VEGF-sustained regulation and hydrogel-activated graded porous titanium scaffold for functional regeneration of the tendon-bone interface

Abstract

Recapitulating the complex coupling mechanisms at the tendon-bone interface, which involve gradients in structure, composition, and mechanics, poses a primary challenge for healing. In this study, a 3D-printed graded porous Ti6Al4V scaffold was designed and fabricated to address this challenge. This scaffold features a gradient pore architecture tailored to the distinct requirements of both soft and hard tissues. The scaffold was further functionalized with a GelMA and silk fibroin hydrogel to activate the interface. Additionally, chitosan-stabilized BSA nanospheres loaded with VEGF were encapsulated within this hydrogel to achieve sustained release. In vitro studies demonstrated that this composite scaffold effectively promoted the migration and proliferation of tendon-derived stem cells, upregulated tenogenic marker genes, and enhanced angiogenic activity and synergistic osteogenesis. Animal model evaluations confirmed that the scaffold promoted the regeneration of type I collagen and Sharpey-like fibers, angiogenesis, and improved osseointegration, ultimately leading to a biologically continuous tendon-bone connection. The ultimate failure load of the repaired interface reached 107.61 ± 5.16 N, restoring 82% of the native enthesis strength. In conclusion, the synergistic strategy of “Structural Adaptation — Interface Activation — Signaling Regulation” presented in this study demonstrated the capacity to facilitate the regeneration of a biologically continuous tendon-bone interface in vivo and significantly improve functional integration. This approach provides a novel solution for enthesis regeneration and holds promise for clinical translation.

Graphical abstract

Schematic illustration of preparation of Ti@GS-VNS scaffold and its mechanism for tendon-bone interface repair.

Highlights

• •Biomimetic strategy guided by the “Structure–Interface–Signal” regeneration paradigm.
• •3D-printed graded titanium scaffold integrated with VEGF-functionalized hydrogel.
• •Spatially coordinated mechanical and biochemical cues mimic native enthesis.
• •Achieved 82% recovery of native failure load, restoring functional integration.
• •Provides a translational platform for orthopaedic and sports medicine repair.

1. Introduction

The native tendon-bone interface (TBI) is a complex structure comprising tendon, non-calcified fibrocartilage, calcified fibrocartilage, and bone. Its continuous gradients in both tissue organization and mechanical properties are pivotal for facilitating efficient load transfer between soft and hard tissues, which is essential for the biomechanical function of the enthesis [1]. In clinical scenarios such as bone tumor resection, severe trauma, or revision surgery, bone defects and TBI injuries are frequently encountered [2]. Current treatment strategies primarily rely on sutures or tunnel fixation to achieve “passive fixation” of the tendon to bone or metallic prostheses. However, the subsequent healing process often results in the formation of biomechanically inferior fibrous scar tissue, leading to suboptimal functional recovery. Consequently, re-rupture rates and the need for secondary revision surgery remain as high as 28%, significantly compromising the ultimate restoration of joint function [3,4]. Therefore, achieving robust integration between metallic implants and the TBI remains a critical, elusive challenge for improving long-term clinical outcomes.

3D-printed porous titanium alloy scaffolds offer a promising pathway for enhancing soft tissue integration. Research has demonstrated that titanium scaffolds with an appropriate pore size can effectively facilitate tendon ingrowth and improve fixation strength [5]. However, scaffolds with uniform pore sizes fail to recapitulate the graded structure of the native TBI, thereby limiting their regenerative efficacy. Consequently, fabricating porous scaffolds with gradient pore sizes has emerged as a key strategy for enhancing tendon-bone integration [6]. Triply Periodic Minimal Surface (TPMS)-based graded porous structures present unique advantages. These multi-scale TPMS structures are characterized by large pores in the peripheral zones to facilitate cell migration and angiogenesis, while transitioning to smaller pores internally to enhance ECM deposition, thus better mimicking the biomechanical environment of the native interface [7,8].

Despite these advances in structural design, the inherent bioinertness of titanium alloys remains a critical bottleneck, limiting their functional integration with tissues [9]. To address this and achieve both enhanced interfacial bioactivity and controlled delivery of biological signals, functional hydrogel coatings have emerged as a pivotal strategy [10]. A composite hydrogel of Gelatin Methacryloyl (GelMA) and silk fibroin (SF) is particularly promising. This system not only recapitulates key characteristics of the native extracellular matrix but also promotes the formation of a β-sheet secondary structure, imparting tunable mechanical properties and degradation behavior to the hydrogel, making it an ideal biological bridging layer [11,12]. Infiltrating the porous scaffold with this GelMA/SF (GS) hydrogel modulates the local microenvironment while providing a platform for the sustained release of bioactive molecules [13].

Within the critical healing window for TBI repair, angiogenesis is indispensable for successful tissue remodeling. Newly formed vasculature delivers essential oxygen and nutrients to the highly metabolically active repair site and serves as a conduit for tendon-derived stem cells (TDSCs) to infiltrate the injured area [14]. Conversely, insufficient blood supply often leads to the formation of weak, fibrotic scar tissue at the interface between the tendon and the scaffold [5]. Research indicates that injured tendon tissue intrinsically upregulates the expression of VEGF isoforms, such as VEGF121 and VEGF165, suggesting an activated VEGF signaling pathway during the natural repair process [15,16]. Furthermore, the integration between a prosthesis and bone relies on VEGF-mediated vascularization, laying the foundation for subsequent osteogenesis [17,18]. However, the clinical application of VEGF is challenging due to its short in vivo half-life, and uncontrolled supraphysiological concentrations can induce aberrant angiogenesis, potentially compromising the mechanical integrity of the repaired tissue [19]. While hydrogels themselves serve as drug carriers, their inherent diffusion mechanisms often lead to an inevitable initial burst release [20]. This study introduces Chitosan-stabilized Bovine Serum Albumin (BSA) nanospheres as secondary carriers to achieve more precise control over VEGF release kinetics [21,22]. Integrating this nanosphere system into the GS hydrogel matrix is expected to establish a composite system that effectively mitigates the risks of initial VEGF burst release.

Therefore, this study proposes an integrated “Structural Adaptation — Interface Activation — Signaling Regulation” strategy. This approach utilizes a graded TPMS porous titanium alloy as the structural foundation, incorporates a GS hydrogel as a bioactive bridging layer, and further embeds VEGF-loaded BSA/CS nanospheres (VNS) to construct a sustained release system. The overarching aim is to synergistically promote coupled regeneration across tissues—encompassing angiogenesis, tenogenesis, and osteogenesis —offering a novel treatment paradigm for achieving robust biological integration between tendon, bone, and metallic prostheses.

2. Materials and methods

2.1. Design of titanium alloy scaffold

Extensive research has demonstrated that the Gyroid structure, a member of the TPMS family, exhibits a mean curvature of zero at any given point [23,24]. This characteristic is conducive to cell migration and bone ingrowth, facilitating early-stage vascularization. The following implicit equation (Equation (1)) mathematically defines the Gyroid unit cell geometry [25]:

$$f\left(x,y,z\right)=\mathit{sin}\left(\alpha x\right)\mathit{cos}\left(\beta y\right)+\mathit{sin}\left(\beta y\right)\mathit{cos}\left(\gamma z\right)+\mathit{sin}\left(\gamma z\right)\mathit{cos}\left(\alpha x\right)=c$$ (1)

In this equation, α, β, γ and c control the model's unit cell size and wall thickness. The Gyroid structure generated by Equation (1) was constructed within the designated scaffold volume using the open-source NS Lattice software. By varying parameters such as unit cell size and wall thickness, scaffolds featuring identical porosity but different pore sizes were designed (Fig. S1).

The continuously graded structure of the native TBI renders a uniform, single-pore-size architecture inadequate for accommodating the growth requirements of different tissue-specific cells [26]. Therefore, in this study, Gyroid porous discs (8 mm diameter, 2 mm thickness) with different pore sizes (350 μm, 500 μm, 650 μm) were fabricated at a constant porosity of 60%, using solid discs as controls. Initial cell culture experiments were conducted to assess the biological responses of MC3T3-E1 pre-osteoblasts and TDSCs to these Gyroid scaffolds with varying pore sizes. The results from these assays informed the subsequent design of a gradient pore structure, intended to better match the TBI's cellular and structural gradient requirements.

2.2. Fabrication and characterization of titanium alloy scaffolds

Titanium alloy scaffolds were fabricated using a Renishaw AM400 printer, following a previously established process [27]. Scaffolds were produced from the STL dataset in an ultra-pure argon atmosphere, using Ti-6Al-4V ELI-0406 metal powder and Selective Laser Melting(SLM) technology. After printing, all samples were sequentially ultrasonically cleaned in acetone, ethanol, and deionized water. Subsequently, the cleaned scaffolds were autoclaved at 120 °C for 20 min and then dried. The surface morphology of the scaffolds was characterized using Scanning Electron Microscopy (SEM). Finally, the scaffold porosity was determined via the weighing method. Additionally, cylindrical porous scaffolds (8 mm diameter, 10 mm height) with varying pore sizes were designed and printed to investigate the effect of pore size on yield strength and elastic modulus.

2.3. Pore size gradient optimization of titanium alloy scaffolds

Subsequent to autoclave sterilization of the scaffolds, the effects of Gyroid scaffold pore size on MC3T3-E1 cells and TDSCs were evaluated through a series of assays, including cell proliferation, live/dead staining, Sirius Red staining, alkaline phosphatase (ALP) activity measurement, and RT-qPCR. Solid titanium discs served as the control group.

2.3.1. Cell culture

The MC3T3-E1 cell line and HUVECs were obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences. Cells were cultured in DMEM supplemented with 10% (v/v) FBS (164210, Procell) and 1% (v/v) penicillin/streptomycin. Primary TDSCs were isolated from rat Achilles tendons following a previously established method [28]. Detailed isolation procedures are provided in the Supplementary Information and Fig. S2.

2.3.2. Live/dead cell assay

Sterilized scaffolds with pore sizes of 350 μm, 500 μm, 650 μm, and solid scaffolds, were placed in a 24-well plate. A cell suspension (2 × 10⁴ cells per scaffold) was seeded onto each scaffold, followed by incubation in a humidified atmosphere at 37 °C with 5% CO₂ and 95% air. After 2 days of culture, the scaffolds were gently rinsed 2-3 times with DPBS to remove residual culture medium. A staining working solution was prepared using a Live/Dead Cell Double Staining Kit under light-protected conditions. Subsequently, 500 μL of the staining solution was added to each well and incubated for 30 min. Finally, the scaffolds were washed three times with DPBS.

2.3.3. Cell proliferation assay

Cell proliferation on the different scaffold groups was assessed using the AlamarBlue assay. One day post-seeding, scaffolds were transferred to a new 24-well plate. At designated time points, a working solution containing 10% (v/v) AlamarBlue reagent in culture medium was added to each well. The plates were incubated at 37 °C with 5% CO₂ for 3 h. The absorbance of the supernatant was measured using a spectrophotometer at 570 nm with 600 nm serving as the reference wavelength.

2.3.4. ALP activity assay

At 24 h post-seeding of MC3T3-E1 cells, the culture medium was replaced with osteogenic induction medium (MUXMT-90021, Cyagen). The medium was replenished every 2 days to maintain optimal differentiation conditions. On day 7, cell lysate supernatants were collected. ALP activity, indicative of osteogenic differentiation, was quantified using an Alkaline Phosphatase Assay Kit (Beyotime Biotechnology, China).

2.3.5. Sirius Red staining

At 24 h post-seeding of TDSCs, the culture medium was replaced with DMEM supplemented with ascorbic acid (50 μg/mL), 10% (v/v) FBS, and 1% (v/v) penicillin/streptomycin. Cells were fixed with 4% paraformaldehyde after 7 days of culture. The fixed constructs were then stained for 1 h using a Sirius Red Staining Kit (BP-DL030, BioChannel Biological Technology). Following staining, the scaffolds were thoroughly rinsed three times with 0.1 M acetic acid to remove excess dye. Subsequently, 300 μL of an alkaline elution buffer (a 1:1 mixture of 0.2 M NaOH and methanol) was added to elute the bound dye from the scaffold surfaces. Absorbance was then measured at 540 nm using a microplate reader.

2.3.6. Real-time quantitative PCR (RT-qPCR)

Following a 7-day culture of MC3T3-E1 cells or TDSCs on different scaffold groups, RT-qPCR was performed to analyze the expression of osteogenesis-specific genes (ALP, OCN, OPN and RUNX2) and tenogenesis-specific genes (COL1, DCN, SCX and TNC). Total RNA was extracted using Trizol reagent. Subsequently, RNA was reverse-transcribed into cDNA using an Evo M-MLV RT Kit. Quantitative PCR was conducted using a SYBR Green qPCR Kit (AG11718, ACCURATE BIOTECHNOLOGY (HUNAN) CO., LTD, Changsha, China) on an RT-PCR system. GAPDH served as the housekeeping gene for normalization. Relative changes in gene expression compared to the control group were calculated using the 2^–ΔΔCT method. The specific primer sequences used are listed in Table S1.

2.3.7. Scaffold pore size gradient optimization

Informed by the cumulative findings from these experiments, the scaffold pore size was optimized to yield a gradient structure. The finalized design models were saved in STL format and submitted for fabrication.

2.4. Preparation and characterization of VEGF-loaded nanospheres

VEGF-loaded nanospheres were prepared using a desolvation method [29]. Briefly, BSA powder was dissolved in 5 mL of 10 mM NaCl solution to obtain a 1 wt% BSA solution. A VEGF solution (10 μg mL⁻¹, 1 mL) was added to the prepared BSA solution, and the mixture was incubated for 4 h to form a BSA-VEGF mixture. Subsequently, 10 mL of ethanol was added to the BSA solution at a rate of 1 mL/min under continuous stirring at room temperature for 12 h (600 rpm), resulting in an unstable suspension of VEGF-loaded BSA (VEGF-BSA) nanospheres. Chitosan was dissolved in 1% acetic acid solution (20 mL) to prepare a 0.05 wt% chitosan solution. The chitosan solution was then added to the unstable BSA nanosphere suspension slowly at 0.5 mL/min. The newly formed mixture was stirred for another 8 h (600 rpm) to obtain a suspension of VEGF-loaded chitosan-stabilized BSA nanospheres (VNS). The nanospheres were collected via high-speed centrifugation at 4 °C and 12,000 rpm, then lyophilized to obtain the final VNS. The VNS were characterized using nanoparticle size and zeta potential analyzer, SEM, transmission electron microscopy (TEM), and Fourier Transform Infrared spectroscopy (FTIR). An ELISA kit (97053ES96; Yeasen, China) was used to determine the encapsulation efficiency and drug loading capacity. In addition, the colloidal stability of VNS was evaluated in simulated body fluid (SBF) at 37 °C. Variations in particle size and Zeta potential were monitored using nanoparticle size and zeta potential analyzer, and the final morphology was observed via scanning electron microscopy (SEM).

2.5. Preparation and characterization of nanosphere-loaded GelMA-SF hydrogel

Sterile PBS was mixed with 6 wt% GelMA and 1 wt% SF. Subsequently, 0.25 wt% photoinitiator LAP was added and stirred thoroughly, followed by the incorporation of the VNS suspension to prepare the VNS-loaded hydrogel precursor solution. Then, 500 μL of the precursor solution was pipetted into a mold and exposed to UV light for 2 min to induce free radical polymerization via photo-crosslinking. After demolding, the formed hydrogel was treated with 70% (v/v) ethanol solution at room temperature for 1 h to obtain the GS composite hydrogel [30]. The swelling ratio of the hydrogel was measured, and its degradation rate was assessed by adding 1 U/mL type II collagenase. To characterize the internal microstructure, lyophilized hydrogels were sectioned to expose the cross-section, sputter-coated with gold, and observed using SEM. To assess the effect of VEGF concentration on TDSCs, GS hydrogels loaded with nanospheres containing 0, 10, 60, and 120 ng/mL VEGF were prepared. A scratch assay was performed to evaluate the effect of different VEGF concentrations on TDSCs migration. Cell proliferation under the different VEGF concentrations was assessed using the AlamarBlue assay. The optimal VEGF concentration for the sustained-release system was screened by detecting the expression of tenogenic-related genes via RT-qPCR. Finally, the VEGF release profile was plotted using an ELISA kit, and hydrogels with directly added free VEGF were evaluated to compare the drug release kinetics of the different loading systems.

2.6. Preparation and structural characterization of composite scaffolds

Sterilized scaffolds were immersed in different hydrogel solutions. After thorough mixing, vacuum degassing was applied to remove air bubbles, followed by low-speed centrifugation to facilitate hydrogel infiltration into the scaffold pores. The hydrogel was fully cured by exposing the scaffold to UV light from multiple angles. The resulting scaffolds were treated with ethanol solution and rinsed thoroughly with DPBS. The scaffolds were divided into the following groups: Gradient pore size Ti scaffold (Ti), gradient pore size Ti scaffold combined with GS hydrogel (Ti@GS), gradient pore size Ti scaffold combined with GS hydrogel and empty nanospheres (Ti@GS-ENS), and gradient pore size Ti scaffold combined with GS hydrogel and VEGF-loaded nanospheres (Ti@GS-VNS). The morphology, pore size, and elemental composition of the scaffolds were analyzed using SEM and EDS. The hydrophilicity of the different scaffolds was assessed by measuring the water contact angle using a contact angle goniometer. Uniaxial compression tests were performed on the Ti scaffolds using a Instron 5982 universal testing machine (USA) at a crosshead speed of 0.5 mm/min. Furthermore, VNS loading on the scaffolds was visualized by labeling BSA with FITC and imaging with a fluorescence microscope.

2.7. Biocompatibility of composite scaffolds

To evaluate the biocompatibility of the scaffolds, TDSCs (1 × 10⁵ cells per scaffold) were seeded onto each scaffold group. The effect of the scaffolds on TDSC viability and proliferation was assessed using live/dead staining and the AlamarBlue assay. Cell adhesion characteristics on the composite scaffolds were evaluated via phalloidin and DAPI fluorescence staining. For the hemolysis assay, the different scaffolds were added to a diluted red blood cell suspension to a final concentration of 5% (v/v). The samples were incubated at 37 °C for 1 h, and subsequently centrifuged at 10,000 rpm for 10 min. The absorbance of the supernatant was recorded at 540 nm using a microplate reader. The hemolysis ratio was calculated using Equation (2):

$$Hemolysisratio=\frac{ODs–ODn}{ODp–ODn}\times 100\%$$ (2)

Where ODs is the OD value of the sample, ODn is the OD value of the negative control (normal saline group), and ODp is the OD value of the positive control (0.1% Triton X-100 solution group).

2.8. In vitro tenogenic differentiation potential of composite scaffolds

TDSC suspensions were added to the composite scaffolds, and the constructs were cultured in tenogenic induction medium to evaluate the in vitro tenogenic potential. Immunofluorescence staining was performed after 7 days of culture. Cells were fixed with 4% paraformaldehyde for 30 min, permeabilized with 0.1% Triton X-100, and blocked with PBS containing 5% (w/v) BSA. After incubation with primary antibodies at 4 °C overnight and washing with PBS, secondary antibodies were applied and incubated for 1 h. F-actin and DAPI staining were performed, followed by imaging using a confocal laser-scanning microscope. Simultaneously, the expression of tenogenic markers in TDSCs treated with the scaffolds was detected using RT-qPCR. After 14 days of culture, collagen deposition on the different scaffold groups was assessed using Sirius Red staining. Western blotting was used to detect tenogenic-related proteins. Cells on the scaffolds were washed three times with PBS and lysed with RIPA lysis buffer containing protease inhibitors. After protein quantification, standard procedures were followed: proteins were loaded onto SDS-PAGE gels, separated by electrophoresis, and transferred to PVDF membranes. After blocking with 5% skim milk, the membranes were incubated with primary antibodies at 4 °C overnight, namely anti-COL1 (1:2500, Proteintech, China), anti-DCN (1:2500, Proteintech, China), and anti-GAPDH (1:4000, Proteintech, China). Secondary antibodies (1:1000, Beyotime, China) were then incubated at room temperature for 1 h. Following washing with TBST for 30 min, the developing reagent was added for signal detection. Analysis was performed using ImageJ software.

2.9. In vitro angiogenic potential of composite scaffolds

To prepare scaffold extracts, scaffolds were immersed in cell culture medium and incubated in a humidified atmosphere containing 5% CO₂ at 37 °C for 72 h. The extracts were utilized to evaluate the pro-angiogenic potential of the composite scaffolds via a tube formation assay. Matrigel matrix, 48-well plates, and pipette tips were pre-cooled. Then, 100 μL of Matrigel was evenly coated onto the bottom of the 48-well plates and incubated at 37 °C for 1 h to promote gelation. Starved HUVECs were resuspended in the extracts from the different scaffolds and seeded onto the surface of the Matrigel at a density of 2 × 10⁴ cells per well. After 6 h of incubation, tube formation was visualized under an inverted microscope and the results were quantified using ImageJ software. Simultaneously, HUVECs (2 × 10⁴ cells per scaffold) were co-cultured with the different scaffold groups, Immunofluorescence staining was performed after 7 days. The expression of angiogenesis-related markers (CD31 and FGF-2) in HUVECs was then assessed using RT-qPCR. Specific primer sequences are listed in Table S1.

2.10. In vitro synergistic osteogenic potential of composite scaffolds

To simulate the effect of scaffolds on osteoblast differentiation in a multicellular co-culture environment mimicking in vivo conditions, MC3T3-E1 cells were co-cultured with HUVECs. MC3T3-E1 cells (4 × 10⁴ cells per scaffold) were seeded onto the scaffolds and placed in 24-well plates. After 24 h of culture, the scaffolds were transferred to Transwell inserts. HUVECs (1 × 10⁴ cells per well) had been seeded in the lower chamber of the Transwell plate 24 h prior to the experiment. After 7 and 14 days of standard osteogenic induction culture, the scaffolds were retrieved and ALP activity was quantified using an ALP detection kit. Simultaneously, osteogenesis-specific gene expression was detected by RT-qPCR. After 21 days of cell culture, mineralization nodules on the scaffolds were assessed via Alizarin Red staining. Subsequently, 10% cetylpyridinium chloride solution was added, and the scaffolds were shaken at 120 rpm for 15 min to elute the dye. The absorbance was recorded at 620 nm using a microplate reader to analyze the formation of calcified nodules.

2.11. Animal experiments

The experimental protocol was approved by the Animal Ethics Committee of Southern Medical University (Approval No. SMUL202409018). All animal care and experimental procedures were conducted in compliance with the Guide for the Care and Use of Laboratory Animals. Twenty-seven healthy adult New Zealand white rabbits (2.5-3.0 kg) were acclimatized for 1 week and then randomly divided into three groups: Ti group, Ti@GS group, and Ti@GS-VNS group. An infraspinatus tendon injury and insertion defect model was established unilaterally in all groups. After anesthesia with sodium pentobarbital, a lateral skin incision was made between the acromion and the greater tubercle. The infraspinatus tendon was transected at its insertion base using a sharp surgical blade, and a bone tunnel (6 mm in diameter, 6 mm in depth) was created in the humeral footprint to simulate the insertion defect. The tendon was fixed into the scaffold groove using a modified Mason-Allen bridging suture technique. After thoroughly irrigating the defect area with saline, the corresponding scaffolds were implanted according to the group. Transosseous sutures were used to secure the scaffold and prevent post-operative displacement. Finally, muscles, subcutaneous tissue, and skin were sutured layer by layer. Post-operative care, including antibiotic injections and wound management, was administered.

2.12. Micro-CT analysis

At 4 and 12 weeks post-implantation, the implanted specimens were harvested, thoroughly rinsed with saline, and fixed in 10% neutral buffered formalin for 24 h to preserve tissue structure. Scanning was performed using a micro-CT system (SkyScan 1172, Bruker, Belgium). Three-dimensional images were reconstructed using CTvox (64-bit), and bone tissue within and around a 100 μm region of interest from the implant was quantitatively analyzed.

2.13. Hard tissue sectioning

After micro-CT scanning, the samples were dehydrated through a graded ethanol series and embedded in polymethyl methacrylate. Hard tissue sections were prepared using a hard tissue microtome and grinding system (EXAKT Vertriebs GmbH, Germany) and subjected to H&E and Sirius Red staining. New bone formation, vascular count, and tendon fiber conditions were observed using an optical microscope. Additionally, collagen fiber types in the tendon were visualized under polarized light microscopy at 12 weeks, and the results were analyzed using ImageJ software.

2.14. Tensile testing

At 12 weeks post-implantation, three fresh infraspinatus tendon-scaffold-humerus complexes were harvested from each group, along with three samples from the healthy contralateral side for biomechanical testing. The biomechanical properties of the tendon-bone interface were evaluated using an electronic universal material testing system (LS1E, LLOYD, USA). A static preload of 2 N was applied for 5 min to precondition the tissue. The ultimate failure load was tested at an elongation rate of 5 mm/min, with tissue rupture defined as the endpoint.

2.15. Immunohistochemical and immunofluorescence analyses

Peri-implant bone tissue samples harvested at 12 weeks post-implantation were fixed in 4% paraformaldehyde, decalcified in EDTA, and embedded in paraffin. Sections (5 μm thick) were prepared using a Leica microtome. For immunohistochemistry, sections underwent antigen retrieval in citrate buffer and peroxidase blocking, followed by sequential incubation with primary anti-CD31 antibody (1:200; Biorbyt, UK) and an HRP-conjugated secondary antibody (1:1000; Abcam, UK). Samples were visualized with DAB and counterstained with hematoxylin. For immunofluorescence, sections were incubated with primary anti-CD31 antibody (1:400) and secondary antibody (1:5000; Hangzhou Fude, China), followed by DAPI counterstaining. Fluorescence intensity was quantified using ImageJ software.

2.16. Statistical analysis

Each experiment included at least three independent replicates. IBM SPSS Statistics 27.0, GraphPad Prism 9, and ImageJ software were used to process and statistically analyze all data. Quantitative data are presented as mean ± standard deviation. Comparisons between multiple experimental groups were performed using one-way analysis of variance (ANOVA), while comparisons between two groups were conducted using an independent t-test. Statistical significance was denoted at ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001; “ns” indicates no significant difference.

3. Results

3.1. Gradient optimization of 3D-Printed porous titanium alloy scaffold

This study fabricated porous titanium discs with a constant porosity of 60% and varying pore sizes of 350 μm, 500 μm, and 650 μm using Selective Laser Melting (SLM) technology (Fig. 1A). The discs measured 8 mm in diameter and 2 mm in thickness, with solid titanium discs as the control. SEM images confirmed that all scaffolds accurately replicated the designed morphology, with some unmelted titanium alloy powder particles visible on the surfaces (Fig. 1B). The measured pore size and porosity data are detailed in Table S2, while the mechanical properties are presented in Table S3. Notably, pore size exhibited a negative correlation with the mechanical properties. Cell co-cultured results indicated good cytocompatibility for all scaffold materials, although fewer cells adhered to the solid titanium discs (Fig. 1C). AlamarBlue proliferation assays revealed that MC3T3-E1 pre-osteoblasts exhibited the highest proliferation on the 650 μm scaffolds. In contrast, Tendon-Derived Stem Cells (TDSCs) showed the strongest proliferation on the 500 μm scaffolds (Fig. 1D and E). These results highlight the differential effects of pore size on the proliferation of distinct cell types. Further investigation into the osteogenic differentiation potential of MC3T3-E1 cells showed that after 7 days of co-culture, the 650 μm scaffold group demonstrated significantly higher ALP activity compared to other groups. Concurrently, osteogenic gene expression was significantly upregulated in the 650 μm group (Fig. 1F–H). Regarding tenogenic differentiation potential, the Sirius Red staining after 7 days of TDSCs co-culture showed significantly higher values in the 500 μm group. Furthermore, the expression of tenogenic genes was significantly upregulated in the 500 μm group, indicating its pronounced advantage in promoting TDSCs tenogenic differentiation (Fig. 1G–I).

Fig. 1.

Optimization of Gradient Pore Sizes in Porous Titanium Alloy Scaffolds for Tendon-Bone Interface Regeneration. (A) (i) Design drawings and (ii) top-view of 3D-printed porous titanium alloy scaffolds with different pore sizes. (B) SEM micrographs showing the surface microstructure of the different scaffold groups. (C) Live (green)/dead (red) staining images of MC3T3-E1 cells and TDSCs co-cultured with the different scaffold groups for 48 h. (D) Proliferation of MC3T3-E1 cells, assessed by AlamarBlue assay. (E) Proliferation of TDSCs. (F) Semi-quantitative analysis of ALP activity in MC3T3-E1 cells. (G) Semi-quantitative analysis of Sirius Red staining in TDSCs. (H) Expression levels of osteogenic differentiation-related genes in MC3T3-E1 cells. (I) Expression levels of tenogenic differentiation-related genes in TDSCs. (J) Schematic diagram of the gradient pore size porous titanium alloy disc for in vitro experiments. (K) Schematic diagram of the gradient pore size porous titanium alloy scaffold for in vivo experiments.

Based on these collective results and the structural characteristics of the tendon-bone interface (TBI), the scaffold was designed with a gradient pore structure. It features a constant porosity of 60%, transitioning from a peripheral pore size of 650 μm to a central pore size of 500 μm, to enhance TBI compatibility. As shown in Fig. 1J, the in vitro test discs were designed with a 10 mm diameter and a 2 mm thickness. The in vivo scaffold design included a cylindrical scaffold measuring 6 mm in height and 6 mm in diameter, featuring a top groove (1.7 mm wide, 2 mm deep) to accommodate the tendon and a central through-hole (1.6 mm diameter) for suture passage during tendon fixation (Fig. 1K).

3.2. Preparation and characterization of nanospheres and composite hydrogel

SEM revealed that the VNS were uniformly spherical, with particle sizes primarily distributed between 390 and 440 nm (Fig. 2A–C). TEM images further revealed the core-shell structure of the VNS and visible internal encapsulates, whereas empty nanospheres (ENS) showed only the core-shell structure (Fig. 2B). Dynamic Light Scattering analysis indicated an average VNS particle size of 416.65 ± 16.19 nm with a narrow distribution (PDI = 0.17 ± 0.08) (Fig. 2D, Fig. S3, Table S4). Zeta potential measurements showed that the surface potential of the CS-coated nanospheres was above +30 mV (Fig. 2F), confirming successful chitosan coating and conferring good colloidal stability to the system. FTIR analysis of VNS showed the characteristic chitosan absorption peak at 1081 cm⁻¹, along with shifts in the amide I band and the symmetric stretching vibration of carboxylate groups, confirming the formation of a composite structure between CS and BSA via electrostatic interactions (Fig. 2E). VEGF's encapsulation efficiency and drug loading capacity, measured using an ELISA kit, were 76.87 ± 0.65% and 2.47 ± 0.13 ng/mg, respectively. The stability of VNS was evaluated in Simulated Body Fluid (SBF) at 37 °C over 7 days. As summarized in Table S5, VNS exhibited a gradual increase in average particle size over time, accompanied by a concurrent declining trend in Zeta potential. Furthermore, SEM analysis on Day 7 (Fig. S8) revealed that the VNS had transitioned from an intact spherical shape into a collapsed or partially dissolved morphology.

Fig. 2.

Preparation and characterization of nanospheres and composite hydrogel materials. (A) (i) SEM image of VNS aggregates and (ii) single ENS and VNS particle. (B) TEM images of VNS and ENS. (C) Gaussian-fitted size distribution of VNS calculated from SEM images. (D) Particle size distribution of different nanospheres measured by dynamic light scattering (DLS). (E) FTIR spectra of BSA, CS, and VNS. (F) Zeta potential values of different nanospheres. (G) SEM image of the freeze-dried GS hydrogel. (H) FTIR spectra of GelMA and GS hydrogels. (I) In vitro degradation rate of GelMA, GS, and GS-VNS materials. (J) Results of the 24-h scratch assay with TDSCs in sustained-release systems containing different VEGF concentrations. (K) Swelling ratio of different hydrogels. (L) Quantitative analysis of scratch assay results. (M) Proliferation of TDSCs in sustained-release systems with different VEGF concentrations. (N) Cumulative VEGF release profiles from different carrier systems. (O) Expression levels of tenogenic-related genes in TDSCs cultured for 7 days in sustained-release systems with different VEGF concentrations.

SEM scanning of the freeze-dried GS composite hydrogel cross-section revealed a typical semi-porous, semi-lamellar microstructure without obvious phase separation (Fig. 2G). FTIR analysis detected characteristic absorption peaks at 1605 cm⁻¹ (amide I), 1516 cm⁻¹ (amide II), and 1224 cm⁻¹ (amide III), attributed to the β-sheet structure, confirming the formation of a stable β-sheet secondary structure by SF within the hydrogel network (Fig. 2H). Performance tests showed that compared to pure GelMA hydrogel, the introduction of SF and the formation of the β-sheet structure enhanced the crosslinking density, leading to significantly lower swelling ratio and degradation rate for the GS hydrogel (Fig. 2I–K).

Regarding biological functionality, scratch assay and cell proliferation results indicated that a 10 ng/mL VEGF concentration was most effective in promoting TDSCs migration and proliferation (Fig. 2J, L, 2M, S4). Simultaneously, this concentration significantly upregulated the expression levels of tenogenic marker genes (Fig. 2O). In terms of release kinetics, the GS-VNS system demonstrated a stable and sustained VEGF release profile over 20 days (cumulative release of 68.48 ± 2.52% over 456 h), whereas VEGF loaded directly into the GS hydrogel exhibited a significant initial burst release (cumulative release of 73.35% ± 1.17% within 120 h) (Fig. 2N). Based on these results, the subsequent experiments utilized the GS-VNS sustained-release system with a VEGF concentration of 10 ng/mL.

3.3. Preparation and characterization of composite scaffolds

The optimized gradient porous titanium alloy scaffold (Ti) was fabricated using SLM (Fig. 3A and B). The study incorporated the composite hydrogel and nanospheres to prepare three types of composite scaffolds: Ti scaffold combined with GS hydrogel (Ti@GS), GS hydrogel and ENS (Ti@GS-ENS), GS hydrogel and VNS (Ti@GS-VNS). SEM showed that the Ti scaffold had pore sizes primarily distributed between 475 and 575 μm (Fig. 3C).

Fig. 3.

Preparation and characterization of composite scaffolds. (A,B) Photograph of the 3D-printed gradient pore size scaffold for in vivo and in vitro experiments. (C) SEM micrographs and EDS elemental mapping of Ti and Ti@GS-VNS scaffolds. (D) Merged fluorescence (green) and bright-field images showing the distribution of FITC-labeled nanospheres within the Ti@GS-VNS scaffold. (E) EDS elemental composition analysis of Ti and Ti@GS-VNS scaffolds. (F) Statistical analysis of water contact angle results for the scaffolds. (G) Typical stress-strain curve of the Ti scaffold. (H) Elastic modulus and yield strength of the Ti scaffold.

SEM images of the Ti@GS-VNS scaffold revealed the presence of hydrogel within its interior. EDS analysis indicated a substantial increase in the proportions of C, N, and O elements in the Ti@GS-VNS scaffold compared to the Ti scaffold, confirming the successful integration of the composite hydrogel with the titanium alloy scaffold (Fig. 3C and E, S5). VNS prepared using FITC-labeled BSA showed a uniform distribution of green fluorescent nanospheres within the scaffold pores under fluorescence microscopy, maintaining a regular spherical shape (Fig. 3D). Water contact angle measurements showed that the Ti scaffold had a contact angle of 97.48 ± 3.79°. After hydrogel incorporation, the contact angle for composite scaffolds was significantly lower than that of the Ti group, indicating a marked improvement in the hydrophilicity of the hydrogel-composite scaffolds (Fig. 3F). Furthermore, mechanical testing showed a yield strength of 119.62 ± 1.81 MPa and an elastic modulus of 2.68 ± 0.12 GPa, closely matching the compressive properties of human bone tissue and effectively reducing stress shielding effects (Fig. 3G and H).

3.4. Biocompatibility testing of composite scaffolds

After 48 h of co-culture with TDSCs, live/dead staining revealed many viable cells across all scaffold groups, with more intense green fluorescence in the Ti@GS-VNS group (Fig. 4A–C). The AlamarBlue assay revealed no significant differences between groups on day 1. However, on days 4 and 7, the Ti@GS-VNS group showed a higher TDSCs proliferation rate than the other groups, highlighting the role of VEGF in further enhancing TDSCs proliferation efficiency (Fig. 4D). As shown in Fig. 4B, cells cultured on Ti exhibited a small, narrow morphology with few pseudopodia. In contrast, cells cultured on Ti@GS-VNS exhibited a larger spreading area, appearing polygonal or spindle-shaped with distinct filamentous pseudopodia. Analysis of the cytoplasm-to-nucleus ratio showed that this ratio was significantly higher for TDSCs on the composite scaffolds compared to the Ti group, indicating a positive impact of hydrogel incorporation on early TDSCs adhesion (Fig. 4E). The hemolysis assay results showed that the hemolysis rates for all four scaffold groups were below 2%, meeting the hemolysis requirements for biomaterials (Fig. 4F). These findings collectively demonstrate the excellent biocompatibility of all scaffold groups.

Fig. 4.

Biocompatibility testing of composite scaffolds. (A) Live (green)/dead (red) staining of TDSCs cultured on scaffolds for 48 h. (B) Fluorescence staining of cytoskeleton (red)/nucleus (blue) of TDSCs seeded on scaffolds. (C) Quantitative analysis of live/dead staining. (D) Proliferation of TDSCs co-cultured with scaffolds, assessed by the AlamarBlue assay. (E) Cytoplasm-to-nucleus ratio calculated from cytoskeleton fluorescence staining of TDSCs on scaffolds. (F) Hemolysis assay results and quantitative analysis.

3.5. Evaluation of the tenogenic potential of composite scaffolds

The incorporation of drug-loaded nanospheres and composite hydrogel into the Ti scaffold aimed to confer tenogenic properties. Therefore, in vitro experiments were conducted to evaluate the tenogenic efficacy of the Ti@GS-VNS scaffold. TDSCs were seeded onto the various scaffolds. Following 7 days of co-culture, immunofluorescence staining results showed that the fluorescence intensities of COL1, DCN, SCX, and TNC in the Ti@GS-VNS group were significantly higher than those in the other groups (Fig. 5A–E). Fig. 5F shows Sirius Red staining of TDSCs following 14 days of tenogenic induction on the various scaffolds, revealing significantly increased collagen deposition in the Ti@GS-VNS group. Semi-quantitative analysis of the eluted dye confirmed that the value for the Ti@GS-VNS group was significantly higher than that of the other groups. While the Ti@GS group showed no significant difference compared to the Ti@GS-ENS group, its value was significantly higher than that of the Ti group (Fig. 5G). RT-qPCR further confirmed that the expression levels of all tenogenic markers in the Ti@GS-VNS group were superior to those in the other groups, with no significant difference observed between the Ti@GS-ENS and Ti@GS groups (Fig. 5H). These results indicate that Ti@GS-VNS significantly promotes tenogenic differentiation of TDSCs in vitro, while the hydrogel in Ti@GS-ENS and Ti@GS mimics the extracellular matrix environment, indirectly facilitating cell differentiation [31]. Tenogenic-related protein expression in TDSCs was detected by Western blot analysis following 14 days of co-culture with the different scaffolds. Representative immunoblot results showed markedly enhanced band intensities for COL1 and DCN in the Ti@GS-VNS group (Fig. 5I). Grayscale semi-quantitative analysis of the bands, normalized to the internal reference protein, confirmed that the expression levels of both COL1 and DCN in the Ti@GS-VNS group were significantly upregulated compared to the other groups (Fig. 5J and K). These results demonstrate that the Ti@GS-VNS scaffold effectively promotes the tenogenic differentiation of TDSCs and significantly enhances the expression of tenogenic-related proteins.

Fig. 5.

Evaluation of the tenogenic potential of composite scaffolds. (A) Representative immunofluorescence images of TDSCs cultured on Ti, Ti@GS, Ti@GS-ENS, and Ti@GS-VNS scaffolds for 7 days. (B-E) Quantitative analysis of fluorescence intensity. (F) Sirius Red staining results of TDSCs seeded on scaffolds for 14 days. (G) Semi-quantitative analysis of stain elution. (H) Relative expression levels of tenogenic-related genes in TDSCs after 7 days of culture on the scaffolds. (I) Western blot bands showing COL1 and DCN protein expression in TDSCs co-cultured with the different scaffold groups. (J, K) Semi-quantitative analysis of protein band intensity.

3.6. Evaluation of the angiogenic potential of composite scaffolds

A tube formation assay showed that the Ti group formed only a few vessel-like structures. HUVECs in the Ti@GS and Ti@GS-ENS groups formed some fragmented, discontinuous tubes, while the Ti@GS-VNS group formed an extensive, continuous vascular network (Fig. 6A). Quantitative analysis indicated that the number of junctions, meshes, nodes, and master segments in the Ti@GS-VNS group was significantly greater than that in the other groups (Fig. 6B–E). The scratch wound assay results also indicated that HUVECs in the Ti@GS-VNS group exhibited superior migratory capacity (Fig. 6F and G, S6). Immunofluorescence experiments showed that the Ti@GS-VNS scaffold significantly promoted the expression of angiogenesis-related factors CD31 and FGF-2 in HUVECs (Fig. 6I–N). To further confirm the positive impact of the composite scaffolds, the expression of angiogenesis-related genes was analyzed via RT-qPCR, revealing significantly elevated expression levels of FGF-2 and CD31 in the Ti@GS-VNS group (Fig. 6K and L). These results indicate that the Ti@GS-VNS scaffold can promote the formation of vascular networks.

Fig. 6.

Evaluation of the angiogenic potential of composite scaffolds. (A) Representative images of the tube formation assay after 6 h. (B-E) Quantitative analysis results of the number of junctions, meshes, nodes, and master segments from the tube formation assay. (F-G) Representative images of the 48 h scratch assay and quantitative analysis of wound healing. (H) Immunofluorescence staining images of HUVECs cultured on scaffolds for 7 days. (I-J) Quantitative analysis of CD31 and FGF-2 fluorescence intensity. (K-L) Relative expression levels of angiogenesis-related genes (CD31 and FGF-2) in HUVECs after 7 days of culture on the scaffolds.

3.7. Evaluation of the synergistic osteogenic potential of composite scaffolds

MC3T3-E1 cells seeded on the various scaffolds were co-cultured with HUVECs to assess the scaffolds' osteogenic potential in a multicellular environment mimicking in vivo conditions. Co-culture results showed that the Ti@GS-VNS group had significantly higher ALP activity on days 7 and 14 than the other groups, while the Ti@GS and Ti@GS-ENS groups did not differ significantly (Fig. 7A and B). Alizarin Red S (ARS) staining after 21 days of co-culture showed a consistent trend, with significantly more mineralized nodules in the Ti@GS-VNS group than in the other groups (Fig. 7C and D). RT-qPCR analysis indicated that after 7 days of co-culture, MC3T3-E1 cells in the Ti@GS-VNS group exhibited significantly higher expression of OPN, RUNX2, and OCN genes than those in the other groups. After 14 days of co-culture, the ALP gene expression in the Ti@GS-VNS group was also significantly higher than in the other groups (Fig. 7E and F). These results demonstrate that the Ti@GS-VNS group exhibits superior performance and possesses significant osteogenic potential.

Fig. 7.

Evaluation of the synergistic osteogenic potential of composite scaffolds. (A, B) Quantitative analysis of ALP activity in MC3T3-E1 cells co-cultured with HUVECs on scaffolds for 7 and 14 days. (C, D) Representative ARS staining images of MC3T3-E1 cells co-cultured on scaffolds for 21 days, and quantitative analysis of mineralization. (E, F) Relative expression levels of osteogenesis-related genes in MC3T3-E1 cells co-cultured on the different scaffolds for 7 and 14 days.

3.8. In vivo animal experiments

This study established a New Zealand rabbit model of an infraspinatus tendon insertion defect to verify the scaffolds' ability to induce tendon-bone integration in vivo. H&E staining of major organs 12 weeks post-operation showed no significant inflammation or pathological changes (Fig. S7). Fig. 8A illustrates the establishment of the animal model and the experimental procedure; implantation did not cause significant scaffold displacement. The micro-CT images showed more newly formed bone tissue on the Ti@GS-VNS scaffolds compared to that on the Ti and Ti@GS scaffolds at both 4 and 12 weeks (Fig. 8B). At 4 and 12 weeks post-operation, the bone volume/tissue volume (BV/TV), trabecular number (Tb.N), and trabecular thickness (Tb.Th) in the Ti@GS-VNS group were significantly higher than those in the other groups (Fig. 8C–E).

Fig. 8.

In vivo TBI repair experiment in New Zealand rabbits. (A) Schematic overview of the animal experiment and surgical procedure. (B) Micro-CT 3D reconstruction images showing bone ingrowth at 4 and 12 weeks post-implantation for each scaffold group (yellow: new bone, white: scaffold). (C-E) Quantitative analysis of bone volume/tissue volume (BV/TV), trabecular thickness (Tb.Th), and trabecular number (Tb.N) within the ROI at 4 and 12 weeks post-implantation. (F) Representative H&E-stained sections showing differences in new bone formation, tendon ingrowth, and tendon-bone integration among the Ti, Ti@GS, and Ti@GS-VNS groups at 4 and 12 weeks post-implantation. Green arrows indicate new bone formation, yellow arrows point to new blood vessels, white arrows indicate tendon ingrowth, and blue arrows indicate Sharpey-like fibers. (G) Semi-quantitative analysis of the new bone-scaffold contact area. (H) Vascular counts within the ROI for each group. (I) Representative images of hard tissue sections stained with Sirius Red under polarized light microscopy at 12 weeks; red/orange indicates Type I collagen, green indicates Type III collagen. (J) Relative area of Type I collagen. (K-L) Schematic of the tensile testing setup and average maximum tensile load for each group.

H&E staining was used to further evaluate tissue integration of the scaffolds. At 4 weeks post-implantation, new bone tissue grew around the scaffolds in the Ti and Ti@GS groups. In contrast, bone ingrowth into the scaffold interior and a small amount of disorganized tendon collagen tissue were visible in the Ti@GS-VNS group. At 12 weeks post-implantation, tendon collagen tissue gradually filled the scaffold pores in the Ti@GS-VNS group, with collagen alignment transitioning from disordered to more ordered. The extent of new bone formation was more pronounced, with more bone tissue growing into the scaffold interior and forming robust osseointegration (Fig. 8F and G). Sharpey-like fibers penetrated deep into the Ti@GS-VNS scaffold and anchored onto the new bone. Vascular count results showed that new blood vessels were present in the bone tissue areas of the Ti@GS-VNS group at both 4 and 12 weeks, and the vessel number was significantly higher than that in the other groups (Fig. 8H).

At 12 weeks, Sirius Red staining under polarized light microscopy distinguished collagen types growing into the scaffold pores. The Ti group showed sparse, fragmented orange and green fibers within the pores. The Ti@GS group exhibited sparse, slender green and orange fibers. The Ti@GS-VNS group displayed abundant, continuous orange fibers filling the pores (Fig. 8I). Semi-quantitative analysis using ImageJ showed that the relative area of Type I collagen in the Ti@GS-VNS group (23.57 ± 1.82%) was significantly greater than that of the other groups (Fig. 8J). At 12 weeks, three tendon-scaffold-humerus complexes from each group and three samples from the healthy contralateral side underwent tensile testing, as illustrated in Fig. 8K. The results showed that the Ti@GS-VNS group exhibited the highest average maximum load of 107.61 ± 5.16 N (Fig. 8L). This load was 2.6 times higher than that of the Ti group and 1.6 times higher than that of the Ti@GS group, reaching 82% of that of the healthy side and effectively restoring the tendon's biomechanical function.

Furthermore, the scaffolds' angiogenic potential was evaluated by immunofluorescence (IF) and immunohistochemistry (IHC). The IF results revealed that CD31 expression in the Ti@GS-VNS group was significantly higher than that in the other groups (Fig. 9A and B). Consistent with this, IHC results showed that the Ti@GS-VNS group exhibited a significantly larger area of CD31-positive staining than that in the Ti@GS and Ti groups. This finding suggests an increase in blood vessel density, further validating the pro-angiogenic efficacy of the Ti@GS-VNS scaffold (Fig. 9C and D).

Fig. 9.

Immunofluorescence and immunohistochemical assessment of angiogenesis. (A) Representative CD31 immunofluorescence staining of peri-implant bone tissue sections at 12 weeks post-implantation (CD31: green; nuclei: blue). (B) Quantitative analysis of fluorescence intensity. (C) Representative CD31 immunohistochemical staining of peri-implant bone tissue sections at 12 weeks post-implantation. (D) Quantitative analysis of the CD31-positive area.

4. Discussion

The tendon-bone interface (TBI) is characterized by a seamless transition from tendon to bone, underscored by gradients in tissue composition and mechanical properties. Surgical reconstruction of this biologically complex interface remains challenging, often resulting in suboptimal outcomes due to scar tissue formation at prosthetic junctions, which compromises clinical efficacy [32,33]. To address these obstacles, this study developed a multifunctional scaffold integrating structural adaptation, interface activation, and signaling regulation to enable effective TBI regeneration. This scaffold comprises a topologically optimized, gradient-porosity Ti-6Al-4V framework combined with a β-sheet-enhanced GelMA/SF (GS) hydrogel network embedded with vascular endothelial growth factor (VEGF)-loaded bovine serum albumin/chitosan (BSA/CS) nanospheres. This hierarchical design fosters precise spatiotemporal coordination across structural, interfacial, and biochemical cues.

Extensive in vitro and in vivo assessments demonstrated that this scaffold confers distinct biological and mechanical advantages. The gradient porous architecture accommodates the differential growth requirements of soft tendon and hard bone tissues. Concurrently, the GS hydrogel matrix enhances cell adhesion and stabilizes the interfacial biochemical milieu, while the VEGF-loaded nanospheres ensure controlled, prolonged cytokine release. The dual-release system prolongs VEGF bioactivity, supporting angiogenic, tenogenic and osteogenic responses. Functionally, the scaffold promotes tenogenic differentiation of tendon-derived stem cells (TDSCs), angiogenesis of human umbilical vein endothelial cells (HUVECs), and osteogenic maturation of MC3T3-E1 pre-osteoblasts. In vivo, the scaffold promotes coordinated regeneration of Sharpey-like fibers, type I collagen, and bone, re-establishing biological continuity and restoring mechanical properties closely approximating those of native TBI.

The foundation of these regenerative functions lies in the scaffold's structural design. Titanium alloy, widely utilized clinically due to its superior mechanical strength and corrosion resistance, serves as the scaffold material. Recent advances in additive manufacturing enable precise control over pore size, distribution, and porosity in 3D-printed porous titanium scaffolds, thereby enhancing tissue integration [34]. Literature indicates an optimal pore size near 527 μm for tendon fiber ingrowth, yet the transitional nature of the TBI necessitates a graded pore architecture meeting diverse tissue-specific requirements [5,35]. Guided by the differential cellular responses elicited by varying pore sizes, we established a “structural adaptation” design strategy. Specifically, the scaffold periphery was configured with a large pore size of 650 μm; the resulting lower elastic modulus helps mitigate stress shielding, while the open space maximizes osseointegration efficiency at the bone interface. Conversely, the central region was designed with 500 μm pores, where the relatively denser structure provides a higher modulus to withstand tensile loads, thereby promoting tenogenic differentiation and soft tissue anchoring within the core. This dual spatial regulation of structure and mechanics aims to achieve synergistic regeneration of the bone-tendon interface within a single construct. Traditional dense metal implants, such as titanium alloys, have an elastic modulus of 110 GPa, while stainless steel and cobalt-chromium alloys range from 200 to 230 GPa. These values are significantly higher than that of natural bone, leading to stress shielding [36]. In this study, the designed scaffold exhibited a strength of 119.62 ± 1.81 MPa and an elastic modulus of 2.68 ± 0.12 GPa. While providing stable mechanical support, its modulus falls well within the reported range of human bone (0.75–20 GPa), effectively minimizing the stress-shielding effect [37,38].

Bone regeneration is critically dependent on vascularization, as new bone formation requires oxygen, nutrients, growth factors, and stem cells delivered through blood vessels. The tightly coupled “angiogenesis-osteogenesis” mechanism underscores the interdependence of vascular and bone tissue growth [18]. VEGF plays a pivotal role by directly promoting the recruitment, proliferation, and differentiation of osteoprecursor cells and indirectly modulating osteoclast function, thereby synchronizing bone formation and remodeling [39]. Similarly, early tendon repair necessitates timely angiogenesis to support metabolically active repair cells and provide essential molecular signals. Inhibition of angiogenesis impairs tendon graft maturation and biomechanical integrity [40]. It is noteworthy that the contribution of VEGF to tendon repair extends beyond its classical angiogenic function. Evidence indicates that VEGF directly upregulates the expression of the key transcription factor Scleraxis (SCX) and Type I Collagen (Col1) in tenocytes, while concurrently downregulating Type III Collagen (Col3), demonstrating a direct pro-differentiation capacity independent of vascular supply [41]. The mechanisms underlying this direct effect are complex. Literature reported that VEGF enhances tendon-bone healing by activating the YAP signaling pathway [42]. Given that YAP is recognized as a core regulator essential for maintaining the stemness and rejuvenation of tendon stem/progenitor cells (TSPCs) and driving functional tendon repair [43,44], it is plausible that VEGF released from the scaffold acts partially through such pathways to stimulate the observed tenogenic differentiation in our study. Thus, based on current biological understanding, VEGF likely synergistically promotes functional tendon-bone healing by improving the microenvironment via angiogenesis and directly stimulating tenogenic differentiation. However, sustained excessive VEGF expression disrupts matrix organization and impairs tendon recovery, emphasizing the necessity for controlled, sustained, low-dose VEGF delivery at the TBI [45].

To achieve controlled delivery, the study developed a GS composite hydrogel integrated into a gradient-porous titanium scaffold, embedding VEGF-loaded nanospheres (VNS) to form a Ti@GS-VNS composite. The desolvation method enabled efficient VNS preparation with an encapsulation efficiency of 76.87 ± 0.65%. Transmission electron microscopy (TEM) confirmed a uniform spherical morphology with bilayer structures and a strong positive surface charge, imparting enhanced colloidal stability. In simulated body fluid (SBF), the nanoparticles exhibited changes in particle size and Zeta potential driven by hydration-induced swelling and deprotonation/charge-screening effects, validating their favorable colloidal stability. Crucially, the degraded morphology observed via SEM indicates that the material poses no risk of long-term in vivo accumulation. This preparation avoided toxic organic solvents, improving biocompatibility and minimizing immunogenicity [46]. The β-sheet GS hydrogel forms a semi-porous, interpenetrating network, verified by FTIR and SEM, which substantially retards hydrolytic degradation compared to pure GelMA hydrogels. The composite Ti@GS-VNS scaffold exhibited controlled VEGF release kinetics, with sustained bioavailability for approximately 456 h, thereby mitigating burst release. Cytological assays revealed that a sustained VEGF concentration of 10 ng/mL most effectively promoted TDSC migration, proliferation, and tenogenic gene expression, consistent with prior findings [41].

The Ti@GS-VNS scaffold exhibited enhanced hydrophilicity and a refined porous architecture, confirmed by water contact angle and SEM analyses, respectively. Cellular viability and proliferation studies verified excellent biocompatibility, with significant increases in cytoplasm-to-nucleus ratios in TDSCs, indicating favorable cell morphology and activity. Immunofluorescence and gene expression analyses revealed substantial upregulation of key tenogenic markers (COL1, DCN, SCX, TNC), while protein expression analyses via Western blot demonstrated elevated COL1 and DCN levels, critical for maintaining collagen stability and tendon functionality [47]. Additionally, angiogenesis assays, including tube formation and gene expression of CD31 and FGF-2, confirmed the scaffold's pro-vascular effects. Osteogenic potential was validated by increased ALP activity, mineralization staining, and upregulation of osteogenic genes (OPN, RUNX2, OCN, ALP).

In vivo, micro-CT imaging demonstrated superior new bone volume and quality around the Ti@GS-VNS scaffold. Histological evaluations showed robust integration between scaffold and surrounding bone, confirmed stable tendon attachment, and revealed enhanced neovascularization. The regenerated tendon tissue exhibited well-organized fibrous structures, forming continuous Sharpey-like fibers that anchored into host bone, restoring structural integrity. Sirius Red staining and polarized light microscopy highlighted improved collagen fiber alignment and elevated type I collagen content, hallmarks of functional matrix maturation [48]. Mechanical testing confirmed recovery of approximately 82% of that of native tendon-bone junction strength. These results collectively underscore the scaffold's capacity to promote bone and tendon regeneration and restore biomechanical function synergistically.

This investigation highlights the efficacy of a “structural adaptation–interface activation–signaling regulation” strategy, where the Ti@GS-VNS scaffold's dual VEGF release converts conventional burst release kinetics into a sustained, low-dose release profile that better corresponds to tissue repair dynamics. This controlled release modality enabled concurrent enhancement of tendon regeneration, osteogenesis, and vascularization within a single scaffold platform, ultimately facilitating biological and mechanical restoration of the TBI.

However, this study has limitations. The quadrupedal rabbit model does not fully replicate human postoperative joint biomechanics, potentially affecting translational relevance. A 12-week observation period limits the evaluation of long-term scaffold performance and final interface maturity. Moreover, TBI healing is governed by complex, orchestrated cellular processes involving multiple signaling pathways. The precise molecular mechanisms through which VEGF mediates interface regeneration in the current system remain to be fully elucidated. Additionally, while functional integration was achieved, fully recapitulating the native four-layer gradient structure within a porous metallic scaffold remains a significant challenge. We did not explicitly quantify the fine-tuned mapping of specific molecular gradients across the interface in this study. Future work should focus on extended in vivo assessments, detailed analysis of these gradient distributions, and the dissection of the underlying pathways to optimize scaffold design and functional outcomes for clinical application.

5. Conclusion

This study developed an integrated TBI repair scaffold that combines structural adaptation, interface activation, and signaling regulation. The gradient porous titanium alloy, combined with the GS hydrogel network and VEGF-loaded chitosan-stabilized BSA nanospheres, enabled differential tissue adaptation, achieving sustained VEGF release and multi-tissue induction effects. In vitro and in vivo experiments demonstrated that this composite scaffold simultaneously promotes angiogenesis, tendon regeneration, and synergistic osteogenesis, reconstructing a biologically continuous tendon-bone connection and significantly enhancing the mechanical strength of the interface. Overall, this composite scaffold shows excellent repair potential for TBI defects, providing a new strategy for addressing the challenge of soft-hard tissue integration and developing personalized regenerative implants.

CRediT authorship contribution statement

Tao Lin: Conceptualization, Data curation, Formal analysis, Methodology, Writing – original draft. Jiaying Li: Conceptualization, Data curation, Formal analysis, Methodology. Lincong Luo: Methodology. Hang Sang: Investigation, Writing – review & editing. Manoj Kumar Vashisth: Data curation, Writing – review & editing. Jianlin Shen: Investigation. Xiang Luo: Data curation. Lin Xu: Formal analysis, Funding acquisition, Investigation. Jing Pan: Conceptualization, Methodology. Wenhua Huang: Conceptualization, Data curation, Funding acquisition.

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:

Data availability

Data will be made available on request.