Three-dimensional-printed shape-memory composite scaffold for meniscus repair

Abstract

Objective

The meniscus has poor intrinsic healing capacity, particularly in avascular regions. Meniscal injury is strongly associated with progressive knee dysfunction, chronic pain, and accelerated osteoarthritis development. Current treatments, such as allografts and partial meniscectomy, are limited by donor scarcity and secondary joint degeneration. Therefore, this study aimed to develop a regenerative meniscal scaffold with integrated biomechanical and biological functionality.

Methods

A biomimetic composite scaffold was fabricated via low-temperature three-dimensional printing using poly(D,L-lactic acid-co-trimethylene carbonate) and bacterial cellulose. Native meniscal architecture was reproduced using micro-computed tomography–based modeling, and interconnected porosity was designed to promote cell infiltration. Material integration employed dichloromethane as a shared solvent, enabling dual-ink (poly(D,L-lactic acid-co-trimethylene carbonate) + bacterial cellulose) co-printing. Scaffold performance was assessed using scanning electron microscopy morphology, porosity and water absorption assays, mechanical testing, Fourier-transform infrared spectroscopy, and in vitro cytocompatibility studies with rat bone mesenchymal stem cells.

Results

The scaffold exhibited a tensile modulus of 16.6 MPa and compressive modulus of 2.97 MPa, closely matching native meniscal mechanics. Porosity reached 63.57% ± 5.72%, supporting cell adhesion, while water absorption exceeded 138% after 7 days. Notably, the scaffold exhibited temperature-responsive shape memory behavior, allowing minimally invasive implantation and anatomic recovery at 37°C. Bone mesenchymal stem cells exhibited 95% viability (live/dead staining), significant proliferation (cell counting kit-8), and spontaneous chondrogenic differentiation (SRY-box transcription factor 9/collagen type II (SOX9/COLII) expression) without exogenous induction.

Conclusion

This three-dimensional-printed poly(D,L-lactic acid-co-trimethylene carbonate)/bacterial cellulose composite scaffold integrates biomimetic mechanics, shape-memory functionality, and pro-chondrogenic bioactivity, offering a promising strategy for meniscal regeneration. Further in vivo studies are warranted to confirm long-term efficacy and clinical translational potential.

Introduction

Although meniscus repair is particularly important in young and athletic populations to prevent osteoarthritis,^1–3 the vascularized area (red zone) comprises less than one-third of the meniscus, and only isolated lacerations confined to the margins may achieve autogenous repair through suturing.^1,4,5 In contrast, partial or total meniscal resection is often unavoidable owing to its low healing capacity (lack of blood flow) in the white and red-white transition zones.^4,6 Currently, allogeneic meniscus transplantation is a viable clinical solution;^7,8 however, it faces challenges such as donor scarcity and immune rejection. ⁹ In recent years, the use of tissue-engineering techniques to create meniscal scaffolds (MSs) for the repair and regeneration of injured menisci has become a popular research topic in orthopedics and sports medicine.

Three-dimensional (3D) printing enables personalized extracellular matrix (ECM)–mimetic scaffolds with uniform cell distribution.^7,10–13 This study aims to develop a novel MS using 3D printing to achieve breakthroughs in meniscal regeneration.

Poly(D, L-lactic acid-co-trimethylene carbonate) (P(DLLA-TMC)) combines shape memory with biocompatibility, making it ideal for use in MSs.^14–16 Furthermore, bacterial cellulose (BC) is known for its high crystallinity, high degree of polymerization, and consistent molecular orientation. It exists as single fibers and exhibits excellent cytocompatibility, providing an ideal environment for cell growth.^17–19

Currently, the most commonly used architectural design for bone scaffolds is the grid structure, characterized by its 0°/90° layering pattern, which provides uniform and highly consistent pore distribution. However, this structure often suffers from poor pore interconnectivity and permeability, negatively impacting continuous cell and tissue ingrowth, as well as stress concentration at the nodal intersections, which compromises its mechanical performance. ²⁰ In recent years, triply periodic minimal surface (TPMS) structures, such as the Gyroid, have attracted increasing attention as an advanced class of porous architectures. These structures are mathematically defined and exhibit excellent pore connectivity, high surface area, and superior permeability, better mimicking the natural bone environment to facilitate cell migration, proliferation, and differentiation. ²⁰ Moreover, the continuous curvature of TPMS geometries helps alleviate stress concentration and promotes more uniform stress distribution, thereby enhancing mechanical behavior.^20,21 Additive manufacturing (AM), particularly fused deposition modeling (FDM), enables the accurate fabrication of such complex TPMS scaffolds, making it a feasible strategy for producing structures with tailored functional and biological properties for bone regeneration applications. ²¹ Building on previous research, we combined P(DLLA-TMC) with BC to construct biomimetic MS using 3D printing. This scaffold had an ideal elastic modulus, excellent biocompatibility, and superior shape memory. These findings provide new insights into the development of simple MSs with excellent biomechanical properties. These findings may facilitate the development of novel therapeutic strategies for meniscal repair and regeneration.

Methods

Ethics approval

Experiments involving New Zealand (NZ) white rabbits were approved by the Shenzhen Peking University-HKUST Medical Center Institutional Animal Care and Use Committee (Approval No: [2022-239]) and were conducted according to the ethical requirements for animal experimentation and in compliance with established guidelines similar to the Helsinki Declaration II for human studies adapted for animal research.

Meniscal tissue was obtained from three euthanized, 3-month-old male New Zealand white rabbits. No live animals were used specifically for this study, thereby adhering to the Replacement and Reduction principles of the 3Rs (Replacement, Reduction, and Refinement) by repurposing existing tissue samples and avoiding additional animal use. Consequently, detailed individual animal weights were not recorded. Efforts to minimize suffering included the use of appropriate anesthesia during surgical procedures and administration of postoperative analgesia (Zoletil 50) as prescribed by the veterinary staff. Humane endpoints were strictly observed, and any animal showing signs of severe distress or complications related to the experimental objectives was humanely euthanized according to the approved protocol.

Materials and reagents

BC (4%) was provided by Professor Zhong Cheng (State Key Laboratory of Food Nutrition and Safety, Tianjin University of Science and Technology). P(DLLA-TMC) was obtained from Jinan Daigang Biological Co. Ltd. (Jinan, China). To prepare 3D-printing Ink I, 3 g of P(DLLA-TMC) was mixed with 6 mL of dichloromethane. After thoroughly mixing and shaking, the mixture was placed in a low-temperature 3D printer (Creality, Shenzhen, China) for further use. Subsequently, an equal proportion of 4% BC was added to Ink I and thoroughly mixed by shaking to create Ink II. After printing, the P(DLLA-TMC) MS was immediately infused with Ink II.

Scaffold design and fabrication

Fresh medial menisci were obtained from 3-month-old male NZ rabbits (Figure 1(a)). To precisely replicate the meniscal structure, the medial menisci of NZ rabbits were scanned using micro-computed tomography. A 3D meniscus model of the scaffold was designed based on the scan data using 3ds Max (Autodesk, San Francisco, CA, USA). The scaffold filaments were designed with a mesh-like structure (Figure 1(b)) to enhance structural strength and cell permeability. Subsequently, Ink I was loaded into a syringe equipped with a 0.21-mm nozzle for 3D printing. Furthermore, the 3D model of the scaffold was imported into dedicated slicing software, and the following printing parameters were set: layer thickness, 0.1 mm; infill spacing, 2.8 mm; and speed, 6 mm/s. This specific infill spacing was critically optimized to achieve a dual objective: it provided a biomimetic porous architecture that closely resembled the native meniscal tissue while simultaneously ensuring complete and robust infusion of the subsequent P(DLLA-TMC)/BC emulsion (Ink II). An excessive gap would compromise the structural integrity and biomimicry, whereas an insufficient spacing would impede the emulsion perfusion, leading to incomplete composite formation. The formulated ink was then loaded into a microextrusion 3D printer, and printing was initiated (Figure 1(c)). The scaffold dimensions were measured after printing using digital calipers (Figure 1(d)) to confirm compliance with the design specifications. Subsequently, the P(DLLA-TMC) scaffold was infused with Ink II and lyophilized by pre-freezing (−80°C, 2 h), followed by 24 h of ethanol sterilization (5%) and deionized water rinsing (10 × 5 min). After sterilization with Cobalt-60, the highly biomimetic MS was stored at room temperature until further use.

Figure 1.

Macroscopic and computerized three-dimensional (3D) reconstruction images of the tissue structure. Panel a depicts the natural morphology and structure of a fresh medial meniscus from the left knee of a New Zealand rabbit. Panel b depicts a 3D model of the rabbit medial meniscus reconstructed using 3D computed tomography scanning technology, accurately replicating the complex geometric morphology of the meniscus. Panel c depicts the mesh-like structure of the meniscus prepared from Poly(D, L-lactic acid-co-trimethylene carbonate) (P(DLLA-TMC)) using the low-temperature deposition 3D printing technique; this meniscus exhibits high biomimicry and structural strength. Panel d depicts the dimensional measurements of the printed mesh-like structure of the P(DLLA-TMC) meniscus, ensuring the precision of the scaffold and its alignment with the design requirements. Panel e depicts the biomimetic scaffold fabricated by infusing P(DLLA-TMC) mixed with 4% bacterial cellulose into the mesh-like P(DLLA- TMC)-printed meniscus structure, followed by lyophilization.

In vitro performance and cellular research

Scanning-electron-microscopy (SEM)–based characterization of scaffold morphology

First, each scaffold was secured using conductive adhesive tape in the top, coronal, and sagittal views to ensure stability and facilitate observation. The scaffolds were then sputter-coated with gold and observed under a scanning electron microscope.

Porosity

To assess the porosity of the scaffolds (21 × 7 × 3 mm), the initial mass of each scaffold (m0) was determined using a high-precision weighing device. The scaffolds were then fully immersed in double-distilled water and subjected to vacuum degassing until no bubbles were observed originating from the inside of the scaffolds, indicating that all pores inside the scaffolds were filled with water. After vacuum degassing, the scaffolds were removed and immediately weighed, and their masses were recorded as m1 (mass of the scaffold plus the water filling the pores). The scaffold porosity (%) was calculated as: [(m1–m0)/ρV] × 100%.

Water-absorption capacity

To assess the water absorption capacity of the scaffolds (10 ×10 × 6.5 mm), the m0 of each scaffold was determined. The scaffolds were then fully immersed in double-distilled water at a constant temperature of 37°C. After 1, 4, and 7 days of soaking, the scaffolds were removed, and water on the scaffold surface was gently absorbed with a towel. The scaffolds were immediately weighed to determine m1. The water absorption rate (%) of the scaffolds was calculated as: (m1 − m0)/m0 × 100%.

Mechanical properties

Uniaxial stretching was performed on scaffolds (40 × 6 × 2 mm) using an electronic universal material testing machine (5982; Instron, Norwood, MA, USA) at a speed of 10 mm/min until fracture. The stress–strain curve corresponding to the stretching process was plotted to calculate the tensile elastic modulus and strength. Similarly, to assess their compression performance, uniaxial compression was performed on scaffolds with dimensions of 4 × 3 × 6 mm at a speed of 1 mm/min. The stress–strain curve was plotted to calculate the compression modulus and strength.

Fourier-transform infrared (FTIR) spectroscopy

FTIR spectroscopy was used to analyze and integrate specific chemical groups into the scaffolds. A Bruker Tensor 27 FTIR spectrometer (Frontier; PerkinElmer) was used to perform the measurements in the reflection mode. All FTIR spectra had a resolution of 1 cm⁻¹ and a measurement range of 4000–500 cm⁻¹. Pure P(DLLA-TMC) and BC and the fabricated composite scaffolds were analyzed separately.

Deformability testing of the scaffolds

Tensile testing

The MSs were printed into strip-shaped scaffolds (length: 24.20 mm; thickness: 6.60 mm) and stretched uniformly under a constant tensile force in a 40°C water bath (DLM6; Xiaoyu Scientific Instruments Co., Ltd., Shanghai, China) until the water reached 20°C. Length was measured after the force was removed. Finally, the scaffolds were returned to 40°C to observe retraction. The length was measured again after complete retraction.

Compression testing

Identical scaffolds were uniformly compressed under constant pressure in a 40°C water bath until the water returned to 20°C. Following the removal of the pressure, the thickness was measured. Finally, the scaffolds were placed in the 40°C water bath to observe their shape recovery. The thickness was measured again after the complete shape recovery.

Material recovery capability testing

Identical scaffolds were stretched as described above. The length was recorded at 5-s intervals until the scaffolds returned to their initial length.

Live/dead staining

Rat bone mesenchymal stem cells (BMSCs) were cultured on the fabricated scaffold surfaces in 24-well plates for 24 h. The medium was then discarded, and the scaffolds were rinsed with phosphate-buffered saline (PBS) for 5 min. Subsequently, the cells were stained with live/dead staining solution (500 μL/well) in the dark at 20°C for 10 min (PerkinElmer, Waltham, MA, USA), followed by another wash with PBS for 5 min. Finally, the cells were observed under a fluorescence microscope (BDS400; Chongqing Aote Optical Instrument Co., Ltd., Chongqing, China; channels used: 488 and 561 nm), and images were captured.

Cell counting kit-8 (CCK-8) assay

A cytotoxicity assay was performed on MSs (2 × 2 × 1 mm) using rat BMSCs. Briefly, 10,000 cells were inoculated onto each scaffold on a tissue culture plate. On days 1, 3, 5, and 7, the medium was discarded, and the scaffolds were washed three times with PBS (5 min each). Subsequently, the cells were incubated with 200 μL of CCK-8 solution for 2 h. Next, 100 μL of the incubated solution was pipetted into 96-well plates, and the absorbance at 450 nm was measured using a multifunctional plate reader.

Chondrogenic differentiation assay

MSs inoculated with BMSCs were removed after 14 days of culture in normal medium (iCell Bioscience Inc., Shanghai, China). Scaffolds from another group—treated identically but supplemented with transforming growth factor-β1 and dexamethasone—were also removed after 14 days of culture. Scaffolds from both groups were blocked with bovine serum albumin (Thermo Fisher Scientific, Waltham, MA, USA) and probed with anti-SRY-box transcription factor 9 (SOX9) and anti-collagen type II (COLII) antibodies (Guangzhou Genxion Biotechnology Co., Ltd., Guangzhou, China). The samples were then stained with phalloidin (Guangzhou Genxion Biotechnology Co., Ltd.) and 4′,6-diamidino-2-phenylindole and observed at 40× magnification under an inverted fluorescence microscope (BDS400).

Statistical analysis

For all in vitro experiments, a sample size of n = 5 independent replicates per group was used to ensure robust statistical power. Quantitative data were expressed as mean ± SD. Statistical analyses were performed using SPSS 27.0 (IBM, Armonk, NY, USA), GraphPad Prism 9 (GraphPad, San Diego, CA, USA), and Origin (OriginLab, Northampton, MA, USA). For the quantitative analysis of fluorescence staining (immunofluorescence assays for SOX9/COLII), images from at least three independent fields of view per sample were processed and analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). The mean fluorescence intensity and/or the percentage of positive staining area for specific markers were quantified. Normally distributed data were expressed as mean ± standard deviation, and an independent-sample Student’s t-test was used for multiple-group comparisons. Statistical significance was defined as p < 0.05.

Results

SEM-based characterization of scaffold morphology

The internal morphological characteristics of the scaffolds are shown in Figure 2. SEM confirmed uniform pore distribution (600–700 µm) and fiber alignment (Figure 2). Panel a.2 shows the outcome after successful perfusion of Ink II into the P(DLLA-TMC) scaffolds (200×). Ink II exhibited a uniform and dense distribution within the pores retained in the scaffolds. Furthermore, the infusion process ensured the structural integrity of the scaffolds with no signs of damage or deformation. The SEM images confirmed uniform pore distribution and fiber alignment (Figure 2). These images present the macroscopic morphology of the scaffolds and reveal their internal microstructural characteristics, thereby facilitating an in-depth understanding of their performance and potential applications.

Figure 2.

Microstructure of the biomimetic meniscal scaffolds under electron microscopy. Panel a.1. Top view of the meniscal scaffolds printed with mesh-like pure Poly(D, L-lactic acid-co-trimethylene carbonate) (P(DLLA-TMC)) scaffolds under scanning electron microscopy (SEM) at 180× magnification. Panel a.2. Top view of the mesh-like pure P(DLLA-TMC) scaffolds infused with Ink II(P(DLLA-TMC) composite with bacterial cellulose) at 200×. Panels b.1–b.2. Coronal and sagittal views of the P(DLLA-TMC)/bacterial cellulose shape-memory composite meniscus scaffolds under SEM at 200×.

Porosity

The scaffold porosity reached 63.57% ±5.72%.

Water absorption

Water absorption rates were 120.54% ±14.1% (Day 1), 137.90% ± 12.5% (Day 4), and 138.58% ± 10.9% (Day 7).

Mechanical properties

Figure 3(a) shows the stress–strain curve of the tensile process collected at a speed of 10 mm/min until the scaffolds fractured. The stress–strain curve of the scaffolds remained nearly straight from the original state up to a strain of approximately 3.4%. During this phase, the stress was proportional to the strain, which is consistent with Hooke’s law, indicating that the material exhibited elastic behavior with an elastic modulus of 16.6 MPa. Tensile testing revealed the presence of elastic (3.4% strain) and plastic deformation phases (Figure 3(a)). During this phase, the stress increased slowly, whereas the strain increased rapidly, indicating the onset of plastic deformation and suggesting that residual deformation occurred once the external force was removed. At tensile strains of 20.5%–36.1%, the material resisted further deformation, and the stress continued to increase as the strain increased. Plastic deformation was predominant during this phase, indicating increased strength and hardness of the material and decreased plasticity and toughness. At a strain of approximately 36.1%, the material exhibits necking; the stress reaches its maximum value and subsequently decreases until the samples fracture. Because of the excellent elastic modulus of this scaffold material, intraoperative stretching allowed the implantation of the scaffolds through small incisions.

Figure 3.

Tensile and compressive stress–strain curves and infrared spectrograms. Panel a shows the tensile stress–strain curve at a speed of 10 mm/min. Panel b depicts the compressive stress–strain curve at a speed of 2 mm/min. Panel c depicts the infrared spectrograms of Poly(D, L-lactic acid-co-trimethylene carbonate) and bacterial cellulose.

Figure 3(b) shows the stress–strain curve during the compression process when the MSs were compressed at a rate of 2 mm/min. The scaffolds exhibited a nearly linear curve from their original state up to a strain of approximately 27.8% with the pressure gradually increasing from 0 to 0.827 MPa. The compression modulus of the scaffolds was 2.97 MPa. As the strain increased beyond 27.8%, the rate of increase in the stress gradually increased. Notably, after the strain reached 55.2%, the increase in the stress became more pronounced.

FTIR spectroscopy

The material structure was analyzed using FTIR spectroscopy. As shown in Figure 3(c), pure BC exhibited two characteristic peaks at 1078 and 1750 cm⁻¹, corresponding to the glycosidic bond and carbonyl group (C = O), respectively. In pure P(DLLA-TMC), two strong absorption peaks were observed at 1085 and 1743 cm⁻¹, corresponding to the carbon–carbon single bond (C-C) and carbonyl group (C = O), respectively. A carbon–carbon single bond appeared in both the stretching and bending vibration modes. Stretching vibrations involve periodic elongation and contraction of the atoms along the bond axis, whereas bending vibrations involve the movement of atoms perpendicular to the bond. The carbonyl group exhibited a stretching vibration mode. The spectra of the composite scaffolds exhibited characteristic absorption peaks for both the components; however, the shapes and intensities of the peaks changed. This indicates that the chemical structure of the composites contained both bonds and functional groups, and the positions and intensities of the absorption peaks were altered due to the interactions between the two components.

Deformability testing of the scaffolds

The results of tensile and compression tests are shown in Figure 4. Tensile and compression tests were conducted on the material.

Figure 4.

Deformability testing of the composite scaffolds. Panel a.1 reveals a scaffold length of 41.20 mm before stretching. Panel a.2 shows the stretching of the material to 51.00 mm. Panel a.3 shows a measured length of 41.48 mm after the material recovered from deformation. Panel b.1 shows the thickness of 6.60 mm before compression. Panel b.2 shows the compression of the material to 3.62 mm. Panel b.3 shows the measured thickness of 5.66 mm after the material recovered from deformation. Panel c shows the length of the material measured every 5 s after stretching until, the initial length was restored after 35 s.

Material recovery capability testing

Identical scaffolds were stretched using the same method, and their lengths were recorded at 5-s intervals.

The results showed that the P(DLLA-TMC)/BC composite material fabricated via low-temperature deposition printing maintained excellent tensile and compressive resistances, further demonstrating the superior shape-memory properties of this material.

CCK-8 assay

A CCK-8 assay was used to further investigate the cell growth characteristics of the scaffolds. The results showed optical density values of 0.20 ± 0.06 (Day 1), 0.50 ± 0.09 (Day 4), and 2.50 ± 0.2 (Day 7), indicating substantial cell proliferation. This demonstrates that the MSs did not exhibit cytotoxicity.

Live/dead staining

The approximate ratios and distributions of live and dead cells in the samples are shown in Figure 5. Approximately 95% of the cultured cells were stained green (live cells), whereas approximately 5% were stained red (dead cells; Figure 5(b)). These results demonstrate that the material was not toxic to the cells.

Figure 5.

Results of the cell counting kit-8 (CCK-8) assay and live/dead staining assay. Panel a shows the results of the CCK-8 assay. Panels b and c depict the results of the live/dead staining assay. In Panel b, live cells are observed with green fluorescence, and in Panel c, dead cells are observed with red fluorescence.

Chondrogenic differentiation assay

After 14 days of culture in normal medium, confocal laser scanning microscopy at 40× magnification revealed that some cells on the scaffold surfaces exhibited round morphology or were aggregated into clusters. Additionally, all cells that aggregated into clusters exhibited higher expression levels of SOX9, a chondrogenic differentiation marker, whereas the spread-out cells exhibited lower expression levels, indicating enhanced chondrogenic differentiation in the aggregated stem cells. Aggregated rat BMSCs cultured in normal medium exhibited high SOX9 expression levels, indicating that these cells differentiated into chondrocytes (Figure 6(a.1) to (a.4)). In contrast, after culturing BMSCs in the chondrogenic differentiation induction medium, chondrogenic differentiation was observed in both aggregated and spread-out cells (Figure 6(b.1) to (b.4)). The spread-out cells also expressed SOX9 in response to chondrogenic differentiation, suggesting the presence of fibrochondrocytes. The aggregated cells within the square areas in the figures indicate hyaline chondrocytes. Additionally, SOX9 expression and chondrogenic differentiation were enhanced upon cell clustering. Similar conclusions were drawn from the results presented in Figure 6(c.1) to (c.4) and Figure 6(d.1) to (d.4). COL2 expression was also observed in cells grown in the normal medium (cells within the dashed circles in Figure 6(c.1) to (c.4)), demonstrating that some of the clustered cells in these areas differentiated into chondrocytes. Chondrogenic differentiation was observed in cells with varying morphologies (Figure 6(d.1) to (d.4)). Certain cells (those within the white circles) exhibited low F-actin expression but high COL2 expression. These results suggest that during induced chondrogenic differentiation, lower F-actin and higher COL2 expression levels are key characteristics of chondrocyte differentiation and subsequent maturation.

Figure 6.

Images of cells in the chondrogenic differentiation assay at 40× magnification. Panels a.1–a.4 and c.1–c.4 depict the expression of SRY-box transcription factor 9 (SOX9) and collagen type II (COLII), respectively in rat bone mesenchymal stem cells (BMSCs) cultured in normal media. Panels b.1–b.4 and d.1–d.4 depict the expression of SOX9 and COL2, respectively, in rat BMSCs cultured in chondrogenic Continued.differentiation media. Good cell growth on this material was observed in normal media, with some clustered/aggregated cells showing signs of cartilage formation. Upon exposure to chondrogenic differentiation inducers, both spread-out cells and aggregated/clustered cells exhibited signs of chondrogenic differentiation. Quantitative analysis (b–c) confirmed this observation, showing a significant upregulation of both chondrogenic markers in the differentiation group (COL2 fluorescence density: 3.110 ± 0.82 vs. 1103 ± 82.69; SOX9 fluorescence density: 34.62 ± 5.63 vs. 675.00 ± 31.35; both p < 0.01, Student’s t-test). However, enhanced cell aggregation was associated with superior differentiation into hyaline cartilage cells, while spread-out cells showed a tendency to differentiate into fibrocartilage cells.

Discussion

Tissue engineering primarily involves the use of growth factors, seed cells, and tissue-engineered scaffolds, with scaffold materials serving as a foundation for tissue repair and regeneration. Sources of MS include natural materials such as silk fibroin, ²² BC, ²¹ and collagen. ²³ Compared with CMI® and Atifit®, our scaffold offers superior mechanical strength and chondrogenic potential.^24–26

Owing to the unique anatomical structure of the meniscus, ²⁷ 3D printing offers significant advantages over other tissue engineering techniques (e.g. freeze–drying, particulate leaching, and electrostatic spinning) ²⁸ for the fabrication of MSs. It provides precise control and high reproducibility in terms of accuracy, pore modulation, spatial structural complexity, and personalized customization, enabling better mimicking of the shape and structure of the natural meniscus.^29–31 Building on this method, we employed composite printing using a combination of Ink I (P(DLLA-TMC) dissolved in dichloromethane) and Ink II (an oil-based mixture created by adding equal proportions of BC to Ink I). After the printing process was completed, the shared solvent, dichloromethane, facilitated mutual integration and inseparable bonding between Ink I and Ink II.

Cell-free scaffolds are preferred for meniscal transplantation. First, these scaffold materials offer straightforward fabrication, avoiding the costs and risks associated with cell manipulation such as bacterial contamination and phenotype loss. In addition, ethical and regulatory constraints related to the use of cells have hindered the translation of these products into clinical practice. ⁹ However, “cell-free” scaffolds do not imply the complete absence of cells. These scaffolds require cell infiltration to produce ECM and enhance their regenerative potential. ³²

Preliminary studies have suggested that BC possesses distinct characteristics such as porosity, strong water absorption, and air permeability. Additionally, it demonstrates high crystallinity, polymerization, and consistent molecular orientation. However, the mechanical properties of BC do not meet the practical requirements for standalone 3D printing materials. Based on these findings, we employed a low-temperature deposition 3D printing technique to fabricate P(DLLA-TMC)/BC composite scaffolds, in which P(DLLA-TMC) provided sufficient mechanical support and shape memory. This approach has been widely employed in orthopedic materials, ³³ as BC has been shown to enhance the cell adhesion properties of scaffolds.

Lee et al. ³⁴ noted that P(DLLA-TMC) scaffolds enhance collagen synthesis, tissue elasticity, tissue regeneration, osteoblast adhesion and proliferation, alkaline phosphatase expression levels, and mineral deposition. Although pure P(DLLA-TMC) scaffolds exhibit a porosity of approximately 70%, ³³ their smooth surfaces are not conducive to cell adhesion, which affects tissue remodeling in vivo. We incorporated BC with higher porosity into pure P(DLLA-TMC) scaffolds, making the scaffolds more favorable for meniscal cell growth. The porosity of our scaffolds reached 63.57%, which met the experimental design requirements.

Komatsu et al. ¹⁴ used P(DLLA-TMC) as the sole raw material to fabricate MSs via 3D printing for tissue engineering. They observed that pure P (DLLA-TMC) scaffolds exhibited a gradual decline in thermal stability, mechanical properties, and molecular weight over time. Therefore, to exploit the excellent biomechanical properties and temperature-dependent shape memory of P(DLLA-TMC), along with properties such as high porosity, crystallinity, polymerization, and consistent molecular orientation of BC, we combined the two materials to fabricate MSs via 3D printing.

We adopted personalized 3D printing to address the variance among different individuals. Low-temperature deposition enables the formation of interconnected micropores that enhance cell signaling.^35,36 More importantly, shape-memory materials facilitate minimally invasive implantation. These materials allow scaffolds to be delivered into the knee-joint cavity through small incisions, and their original shape can be restored via temperature modulation. Murphy et al. ³⁷ noted that accurately placing scaffolds at defect sites using arthroscopic approaches is challenging and that significant technical challenges are encountered during steps such as trimming, suturing, and securing and fixing synthetic implants. As the P(DLLA-TMC)/BC composite MS is a biomaterial with shape memory, we predicted that in subsequent animal experiments, biomimetic MSs printed in a personalized manner could be stretched into strips in a 40°C water bath, followed by the application of a constant tensile force until the water temperature returned to room temperature, to help transform them into strip-shaped materials, thus allowing them to be implanted through small incisions. At body temperature (approximately 37°C), the biomimetic scaffold deformed to its original shape after implantation, which aligned with our expectations.

The primary function of the meniscus is to distribute vertical pressure on the knee joint, thereby reducing wear and tear of the articular cartilage. Stocco et al. ⁹ noted that the polymerization modulus of a natural meniscus, which enables it to withstand axial compression forces, ranges from 100 to 150 kPa. This property allows the meniscus to bear a load up to five times the body weight of the knee. The polymerization modulus of our bio-scaffold reached 2.97 MPa, allowing the knee joint to bear a load of almost 100 times the body weight and dispersing pressure more effectively, reducing wear and tear on the articular cartilage. This significantly satisfies the clinical needs.

In the field of meniscal tissue engineering, achieving mechanical properties comparable to those of native tissue while avoiding stress shielding effects represents a core design objective. Excessively high compressive modulus and an inappropriate difference in tensile modulus are among the fundamental causes of stress shielding. Selecting implant materials that match the modulus of bone is key to mitigating this effect and promoting normal bone healing and long-term health. Our scaffold exhibited a tensile modulus of 16.6 MPa and compressive modulus of 2.97 MPa, closely approximating native meniscal mechanics, which is a design objective that aligns with prevailing research data aimed at promoting physiological load-sharing and integration.^38,39 This mechanical compatibility stems from the following strategy: polycaprolactone provides initial mechanical support, while the bioinspired hydrogel replicates the regional anisotropy of the natural meniscus.

Notably, although the scaffold’s initial compressive modulus is relatively high, its biodegradable nature ensures that short-term mechanical support does not lead to long-term stress shielding. Vakili et al. ³⁹ further validated this through degradation experiments: polyvinyl alcohol/chitosan/MgCl₂ scaffolds showed a weight loss of up to 49% after immersion in phosphate buffer solution for 4 weeks, indicating that load transfer to newly formed tissue occurs progressively, promoting physiological load-sharing. Moreover, Sun et al. ⁴⁰ demonstrated that functional regeneration requires biomechanical anisotropy rather than simply high modulus, and their composite strategy successfully replicated native zone-specific properties while preventing stress-shielding.

Similarly, the scaffold demonstrated good biological properties, as chondrocytes were detected even without chondrogenic differentiation inducers (after in vitro implantation of BMSCs on the scaffold). Asami et al. ⁴¹ reported that pure P(DLLA-TMC) exhibited biocompatibility, primarily by promoting osteoblast proliferation and differentiation. In our MSs, direct inoculation with BMSCs resulted in favorable cell growth, with cells either spreading or aggregating into round shapes. Among these cells, the aggregated spherical stem cells exhibited higher COL2 and SOX9 expression levels, indicating a higher propensity for chondrogenic differentiation in the corresponding areas. When the BMSCs were inoculated after the scaffolds were exposed to chondrogenic differentiation inducers, the cells that aggregated into clusters exhibited higher SOX9 and COL2 expression levels. However, cells at the same location exhibited lower F-actin expression, indicating that the aggregated cells differentiated into hyaline cartilage cells. The spread-out cells also expressed SOX9 and COL2, suggesting that these cells differentiated into fibrochondrocyte-like cells. High F-actin and low SOX9 expression levels were observed, indicating that the spreading cells had differentiated into fibrocartilage cells. This finding further demonstrated that aggregated spherical stem cells tend to differentiate into cartilage cells. These results verified the biocompatibility of the composite scaffolds and their ability to promote stem cell differentiation into cartilage cells.

For improving integration with the host tissue at the suture interface, an additional strategy could be the in situ formation of a polydopamine (PDA) coating at the scaffold–host meniscus junction. Inspired by its strong wet-adhesion properties, PDA can serve as a bioadhesive to enhance binding strength, supplementing or even reducing the reliance on sutures. This approach may improve surgical handling, minimize micromotion, and promote seamless biological integration, thereby supporting long-term functional restoration.

Notwithstanding these promising in vitro results, we acknowledge that this study represents a foundational investigation. To further advance this scaffold toward clinical application, several critical lines of inquiry are planned as immediate future work. First, comprehensive in vivo evaluations in established animal models of meniscal injury are essential for rigorously assessing the scaffold’s functionality, biocompatibility, degradation kinetics, and, most importantly, its efficacy in promoting meniscal regeneration and protecting articular cartilage in a physiological environment. Second, a more profound understanding of the scaffold’s mechanical performance under dynamic conditions is required. Therefore, future research should focus on evaluating the dynamic viscoelastic and fatigue properties of the scaffold through methods such as dynamic mechanical analysis (DMA) to simulate physiological loading conditions. Furthermore, the rheological properties of the precursor inks can be characterized to refine the printability and mechanical performance of the scaffolds. Third, although immunofluorescence staining provided crucial spatial protein distribution data, a deeper investigation into the specific molecular mechanisms is warranted. Quantitative techniques such as reverse transcription–quantitative polymerase chain reaction and western blot analysis should be employed to thoroughly validate the chondrogenic differentiation at both the genetic (mRNA) and protein levels, providing complementary data to elucidate the signaling pathways through which our scaffold promotes chondrogenesis.

Conclusions

We combined P(DLLA-TMC) and BC using composite 3D printing to create novel scaffolds with high mechanical strength and bioactivity. P(DLLA-TMC) exhibits shape memory and a high polymerization modulus, which significantly enhanced the loading capacity of the knee joint, whereas BC promoted fibrocartilage differentiation by improving cell adhesion and migration because of its high porosity. The low-temperature deposition 3D printing technique allows for personalized customization and the development of precise biomimetics of complex structures. In addition, minimally invasive implants simplify surgical procedures because of their temperature-responsive properties. In vitro experiments demonstrated that the scaffold supported the differentiation of BMSCs into cartilage cells without the need for inducers and that the novel scaffold guided fibrocartilage formation by modulating cell morphology. However, the in vivo degradation kinetics, long-term biomechanical stability, and clinical suitability of these scaffolds require further validation. In future, combining dynamic mechanical stimulation with multifactorial synergistic modulation is essential for advancing meniscal regeneration from structural to functional biomimicry.