime trajectory inferred by RNA velocity on t-SNE embeddings. G) Monocle pseudotime trajectory showing the progression of clusters. H) Pseudotime heatmap of DIBS. The left side is the Immunoregulatory BMSCs subset, and the right side is the Osteo progenitors subset. I) Pseudotime kinetics of the CCL2 gene during the development of Immunoregulatory BMSCs cluster to Osteo progenitors cluster in DIBS. J) Pseudotime kinetics of the VEGF gene during the development of Immunoregulatory BMSCs cluster to preosteoblasts cluster in DIBS. K) Pseudotime heatmap of the HA group. The left side is the Immunoregulatory BMSCs subset, and the right side is the Osteo progenitors subset. L) Pseudotime kinetics of the CCL2 gene during the development of Immunoregulatory BMSCs cluster to Osteo progenitors cluster in the HA group. M) Pseudotime kinetics of the VEGF gene during the development of Immunoregulatory BMSCs cluster to preosteoblasts cluster in the HA group.

Moreover, we analyzed the interaction between different cell subsets in HA and DIBS groups via CellChat, a cell ligand/receptor pairing-based database, based on the scRNA-seq database. Results demonstrated significantly stronger interactions among all cell subsets in DIBS (Fig. 6A). Notably, the immunoregulatory BMSCs exhibited the highest number and intensity of ligand/receptor interactions with osteo progenitors and early preosteoblasts (Fig. S20, Supporting Information), indicating that the DIBS with hierarchical structures enhanced intercellular communication, thereby promoting cascaded regeneration. Additionally, bulk RNA sequencing results revealed that, compared to the HA scaffold, differentially expressed genes (DEGs) in DIBS were significantly enriched in wound healing, immunomodulation, angiogenesis, and bone regeneration (Fig. S21, Supporting Information). Representative DEGs are shown in Fig. S22. Collectively, single-cell transcriptomics revealed a previously unrecognized heterogeneity in BMSCs cultured within DIBS, including a subpopulation with enriched immunomodulatory gene expression. This observation prompted us to hypothesize that DIBS might enhance MSC-mediated immunomodulation.

Fig. 6.

Immune regulation and endogenous bone regeneration mechanism investigation. A) Network diagram showing the number of interactions between six subclusters. B) KEGG enrichment analysis of the upregulated DEGs in DIBS group compared to the HA group. C) Circular visualization of related pathway–gene enrichment analysis. D) Heatmap of key gene regulation in specific pathways. E) qRT-PCR validation for key gene expression in specific pathways. F) The interaction networks showing the correlation of representative immunomodulatory genes (CCL2, CCL20, Sfrp1, and Stat3, etc.) with angiogenesis/osteogenesis and macrophage regulation gene sets. G) Flow cytometry analysis and quantification of CCR2 F4/80 macrophage in peripheral blood. H) Immunofluorescence staining analysis of macrophage polarization inside scaffolds (one week after intramuscular implantation). I) Macrophage proliferation assay in a CCR2-dependent manner. J and K) Macrophage polarization assay in a CCR2-dependent manner. Data are represented as means ± SD, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ^$p < 0.05 (vs group without inhibitor), ^$$p < 0.01 (vs group without inhibitor), ^$$$p < 0.001 (vs group without inhibitor), ^$$$$p < 0.0001 (vs group without inhibitor). ns, not significant.

2.5. CCL2-mediated macrophage M2 polarization and osteogenesis, angiogenesis in a CCR2-Dependent manner by DIBS

We next investigated how DIBS orchestrated cascaded bone repair by training the immunomodulatory function of BMSCs. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis further provided valuable insights into the functional roles of upregulated DEGs. The top enriched upregulation-related pathways in DIBS group versus HA group were exhibited in Fig. 6B, which could be divided into three components: i) positive regulation of cell energy metabolism and differentiation: carbon metabolism, glycolysis/gluconeogenesis, and phosphatidyl-inositol 3-kinase/serine-threonine kinase (PI3K-Akt) signaling pathway; ii) immune relevant event: cytokine-cytokine receptor interaction, chemokine signaling pathway, and IL-17 signaling pathway; and iii) promotion of vascularization and osteogenesis: hypoxia-inducible factor-1 (HIF-1) signaling pathway and mineral absorption. These upregulated pathways were also significantly enriched in subclusters such as Immunoregulatory BMSCs (Fig. S23, Supporting Information) and Osteo progenitors (Fig. S24, Supporting Information), respectively. Multiple biological processes were coordinated across these subclusters to plays a critical role in regenerative repair process within the instructive niche. The specific gene-pathway interaction network and related gene expression levels for selected key pathways are displayed in Fig. 6C and D. Notably, CCL2, as a crucial regulatory growth factor, was involved in multiple signaling pathways including chemokine signaling pathways, cytokine-cytokine receptor interaction and IL-17 signaling pathways, appearing to be a central mediator of endogenous bone regeneration initiated by domesticating the immune regulation function. The qRT-PCR was performed to validate the expression of eight key genes involved in these pathways. Significant intergroup differences were observed for seven genes, corroborating the RNA sequencing results (Fig. 6E). The interaction network analysis revealed strong interactions of representative immunomodulatory genes (CCL2, CCL7, Stat3, etc.) in the DIBS group with the osteogenic/angiogenic and macrophage-regulatory gene sets (Fig. 6F). This further underscored chemokines, particularly CCL2, as pivotal immunomodulatory factors promoted regenerative repair. Besides, CCR2 is the only receptor for CCL2. Research indicates that the CCL2/CCR2 axis participates in macrophage proliferation and M2 phenotypic activation [[55], [56], [57]], indicating its critical role in endogenous tissue regeneration induced by immune regulation. To corroborate our transcriptomic findings at the protein level, we performed immunofluorescence staining for CCL2 on BMSCs cultured on the different scaffolds for 14 days (Fig. S25, Supporting Information). In line with the scRNA-seq data, BMSCs cultured on DIBS displayed markedly higher CCL2 protein expression compared to those on HA or OCHA scaffolds. Therefore, we hypothesize that DIBS may mediate subsequent bone repair processes by secreting chemotactic factors and cross-talking with macrophages.

To further elucidate the regulatory mechanism of the CCL2/CCR2 pathway between ESCs and macrophages, we first investigated the M2 macrophage activation in the intramuscular implantation model of CCR2-inhibited mice. RS504393, a highly selective CCR2 chemokine receptor antagonist [58], was administered via daily intraperitoneal injection for the first 7 days post-implantation (and continued thereafter) to inhibit CCL2/CCR2 signaling. After 7 days of inhibition, the proportion of CCR2⁺ F4/80⁺ macrophages in peripheral blood decreased significantly from 27.5% (normal group, no intervention) to 8.24% (inhibitor group), representing a remarkable reduction of 19.26% (p < 0.0001), confirming successful CCR2 inhibition (Fig. 6G). Following 7 days of intramuscular implantation, significantly more M2 macrophages infiltrated the hierarchical porous structures of DIBS compared to the HA group. However, M2 macrophage infiltration was drastically reduced after CCR2 inhibition, demonstrating that CCR2 signaling was crucial for the M2 macrophage activation induced by DIBS (Fig. 6H). Subsequently, we employed the BMSC-macrophage co-culture model to investigate the spatiotemporal regulation of macrophage proliferation and polarization by BMSCs via the CCL2/CCR2 pathway under CCR2 inhibition. Cell proliferation assays revealed that DIBS promoted macrophage proliferation in a CCR2-dependent manner. This pro-proliferative effect was most pronounced on day 3 and was significantly attenuated by CCR2 inhibition. A similar dependency was observed in groups treated with exogenous CCL2 (40 ng mL⁻¹) or CCL2 plus inhibitor (Fig. 6I). Similarly, DIBS significantly upregulated the M2 phenotype marker (CD206) in macrophages in a CCR2-dependent manner. Although the induction level was lower than that achieved with exogenous CCL2 (p < 0.0001), the upregulation almost disappeared after CCR2 inhibition (p < 0.0001) (Fig. 6J). Furthermore, DIBS significantly downregulated the expression of the pro-inflammatory marker (iNOS), an effect that persisted regardless of CCR2 inhibition (Fig. 6K).

M2 macrophages can promote endothelial cell migration and stem cell osteogenic differentiation by secreting various growth factors [59]. DIBS significantly increased macrophage secretion of bone morphogenetic protein-2 (BMP2) and vascular endothelial growth factor (VEGF) (Fig. S26, Supporting Information). Subsequently, we investigated the effects of CCR2-dependent M2 macrophage activation-mediated immune-osteogenic cascade regulation on osteogenic differentiation and angiogenesis. As illustrated in Fig. 7A, experiments utilized conditioned medium derived from macrophages and scaffold-BMSC constructs cultured under different conditions. Compared to the HA group, the M2 macrophage polarization induced by DIBS via the CCL2/CCR2 axis significantly promoted the proliferation of both HUVECs and BMSCs (Fig. 7A and B). Functional assays further demonstrated that M2 macrophage polarization substantially enhanced HUVEC migration, increased the number of tube-like structures, and improved cellular junction integrity (Fig. 7C and D). This pro-angiogenic effect was more pronounced under CCL2/CCR2-dependent activation. However, CCR2 inhibition resulted in insufficient vascularization induced by DIBS. Similarly, CCL2-induced M2 macrophages were conducive to BMSC osteogenic differentiation (Fig. 7E and F). It is noteworthy that Sr²⁺ itself possesses direct osteogenic and pro-angiogenic effects, potentially contributing to overall regeneration. Following the addition of inhibitors, DIBS still maintained certain angiogenic and osteogenic activities. This suggests that, in addition to DIBS-mediated immunomodulation, the physicochemical properties of materials such as Sr²⁺ may exert synergistic effects on regenerative repair. Our findings support a model wherein DIBS orchestrates a multi-faceted regenerative response that includes both direct Sr²⁺-mediated osteogenesis and CCL2/CCR2-dependent immunomodulation; the latter appears to be a critical upstream event that amplifies the regenerative microenvironment.

Fig. 7.

Revascularization and osteogenesis are reinforced by M2 macrophage activation via the CCL2/CCR2 pathway. A and B) HUVECs and BMSCs proliferation assay under M2 macrophage activation. Created with BioRender.com. C) Migration assay and quantification of HUVECs. D) Tube formation assay and quantification of HUVECs. E and F) Early and later osteogenic differentiation of BMSC influenced by macrophage-induced microenvironment. Data are represented as means ± SD, ∗p < 0.05 (vs Control), ∗∗p < 0.01 (vs Control), ∗∗∗p < 0.001 (vs Control), ∗∗∗∗p < 0.0001 (vs Control); ^$p < 0.05 (vs group without inhibitor), ^$$p < 0.01 (vs group without inhibitor), ^$$$p < 0.001 (vs group without inhibitor), ^$$$$p < 0.0001 (vs group without inhibitor). ns, not significant.

Following early inflammatory regulation, DIBS further enhances the immune function of MSCs by inducing secretion of the key immunomodulatory factor CCL2 to activate the M2 macrophage phenotype, thereby promoting angiogenesis and enhancing ESC osteogenic differentiation. Compared to conventional immunomodulatory strategies primarily focused on the initial macrophage response [10,11,60,61], DIBS incorporates the immunoregulatory functions of stem cells and their crosstalk with macrophages, thereby achieving multi-cellular immune regulation across a temporal sequence. This approach holds promise for promoting functional regenerative repair in critical-sized bone defects.

2.6. DIBS induces endogenous bone regeneration in a rat cranium defect model

The in vivo bone regeneration capability of DIBS was evaluated using a rat model of critical-sized cranial defects (φ = 8 mm) (Fig. 8A). Comprehensive histopathological evaluation of critical organs, including heart, liver, spleen, lung, and kidney, after 12 weeks of implantation showed no anomalies, confirming excellent in vivo biocompatibility of DIBS (Fig. S27, Supporting Information). After 4 weeks post-implantation, as evidenced by the percentage of new bone volume to total volume (BV/TV), DIBS exhibited the most pronounced bone repair effects (6.864 ± 0.28%) (Fig. S28D, Supporting Information), with evident new bone initiating growth from the edges of the scaffold along the multilevel micropore structure toward the center of the scaffolds (Fig. S28B, Supporting Information). DIBS also showed the statistically highest trabecular number (Tb.N) and the lowest trabecular separation (Tb.Sp), indicating the amount of new bone was high and dense (Fig. S28E and F, Supporting Information). In contrast, due to the lack of ability to rapidly recruit endogenous stem cells and to modulate the inflammatory microenvironment during the critical early osteogenic stage (Fig. 3G–I), insufficient bone healing was observed in the HA (BV/TV = 1.16 ± 0.23%) and control (BV/TV = 3.66 ± 0.22%) groups, with only scattered bone islands within the defect area.

Fig. 8.

DIBS induces endogenous bone regeneration in a rat cranium defect model. A) Schematic diagram of the establishment of critical-sized bone defects and the following experiment design. Created with BioRender.com. B) Representative micro-CT images of rat calvaria in different groups at week 12 after implantation. (Blue dotted circles: bone defect boundary; Yellow arrows: new bone bridging the scaffold edges; Red arrows/red dotted circles: new bone growing toward the center of scaffolds). C-E) Quantitative analysis of new bone calculated from micro-CT data, including BV/TV (C), Tb. N (D) and Tb.Sp (E). F) H&E and Masson staining at week 12 after implantation. (NB: new bone, V: vessels). Representative immunohistochemistry staining and semi-quantification of G, H) Runx2 cells, I, J) OPN cells, and K, L) OCN cells at week 12 after implantation. M, N) Immunofluorescence staining and quantification of type H vessels (CD31⁺ EMCN+) at week 12 after implantation. O, P) Immunofluorescence staining and quantification of Osterix + cells and EMCN + vessels at week 12 after implantation. Data are represented as means ± SD, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. ns, not significant.

After 8 weeks and 12 weeks post-implantation, three-dimensional micro-CT reconstructions (Fig. S28C, Supporting Information) revealed that DIBS induced bone tissue to grow deeper into the scaffold at week 8, evolving into a bone-scaffold complex with nearly complete defect restoration by week 12, surpassing the OCHA group in bone regeneration. Coronal and sagittal views confirmed the formation of a complete osseous bridge in DIBS (Fig. 8B). In contrast, the control and HA groups showed minimal new bone formation, primarily at the defect periphery, at weeks 8 and 12. Bone regeneration parameters, including BV/TV, Tb. N, and Tb. Sp (Fig. S28D–F, Supporting Information), corroborated the micro-CT results. By week 12, the BV/TV ratio of DIBS was 23.97 ± 1.54% (Fig. 8C), significantly higher than that of other groups (p < 0.01). Furthermore, DIBS exhibited the significantly highest values for Tb. N (1.35 ± 0.04 mm⁻¹), and the lowest values for Tb.Sp (0.50 ± 0.01 mm) (p < 0.01) (Fig. 8D and E).

Hematoxylin and Eosin (H&E) and Masson's trichrome staining revealed that at both 8 and 12 weeks post-implantation, the control group (no scaffold) still exhibited significant bone defects. The HA group showed minimal new bone formation, primarily at the defect periphery. This demonstrated the inherently limited regenerative capacity of the body for large-scale bone defects and highlighted the insufficient restoration potential of HA scaffolds, which poses challenges for self-healing after trauma [62]. In contrast, histological analysis at 8 weeks demonstrated that the OCHA and DIBS, leveraging their unique architecture of the dynamic hydrogel hierarchical microporous network embedded within ceramic macropores, guided bone tissue ingrowth deep into the scaffold—mature bone tissue formed in regions approaching the defect center (Fig. S29, Supporting Information).

The immunophenotypic analysis results demonstrate that DIBS maintains a significantly higher M2/M1 ratio at 8 weeks compared to HA and OCHA groups. Moreover, we quantified CD105⁺ MSCs co-localized with CCL2 at this time points. CCL2 expression remained elevated at 8 weeks in the DIBS group, confirming the mid-phase immunoregulatory function of MSCs as hypothesized (Fig. S30–31, Supporting Information). By week 12, new bone formation increased most significantly, with larger bone surface areas growing towards the scaffold center and establishing robust osseointegration. This effect was particularly pronounced in DIBS (Fig. 8F). Furthermore, it was observed that abundant new bone attached to the inner surfaces of the ceramic scaffold macropores and infiltrated the hierarchical micropores of the dynamic hydrogel, effectively utilizing the hierarchical porous structures of DIBS to achieve multilevel osseointegration. Compared to the HA and OCHA groups, the functionalization with Sr²⁺ ions significantly enhanced this multilevel osseointegration process, further underscoring the advantage of inorganic bioactive elements in osteogenic differentiation. Immunohistochemical analysis targeting Runx2, OPN, and OCN at week 12 (Fig. 8G–L) demonstrated that DIBS exhibited the most intense staining for these proteins, surpassing the HA and OCHA groups, indicating superior osteogenic activity.

Recent studies have revealed that Type H vessels, characterized by high co-expression of Endomucin (Emcn) and CD31, mediated angiogenesis-osteogenesis coupling during bone repair. These vessels exert paracrine effects on the proliferation and differentiation of perivascular osteoprogenitors (highly expressing Osterix), thereby creating a vascularized microenvironment to accelerate osteogenesis [63]. IF analysis confirmed significantly increased CD31⁺Emcn⁺ Type H vessels in DIBS at 12 weeks (Fig. 8M and N). Besides, we observed that Osterix osteoprogenitors surrounded Type H vessels. Notably, DIBS exhibited dramatically greater abundance of Emcn⁺ vessels and associated Osterix⁺ osteoprogenitors in regenerated bone compared to that of the HA and OCHA groups (Fig. 8O and P). These results demonstrate that synergistic hierarchical structures and bioactive cues enabled vascularized new bone to infiltrate into the depth of the scaffold progressively. The generation of blood vessels could facilitate the further maturation and functional remodeling of bone tissue.

2.7. DIBS induces functional bone regeneration in a rabbit femoral defect model

DIBS not only directly regulates macrophage polarization towards the M2 phenotype during early inflammatory regulation (Fig. 3, Fig. 4), but also enhances the immunomodulatory capacity of MSCs, as demonstrated by in vitro and in vivo functional experiments (Fig. 6, Fig. 7). This is attribute to the pH-responsive degradation design aligning with the inflammatory phase (acidic environment). The release of Sr²⁺ and PA is sustained over the 30-day in vitro study and is expected to continue at lower levels in vivo (Fig. 2K and L), which aligns well with the timeframe for immunomodulation, ESC recruitment and immune phenotype reprogramming, and early osteogenesis (weeks 2-4). To further investigate the potential of DIBS-induced immunomodulatory microenvironment in promoting functional bone regeneration, DIBS (4 × 15 mm) were implanted into rabbit femoral bone marrow cavity models. The 12-week immunofluorescence staining in vivo revealed significant co-expression of CCL2 and CD105 in DIBS-treated regenerated bone, highlighting the enhanced immunomodulatory function of MSCs from DIBS (Fig. S32, Supporting Information). Compared to the HA group, DIBS exhibited lower expression of CD86, a marker of M1-type macrophages, while the expression of CD206, a representative marker for M2-type macrophages, was markedly upregulated. Moreover, the expression levels of osteogenic protein OCN and angiogenic factor VEGF are closely associated with the increase in M2-type macrophages. These findings indicate that DIBS not only directly induces macrophage polarization toward the M2 anti-inflammatory phenotype in the early stage but also simultaneously enhances the immunomodulatory capacity of BMSCs during the osteogenic phase. This supports its role in regulating the immune microenvironment across the bone healing process to promote bone regeneration.

As revealed by Micro-CT, DIBS exhibited superior bone regeneration with extensive new bone formation covering the large-sized defects (Fig. 9A). Bone regeneration parameters, including BV/TV, Tb.Sp, Bone surface area/bone volume (BS/BV), trabecular thickness (Tb.Th), connectivity density (Conn.D), corroborated the micro-CT results (Fig. 9B). Sequential fluorescent labeling with tetracycline and calcein demonstrated that DIBS had a new bone mineralization rate of 6.59 μm/day, which was 3.7 times higher than that of HA scaffolds, emphasizing its strong osteoinductive capacity during late-stage osteogenesis (Fig. 9E). H&E and Masson's trichrome staining revealed more complete, extensive, and dense mature bone tissue formation within the core region of DIBS (Fig. 9C–F). Safranin O/Fast Green staining further showed distinct Safranin O-positive staining areas in both groups (Fig. 9C), indicating bone repair proceeded through endochondral ossification [10]. This mechanism is attributed to sustained release of Ca²⁺ and PO₄³⁻ ions from the HA ceramic framework (Fig. S9A, Supporting Information), facilitating biomineralization of the cartilaginous matrix [10]. Directional analysis of Masson-stained sections revealed disorganized collagen matrix orientation in the regenerated bone of the HA group. In contrast, DIBS orchestrated an orderly directional growth of the collagen matrix (Fig. 9D): well-aligned lamellar bone formed in macropores parallel to the femoral long axis (Fig. 9G). Meanwhile, the regenerated collagen matrix perpendicular to the femoral long axis exhibited parallel orientation extending toward the scaffold center (Fig. 9H). This highly organized bone microstructure is expected to enhance osseointegration and mechanical strength [64]. In summary, DIBS effectively remodels the osteoimmune microenvironment, and in concert with the direct osteogenic activity of its components (e.g., HA framework, Sr²⁺ ions), induces organized collagen arrangement, facilitating functional bone tissue repair.

Fig. 9.

Functional bone regeneration is enhanced by DIBS in vivo. A) Micro-CT assessment and Calcein/Tetracycline staining at 12 weeks post-surgery for each group. B) Quantitative analysis of new bone calculated from micro-CT data. C) H&E, Masson's trichrome, and Safranin-O/Fast Green (SO/FG) staining at 12 weeks postoperatively. D) Trichrome staining discovered the newly formed collagenous fiber orientation of the ingrowth bone tissue in the HA and DIBS groups at 24 weeks postoperatively. The false color images indicate the orientation of collagenous fibers. E) Analysis of mineral deposition rates in the HA and DIBS groups at 12 weeks postoperatively. F) New bone area in the HA and DIBS at 24 weeks postoperatively. G, H) Angle distribution of the newly formed collagenous fibers after 12 weeks of operation: (G) the angle distribution of collagen fibers parallel to the femoral long axis and (H) the angle distribution of collagen fibers perpendicular to the femoral long axis. Data are represented as means ± SD, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Although DIBS achieves functional bone tissue regeneration throughout the entire repair cycle via a multi-cell synergistic immunomodulatory strategy, it is essential to consider the limitations inherent in current research.

First, while our results delineate a clear temporal and functional association between DIBS, the emergence of an immunoregulatory MSC phenotype, and subsequent M2 macrophage polarization via CCL2/CCR2, we acknowledge that this represents a correlative regulatory framework rather than a definitively proven causal sequence. The remarkable bone regeneration observed with DIBS is likely the result of a synergistic interplay between multiple factors (e.g., immunomodulation, HA framework, Sr²⁺ ions). Future studies employing targeted loss-of-function models, such as MSC-specific CCL2 knockout or macrophage depletion, are warranted to definitively dissect the non-redundant role of each component within this cascade. Second, we have not conducted more systematic investigations into the long-term degradation and absorption of DIBS in vivo. In the future, we will further explore its correlation with new bone formation and remodeling. Third, although CCL2 was validated as a pivotal immunomodulator coordinating stem cell-macrophage crosstalk and osteogenesis, other significantly upregulated immunomodulatory genes (e.g., Sfrp1, CXCL1, CCL7) warrant investigation into their mechanistic roles in bone repair. Moreover, the effects of BMSCs immunoregulation need to be clarified in more microenvironments, such as osteoporosis, autoimmune diseases, and tumorigenesis [65]. Fourth, the translation of DIBS to clinical practice would require large animal studies with functional loading and long-term safety assessment. Finally, the potential for excessive immune activation leading to adverse effects must be carefully monitored in future preclinical evaluations.

On the other hand, the exceptional anti-inflammatory and immunomodulatory properties of DIBS render it suitable not only for critical-sized bone defects caused by trauma, but also for challenging clinical scenarios with compromised healing, such as osteoporotic bone defects, infected non-unions, and bone defects in elderly patients, where the inflammatory microenvironment is particularly hostile. Moreover, the modular design of DIBS permits the exchange of metal ions (such as Mg²⁺ and Zn²⁺) to tailor the scaffold for specific disease microenvironments, thereby broadening its potential clinical utility. Bone organoid-based strategies represent a promising frontier for viable bone tissue engineering [66]. The organic/inorganic composition (dynamic hydrogel/HA bioceramic) and hierarchical structure of DIBS align closely with the unique matrix requirements for bone organoids. Thus, future integration with organoid technology may enable breakthrough progress toward complete functional regeneration of large-scale bone defects. Through tackling these challenges and opportunities, we believe that the DIBS holds considerable translational potential for clinical bone tissue engineering.

3. Conclusion

In summary, we have engineered a dual-immunomodulatory bone scaffold (DIBS) that sequentially programs the immune-osteogenic cascade to drive functional regeneration of critical-sized bone defects. By integrating a pH-responsive immunomodulatory hydrogel containing Sr-PA nanoparticles within a 3D-printed macroporous HA bioceramic framework, DIBS achieved spatiotemporal control over immunomodulator and calcium and phosphate release, effectively coupling inflammatory resolution with sustained regenerative modulation. DIBS orchestrated a two-phase immunomodulatory strategy where OXG initially directed macrophages toward an anti-inflammatory M2 phenotype, and subsequently, the acidic microenvironment triggers the release of Sr²⁺ and PA from nanoparticles to reprogram homing MSCs into an immunoregulatory phenotype, thereby amplifying the regenerative immune microenvironment. The single-cell transcriptomics and validation demonstrate that DIBS-educated MSCs communicate with macrophages via the CCL2/CCR2 axis to reinforce M2 polarization, enhance angiogenesis, and promote osteogenic differentiation, enabling precise regulation of the healing process during bone repair. Compared to conventional immunomodulatory scaffold materials, the designed scaffold exhibited dual immunoregulatory functions that sequentially regulated macrophage phenotypic polarization and enhanced MSCs immunoreactivity. Simultaneously, the hydrogel network and the 3D-printed macroporous HA framework form an interpenetrating-phase composite structure, which mimicked the bone tissue hierarchical architecture and components to support cell recruitment, migration, and efficient differentiation. Through regulating immune microenvironment throughout the entire cycle, the scaffold achieved robust functional bone regeneration in rat critical-sized cranial defects and rabbit femoral defects. This study highlights the critical roles of temporal immune modulation and MSCs immunoreprogramming in endogenous bone regeneration, offering a novel pathway for designing advanced immunomodulatory biomaterials for various immunomodulatory and tissue regeneration applications.

4. Experimental section

4.1. Material

Xyloglucan (XG, Mw = 662,990 Da) was purified from tamarind gum (Tokyo Chemical Industry, Japan) [49]. Carboxymethyl chitosan (CMCS, degree of substitution ≥80%, Mw = 311,231 Da) was purchased from Macklin (China). Strontium chloride (SrCl₂·6H₂O, 99%), 3,4-dihydroxybenzaldehyde (98%, also known as protocatechualdehyde (PA)) were purchased from Yuanye Bio-Technology (China). Sodium periodate (NaIO₄, AR) and ethylene glycol (>99%) were purchased from Sinopharm Chemical Reagent (China). The powdered materials that were used were 15 μm and 200 μm HA (Macklin, China). The photosensitive resins (HDDA, HEMA, and TMPTA) were produced by Si Cheng New Materials Co., Ltd., China. The photoinitiator that was used was diphenyl (2,4,6-trimethylbenzoyl) phosphine oxide produced by BASF (Germany). Fetal bovine serum (FBS), penicillin/streptomycin (PS), trypsin solution, and DMEM medium were purchased from Gibco (USA).

4.2. Preparation and characterization of Sr-PA nanoparticles

PA were used to modify SrCl₂ (Sr-PA) through a hydrothermal reaction. A hydrothermal reactor with a capacity of 50 mL was charged with 20 mM SrCl₂ and 40 mM PA for the synthesis of Sr-PA (the molar ratio of Sr²⁺ to PA was 1:2). Gradually adjust the pH to alkaline (9–10) after mixing the solution well under acidic or neutral conditions. The reactor was sealed and heated at 90 °C for 12 h. After the hydrothermal reaction, the crude product was collected by centrifugation at 8000 rpm for 5 min. Further purification was achieved by washing the product with ethanol and water for at least three cycles. Finally, the purified Sr-PA was obtained through freeze-drying.

The scanning electron microscopy (SEM) images and energy dispersive (EDS) data of Sr-PA nanoparticles were obtained using a Hitachi S-4800 microscope from Japan. Fourier transform infrared spectroscopy (FT-IR, SHIMADZU-IRTracer-100, Japan) was employed with a 2 cm⁻¹ resolution in the wavenumber range of 400–4000 cm⁻¹. Thermogravimetric analysis (TGA) was performed using a thermogravimetric analyzer (NETZSCH STA 2500) from Nai, Germany, under an N₂ atmosphere with a heating rate of 10 °C per minute.

4.3. Preparation and characterization of pH-responsive dynamic hydrogel

The Oxidized xyloglucan (OXG) was synthesized via an oxidation reaction based on our previous method. To prepare the hydrogels, 5%(w/v) CMCS solution, 3%(w/v) OXG solution and 2%(w/v) Sr-PA nanoparticles solution using DI water as a solvent. Then, hydrogels were prepared by mixing the CMCS solution with OXG, or OXG/Sr-PA solutions at optimal concentrations and ratios. The cross-sectional morphology of the samples was observed using a scanning electron microscope (SEM, S-4800, Hitachi, Japan). The changes in the elemental composition and chemical bonding of the hydrogels were analyzed using XPS (Fisher-K-Aipha, Thermo, UK) with a monochromatic Al Kα source.

The rheological properties of the hydrogel were measured using a Discovery HR-2 (TA Instruments, USA) dynamic rheometer with a 20 mm plate (stainless steel). Different models, including time-sweep (600 s, strain of 10%, frequency of 10 rad s⁻¹), oscillation strain-sweep (strain of 1-1000%, 10 rad s⁻¹), and alternate strain-sweep test (strain of 1% and 250% for four cycles, each time is 100 s, strain of 10%, 10 rad s⁻¹).

4.4. Preparation and characterization of hydroxyapatite (HA) framework

The details on the preparation of porous HA framework can be found in the supporting information.

4.5. Preparation and characterization of DIBS

The 3 wt% PA solution was prepared by dissolving PA in distilled water at 70 °C. HA frameworks were immersed in the PA solution under continuous stirring in the dark for 24 h. During the reaction, 1 M NaOH aqueous solution was added dropwise to adjust the pH to 9.8–10. Following 24 h incubation, HA frameworks were thoroughly rinsed with distilled water to remove residual reagents and unreacted PA. The resulting PA-modified HA frameworks were subsequently lyophilized overnight. Fourier transform infrared spectroscopy (FT-IR, SHIMADZU-IRTracer-100, Japan) was employed with a 2 cm⁻¹ resolution in the wavenumber range of 1000–4000 cm⁻¹. The changes in the elemental composition and chemical bonding of the scaffolds were analyzed using XPS (Fisher-K-Aipha, Thermo, UK) with a monochromatic Al Kα source. The PA-modified and unmodified HA discs were bonded with the hydrogel, and the interfacial bonding strength was measured using a universal testing machine (UTM5305H, China).

Dynamic hydrogel precursors were prepared as follows: Precursor A contained OXG and Sr-PA, while precursor B consisted of CMCS. Both components were vortex-mixed for 30 s and loaded into separate chambers of a dual-barrel syringe. The precursors were co-injected into PA-modified porous HA frameworks. Within 2 min before gelation, three reciprocating injections were performed to ensure homogeneous infiltration. The negative-pressure infiltration system (YUWELL) was subsequently applied to reinforce the interfacial integration between the dynamic hydrogel and HA framework Fig. S1. According to experimental designs, two scaffolds were fabricated: without Sr-PA nanoparticles (denoted OCHA) and Sr-PA nanoparticles incorporated (DIBS), with unmodified HA scaffolds serving as controls.

The morphology and EDS elemental mapping of different scaffolds by scanning electron microscopy (SEM, S-4800, Hitachi, Japan). The pore size of scaffolds in the SEM images was analyzed using ImageJ software (NIH, Bethesda, MD). Gross images of different scaffolds were taken by a stereo microscope (Leica, Germany). The compressive strength of scaffolds was tested by a universal testing machine equipped with a 1000 N load sensor (Zwick Roell-Z100, Germany). The compression strength of scaffolds was measured with a crosshead speed of 0.5 mm/min at room temperature. The toughness of the ceramic scaffolds was obtained by calculating the area under the compressive stress-strain curve, and the elastic modulus of the scaffolds was obtained by calculating the slope of the initially linear region of the stress-strain curve. To measure the release of ions, different scaffolds (8 mm diameter × 4 mm height) were immersed in 10 mL of PBS solution and placed on a 37 °C shaker (n = 3 independent samples per group). At each time interval, the supernatant was collected by centrifugation and replaced with 10 mL of fresh PBS solution. Ca and P ion concentrations were detected by an inductively coupled plasma atomic emission spectroscopy (ICP-OES, Agilent-7900, USA).

To simulate the release behavior of Sr-PA from DIBS responding to an acidic inflammatory microenvironment, we placed sterilized DIBS in buffers with pH values of 7.4 and 6.0. At each predetermined time interval, 1.0 mL of solution was removed from the buffer and replaced with an equal volume of fresh buffer to maintain constant volume. Released Sr²⁺ was quantified using an inductively coupled plasma emission spectrometer (ICP-OES, Agilent-7900, USA). The concentration of PA in the release medium was measured at 280 nm wavelength using a UV/visible spectrophotometer (Shimadzu 2600i, Japan). Each sample underwent three parallel measurements. This experimental setup enabled monitoring of the release profiles of S²⁺ and PA over time in the DIBS.

In vitro degradation testing involved immersing DIBS in sterile buffers at pH 6.0 and pH 7.4, maintained at 37 °C. The initial weight (W₀) of the lyophilized scaffold was recorded before immersion. The residual weight (W_n) of the lyophilized scaffold was recorded on day n (n = 4, 8, 12, 16, 20, 30) after immersion. The mass remaining (%) was calculated using the following formula:

$$\text{Mass}\text{remaining}(\%)=\frac{Wo-Wn}{Wo}\times 100\%$$ (1)

4.6. Molecular docking experiments

The structures of OXG and CMCS were optimized using MD simulations. Ligand preparation was performed with AutoDock Raccoon to generate PDBQT files. The fibronectin receptor (PDB ID: 1FNH) was retrieved from the RCSB Protein Data Bank. The protein structure was processed in AutoDock by removing water molecules, adding hydrogen atoms, and assigning Gasteiger charges. Molecular docking between ligands and the receptor was executed using AutoDock Vina to compute binding affinity energies. Resultant binding poses were visualized and analyzed in Discovery Studio.

4.7. Cell culture

BMSCs were isolated from the femurs of 3-week-old SD rats by flushing the bone marrow with low-glucose Dulbecco's Modified Eagle Medium (DMEM, Gibco). Cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37 °C in 5% CO₂. Non-adherent cells were removed after 24 h, and the medium was changed every 3 days. Passage 3–5 cells were used for all experiments.

Mouse macrophage cells (RAW264.7) (CX0022, BOSTER) were maintained in high-glucose DMEM (Procell, Wuhan, China) with 5% FBS and 1% PS. Second passage RAW264.7 cells were used for subsequent assays.

Human umbilical vein endothelial cells (HUVECs) (iCell-h110, icellbioscience) and endothelial cell medium (ScienCell) supplemented with 5% FBS, 1% PS, and 1% endothelial cell growth supplement were used. Second passage HUVECs were employed for experiments.

4.8. In vitro biocompatibility assessment

Cell viability was evaluated via CCK-8 assays and live/dead staining. For cytotoxicity assays, 1 × 10⁵ BMSCs were seeded on scaffolds in 24-well plates. After 24 h, 500 μL CCK-8 working solution (Beyotime, China) was added per well. Following 40 min incubation at 37 °C, optical density (OD) was measured using a microplate reader (NanoQuant M200 Pro, Tecan). Cell viability (%) = [(OD_exp - OD_blank)/(OD_neg - OD_blank)] × 100%, where complete medium and 10% DMSO served as negative and positive controls, respectively. For live/dead cell staining, BMSCs were stained with 2 μM Calcein-AM and 2 μM Propidium Iodide (Beyotime, China) for 15 min at 37 °C. After PBS washing to remove unbound dye, fluorescence images were captured using a fluorescence microscope (IX73, Olympus, Japan). Cell proliferation was assessed on days 1, 3, and 5 using CCK-8 (n = 3 per group). At each time point, after removing the culture medium and washing the wells with PBS three times, each well was treated with serum-free medium containing 10% CCK-8 solution. After incubation at 37 °C for 40 min, absorbance at 450 nm for each group was measured using a microplate reader.

To evaluate cell migration, proliferation, and distribution on different scaffolds, cytoskeletal staining was performed. After 3 and 5 days of culture, cell-seeded scaffolds were fixed with 4% paraformaldehyde (15 min), permeabilized with 0.1% Triton X-100 (10 min), and blocked with 1% BSA (1 h). F-actin was stained with FITC-phalloidin (30 min), followed by DAPI nuclear counterstaining (5 min). After thorough PBS washing, the cell 2D spreading morphology was imaged using the confocal laser scanning microscope (CLSM, Zeiss LSM880, Germany), and the cell number was counted. For cell 3D distribution analysis, z-stack images with a stack of 2 μm and a height of 200 μm were acquired at day 7, and 3D images were reconstructed using ZEN software (Zeiss, United Kingdom).

4.9. In vitro immunomodulation analysis

RAW264.7 macrophages were cultured in DMEM supplemented with 10% FBS and 1% PS and seeded on scaffolds or in 24-well plates for 7 days (n = 3). Macrophage polarization was evaluated by flow cytometry to assess the levels of M1 (CD86) and M2 (CD206) phenotypes. Immunofluorescence was employed to examine the anti-inflammatory (M2) and pro-inflammatory (M1) phenotypes and to observe cell morphology. Subsequently, the expression of inflammation-related genes and proteins (M1: iNOS and TNF-α; M2: CD206 and Arg-1) was further analyzed using quantitative real-time polymerase chain reaction (qRT-PCR) and Western blotting (WB). The primers (5′-3′) used in this study are listed in Table S3. Cytokine levels (TNF-α, IL-6, TGF-β, and IL-10) were measured using commercial ELISA kits (Elabscience), following the manufacturer's instructions.

4.10. Enzyme-linked immunosorbent assay (ELISA)

Depending on the experimental design, harvest different culture medium supernatants and centrifuge at 300 g for 20 min. Then cytokines were measured using an ELISA kit (Shanghai Enzyme-linked Biotechnology Co., China).

4.11. Immunofluorescence staining

Fixed cells, cell-seeded scaffolds, or slides were blocked with a blocking buffer (ZSGB-BIO, China) for 1 h at room temperature. Subsequently, primary antibodies were then applied and incubated overnight at 4 °C: CD105 (1:500, Proteintech, 10862-1-AP), CD44 (1:400, Proteintech, 60224-1-Ig), CD86 (1:200, Santa, sc-19617), iNOS (1:200, ABclonal, A3774), and CD206 (1:200, Abcam, ab64693), Vinculin (1:200, HUABIO, ET1705-94). After three washes with PBS, species-specific secondary antibodies conjugated to ABflo®488 (ABclonal, AS073/AS076) or ABflo®594 (ABclonal, AS054/AS039) were diluted 1:200 and incubated with samples at 37 °C for 1 h. Following thorough PBS washes, cell nuclei were counterstained with DAPI for 5-10 min. Images were captured using a confocal laser scanning microscope (CLSM, Zeiss LSM880, Germany). The positive cell ratio and mean fluorescence intensity (MFI) were quantified via ImageJ software. Detailed specifications of primary and secondary antibodies are provided in Tables S1 and S2 (Supporting Information).

4.12. Western blot (WB)

Total proteins from RAW264.7 were extracted using RIPA lysis buffer (Beyotime, China) containing 1 mM PMSF (Solarbio, China) on ice. Protein concentrations were determined via a bicinchoninic acid (BCA) assay kit (Beyotime, China). After denaturation, proteins were separated on 10% SDS-polyacrylamide gels and transferred to the polyvinylidene fluoride (PVDF, Sigma, USA) membranes. Membranes were blocked with 5% defatted milk for 1 h at room temperature, followed by overnight incubation at 4 °C with primary antibodies (1:1000 dilution). Subsequently, membranes were incubated with HRP-conjugated secondary antibodies (1:50,000, Beyotime, A0208 & A0216) for 1 h at room temperature. Protein bands were visualized using an automatic chemiluminescence image analysis system (Tanon 4800, China). GAPDH or Actin served as the loading control, and relative expression levels were quantified. Band intensities were analyzed using ImageJ software. Primary and secondary antibody details are provided in Tables S1 and S2 (Supporting Information).

4.13. Quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was extracted using a Total RNA kit (NOBELAB, China) according to the manufacturer's protocol, and the concentration of RNA was determined with a spectrophotometer (NanoDrop 2000, Thermo Fischer, USA). Then, RNAs were reverse transcribed into complementary DNA (cDNA) using the ReScript™II RT SuperMix kit with gDNA Eraser (NOBELAB, China). qRT-PCR was conducted using SYBR® Premix UrTaq™ II (NOBELAB, China) on a Real-Time PCR system (Quant Studio 5, Thermo Fischer, USA). GAPDH or Actin was utilized as a normalization control, and the fold change of gene expression was calculated by the 2^−ΔΔCt method (n = 3 independent samples per group). The specific primer sequences are shown in Table S3 in the Supporting Information.

4.14. Flow cytometry analysis

The surface markers of macrophages were identified using flow cytometry. Macrophages were collected to obtain a single-cell suspension. Then, cells were stained with APC-conjugated rat monoclonal anti-mouse CD86 antibody (BioLegend, USA) and Brilliant Violet 421™-conjugated rat monoclonal anti-mouse CD206 antibody (BioLegend, USA) for 15 min at room temperature. Finally, the cell suspension was then assessed on a CytoFLEX LX flow cytometer system (BECKMAN Coulter, USA) and analyzed using FlowJo software.

4.15. Single-cell RNA sequencing and data analysis

Given the pivotal role of the immunomodulatory function of MSCs in regulating macrophage polarization and maintaining immune homeostasis during the subsequent regenerative phase, the effects of DIBS on the immunomodulatory function of BMSCs were further investigated. After 14 days, BMSCs were dissociated from scaffolds for single-cell suspension preparation. Isolated cells were subsequently subjected to scRNA-seq analysis using the 10x Genomics platform, followed by sequencing on the Illumina NovaSeq™ X Plus to generate a comprehensive dataset (Hangzhou Kaitai Biotechnology Co., Ltd., Hangzhou, China). For specific analysis methods, see the Support Information.

4.16. Intramuscular implantation in CCR2-inhibited mice

Thirty male BALB/c mice (6 weeks old, ∼20 g) were utilized in this trial. Throughout the experimental period, sustained CCR2 inhibition was achieved via daily intraperitoneal injections (2 mg/kg) of the highly selective CCR2 inhibitor RS504393 (Cat. No. HY-15418, MCE). Before each administration, RS504393 (10 mM stock) was diluted in sterile corn oil (Cat. No. HY-Y1888, MCE) to the target concentration. RS504393 was administered via daily intraperitoneal injection from day 1 post-implantation until the day of euthanasia (day 14), ensuring continuous CCR2 blockade throughout the implantation period. The Control group received equivalent injections of inhibitor-free solvent, and the normal group received no processing. On day 7, peripheral blood was collected, followed by red blood cell lysis (Lysing Solution, BD) and staining with PE-conjugated anti-mouse F4/80 antibody (BD Pharmingen, USA) and APC-conjugated anti-mouse CD192 antibody (BioLegend, USA). Flow cytometry was performed to evaluate the CCR2 inhibition effect in macrophages. After anesthesia, scaffolds (diameter = 5 mm, height = 2 mm) were implanted into the bilateral thigh muscles. Implants were harvested 14 days post-surgery (n = 3) for macrophage phenotypic analysis.

4.17. Proliferation and polarization analysis of macrophages in a CCR2-inhibited model

An in vitro co-culture model of BMSCs and macrophages was established. Specifically, BMSCs (1 × 10⁵ cells) seeded on scaffolds were placed in the upper chamber of the transwell insert, while macrophages (1 × 10⁵ cells) were cultured in the lower chamber. After 3 days of co-culture, CCK-8 assays were performed on days 1 and 3 to assess proliferation. For polarization analysis, mRNA expression of M1 (iNOS) and M2 (CD206) phenotypic markers in macrophages was quantified by qRT-PCR. For comparison, macrophages were stimulated with 40 ng/mL CCL2 or 1 μg/mL RS504393 in the transwell model and served as positive or negative controls.

4.18. HUVECs and BMSCs biological functions analysis under the effect of CCL2/CCR2-mediated M2 macrophage activation

In this experiment, scaffold-BMSC constructs were co-cultured with macrophages to obtain conditioned supernatants. Supernatants supplemented solely with CCL2 or RS504393+ CCL2 served as positive or negative controls, respectively. Conditioned media for culturing HUVECs and BMSCs were prepared by mixing the supernatants with basal medium at a 1:1 ratio. For proliferation analysis, HUVECs and BMSCs were cultured in their respective conditioned media for 3 days, and CCK-8 assays were performed on days 1 and 3. To assess HUVEC migration, cells (1 × 10⁵ per well) were seeded into the upper chambers of 5-μm pore transwell inserts, while HUVEC-conditioned media was added to the lower chambers. After 24 h, migrated cells were quantified. For tube formation analysis, HUVECs (5 × 10⁴ cells) were seeded on matrigel-coated plates and treated with HUVEC-conditioned media. Tube structures were imaged under an inverted microscope after 6 h, and parameters (including tube meshes and junctions) were quantified using ImageJ. For the analysis of osteogenic differentiation of BMSCs, BMSCs were cultured in conditioned medium, and after 7 days, ALP was stained and ALP activity was measured. After 21 days, calcium nodules were stained and semi-quantitative analysis was performed.

4.19. Animal surgery

Eighteen male Sprague-Dawley (SD) rats (≈6 weeks old, weight ≈180–200 g) were used. Briefly, critical-sized calvarial defects (diameter = 8 mm) were created at the midline of each rat's cranium using a trephine drill under anesthesia induced by intraperitoneal injection of 1% pentobarbital sodium (50 mg kg⁻¹, Sigma). All scaffolds were sterilized by ethylene oxide gas before implantation. Then sterilized cylindrical scaffolds (diameter 8 mm, thickness 1 mm) were implanted into the defects, and the incisions were sutured correctly. At different time points after implantation, rats were euthanized with an overdose of pentobarbital sodium, and the scaffold-containing defect tissues were harvested for immunofluorescence analysis.

Forty-eight male SD rats (≈6 weeks old, weight ≈180–200 g, n = 4 per group per time point) were randomly divided into four groups: control (defect only), HA (defects received HA scaffold), OCHA (defects received OCHA scaffold), and DIBS (defects received DIBS scaffold). After anesthesia via intraperitoneal pentobarbital sodium (50 mg kg⁻¹), critical-sized calvarial defects (d = 8 mm) were created, followed by implantation of sterilized cylindrical scaffolds (diameter 8 mm) and suture closure. Rats were euthanized at 4, 8, and 12 weeks after implantation, and the defect regions with scaffolds were harvested for characterization.

Twelve male New Zealand rabbits (≈10 weeks old, weight ≈3 kg, n = 6 per group per time point) were randomly assigned to two groups: HA (defects received HA scaffold, n = 6) and DIBS (defects received DIBS scaffold, n = 6). Anesthesia was induced by intramuscular injection of xylazine (0.1 mL kg⁻¹), followed by ear vein injection of 3% pentobarbital sodium (0.1 mL kg⁻¹) after disinfection with 75% ethanol. A bone defect (diameter = 4 mm, depth = 15 mm) was created at the knee joint center using a trephine. Sterilized cylindrical scaffolds (diameter 4 mm, height 15 mm) were implanted and sutured. At 12 weeks post-implantation, rabbits were euthanized, and the defect regions with scaffolds were harvested for characterization.

4.20. Fluorescent sequential labeling

To investigate the role of DIBS in inducing new bone formation and mineralization rates, sequential fluorescence labeling was employed. On postoperative weeks 10, and 12, different fluorescent labels were administered via intramuscular injections: tetracycline hydrochloride (25 mg/kg) and calcein (50 mg/kg). Deposition Rate: Measure the distance between consecutive fluorescent labels (representing bone formation over time) and divide by the time interval between label injections to determine the rate of mineral deposition (μm/day).

4.21. Micro-CT scanning

The harvested specimens were fixed in 4% paraformaldehyde solutions for 7 days and then scanned by a high-resolution micro-CT scanner (Quantum GX2, PerkinElmer, Japan). The acquired data were imported into CTvox software (Bruker, Germany) for visualization and analysis. Bone parameters of the interested region (ROI) for each sample were calculated using CTAn software (Bruker, Germany).

4.22. Histological analysis

Hard Tissue Processing. Specimens were fixed in 4% paraformaldehyde, dehydrated through a graded ethanol series, and embedded in methyl methacrylate resin without decalcification. Undecalcified histological sections were prepared using the hard tissue cutting grinding system (EXAKT Vertriebs GmbH, Germany) and stained with hematoxylin and eosin (H&E), Masson's trichrome, and Van Gieson. They were then evaluated for bone repair, new bone formation, and osseointegration of the scaffold using the digital slide scanner (Hamamatsu NanoZoomer, C13210-01, Japan).

Decalcified Tissue Processing. Bone specimens were fixed in 4% paraformaldehyde for 72 h. Bone tissues were decalcified in 10% (w/v) EDTA (pH 7.4) for 30-50 days at 4 °C with daily solution renewal. Then the decalcified tissues were dehydrated and paraffin-embedded. Frozen sectioning was then performed at 3 μm thickness. Staining protocols included H&E, Masson's trichrome, and Safranin O/Fast Green. All slides were digitized using a slide scanner (Hamamatsu NanoZoomer, C13210-01, Japan). New bone ingrowth into scaffolds was semi-quantitatively analyzed via ImageJ. The directionality of aligned collagen matrix and new bone matrix was analyzed using the “Directionality” and “OrientationJ” plugins in Fiji software. To further evaluate stem cell recruitment, immune regulation, bone tissue formation, and angiogenesis, immunofluorescence and immunohistochemical staining were performed.

4.23. In vivo safety analysis

Twelve weeks after surgery, some rats were euthanized using the same method to obtain primary organ samples. The collected key organs, including heart, liver, spleen, lung, and kidney samples, were fixed, sectioned at 5 μm, and stained with H&E. Panoramic images were acquired for pathological evaluation using the digital slide scanner (Hamamatsu NanoZoomer, C13210-01, Japan).

4.24. Statistical analysis

The statistics were conducted using GraphPad Prism software (GraphPad Software Inc.) and Origin 2018 (OriginLab Corporation, USA), employing the Student's t-test (unpaired and two-tailed) and one-way ANOVA, followed by Tukey post hoc test. Each group of data undergoes no less than three tests under the same conditions. All data were expressed as mean values with standard deviation (SD) unless otherwise stated. P < 0.05 indicated that the difference was statistically significant.

CRediT authorship contribution statement

Hongyu Zhao: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Qing Wang: Visualization, Methodology, Formal analysis, Data curation. Wei Lin: Software, Methodology, Investigation. Rui Qiao: Visualization, Software, Investigation. Yaning Wang: Methodology, Formal analysis. Teng Xu: Methodology. Hongyu Xing: Software, Resources. Yen Wei: Writing – review & editing, Project administration. Jing Chen: Writing – review & editing, Resources, Project administration, Funding acquisition. Qingguo Lai: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition.

Data availability statement

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

Ethics approval and consent to participate

All animal experiments were approved by the Institutional Animal Ethics Committee of the Second Qilu Hospital of Shandong University (Approval No. KYLL2024961).

Declaration 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: