Nano-to-macroscale dual core-shell structured bioartificial ligament improves outcomes of anterior cruciate ligament reconstruction in rats: A proof-of-concept study

Abstract

Over 400,000 anterior cruciate ligament (ACL) reconstructions are performed worldwide annually. The ability of patients to return to their pre-injury or adapted level remains limited. Although autografts and allografts are recognized as good choices for ACL reconstruction, postoperative complications are frequently reported. The emergence of artificial ligaments offers a promising strategy to solve these challenging medical conditions. However, artificial grafts composed of different materials have been associated with complications such as chronic effusions, synovitis, and graft failure to varying degrees. Although great progress has been made in modifying artificial grafts, the optimal strategy remains to be explored. Magnesium and recombinant spidroin have been proven to be promising implantable materials that can facilitate osteogenesis both in vitro and in vivo, thus exhibiting good potential for orthopaedic implants or grafts. In this study, a novel nano-to-macroscale dual core-shell structured bioartificial ligament integrating magnesium and recombinant spidroin has been developed using conjugate coaxial electrospinning. This innovative graft promoted graft-bone integration by facilitating early expression of osteogenic, neurogenic, and cell adhesion markers (e.g., BMP2, RUNX2, CGRP, and ITGA5) at the interface and accelerating new bone mineralization around the bone tunnel. These effects resulted in improved healing outcomes after ACL reconstruction in a rat model, suggesting significant clinical potential for future ACL applications.

Graphical abstract

Highlights

• •A hierarchical bioartificial ligament was made by conjugate coaxial electrospinning.
• •The bioartificial ligament significantly enhanced early graft-bone integration.
• •It boosted functional recovery and tissue regeneration over pure PET graft.

1. Introduction

Anterior Cruciate Ligament (ACL) injuries are among the most prevalent sports-related, with an estimated incidence of 120,000 cases in the United States every year [1]. Furthermore, over 400,000 individuals undergo ACL reconstruction (ACLR) surgery annually [2]. However, reported rates of successful return to sports after surgery vary between 55% and 74% [3,4], suggesting the need for further improvement in clinical outcomes. Autografts have long been recognized as the “gold standard” for ACLR due to high cost-effectiveness and low immunogenicity. However, associated donor-site morbidity is not uncommon [[5], [6], [7]]. Concurrently with progress in material science and tissue engineering, artificial ligaments have been increasingly accepted as an alternative to autografts. Since the 1970s, various materials, including polypropylene (PP), polytetrafluoroethylene (PTFE), and polyethylene terephthalate (PET), have been employed in the artificial ligament fabrication [[8], [9], [10]]. Among them PET has become the most commonly utilized material for ACLR currently [11]. While PET fibers exhibit sufficient mechanical strength, their marked inertness is associated with a series of postoperative complications, such as bone tunnel enlargement, chronic knee inflammation, and graft fibrosis, which can eventually result in graft failure [[12], [13], [14], [15], [16], [17]]. Motivated by the need to improve clinical outcomes after ACLR, numerous studies have explored surface modification or bioactive coatings on PET fibers to counteract their inert nature, with encouraging progress [[18], [19], [20], [21]]. Nevertheless, the optimal strategy for artificial ligament design and application remains to be elucidated.

Magnesium (Mg), as the fourth most abundant element in the human body, is deeply involved in the construction of the musculoskeletal system [22,23]. Concurrently, Mg-based orthopaedic implants have been extensively developed and comprehensively investigated. Mg, along with its compounds, including magnesium oxide (MgO) and magnesium chloride (MgCl₂), have been demonstrated to exert diverse beneficial effects on bone tissue regeneration [[24], [25], [26]]. These include recruiting endogenous stem cells [27,28], regulating macrophage polarization [29], enhancing angiogenesis [30], and adjusting the metabolic balance between osteoblasts and osteoclasts [31], among other functions [32]. Spidroin, commonly referred to as spider silk protein, is the primary component of spider silk. Recombinant spidroins (rSps), produced through genetic engineering, are increasingly recognized as an ideal biomaterial due to their low immunogenicity and high cell affinity [[33], [34], [35]]. Numerous studies have demonstrated the efficacy of spidroins in the field of orthopaedic regeneration, yielding promising outcomes [[36], [37], [38]]. Recombinant spidroins exhibit considerable potential for broad application in orthopaedics and sports medicine, particularly in the fabrication of artificial ligaments.

Electrospinning is a technique that utilizes electrical force to generate nanofibers with diameters ranging from nanometers to micrometers [39,40]. The conjugate electrospinning technique is commonly employed to produce yarns comprising nanofibers [41]. The fundamental principle involves applying a positive high voltage to one needle and a negative high voltage to another, causing fibers emitted from both needles to converge and intersect midway, thereby forming yarns, as depicted in Fig. 1. Coaxial electrospinning, first applied and reported by Sun et al., in 2003 [42], enables the combination of two or more polymer solutions, thus yielding greater morphological diversity of electrospun fibers [43,44]. Conjugate coaxial electrospinning integrates these two methods to fabricate a hierarchical-structured artificial ligament with controlled degradation rates and sufficient mechanical strength. To the best of our knowledge, no previous study has combined conjugate coaxial electrospinning with bioactive components to create an artificial ligament with a nano-to-macro hierarchical structure. This approach, which integrates structural simulation and functional regeneration promotion, holds a promising pathway in the research and development of artificial implants for the field of sports medicine.

Fig. 1.

Schematic diagram showing the fabrication of dual core-shell bioartificial ligaments via conjugate coaxial electrospinning and the application in rat anterior cruciate ligament reconstruction to improve graft-bone integration through promoting the adhesion and osteogenic differentiation of bone marrow mesenchymal cells (BMSCs).

The aim of this investigation is applying the conjugate coaxial electrospinning method to fabricate a dual core-shell structured bioartificial ligament, which couples hierarchically nano-to-macro structure and biological functions to enhance graft-bone integration at early stage. The hierarchical structure mimicked the native bone-fibrocartilage-ligament interface gradients and the combination of MgO and rSp boosted cell adhesion and osteogenic differentiation at the graft-bone interface after ACL reconstruction in rats, resulting in superior functional recovery and tissue outcomes over pure PET graft. We first investigated the feasibility of fabricating core-shell structured bio-membranes and performed in vitro experiments on these bio-membranes to assess their properties. Subsequently, dual core-shell structured bioartificial ligaments, incorporating identical bioactive components, were fabricated using the conjugate coaxial electrospinning method. Both in vitro and in vivo experiments were performed to evaluate the regenerative potential of these bioartificial ligaments following ACLR in a rat model. We hypothesized that the dual core-shell structured Mg- and rSp-incorporated bioartificial ligament would promote the graft-bone integration at the early stage and accelerate the new bone mineralization around the bone tunnel, thereby improving healing outcomes after ACLR in a rat model.

2. Materials and methods

2.1. Fabrication and characterization of bio-membranes

First, to determine the optimal concentrations of MgO and rSp for facilitating cell viability and osteogenesis in vitro. MgO nanopowder was dissolved at varying weights in 15%w/v poly (D, L-lactide-co-glycolide) (PLGA, Mw: 76000-115000, Aladdin)/hexafluoroisopropanol (HFIP, Aladdin) solution to generate gradient concentrations (3, 6, 9, 12, and 15% w/v, respectively). In the meantime, rSp was dissolved at different weights in 15%w/v PLGA/HFIP solution to achieve gradient concentrations (3, 6, and 9% w/v, respectively). The construction, expression, and purification steps of rSp are detailed in supplementary files (Supplementary file S1). Then each membrane was fabricated via electrospinning. Solutions were loaded into a 10 mL syringe with a #22 blunt-end needle and ejected at 1 mL/h using a syringe pump. Nanofibers were collected on a mandrel rotating at 1500 rpm under +11 kV applied voltage, with the acceptance distance of 12 cm. Following drying and overnight UV sterilization, three identical circle-shaped membranes (Ø = 25 mm) were sectioned from each membrane. Each membrane was placed in a 12-well plate, rinsed with phosphate buffered saline (PBS) solution, and secured under a 316L stainless steel ring (Ø = 25 mm). Passage 3 rat bone marrow mesenchymal stem cells (P3 rBMSCs; 20,000 cells/mL) were seeded onto membranes, and Dulbecco's Modified Eagle Medium (DMEM) was added to a total volume of 2 mL/well. Cell viability was quantified using a CCK-8 kit at 24, 48, and 72 h post-seeding. Optimal concentrations of MgO and rSp were determined based on viability results.

Core-shell structured bio-membranes were subsequently fabricated via coaxial electrospinning. Using optimized MgO and rSp concentrations (6% w/v each, Fig. 2a–f), shell solutions were prepared by supplementing 15% w/v PLGA/HFIP as follows: 15% PLGA/6% rSp (PR), 15% PLGA/6% MgO (PM), and 15% PLGA/6% MgO/6% rSp (PMR). The core solution comprised 15% w/v polycaprolactone (PCL, Mw:80000, Sigma)/HFIP. Four experimental groups were established based on core-shell composition: 15% w/v PLGA/15% w/v PCL (PP group); 15% w/v PLGA + 6% rSp/15% w/v PCL (PRP group); 15% w/v PLGA + 6% MgO/15% w/v PCL (PMP group); 15% w/v PLGA + 6% MgO + 6% rSp/15% w/v PCL (PMRP group). Core solution and shell solution were ejected at 0.2/0.8 mL/h respectively using a coaxial blunt-end needle. Nanofibers were collected on a mandrel rotating at 2000 rpm under +15 kV applied voltage, with the acceptance distance of 12 cm.

Fig. 2.

Determination of the optimal concentration of MgO and rSp, as well as cell characterization of different groups of bio-membranes. a-c Viability of rBMSCs seeded on different ratio of MgO e-spun membranes after 24, 48, and 72 h measured by CCK8 Kit, respectively (n = 3 per group). d-f Viability of rBMSCs seeded on different ratio of rSp e-spun membranes after 24, 48, and 72 h measured by CCK8 Kit, respectively (n = 3 per group). g-j Images of the elements analysis of nanofibers from different groups. Representative (k,l/n,o/q,r/t,u) SEM and (m,p,s,v) TEM images of e-spun nanofibers of PP, PRP, PMP, and PMRP groups, respectively. w The 28-day degradation profiles of different bio-membranes in PBS solutions (n = 3 per group). x The 28-day accumulated release profiles of Mg ion from PMP and PMRP bio-membranes (n = 3 per group). k, n, q, t scale bar: 10 μm; l, m, o, p, r, s, u, v scale bar: 1 μm.

The overall and detailed morphology of bio-membranes were characterized by scanning electron microscopy (SEM), with nanofibril diameters quantified (n = 100 per group). Fine structure features were further analyzed using transmission electron microscopy (TEM). For degradation assessment, twelve square specimens (2 × 2 cm) per group were sectioned. Initial dry weight was recorded prior to immersion on 10 mL phosphate-buffered saline (PBS) within 15 mL tubes (37 °C incubation). At weeks 1, 2, 3, and 4, triplicate specimens per group were removed, thoroughly dried, and reweighed. Mg ion release kinetics were evaluated for PMP and PMRP groups. Triplicate specimens (2 × 2 cm) were immersed in 10 mL Mg²⁺-free deionized water (15 mL tubes, 37 °C). Cumulative Mg²⁺ concentrations were measured via inductively coupled plasma mass spectrometry (ICP-MS) at 12h, 1d, 3d, 7d, 10d, 14d, 21d, and 28d. For the measurement of MgO loading efficiency, three electrospun membranes weighing 10 mg each were taken from the PMP and PMRP groups, respectively, and placed into 1 mL of nitric acid. The samples were dissolved in an oven at 65 °C overnight, and then centrifuged at 15,000 rpm at room temperature for 10 min. Subsequently, 10 μL of the supernatant was mixed with 90 μL of MilliQ water, and the Mg²⁺ concentration was determined using ICP-MS. For the quantification of rSp loading efficiency, three electrospun membranes weighing 20 mg each were taken from the PRP and PMRP groups, respectively, and placed into 1.5 mL tubes. Each membrane was dissolved in 1 mL of chloroform with vigorous shaking until complete dissolution. The tubes were then centrifuged at 15,000 rpm and 4 °C for 10 min. The supernatant was removed, and the pellet was washed twice with 1 mL of chloroform under the same centrifugation conditions. After removing the supernatant, the residual solvent was evaporated. Subsequently, 1 mL of 0.1 M NaOH/5% SDS solution was added to the tube, and the mixture was shaken overnight at 4 °C. The sample was centrifuged at 15,000 rpm and 4 °C for 10 min, and 20 μL of the supernatant was mixed with 80 μL of MilliQ water. The resulting solution was thoroughly mixed and used for protein quantification using the BCA assay (ThermoFisher, A55865).

2.2. Cell viability, alkaline phosphatase, and alizarin red S staining of bio-membranes

P3 rBMSCs were seeded on bio-membranes with a density of 20,000 cells/mL in DMEM. Cell viability was assessed using a CCK-8 Kit at 24, 48, and 72 h post-seeding by calculating the OD₄₅₀ value for each group. At 24 h, DMEM was replaced with osteogenic induction medium. Triplicate membranes per group underwent alkaline phosphatase (ALP) staining at day 5 and alizarin red S (ARS) staining at day 14 post-induction.

2.3. Cell adhesion observation and gene expression measurement

P3 rBMSCs (20,000 cells/mL in DMEM) were seeded on bio-membranes. After 12h culture, membranes were washed thrice with PBS and fixed in 4% buffered formalin solution for 30 min. Gradual dehydration was performed with ethanol series (70%, 80%, 90%, and 100%; 15min each). After drying, samples were gold-coated for SEM imaging.

A targeted screening of cell adhesion-related genes was conducted prior to comprehensive gene expression analysis (Supplementary file S2). For gene expression measurement, total RNA was extracted using RNiso Plus kit (Takara, 9109) per manufacturer's protocol, with concentration measured by NanoDrop 2000. Reverse transcription of 1 μg total RNA utilized PrimeScript RT reagent kit (Takara, RR047A). Quantitative reverse transcription polymerase chain reaction (RT-qPCR) employed TB Green™ Premix Ex Taq™ (Takara, RR420A) in 20 μL reactions containing: 10 μL premix, 4 μL cDNA, 1 μL forward primer, 1 μL reverse primer, and 4 μL diethylpyrocarbonate (DEPC) water. Gene expressions of Itga5, Itgav, Itgb1, Col1a1, Bmp2, Runx2, Sox9, Opn, and Vegfa were analyzed, with β-actin as internal control. Primer sequences are listed in Supplementary file (Table S1). Reactions were run on Applied Biosystems™ QuantStudio™ 12K Flex (Thermo Fisher Scientific) under cycling conditions: 95 °C for 10 min; 40 cycles of 95 °C for 15 s and primer-specific annealing temperature for 1 min. All samples were tested in triplicate. Relative gene expression was calculated using the 2^-ΔΔCt method [45]:

• 1.Normalize target gene Ct values to β-actin Ct values
• 2.ΔCt _test = Ct _{target, test} – Ct _{β-actin, test}; ΔCt _control = Ct _{target, control} – Ct _{β-actin, control}
• 3.ΔΔCt = ΔCt_test – ΔCt_control
• 4.Fold change = 2^-ΔΔCt

2.4. Fabrication and characterization of bioartificial ligament

Polyethylene terephthalate (PET) fibers served as the core to ensure mechanical strength in core-shell structured bioartificial ligaments. Coaxial electrospinning was employed for the outer shell layer to generate core-shell nanofibers with controlled degradation kinetics. Core and shell solutions were identical to Section 2.1. Conjugate electrospinning was utilized to fabricate macroscale core-shell bioartificial ligaments. Core solution and shell solution were loaded separately in 10 mL syringes and were ejected at 0.2/0.8 mL/h respectively via coaxial blunt-end needles. The inter-needle distance was maintained at 30 cm, with the distance between the needles and the collecting drum set at 18 cm. The positive and negative voltage were set at ± 15 kV, respectively. The rotating speed of the collecting drum was set at 50 rpm, and the winder speed was set at 10 rpm. Ligament morphology was characterized by SEM. Failure load was determined using a Hounsfield H25K-S testing system (Hounsfield Test Equipment Ltd), comparing experimental groups with pure PET ligaments and rat flexor digitorum longus (FDL) tendons.

2.5. ACLR surgery in rat model and postoperative gait analysis

All animal experiments were approved by the Animal Ethics and Experimentation Committee (No: 23-010-MIS). One hundred fifty male Sprague-Dawley rats (Aged 12-13 weeks) were randomly allocated to five groups (n = 30 per group): PP, PRP, PMP, PMRP, and PET. The ACLR surgery was performed on the right knee in every rat after general anesthesia (Ketamine: Xylazine = 75: 10 mg/kg) via intraperitoneal injection. Ligament grafts were UV-sterilized overnight and moistened by 0.9% saline before implantation. A 2-cm longitudinal incision was made in the medial side of the knee joint. After opening the joint capsule and pushing the patella to the lateral side, the knee joint cavity was exposed and the synovial tissue in front of the ACL was removed. The native ACL was excised and confirmed by a positive anterior drawer test. Subsequently, the knee joint was fixed at 90° flexion, and a 1.2 mm-diameter drill was used to create the bone tunnels from the distal medial tibia to the proximal lateral femoral condyle. Tunnels were irrigated with 10 mL saline and group-specific ligaments were implanted. The composite core-shell structure maintained integrity during tunnel passage, resisting intraoperative shear forces. Successful reconstruction was confirmed by negative anterior drawer test. Ligament ends were anchored to bone cortex at 30° flexion. The joint was irrigated with 5 mL saline and closed in layers. Postoperative analgesia (0.05 mg/kg Temgesic subcutaneously, twice daily) was administered. Fluorochrome labeling sequenced calcein green (10 mg/kg subcutaneously, Sigma-Aldrich) at postoperative weeks 1/4/10 and xylenol orange (30 mg/kg subcutaneously, Sigma-Aldrich) at weeks 2/5/11, with sacrifice at weeks 3/6/12, respectively.

Gait analysis (CatWalk XT, Noldus) was performed at weeks 3, 6, and 12. A calibrated region of interest (20 × 8 cm) and dual infrared beams (60-cm spacing) recorded walks under dark conditions. Rats were put on and allowed to traverse the glass plate freely. Compliant runs (speed variation <30%) were collected (≥3 runs/rat) and automated limb identification (RF: right front; RH: right hind; LF: left front; LH: left hind) was manually verified. Parameters, including maximum contact area, maximum contact intensity, and stride length, were recorded. Before comparison between different groups, data normalization was performed. The data of LH were used as control data for every run of every rat, and the ratio of RH/LH was calculated for each parameter. Comparisons were then performed between groups.

2.6. Micro-CT analysis and mechanical testing

Rats were sacrificed at postoperative weeks 3, 6, and 12. Right knees were fixed in 4% buffered formalin for 48 h and stored in 70% ethanol. Bone tunnel mineralization and diameter were quantified using high-resolution micro-CT (vivaCT, Scanco Medical). Samples were secured in 50-mL tubes and scanned (70 kV, 114 μA, 17.5-μm voxel size; Gaussian filter σ = 0.8, support = 1). Images were reoriented vertically by Z-axis alignment to present a regular circle of bone tunnels (Fig. 7a and b). One hundred consecutive slices from the femoral tunnel midportion were analyzed (Scanco software). A circular ROI (Ø = 1.2 mm) matching drill diameter was applied. Mineralized tissue thresholds were set at >165 HU for weeks 3 and 6, as well as >220 HU for week 12. Bone volume fraction (BV/TV) and bone mineral density (BMD, mg/cm³) were measured by two individual investigators (H.Z., Y.A.).

Fig. 7.

a Procedures of micro-CT evaluation: the irregular shape (yellow dotted line) of the bone tunnel before Z-axis alignment of micro-CT images, the round shape (green circle) of the bone tunnel after Z-axis alignment, and ROI (a total of 100 slices between the two red dotted lines) selection of the femoral bone tunnel. b-d Results of micro-CT analysis, including the BMD and BV/TV ratio, at 3, 6, and 12 weeks after surgery, respectively (n = 3 per group). e Removal of all tissue except the ACL, mechanical testing diagram, and representative failure mode of pull-out test. f-h Results of mechanical testing, including ultimate failure load and stiffness at week 3, 6, and 12 postoperative, respectively (n = 6 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed.

For mechanical testing, knees were thawed at room-temperature water for 30 min. Periarticular tissues, including muscles, patellar tendon, collateral ligaments, menisci, posterior cruciate ligament, and fixation sutures, were meticulously dissected. Femoral and tibial segments were embedded in 2 mL tubes by polymethyl methacrylate (PMMA). Specimens were mounted on a Hounsfield H25K-S testing machine, following 5 cycles of preconditioning at the maximum displacement of 0.5 mm, and loaded to failure with a displacement rate of 20 mm/min. All tested samples were moistened with saline to avoid desiccation. The load-displacement curves were recorded (n = 6 per time point per group).

2.7. Histological evaluations

2.7.1. Histological evaluations for decalcified bone tissues

The knee joints were dissected into femoral and tibial portions, and both portions were randomly allocated into decalcified group and undecalcified group. Samples in decalcified group were soaked into 12.5% EDTA (pH 7.4) with daily solution changes for 3 weeks. The samples were then dehydrated with graded ethanol (70%, 90%, and 100%), cleared with xylene for 24 h, and embedded with paraffin. Next, 5-μm sections were stained with Hematoxylin and Eosin (H&E) and toluidine blue for peri-tunnel tissue evaluation. For immunohistochemistry (IHC), sections were deparaffinized, rehydrated, then sequentially treated with 10 mM sodium citrate solution at 60 °C for 30min, 3% H₂O₂ for 10min, and blocking solution (5% FBS + 1% BSA in PBS) for 20min. After washed with double-distilled water twice, the sections were incubated with primary antibodies at 4 °C overnight. The primary antibodies used were ITGA5 (1:150, ab150361, Abcam), ITGB1 (1:150, ab183666, Abcam), CGRP (1:150, ab47027, Abcam), RUNX2 (1:150, L0710, Santa Cruz), and BMP2 (1:150, ab214821, Abcam). After thoroughly washing with double-distilled water, the sections were incubated with horseradish peroxidase (HRP) conjugated secondary antibody (1:300, ab6741; ab6721, Abcam) at room temperature for 1 h, then treated with the DAB Solution Kit (TA-060-QHDX, Thermo Fisher) for visualization. Finally, hematoxylin staining was performed to counterstain cell nuclei, and sections were dehydrated and mounted with distyrene-plasticizer-xylene (DPX). For data analysis of IHC staining, three sections were randomly chosen. The positive cells were identified by ImageJ, and the ratio of positive cells to total cells was calculated.

2.7.2. Histological evaluations for undecalcified bone tissues

For undecalcified sections, samples were treated with graded ethanol and xylene, then embedded in methylmethacrylate (MMA) based on our published protocol [46]. A diamond band-cutting machine (EXAKT Advanced Technologies) was utilized to cut the samples into 150 μm-thick sections. Next, sections were ground to <100 μm and polished. A ROI around the bone tunnel area was determined to assess the new bone formation. The mineralization activity of bone was evaluated by the xylenol/calcein ratio with Image J software. After fluorescent imaging assessment, the MMA sections were treated with the modified Stevenel's Blue-Van Gieson-ARS staining according to our established protocol [47]. Briefly, the MMA sections were soaked with Stevenel's Blue solution in a 65 °C oven for 30 min and were rinsed with 65 °C double-distilled water for 2 min. Next, the sections were immersed into Van Gieson's solution at room temperature for 5 min and the solution was removed by using absorbent paper. Thereafter, ARS solution was dipped onto the MMA sections at room temperature for 5 min and the sections were then washed with double-distilled water thoroughly.

All the sections were observed and recorded via a Leica DM5500 microscope (Leica Microsystems).

2.8. Statistical analysis

Quantitative data were presented as mean ± standard deviation (M ± SD), and one-way analysis of variance (one-way ANOVA) with LSD post-hoc analysis was performed for comparison between every group at each time point. The significance level was set at P < 0.05. All the data were processed by SPSS (Version 25.0, SPSS, IL, USA).

3. Results

3.1. Characterization of bio-membranes

CCK-8 assay (OD₄₅₀) demonstrated enhanced cell viability with increasing MgO concentrations (3-12% w/v) at 24, 48, and 72 h post-seeding, with peak viability at 6% w/v. Similarly, rSp concentrations (3-9% w/v) enhanced viability at 48/72 h (no significant difference vs. control at 24 h for 9% rSp), peaking at 6% w/v. Thus, 6% w/v was selected for both MgO and rSp (Fig. 2a–f). SEM images of fiber morphology revealed well-formed nanofibers across groups. The diameters of 100 nanofibers were counted in each group to measure the distribution, and the distribution ranges of all groups conformed to normal distribution, with PP group showing the largest fiber diameter on average (≈600 nm), and PMP group showing the smallest one (≈400 nm) (Fig. 2g–j). Meanwhile, the TEM images of fibers indicated the core-shell structure in each group, with MgO nanoparticles distributed in the shell layer (Fig. 2k–v).

Degradation assessment of bio-membranes over 28 days revealed that the PMP group exhibited the most rapid degradation kinetics. The degradation rate of the PMRP group was slower than PMP but faster than PRP, while the PP group demonstrated the slowest degradation rate (Fig. 2w). Concurrently, Mg ion release profiles for PMP and PMRP groups indicated that PMP displayed both higher cumulative release and faster release kinetics than PMRP. Both groups exhibited sustained release patterns rather than burst release, characterized by relatively rapid release during the initial 10 days followed by substantially slower release from days 10 to 28 (Fig. 2x). The loading efficiency of MgO in PMP and PMRP group was 92% and 94%, respectively, and the loading efficiency of rSp in PRP and PMRP group was 81% and 78%, respectively.

3.2. Cell viability and osteogenic effects on bio-membranes in vitro

Cell viability assessment demonstrated significantly higher values in PRP, PMP, and PMRP groups compared to the blank control group at all time points (all P < 0.05). No significant difference existed between the blank control and PP groups at 48 h and 72 h. Among the experimental groups, PRP, PMP, and PMRP showed comparable viability without significant intergroup differences (Fig. 3a–c). ALP staining intensity was most pronounced in the PMRP group, followed by PMP. Both PMRP and PMP groups exhibited stronger staining than PRP and PP groups (Fig. 3d). Correspondingly, ARS staining revealed enhanced calcium nodule formation in PMP and PMRP groups relative to other groups, with the PMRP group displaying the greatest calcium deposition (Fig. 3e).

Fig. 3.

Cell viability, alkaline phosphatase, and alizarin red S staining of bio-membranes. a-c Viability of rBMSCs seeded on different bio-membranes after 24, 48, and 72 h measured by CCK8 Kit, respectively (n = 3 per group). d ALP staining (n = 3 per group) and e Alizarin red staining (n = 3 per group) of rBMSCs seeded on different bio-membranes. ∗:P < 0.05 compared to the blank control group. #:P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 100 μm.

3.3. Cell adhesion and gene expression assessment on bio-membranes in vitro

SEM evaluation at 12 h post-seeding corroborated these gene expression findings, demonstrating that the PMRP group attained the highest density of adherent cells while virtually no cells were detectable on the PP group membranes, presented in Fig. 4a–h. Following 14 days of osteogenic differentiation in P3 rBMSCs, quantitative analysis confirmed that adhesion-related genes including Itga5, Itgb1, and Opn exhibited significant upregulation in the PRP, PMP, and PMRP groups relative to the blank control group. Concurrently, the angiogenic gene Vegfa and osteogenic genes Bmp2, Runx2, and Sox9 also showed statistically significant elevation in these three experimental groups compared to controls. No expression differences were observed between the blank control and PP groups for any of these genes, and Col1a1 expression remained consistent across all experimental groups, with complete data illustrated in Fig. 4i.

Fig. 4.

(a, c, e, and g) Macroscopic and (b, d, f, and h) microscopic images of rBMSCs adhesion on PP, PRP, PMP, and PMRP bio-membranes, respectively. ∗:P < 0.05 compared to the blank control group. ANOVA with LSD post-hoc analysis was performed. i Relative mRNA expression of adhesion, angiogenesis, and osteogenesis-related genes in rBMSCs, including Itga5, Itgb1, Opn, Vegfa, Col1a1, Bmp2, Runx2, and Sox9. a, c, e, g scale bar: 100 μm; b, d, f, h scale bar: 10 μm.

3.4. Characterization of bioartificial ligaments

SEM imaging confirmed that conjugate coaxial electrospinning successfully fabricated bioartificial ligaments exhibiting dual core-shell structures, characterized by hierarchical fiber organization spanning nanoscale outer layers to micro-scale PET core fibers (Fig. 5a–d). Surface topography analysis further indicated predominant alignment of surface fibers parallel to the longitudinal ligament axis, shown in Fig. 5e–h. The thickness of bio-membranes measured approximately 120 μm for single-strand constructs, and nine-strand ligaments utilized for mechanical testing and surgical implantation averaged 1.2 mm in diameter.

Fig. 5.

a Fabrication of dual core-shell bioartificial ligaments via conjugate coaxial electrospinning. b The cross-sectional and c top view of a nine-strand bioartificial ligament. d The cross-sectional view of a single strand of ligament. e-h Detailed images illustrating the superficial fiber morphology of PP, PRP, PMP, and PMRP groups, respectively. i Mechanical testing diagram and j failure load comparisons between different groups of bioartificial ligaments (n = 3 per group). ∗P < 0.05 compared to FDL tendon; #P < 0.05 compared to PET ligament. ANOVA with LSD post-hoc analysis was performed. b, c, d scale bar: 100 μm; e, f, g, h scale bar: 10 μm.

Mechanical testing results established that pure PET ligaments achieved the highest failure load at 168 ± 13.11 N, whereas rat flexor digitorum longus (FDL) tendons exhibited the lowest failure load at 25.37 ± 1.59 N. Intermediate failure loads were recorded in the experimental groups: PP at 99.07 ± 2.85 N, PRP at 106.33 ± 6.65 N, PMP at 110.33 ± 3.21 N, and PMRP at 110.67 ± 2.08 N, detailed in Fig. 5i and j. Although all experimental groups demonstrated significantly lower failure loads than the PET group, their mechanical performance substantially exceeded that of the commonly used autograft (FDL tendon) group, confirming their mechanical suitability for ACLR in the rat model.

3.5. Postoperative gait analysis

Gait analysis revealed minimal intergroup differences during early postoperative stages, with statistically significant variations observed only in specific comparisons: At week 3, the PMP group demonstrated greater maximum contact area than the PET group, while the PP group exhibited shorter stride length compared to both PET and PMRP groups. At week 6, stride length remained reduced in the PP group relative to the PET group. By week 12, the PMRP group achieved superior maximum contact area and contact intensity measurements compared to all other experimental groups, indicating enhanced weight-bearing capacity on the reconstructed limb and suggesting improved long-term functional recovery (Fig. 6b–g). The emergence of pronounced intergroup differences at this late timepoint may reflect progressive joint degeneration in the PET group, whereas bioactive groups maintained superior joint integrity.

Fig. 6.

a Procedures of ACL reconstruction surgery in rats: exposure and transect of ACL, drilling bone tunnels with a diameter of 1.2 mm from the distal medial tibia to the proximal lateral femur, guiding the ligament through the bone tunnels by a needle, and anchoring the ligament at both exists. b-d Representative images of footprint area at week 3, 6, and 12 postoperative and e-g quantitative analysis of max contact area, max contact intensity, and stride length by Catwalk gait analysis at week 3, 6, and 12 postoperative, respectively.

3.6. Micro-CT evaluations and mechanical testing in vivo

Micro-CT quantification consistently demonstrated that the PMRP group attained the highest BMD and BV/TV values at all postoperative intervals (83.76 ± 12.57 mg HA/ccm, 0.04 ± 0.01%; 97.68 ± 15.78 mg HA/ccm, 0.08 ± 0.02%; 167.6 ± 70.47 mg HA/ccm, 0.12 ± 0.09%), while the PET group consistently yielded the lowest measurements (28.1 ± 8.22 mg HA/ccm, 0.003 ± 0.001%; 42.8 ± 1.07 mg HA/ccm, 0.003 ± 0.001%; 64.88 ± 7.3 mg HA/ccm, 0.001 ± 0.001%). When comparing the PET group to bioactive experimental groups, distinct patterns emerged: The PMP group exhibited significantly elevated BMD to PET at weeks 3 and 6 (62.95 ± 1.51 mg HA/ccm; 74.46 ± 6.55 mg HA/ccm). Similarly, the PRP group showed significantly greater BMD than PET at 3 and 6 weeks (53.49 ± 2.7 mg HA/ccm; 63.61 ± 3.08 mg HA/ccm). For the PP group, only BMD measurement at 3 weeks significantly exceeded that of the PET group (46.38 ± 1.83 mg HA/ccm) (Fig. 7b–d).

The results of mechanical testing are presented in Fig. 7f–h. PMRP group exhibited the highest failure load and stiffness at each tested time point (5.61 ± 0.93 N, 2.17 ± 0.34 N/mm; 7.39 ± 0.23 N, 3.1 ± 0.72 N/mm; 14.73 ± 1.67 N, 5.7 ± 1.28 N/mm), followed by PMP group (4.58 ± 0.49 N, 1.97 ± 0.54 N/mm; 7.21 ± 0.28 N, 2.56 ± 0.59 N/mm; 12.11 ± 1.04 N, 4.38 ± 0.86 N/mm) and PRP group (4.1 ± 0.46 N, 1.62 ± 0.53 N/mm; 6.88 ± 0.59 N, 2.12 ± 0.23 N/mm; 11.21 ± 0.72 N, 3.3 ± 0.58 N/mm). PP group showed the lowest stiffness at weeks 3 and 6 (1.44 ± 0.29 N/mm; 1.9 ± 0.2 N/mm), and the PET group showed the lowest failure load at all time points (3.06 ± 0.65 N; 5.4 ± 0.6 N; 7.06 ± 0.73 N), as well as the lowest stiffness at week 12 (2.64 ± 0.39 N/mm). While little significant difference was detected between the PMP and PRP groups. In addition, the typical failure model of tested samples was pull-out of the graft from the bone tunnel (Fig. 7).

3.7. Histological evaluations

The H&E staining across timepoints demonstrated distinct interfacial microstructures (Fig. 8). The in vivo results showed that at week 3 after surgery, the PMP group exhibited the fastest degradation rate. At week 6, the PMRP group also underwent complete degradation. At week 12, only the PP group did not degrade completely, and in other groups, the space left by the material was filled with regenerated tissue post-degradation. The PET group exhibited comparatively looser histological organization versus bioactive groups. Toluidine blue staining revealed accelerated chondrogenesis surrounding bone tunnels in PRP, PMP, and PMRP groups at 3-6 weeks, whereas PET and PP groups showed very limited chondrocyte activation during this early phase. By 12 weeks, chondrogenesis became evident in PET and PP groups but with reduced chondrocytes around the ROI in other three groups.

Fig. 8.

a, b Images of H&E and toluidine blue staining around the bone tunnel site of each group at 3, 6, and 12 weeks after surgery, respectively. Scale bar: 100 μm.

Immunohistochemical analysis documented temporal expression patterns (Fig. 9, Fig. 10, Fig. 11, Fig. 12). For the results of ITGA5, PRP and PMRP groups displayed significantly higher expression levels than both PMP and PET groups at 3 and 6 weeks, though all groups achieved comparable expression by week 12. For CGRP expression, PMRP demonstrated maximal expression intensity exceeding other groups and significantly higher than that of PET, PP, and PRP at week 3, while no intergroup differences emerged at later timepoints. For the expression of RUNX2, PMRP maintained superior expression versus all groups at 3 and 6 weeks, while PP showed significantly reduced expression relative to PET at week 12. As for BMP2, all bioactive groups (PRP, PMP, and PMRP) exhibited persistently and significantly elevated expression over PET throughout the study, with PMRP showing the highest intensity among bioactive formulations at all time points.

Fig. 9.

Results of IHC staining of ITGA5 in the bone tunnel site of each group, at 3, 6, and 12 weeks after surgery, respectively. (Left) All the pictures have been oriented. In each image, the structure shown from left to right are graft, interface, and bone, respectively. At each time point, the images in the row below are magnified details of the images in the row above. (Right) Comparisons between each groups regarding the ITGA5-positive cell ratio (n = 6 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 100 μm.

Fig. 10.

Results of IHC staining of CGRP in the bone tunnel site of each group, at 3, 6, and 12 weeks after surgery, respectively. (Left) All the pictures have been oriented. In each image, the structure shown from left to right are graft, interface, and bone, respectively. At each time point, the images in the row below are magnified details of the images in the row above. (Right) Comparisons between each groups regarding the ITGA5-positive cell ratio (n = 6 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 100 μm.

Fig. 11.

Results of IHC staining of BMP2 in the bone tunnel site of each group, at 3, 6, and 12 weeks after surgery, respectively. (Left) All the pictures have been oriented. In each image, the structure shown from left to right are graft, interface, and bone, respectively. At each time point, the images in the row below are magnified details of the images in the row above. (Right) Comparisons between each groups regarding the ITGA5-positive cell ratio (n = 6 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 100 μm.

Fig. 12.

Results of IHC staining of RUNX2 in the bone tunnel site of each group, at 3, 6, and 12 weeks after surgery, respectively. (Left) All the pictures have been oriented. In each image, the structure shown from left to right are graft, interface, and bone, respectively. At each time point, the images in the row below are magnified details of the images in the row above. (Right) Comparisons between each groups regarding the ITGA5-positive cell ratio (n = 6 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 100 μm.

Undecalcified MMA sections with xylenol and calcein fluorochrome labeling (Fig. 13) demonstrated that at week 3, PMRP achieved superior new bone formation ratios compared to all groups. By week 6, both PRP and PMP groups significantly exceeded the bone formation capacity of the other three groups. At the 12-week endpoint, bioactive groups (PRP, PMP, and PMRP) maintained comparable high formation ratios that consistently outperformed PET, whereas PET and PP groups remained statistically indistinguishable.

Fig. 13.

a-c Images of new bone formation measured by MMA slides with xylenol and calcein fluorescent staining in the bone tunnel site of each group at week 3, 6, and 12 after surgery, respectively. d-f Quantitative comparisons between each group regarding the xylenol/calcein ratio (n = 3 per group). ∗P < 0.05 compared to the PET group; #P < 0.05 compared to the PMRP group. ANOVA with LSD post-hoc analysis was performed. Scale bar: 250 μm.

The interface between the bone and graft was also evaluated by the modified Stevenel's Blue-Van Gieson-ARS staining (Fig. 14). PRP, PMP, and PMRP groups developed Sharpey's-like fibers (yellow arrows) at the interface at week 3, with PMRP group exhibiting the most extensive fiber networks. These connective structures persisted throughout the postoperative period. By contrast, PP group initiated Sharpey's-like fibers formation only at week 6, while PET group displayed delayed Sharpey's-like fibers formation at week 12.

Fig. 14.

Representative images of interface between the bone and graft measured by MMA slides with modified Stevenel's Blue-Van Gieson-ARS staining. The fibrosis area was stained in blue, the Sharpey's-like fiber area was stained in purple, and the bone area was stained in red. The yellow arrows indicate the alignment of the Sharpey's-like fibers. Scale bar: 100 μm.

4. Discussion

In this study, a dual core-shell structured bioartificial ligament was fabricated from nano to macroscale using conjugate coaxial electrospinning. The bioartificial ligament significantly enhanced early graft-bone integration after ACL reconstruction in rats, and it boosted functional recovery and tissue regeneration over pure PET graft.

As one of the most popular materials utilized in fabricating artificial ligaments, PET exhibits reliable initial mechanical properties postoperatively, as well as immediate and fast rehabilitation after surgery [48]. However, as a non-degradable polymer, this very permanence constitutes its fundamental limitation: chemical inertness and low hydrophilicity impede cellular adhesion and remodeling, thereby obstructing graft-bone integration [9]. Histological analyses consistently demonstrate fibrous scar tissue formation at the graft-bone interface, a primary contributor to graft slippage within bone tunnels [49]. This scar tissue represents the biomechanically weakest region, providing insufficient resistance to graft pull-out forces during mechanical stress, ultimately predisposing to graft failure [50,51]. These inherent limitations have spurred research into bioactive surface modifications to enhance the biological performance of artificial ligaments.

Traditional calcium/phosphorus/silicon-based coatings, such as hydroxyapatite (HAp) deposited via surface soaking, demonstrate osteogenicity at 8 weeks [52]. Calcium-silicate modifications similarly established peri-graft trabecular networks at 12 weeks [53]. Hyaluronic acid/chitosan [54] and HAp/cationized gelatin soakings [55] on PET fibers could suppress inflammation and promote formation of graft fibers. However, surface soaking is not an ideal way for modification of PET fibers, for the precipitated components might agglomerate locally. Such uneven and unstable coating restricts the biological activity in a limited area. Besides, the insufficient coating ratio as well as lack of precise control could also weaken the biological effects on bone-graft integration [56]. Cai et al. constructed a more homogeneous and stable calcium phosphate (CaP) layer on PET fibers using electrodeposition (ED) than using biomineralization (BM) method and yielded significantly stronger osteogenesis in vitro [57]. This research supported that a stable and homogeneous microenvironmental surface was more favorable to the osteogenesis of cells. In our study, the dual core-shell structured PMRP bioartificial ligaments exhibited several biophysical advantages over pure PET controls and traditional surface modification methods. First, the electrospun nanomembranes provided topography cushioning ligament-bone interface gradients, greatly expanding surface area versus pure PET for stress distribution and cell adhesion. This nanotextured architecture establishes micro-niches enhancing hydrophilicity and surface energy [58], while spatially organized mechanical forces regulate cellular polarization-contraction dynamics via stress-driven mechanisms [59,60]. The well-organized gradients of different types of cells and ECM components were observed at the typical ligament-bone interface, thus mimicking the topography of the gradient interface is a potential solution [61]. In the meantime, hierarchical nano-to-macro structuring further potentiates osteogenesis. The hierarchical carbon nanotube (CNT) ligament developed by Wang et al. (10 nm fibers to micro/milli-assemblies) could activate TGF-β, WNT, MAPK, and integrin pathways [62], and a gradient electrospun scaffold with BMP-2/nanoHA fabricated by Jiang et al. [63] could upregulate OPN and OCN to facilitate osteogenesis. These findings collectively establish that spatially controlled nanoarchitectures, not chemical composition alone, dictate osteogenic efficacy.

Besides plausible biophysical advantages, biochemical cues are also important. Magnesium plays indispensable roles in musculoskeletal homeostasis [22], especially in coordinating bone regeneration through tripartite mechanisms: metabolic regulation, mineralization enhancement, and inflammation modulation [64]. Specifically, during the osteogenic differentiation of BMSCs, Mg²⁺ significantly facilitates the maturation stage from preosteoblasts to osteoblasts. It upregulates IGF-1 in preosteoblasts, activating the PI3K/AKT/mTOR signaling pathway to accelerate differentiation [65]. Concurrently, Mg²⁺ inhibits osteoclast differentiation by blocking calcium influx and suppressing the NF-κB/NFATc1 pathway [31]. In the meantime, Mg²⁺ promotes macrophage polarization toward pro-regenerative M2 phenotypes (Supplementary file S5) while attenuating pro-inflammatory M1 responses, thereby shifting the M2/M1 ratio toward tissue regeneration [66,67]. Furthermore, Mg²⁺ regulates bone integration by modulating the CGRP related pathway. Mg²⁺ permeates dorsal root ganglia via magnesium transporter 1, stimulating the release of CGRP from periosteal and endosteal nerve terminals [24]. The released CGRP then enhances osteogenesis by activating distinct pathways in different progenitor cells. In periosteum-derived stem cells (PDSCs), it initiates cAMP/CREB signaling that upregulates osterix expression [24]. In BMSCs, CGRP treatment upregulates the expression of its receptors (RAMP1 and CRL) [68,69] and activates the Wnt/β-catenin signaling pathway (elevating Lef1, Tcf7, cyclin D1, c-myc) [70], thereby stimulating the osteogenic process. On the other hand, elevating CGRP inhibits osteoclastogenesis via suppressing RANKL-induced activation of NF-κB when cultured with bone marrow macrophages [71], thereby inhibiting osteoclast differentiation and bone resorption. Moreover, a study showed that in the mice ACL reconstruction model, CGRP significantly promoted osteogenic differentiation of BMSCs via PKA/CREB/JUNB signaling pathway and elevated SHH (Sonic Hedgehog) expression in the meantime, thus enhancing the nerve regeneration and type-H vessel formation at the tendon-bone interface [72]. Recombinant spidroins demonstrate complementary bioactivity through integrin-mediated mechanisms: Widhe et al. documented FNcc-silk-supported keratinocyte adhesion via α5β1 integrin engagement [73,74], while Martín-Moldes et al. observed silk-silica chimeras triggering mesenchymal stem cell osteogenesis through αvβ3/MAPK/Runx2 activation [37]. Our in vitro analyses demonstrate that the PMRP bio-membrane orchestrates tripartite regenerative capabilities: Firstly, it establishes a favorable microenvironment for BMSCs adhesion and proliferation. Secondly, it significantly potentiates osteogenic differentiation of BMSCs. Thirdly, comprehensive gene expression profiling reveals upregulation of adhesion-related (Itga5, Itgav, and Opn), angiogenic (Vegfa), and osteogenic markers (Bmp2, Runx2, and Sox9), corroborating established biological mechanisms. The rSp alone enhanced Itga5 and Itgb1 expression—an effect magnesium further amplified without achieving statistical significance. Notably, PMRP significantly elevated expression of early (Bmp2) and intermediate (Sox9, Runx2), osteogenic genes alongside the angiogenic regulator Vegfa. These findings align with Cheng et al.'s observations that magnesium implants upregulate BMP2/VEGF/SOX9 cascades during fibrocartilaginous integration in rabbit ACL reconstruction [75], confirming in vitro osteogenic potential of PMRP bio-membrane and warranting further in vivo validation.

The in vivo results comparing the single-component groups (PRP vs. PMP) reveal distinct and complementary advantages: In the PRP group, ITGA5 expression was significantly upregulated at both week 3 and week 6 post-surgery, indicating that rSp promotes sustained cell adhesion efficacy over a longer duration. In the PMP group, CGRP and BMP2 were significantly higher at week 3, showing that MgO provides early osteogenic and neurogenic stimulation, but its adhesion-promoting effect is more transient. Thus, each single component has its own strength: rSp confers durable pro-adhesive properties, while MgO delivers rapid osteogenic and neural-related signals. Neither alone achieves the full spectrum of desirable early events. The PMRP group leverages both advantages. It promotes necessary cell adhesion during the early implantation phase and subsequently initiates robust osteogenic effects. This complementarity is the basis of synergy: the two components target different but inter-related biological processes, and their co-delivery yields a more comprehensive and timely healing response than either alone. Moreover, the early upregulation of ITGA5 in the PMRP group may underlie the concurrent early upregulation of BMP2, RUNX2, and CGRP, thereby accelerating new bone formation around the bone tunnels as early as 3 weeks postoperatively. In contrast, the PET group exhibited a comparatively slower increase in ITGA5, BMP2, RUNX2, and CGRP expression, accounting for the substantially delayed healing outcomes observed in this group. The biomechanical outcomes align consistently with the behavioral, radiological, and histological findings observed in vitro and in vivo. Specifically, the PMRP group demonstrated superior mechanical properties, exhibiting the highest mean failure load and stiffness values across all time points. This advantage likely stems from the capacity of PMRP graft to promote rapid cell adhesion and enhance osteogenic differentiation of BMSCs during the early postoperative phase. This accelerated osteogenesis was evidenced at postoperative week 3 by xylenol/calcein double labeling, which revealed the most rapid new bone formation in the PMRP group. Furthermore, robust Sharpey's-like fibers were formed early at the graft-bone interface in this group. In contrast, significant new bone formation observable by xylenol/calcein labeling, along with clearly discernible Sharpey's-like fibers at the interface, were not evident in the PET group until week 12 postoperatively. The results of gait analysis, a series of parameters to evaluate the behavioral changes postoperatively, showed minimal differences between groups at weeks 3 and 6, with the primary advantage appearing only at week 12. In this study, tissue repair and regeneration began at the microscopic level early after surgery, and it is only with the accumulation of time that more pronounced effects become evident in later-stage behavioral outcomes. Similarly, in degenerative cartilage lesions, patients typically exhibit microscopic histological changes in cartilage first, and only over time do alterations in gait behavior emerge, which aligns with the observations from our animal experiments.

Several limitations are associated with the current study. First, we did not compare the effect between the bioartificial ligaments and commercialized PET ligaments directly. The commercialized PET ligaments such as LARS ligament are not suitable for current little animal studies due to its human-based size. Further investigation will be conducted to directly compare commercialized artificial ligaments and the bioartificial ligaments using big animal models like goats with extended follow-up time such as 24 weeks. Second, our investigation focused solely on the regenerative status pertaining to graft-bone integration within the bone tunnels. The process of ligamentization within the intra-articular portion was not assessed in this study and will be evaluated using big animal models. Third, the results presented here primarily document phenotypic alterations within the ACL reconstruction model; the underlying mechanisms, such as whether and how Itga5 regulates Bmp2 or other osteogenic genes in a temporally and spatially specific manner, remain to be elucidated. Further experiments should be carried out to tackle these problems in future investigations.

5. Conclusion

In this proof-of-concept study, a dual core-shell structured bioartificial ligament was engineered and applied to promote the early graft-bone integration after ACL reconstruction. The refined biophysiochemical features of the ligament, including hierarchical structure that mimics the native graft-bone interface, the proficiency in facilitating cellular adhesion and osteogenic differentiation that MgO and rSp exert, as well as superior histological and functional recovery over pure PET graft, laid a solid foundation for the research and development of bioartificial ligament as well as for its potential clinical translation.

CRediT authorship contribution statement

Haozhi Zhang: Writing – original draft, Methodology, Investigation, Data curation, Conceptualization. Yuanming An: Writing – original draft, Investigation. Ying Luo: Writing – original draft, Methodology, Investigation. Lei Lei: Validation, Methodology. Zhenkang Wen: Methodology, Investigation. Xin Chen: Methodology, Investigation. Linlong Li: Methodology, Investigation. Michael Tim-Yun Ong: Writing – review & editing, Methodology. Lizhen Zheng: Methodology, Investigation. Wenxue Tong: Methodology, Investigation. Yu-Sheng Hsueh: Validation, Methodology. Bingyang Dai: Methodology, Investigation. Shian Zhang: Methodology, Investigation. Bruma Sai-Chuen Fu: Methodology, Investigation. Jiali Wang: Writing – review & editing, Supervision. Jiankun Xu: Writing – review & editing, Supervision, Methodology, Investigation. Patrick Shu-Hang Yung: Writing – review & editing, Supervision. Ling Qin: Writing – review & editing, Supervision.

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.

Ethics approval and consent to participate

All animal experiments were approved by the Animal Ethics and Experimentation Committee of The Chinese University of Hong Kong (No: 23-010-MIS).

Appendix A. Supplementary data

The following are the Supplementary data to this article: