The treatment of meniscus injuries has recently shifted toward the field of tissue engineering (TE). In this work, bovine menisci were characterized, and the regionally-dependent mechanical properties were analyzed. Three-dimensional (3D) printing technology was employed to produce a scaffold that mimicked the mechanical properties of the meniscus. A polycaprolactone (PCL) meniscus scaffold was 3D printed, allowing for the deposition of fibers mimicking the internal architecture of the native meniscus, while achieving regional and variable mechanical stiffness, varying from 2.74 to 0.88 MPa. The PCL scaffold was infiltrated with extracellular matrix (ECM)-like hydrogels composed of gelatin methacrylate (GelMA) and glycosaminoglycans (GAGs), such as hyaluronic acid (HA) and chondroitin sulfate (CS), and subsequently freeze-dried. Human mesenchymal stem cells were seeded onto the scaffolds, and the infiltrated cells were observed to produce ECM components of the native meniscus. Collagen and GAGs production was successfully established. The synthesis of a new matrix reportedly enhances the mechanical properties of the hydrogel over time. Additionally, the circumferential PCL fibers within the scaffold guided the newly synthesized matrix, facilitating replication of the native tissue structure. These results indicate that the ECM-infiltrated 3D-printed PCL scaffold is a promising solution for meniscus repair.

1. Introduction

The menisci are fibrocartilage structures found within the knee joint, which are susceptible to injury due to their location and the extreme forces they endure (up to 24× body weight during high-impact activities), with the meniscus bearing up to 50–70% of these joint loads.¹ Due to their limited vascularity and poor healing potential, meniscectomy is a common treatment for meniscus injury.² However, this treatment leads to a loss in chondroprotective function, which can result in articular cartilage degeneration and the development of osteoarthritis over time.³ Therefore, there is an urgent need to develop scaffolds to adequately support tissue regeneration in damaged menisci. Designing a structural and hierarchical scaffold that mimics the architectural and mechanical features of the native meniscus for supporting tissue regeneration is a significant challenge for researchers, owing to the forces and mechanical demands the native meniscus endures. Along with the load-bearing capabilities of the meniscus, its unique structure is essential to maintain congruency between the femoral condyles and tibial plateau.⁴ The meniscal collagen structure is essential for reinforcing the tissue to withstand the loads experienced across the knee joint, as it transfers forces between the femoral and tibial joint surfaces by developing hoop stress. Radial fibers also play a significant role, as they provide resistance to the lateral separation of circumferential fiber bundles when an axial load is applied.⁵ Therefore, the development of a biomimetic scaffold that replicates the specific threedimensional (3D) microstructure of the meniscus is crucial for effective meniscus tissue engineering.

Research has already been conducted on creating anatomically shaped meniscus scaffolds^6,7 and replicating the internal collagen structure of the native meniscus.^4,6,8 However, there has been limited work on developing regional variations within the scaffold to match the heterogeneous mechanical properties of the native meniscus.

Synthetic materials have been studied to develop a structural meniscus scaffold with appropriate mechanical properties, including polyurethane (PU),⁹ polycarbonate urethane (PCU),¹⁰ poly-L-lactic acid (PLLA),¹¹ and polycaprolactone (PCL).¹² Among these, PCL is one of the most promising materials for meniscus tissue engineering. It plays an important role in matrix organization, augments matrix content, and possesses unique physical strength suitable for applications in hard tissue engineering, such as the meniscus or bone.^13,14 Furthermore, PCL is an attractive material due to its approval by the United States (US) Food and Drug Administration (FDA).⁴ Although synthetic PCL is biocompatible, it does not readily promote cell proliferation. Therefore, in addition to using PCL for the structural framework of the meniscus scaffold, a natural material should be incorporated into the PCL framework to promote cell growth, proliferation, and differentiation, and ultimately facilitate new tissue formation. Hydrogels are widely used in tissue engineering to promote cell infiltration and proliferation.^15,16 The use of hydrogels supports nutrient diffusion and can provide adhesion sites and signaling cues that guide cell growth and the formation of desired tissue.¹⁷ Hydrogels can also be 3D printed with encapsulated cells, known as a bioink, or used as acellular biomaterial inks, where cell attachment and proliferation occur within the scaffold after implantation.¹⁸ Typical hydrogels used in meniscus regeneration include alginate,¹⁹ collagen,²⁰ gelatin,²¹ and gelatin methacrylate (GelMA).²² GelMA scaffolds have gained attention in recent years for tissue engineering applications due to their retention of biological properties, such as cell adhesion domains and enzymatic degradability.^23,24 Moreover, GelMA is easily crosslinked, with potential in cartilage tissue engineering.^23,25

However, currently used biomaterial inks in meniscus tissue engineering typically lack the major compositional units of the native extracellular matrix (ECM), such as glycosaminoglycans (GAGs). GAGs, such as hyaluronic acid (HA) and chondroitin sulfate (CS), are abundant components of the ECM in cartilage and meniscus tissue, participating in numerous biological processes.²⁶ The inclusion of CS in a scaffold may promote chondrogenesis and enhance its mechanical properties, displaying promising results for cartilage tissue engineering.²⁷ In contrast, HA is involved in many critical biological functions, such as regulating cell adhesion and influencing cell proliferation and differentiation.¹⁵

This study has two main objectives. Firstly, we aim to create a novel biomimetic PCL meniscus scaffold for partial tissue regeneration following a traumatic injury. The scaffold has an internal structure inspired by the native meniscus, consisting of circumferential and radial fibers. Furthermore, the unique structure of the scaffold varies by region to match the distinct region-dependent mechanical properties of the native tissue. Secondly, we investigated various GelMA/GAG composite hydrogels to develop a biomaterial ink that mimics the natural ECM of the meniscus and infiltrates into the PCL framework, followed by comprehensive in vitro analysis.

2. Materials and methods

2.1. Materials

All PCL scaffolds were printed by fiber deposition of molten PCL (molecular weight [MW]: 50 kD; Capa 6500D; Perstorp, Sweden), using a pneumatic-based BioBots bioprinter (Allevi, United States of America [USA]). GelMA and the photoinitiator, lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP), were purchased from Allevi (USA). GelMA was derived from Type A, 300 Bloom, porcine skin and fabricated to achieve a final degree of methacrylation of 50%. GAGs used to create the composite biomaterials inks included HA (MW: 0.1 MDa; Shanghai Easier Industrial Development, China), methacrylated HA (HAMA; MW: 0.1 MDa; Advanced Biomatrix, USA), and CS derived from shark cartilage (C4384; Sigma-Aldrich, USA).

2.2. Scaffold design

To develop an appropriate 3D architecture for the meniscus scaffold, 10 × 10 × 2 mm porous PCL scaffolds were 3D printed with varying fiber spacing and offsets. Two different fiber spacings (1 and 2 mm) and three different offsets (0 offset, 1 offset, 2 offsets) were investigated in this study (Figure 1). The printed scaffolds had fibers alternating 90° between each layer. The effect of doubleprinting layers was also investigated, whereby the same fiber orientation was printed twice and then alternated 90°. Fiber spacing was defined as the distance between the edges of each strand. Scaffold formation was characterized using an optical microscope (Olympus, Japan).

2.3. Porosity of the printed scaffolds

The theoretical porosity of the scaffolds was calculated using a cell within the scaffold, employing the following geometry-based equation derived from the void space between fibers and their cross-sectional area:

where V_cell is the volume of one cell and V_fiber is the solid fibers in the cell; L is the length and W is the width of the cell; and D is the diameter of the fiber and is twice the height of the cell.

2.4. Compression testing

To develop a scaffold with a heterogeneous structure that mimics the regional-dependent mechanical properties of the native tissue, a standard mechanical testing criterion was applied for scaffold designs and bovine meniscus samples. Mechanical characterization was performed using uniaxial, unconfined compression testing. Firstly, to determine the regional-dependent mechanical properties of the native tissue, the menisci were sectioned into various regions, including three layers (femoral, middle, and tibial) and two regions (inner and peripheral), as previously outlined.²⁸ A 6-mm punch (Sigma Aldrich, USA) was used to obtain cylindrical discs from the center of each region, and a micrometer (VWR, USA) was used to determine the thickness. To evaluate the influence of fiber architecture on the 3D-printed PCL scaffolds, various designs (illustrated in Figure 1) were printed and mechanically characterized. The height was recorded using a micrometer, and a 9-mm punch was used to obtain cylindrical discs from the center of the scaffolds. For both meniscus and scaffold samples, compression data were obtained using a Tinius Olsen testing machine (H25KS; Tinius Olsen, USA), equipped with a 100 N load cell. The samples were preloaded with a load of 0.5 N to ensure full contact with the sample. The samples were tested at a strain rate of 0.5 mm/min until 20 N was reached. The change in thickness and increase in load were monitored. The thickness and load measurements were converted to stress–strain data using the following equations:

where ε_s is the strain on the scaffolds, Δh is the change in height recorded by the software, and h_o is the original height of the scaffold; δ_s is the stress, F is the load cell force reading, and πr² is the initial cross-sectional area of the scaffold, where r is the radius. From the stress and strain data, the compressive modulus (E) was obtained from the slope of the linear region of the curve.

2.5. Hybrid scaffold fabrication

In this study, three ECM scaffold compositions were produced: PCL+GelMA, PCL+GelMA/CS/HA, and PCL+GelMA/CS/HAMA. The PCL framework was 3D-printed (MW: 50 kD; Capa 6500D; Perstorp, Sweden) using a pneumatic-based BioBots bioprinter (Allevi, USA) with print parameters as outlined in Table 1. For the cell study, the PCL scaffolds were simplified. The PCL was 3D-printed with 1.5 mm fiber spacing, without offsets, and the fibers were printed straight rather than in the circumferential orientation. The scaffold dimensions were 5.3 × 5.3 × 2 mm and 10 layers in height.

Table 1. Polycaprolactone (PCL) printing parameters

Once the PCL framework was printed, it was infiltrated with ECM-like hydrogels. Three different hydrogel formulations were investigated in combination with PCL to determine which variation was most effective for meniscus regeneration. The combinations studied included GelMA, GelMA/CS/HA, and GelMA/CS/HAMA, with the concentration of each material outlined in Table 2. Briefly, CS, HA, and HAMA were dissolved in phosphate-buffered saline (PBS; Sigma Aldrich, Ireland) overnight at 25°C. LAP and GelMA (Allevi, USA) were dissolved in PBS at 60°C. Once GelMA was dissolved, it was added to CS/HA and CS/HAMA solutions to reach a final concentration (indicated in Table 2). While the solution was still warm, it was pipetted into 15 mL centrifuge tubes and vortexed for 1 min, followed by centrifugation at 1000 rpm for 3 min to remove air bubbles. The PCL scaffold was placed into a 6 × 6 mm Teflon mold, and 90 µL of each ECM-like hydrogel solution (37°C) was injected into the PCL scaffold between the fibers and allowed to slightly overfill the mold. The hydrogels were crosslinked under ultraviolet (UV) light (wavelength: 365 nm; EA-160/FBE; Spectroline, USA) for 15 min.

Table 2. Hydrogel compositions investigated to develop an extracellular matrix (ECM)-like material

Abbreviations: CS: Chondroitin sulfate; GelMA: Gelatin methacryloyl; HA: Hyaluronic acid; HAMA: Methacrylated hyaluronic acid.

2.6. Freeze-drying process optimization

To determine the effects of pre-freeze temperature on pore structure, three pre-temperatures were investigated, as follows: (i) freezing at a rate of 1°C/min to −20°C for 2.5 h, then transferred to -80°C; (ii) freezing at a rate of 1°C/min to −80°C; and (iii) snap-freezing using liquid nitrogen. The ECM/PCL hybrid hydrogels were developed as outlined above. Once the samples were crosslinked, each construct was placed in a 1.5 mL microtube and placed in a Mr. Frosty^TM (Thermo Scientific, USA) freezing container to achieve a freeze rate of 1°C/min. Mr. Frosty^TM was either placed in a (i) −80°C freezer or (ii) −20°C freezer for 2.5 h and then transferred to −80°C. Alternatively, samples were placed in a cryovial and snap-frozen in liquid nitrogen followed by storage at −80°C. Samples were then freeze-dried using an adapted method reported by Murphy et al.²⁹ in an attempt to increase pore size. The freeze dryer (Labconco, USA) was set to −10°C, and prefrozen constructs were placed into the freeze dryer for 3 h. This is followed by annealing, where the temperature was dropped to −30°C for 1 h, then increased to a drying temperature of −10°C, and held for 18 h under vacuum (0.2 mBar). Finally, the temperature was increased to 25°C for 2 h, while remaining under vacuum.

2.7. Morphological characterization

To create samples for morphological analysis, each hydrogel precursor composition was injected into a 6 × 2 mm Teflon mold, and a glass cover slip was placed on top and crosslinked under UV light (wavelength: 365 nm; EA-160/FBE; Spectroline, USA) for 15 min. Samples were subsequently freeze-dried as previously outlined. Constructs were cross-sectioned using a scalpel, sputter coated with gold (K550; Emitech, France), and imaged using scanning electron microscopy (SEM; TM-1000; Hitachi, Japan). The SEM images were analyzed using Image J software, and 10 pores were averaged to determine the pore size.

2.8. Cell seeding and culture

Human mesenchymal stem cells (hMSCs) derived from bone marrow (C-12974; PromoCell, Germany) were used in this study (P3) and cultured at 37°C in 5% CO₂. Cells were expanded in high-glucose Dulbecco’s Modified Eagle Medium (hgDMEM; GlutaMAX^TM; Biosciences, USA), supplemented with 10% fetal bovine serum (FBS; Biosciences, USA), penicillin (100 U/mL)–streptomycin (100 μg/mL) (Gibco, Ireland), and 0.25 μg/mL amphotericin B (Sigma-Aldrich, USA). Cells were cultured until they reached confluency, with the media being changed every 2–3 days. Prior to cell seeding, scaffolds were sterilized using ethylene oxide (EtO), followed by aeration for 4 days to ensure complete removal of any EtO trapped in the scaffold. Directly before cell seeding, three pinholes were created on the scaffold surface to facilitate cell infiltration, as a “skin” was observed. Scaffolds were placed in a 24-well plate, whereby each well bottom was first filled with 0.5 mL of 3% agarose solution to prevent cells from attaching to the well (Sigma-Aldrich, USA). At the time of seeding, scaffolds were placed in the well and seeded with a cell density of 5 × 10⁵ cells. Once scaffolds were seeded, cells were allowed to attach for 30 min, followed by the addition of a chemically defined chondrogenic medium (CDM). The CDM consisted of DMEM-GlutaMAX, supplemented with penicillin (100 U/mL)–streptomycin (100 μg/mL), 0.25 µg/mL Amphotericin B, 100 μg/mL sodium pyruvate, 40 μg/mL L-proline, 50 μg/mL L-ascorbic acid-2-phosphate, 1.5 mg/mL bovine serum albumin (BSA), 1 × insulin-transferrin-selenium, 100 nM dexamethasone, 4.7 µg/mL linoleic acid (all from Sigma-Aldrich, USA), 10% FBS (Biosciences, USA ), and 10 ng/mL recombinant human transforming growth factor-β3 (TGF-β3; R&D Systems, USA). Scaffolds were maintained in CDM for 21 days under hypoxia conditions (5% CO₂; 5% O₂; 37°C). A complete media change was performed twice a week.

2.9. Live/Dead assay

Cell survival and distribution were assessed using live/dead assay. Constructs were cross-sectioned, washed with PBS, and transferred to a 24-well plate. Thereafter, 20 µL ethidium homodimer-1 (EthD; BT40014; Biotium, USA) and 5 µL calcein (BT80011-1; Biotium, USA) were added to 10 mL PBS and vortexed for 3 min. Live/Dead solution (0.5 mL) was added to each well and incubated for 1 h. The solution was then removed, and samples were washed 3 times with 0.5 mL PBS. Samples were imaged using a confocal microscope (FV-1000 Point Scanning Microscope; Olympus, Japan) at the following excitation/emission wavelengths: calcein: 495 nm/515 nm; EthD: 495 nm/635 nm.

2.10. Actin/DAPI staining

Cell morphology was assessed using actin/DAPI staining. To evaluate cell morphology, the phalloidin conjugate was used to stain actin filaments, and DAPI (4’,6-diamidino-2-phenylindole dihydrochloride) was used to stain cell nuclei. Samples were fixed by immersing in 4% paraformaldehyde (PFA) and incubating overnight at 4°C. Prior to staining, the scaffolds were cross-sectioned and permeabilized in 0.5% Triton-X for 20 min at room temperature. For actin staining, scaffold cross-sectional slices were incubated with the fluorescent agent rhodamine-conjugated phalloidin (dilution 1:40; VWR, USA) for 1 h, followed by incubation with DAPI (dilution 1:50; VWR) for 10 min under light protection. Samples were imaged using a confocal microscope (FV-1000 Point Scanning Microscope; Olympus, Japan) at the following absorption/emission wavelengths: rhodamine-phalloidin: 540 nm/565 nm; DAPI: 358 nm/461 nm.

2.11. Histology and immunohistochemistry

Histology and immunohistochemistry analysis were performed for day 1 (D1) and day 21 (D21) constructs to confirm fibro-chondrogenic differentiation of the seeded cells. D1 samples display the background staining of the hydrogel and serve as a reference. Constructs were washed with PBS and then fixed by immersing in 4% PFA and incubating overnight at 4°C. Samples were removed from PFA, washed, and stored in PBS at 4°C. At the time of the analysis, samples were dehydrated using a series of ethanol (30%, 50%, 70%, 80%, 90%, 100%, xylene) and wax embedded thereafter. Paraffin wax-embedded constructs were sectioned using a microtome (Leica, Germany) to produce 6-µm-thick slices and mounted on microscope slides. Samples were then deparaffinized and rehydrated using a guided series of xylene and alcohol baths. New tissue formation was determined by staining with hematoxylin and eosin (H&E). Sulfated GAG (sGAG) deposition was stained using Alcian blue, and secreted collagen was stained using Picro Sirius Red (all from Sigma-Aldrich, USA). Following staining, the samples were imaged using an Olympus CKX53 microscope (Olympus, Japan). Immunohistochemistry was conducted using the antigen retrieval method on constructs after 21 days of in vitro culture to determine the type of collagen synthesized. Sections were deparaffinized and rehydrated using a series of xylene and alcohol baths. Briefly, sections were treated with peroxidase, followed by treatment with chondroitinase ABC (0.25 U/mL) (Sigma-Aldrich, USA) for 1 h at 37°C. Sections were incubated with goat serum to block non-specific sites. Primary antibodies, anti-collagen type I (mouse monoclonal IgG; 90395; Abcam, UK), were diluted to 1:400 in blocking buffer; anti-collagen type II (mouse monoclonal IgG; 3092; Abcam, UK) was diluted to 1:100 in blocking buffer. The sections were incubated with the antibody overnight at 4°C. The sections were then washed in PBS and immersed in diluted hydrogen peroxide solution (3%; Sigma-Aldrich, USA) to block endo-peroxidase activity. Sections were washed, followed by incubation with a secondary antibody (Anti-IgG mouse; B7151; Sigma-Aldrich, USA), diluted to 1.5:200 using blocking buffer, for 1 h at room temperature. Sections were washed again and incubated with ABC reagent (Elite kit Vectastain PK-6100; Vector Labs, USA). After washing with PBS, DAB (3,3-diaminobenzidine; Vector Labs, USA) was used to detect the specific antibody reaction, indicated by brown staining. Each section was washed with cold water to stop the reaction, and sections were dehydrated through an alcohol series and soaked in xylene. Stained sections were imaged using a microscope (BX60; Olympus, Japan).

2.12. Biochemical assays

For biochemical analysis, constructs were digested with papain (0.1 mg/mL; pH 6.4) in a sodium phosphate buffer, containing 0.1 M sodium acetate, 5 nM L-cysteine HCl, and 0.05 M ethylenediaminetetraacetic acid (all Sigma-Aldrich), overnight at 60°C with shaking at 100 rpm. However, it was found that samples containing HAMA were not digested using papain alone. Therefore, a hyaluronidase solution was subsequently used for the digestion of constructs containing HAMA, and an adapted protocol by Beck et al.³⁰ was used. Briefly, a 1000 U/mL hyaluronidase (Sigma-Aldrich, USA) solution was prepared using 0.02 M phosphate buffer containing 77 mM sodium chloride and 0.01% BSA. Samples were allowed to digest overnight at 37°C. Fresh papain solution was then added, and constructs were allowed to digest overnight at 60°C. Both the first and second digestion solutions were stored at −20°C until further analysis. Digests were analyzed on days 1 (D1) and 21 (D21), and media samples were taken and analyzed across cell culture time points (days 1 [D1], 7 [D7], 10 [D10], 14 [D14], 17 [D17], and 21 [D21]) for GAG and collagen content. The sGAG content in each sample was quantified using a dimethylmethylene blue dye-binding assay (Blyscan™ Glycosaminoglycan Assay; Biocolor, UK), as per the manufacturer’s instructions, followed by measuring the absorbance using a microplate reader (Synergy Mx; BioTek, USA) set at 656 nm. To obtain the total biochemical content for each sample, the two digestions (papain and hyaluronidase) and media were quantified and added together. Media samples were taken at days of media change and were analyzed with the same assays to track the release of biochemicals to the media.

2.13. Mechanical testing

To determine the effect of cell growth on the mechanical properties of the ECM-like biomaterial within the construct, mechanical testing was performed using microindentation between the PCL fibers. Stress relaxation studies were performed using a standard testing machine equipped with a 5 N load cell. For day 1 and day 21 samples, a preload of 0.01 N was applied at a rate of 0.01 mm/s and held for 30 s to ensure direct contact between the plate and the construct. The cell-seeded rehydrated hydrogels were compressed at a speed of 0.005 mm/s until 20% strain was reached. The strain was then held for 30 min. The compressive modulus was calculated from the initial linear region of the stress–strain curve, and the compressive equilibrium modulus was determined by dividing the stress at the end of the cycle by the applied strain (20%).

2.14. Statistical analysis

GraphPad Prism 8 (GraphPad Software, USA) was employed for statistical analysis. Printing parameters (i.e., pressure and speed) and compressive properties (i.e., fiber spacing and offsets present) were analyzed with a two-way analysis of variance (ANOVA), followed by Tukey’s post hoc method, to identify any significant differences. One-way ANOVA was performed, followed by Tukey’s post hoc method, to determine any significant differences for other parameters. Differences were considered significant at p < 0.05.

3. Results and discussion

3.1. Microstructure of the native meniscus

One of the key issues in meniscus research is inconsistency in the mechanical testing performed on the native tissue. Studies are based on a variety of different testing conditions and criteria, with varying properties being reported.²⁸ In this study, standard criteria for compressive mechanical testing were followed, enabling a direct comparison of the native meniscus and different scaffold designs. Figure 2A displays the compressive properties of different zones within the bovine meniscus. It can be observed that the compressive properties vary with location within the tissue. For layers 1 and 3, it was found that the compressive properties are lower in the inner region than in the peripheral region. However, in layer 2, it was found that there was no significant difference between the peripheral and inner regions. To develop a scaffold with regional mechanical properties, these zones within the meniscus were subdivided into two groups depending on their compressive properties. Zones with a compressive modulus between 0.5 and 1.1 MPa were categorized as Group A, and zones with a compressive modulus > 1.1 MPa were categorized as shown in Figure 2A (ii).

Therefore, the inner and peripheral regions of layer 1 and the inner region of layer 3 are categorized as Group A; the inner and peripheral regions of layer 2 and the peripheral region of layer 3 are categorized as Group B.

3.2. 3D printing parameters of PCL

Polycaprolactone (PCL) is a promising material for meniscus tissue engineering due to its semi-crystalline nature and resorbable aliphatic properties, enabling it to degrade in the body through hydrolysis of its aliphatic ester linkage.³¹ Additionally, PCL retains its molecular weight longer than other aliphatic polyesters,³² which is beneficial in meniscus regeneration as the scaffold can maintain its mechanical properties longer until tissue ingrowth has occurred. PCL also plays an important role in organizing the matrix and enhancing matrix content.³³ Traditionally, synthetic scaffolds for tissue regeneration are produced using electrospinning. However, electrospun meshes can display both mechanical and biological disadvantages. For example, the fibers may slide under compression loads because they are not fused together. Furthermore, scaffolds formed from electrospinning have small pore sizes and thus may be too dense for cell infiltration.³⁴ The key advantages of 3D printing compared to other manufacturing techniques include high resolution and good control of fiber thickness, orientation, and pore size.

To achieve a scaffold with regional mechanical properties, various 3D-printed PCL architectures with varying porosity were studied. Two different fiber spacings (1 and 2 mm), three different offsets (0, 1, and 2 offsets), and two different layer orientations (single and double) were investigated in this study. Each scaffold was 10 layers high and printed in various architectures (Figure 1), with controlled internal fiber deposition and pore size. The fabrication time for each scaffold was approximately 30 min. An example of printed offsets is displayed for 1 mm fiber spacing single-layered scaffolds in Figure 3A.

Key considerations for tissue engineering scaffolds include pore size, porosity, and interconnectivity, all of which are determined by the architecture of the scaffold.³⁵ 3D printing offers an advantage over other production techniques, as pore interconnectivity can be designed directly into the scaffold model. With an increase in fiber spacing from 1 to 2 mm, the theoretical pore size increased. As expected, as offsets are introduced to the scaffold, the pore size decreases. This is attributed to a fiber crossing the pore when the offset layer is printed. The maximum theoretical porosity of 93% was achieved at 2 mm fiber spacing, 0 offsets, and at a pore size of 4 mm², while a minimum porosity of 74 % was achieved at 1 mm fiber spacing, 2 offsets, and a pore size of 0.16 mm² (Figure 3B and C). Porosity may also be increased by decreasing the needle’s inner diameter; however, it was found that attempts to produce thinner fibers using a 30 Ga needle made inconsistent prints that were susceptible to defects.

3.3. Mechanical properties of 3D-printed PCL scaffolds

For cartilage regeneration, scaffolds with porosity > 70% are deemed suitable for cell attachment and matrix deposition.³⁶ However, scaffold porosity affects the mechanical properties, whereby high porosity compromises the structural integrity of the scaffold.³⁷ When the compressive modulus is too low, it may result in deformation and failure of the implant, subsequently leading to failure of the regenerated tissue. The compressive modulus was found to significantly decrease with increasing fiber spacing from 1 to 2 mm as shown in Figure 2B.

When layers were printed in double format, i.e., the same layer orientation printed sequentially, the compressive properties increased compared to singlelayer orientations. A maximum compressive modulus of 2.74 MPa was found for 1 mm fiber spacing, 0 offsets, and double-layer orientations. Simply changing the fiber spacing to 2 mm, while maintaining 0 offsets and doublelayer orientations, decreased the compressive modulus to 1.73 MPa. Compared to the double layer, single-layer orientations exhibited lower compressive properties: 2.63 MPa for 1 mm fiber spacing and 0 offsets; and 0.88 MPa for 2 mm fiber spacing and 0 offsets. The significant decrease in compressive properties is due to the increase in pore size from 1 to 4 mm² and porosity from 87% to 93%. In comparison to that of the human meniscus, the results are comparable at 0.2, 0.23, and 0.28 MPa in the anterior, central, and posterior parts of the meniscus, respectively.³⁸ As the meniscus is an anisotropic material, the mode or direction of the analysis and the depth of the sample tested will all affect the reported properties.

Compressive mechanical properties of PCL scaffolds manufactured with varying porosities have been previously reported in the literature. However, the reported porosities are generally much lower than this study, corresponding to higher compressive properties. Through fused deposition modeling (FDM), a compressive modulus of 4 and 77 MPa at porosities of 77% and 48%, respectively, has been reported for porous PCL.³⁹ Other deposition techniques have reported a compressive modulus of 59 MPa at 65% porosity⁴⁰ and 21.4 MPa at 69.6% porosity.⁴¹ Likewise, the selective laser sintering (SLS) technique has reported a compressive modulus of 47 MPa for bulk PCL and 6 MPa for scaffolds at 55% porosity.⁴²

The addition of offsets to the internal structure had the most significant effect on the mechanical properties (Figure 2B). Across all fiber spacings and layer orientations (single or double), there was a significant decrease in compressive properties when an offset was introduced and an even further decrease in mechanical properties with the addition of 2 offsets. For a 1 mm fiber spacing, single-layer compressive properties were found to be 2.63 MPa, 0.75 MPa, and 0.40 MPa for 0, 1, and 2 offsets, respectively. This decrease in compressive properties with the addition of offsets can be explained by the reduction of support within the scaffold at the fiber intersections.⁴ When no offsets are present within the scaffold, there is direct contact with neighboring fibers, whereby the fiber intersections are supported from above and below. However, with the addition of offsets, this support is removed, and the intersection is suspended, enabling a three-point bending of the fibers under the same compressive loads (Figure 2C).

Most studies using PCL create anatomically shaped meniscus scaffolds that lack zonal differences in the PCL architecture, with the scaffolds possessing mechanical properties far superior to the native tissue.^4,6,7 Many scaffold architectures match the criteria of Group A. However, none of the scaffold architectures match Group B. A 1 mm fiber spacing resulted in a scaffold with compressive properties > 1.5 MPa, and a 2 mm fiber spacing led to compressive properties < 1. To determine the optimal fiber spacing, straight fiber scaffolds were modified to include circumferential and radial fibers. When preparing radial tie-fiber orientation, the angle between the adjacent Y-direction fibers was set to 4.5°. This led to a radial tiefiber spacing of 1.6 mm in the peripheral region, which tapered inward until 0.81 mm. The conversion of straight fibers to curved fibers did not significantly increase the mechanical properties of the scaffold. However, the introduction of radial fibers and circumferential fibers significantly enhanced the mechanical properties (Figure 3D). This increase can be attributed to more fibers being supported by the fiber layer below, thereby increasing the stability of the construct.

3.4. Printing a PCL scaffold that mimics the circumferential and radial fibers of native meniscus

Optimized scaffold architectures with circumferential and radial fibers resulted in properties that matched each grouped zone within the native meniscus tissue. The two internal architectures selected to develop a full-sized biomimetic scaffold were: circumferential 2 mm fiber spacing and 1 offset for Group A; and circumferential 1 mm fiber spacing and 1 offset for Group B. Both regions consisted of radial fibers with an angle between the adjacent fibers set to 4.5°. To design the meniscal scaffold replicating the native architecture, the overall measurements of the meniscus were based on the dimensions of the medial meniscus in a study by McDermott et al.⁴³ The meniscal length was set at 45.7 cm, and the meniscus body width was set at 9.4 cm.

As displayed in Figure 4A–C, the meniscal scaffold was successfully printed. Good control of the oriented fibers via the 3D printing process was achieved, demonstrating the capability of the 3D printing technique to build a 3D scaffold with complex internal structures. Both circumferential and radial fibers were printed for the replication of the specific 3D collagen microstructure of the meniscus. Control over fiber spacing enabled the establishment of region-dependent mechanical properties. The overall diameter of the printed fibers was approximately 200 µm, which is lower than previously printed PCL meniscus scaffolds.^4,6 To fabricate a wedgeshaped meniscus, the number of circumferential fibers for each layer was progressively decreased. The 3D-printed scaffold mimicked the native meniscus, both in external shape and internal microstructure, and displayed better mechanical properties than previously reported, thus suggesting its potential for meniscus repair.

3.5. Development of an ECM-infiltrated PCL scaffold

Although the PCL scaffold provides the macrostructure architecture and mechanical properties of the native meniscus, a natural-based biomaterial ink should be incorporated into the scaffold to enhance cell infiltration and proliferation of cells into the PCL scaffold. Formulating a biomaterial ink is challenging, as it must also provide an ideal environment for cells to attach, proliferate, and differentiate while possessing gelation, mechanical, and rheological properties that facilitate 3D printing. GelMA was selected as the major component of the biomaterial ink formulation due to its potential in biomedical applications. As HA and CS are abundant in cartilage ECM^28,38 and have been investigated for their effect in enhancing chondrogenesis,²⁷ the addition of GAGs to a GelMA matrix was investigated for a fibro-chondrogenic effect. GelMA was the main component of each hydrogel, accounting for at least 87% of the dry content mass. A simplified hybrid structure was developed that consisted of a PCL framework and was imbedded in three different hydrogel combinations: GelMA, GelMA/CS/HA, and GelMA/CS/HAMA. Using an optical microscope, the overall morphology of the constructs was determined (Figure 5A). After the freeze-drying process, all scaffold compositions maintained structural robustness, and PCL fibers retained their structure.

3.6. Morphological analysis of the ECM-infiltrated PCL scaffold

The pore architecture of scaffolds significantly affects their physical properties, as well as cellular activity and distribution.^29,44,45 Controlling the pore size and interconnectivity is essential for successfully creating porous biomaterials and scaffolds.^46,47 Cell functions and the regeneration of new tissue are reliant on pore size.^29,48 It has been established that the pore diameter must be large enough to allow infiltration of the cells, but small enough to present a large scaffold surface area for cellular attachment.⁴⁶ Hybrid scaffolds were developed as described above. The effect of pre-freeze temperature on the porous structure was investigated by pre-treating scaffolds to −20°C, −80°C, or flash-freezing in liquid nitrogen. Scaffolds were fabricated using a freeze-drying protocol adapted from Murphy et al.,³⁵ whereby an annealing step was introduced to the freeze-drying cycle for producing scaffolds with larger pores.

The pore sizes varied considerably in this study with changes in the pre-freeze temperature. From the SEM images (Figure 5B and C), the pore size for GelMA was found to significantly decrease (from 160 to 99 µm) when the pre-freeze temperature was altered from −20 to −80°C. A similar trend was observed with the addition of CS/HAMA, whereby the pore size decreased from 163 to 98 µm as the pre-freeze temperature changed from −20 to −80°C. Incorporating CS/HA into a GelMA hydrogel caused a significant decrease in pore size compared to GelMA and GelMA/CS/HAMA at −20°C. Flash-freezing hydrogels caused a significant decrease in pore size across all groups. Overall, pre-freezing to −20°C resulted in pore sizes of 81–163 µm; pre-freezing to −80°C resulted in pore sizes of 68–99 µm; and flash-freezing resulted in a significant decrease in pore size to 17–50 µm. The most significant change in structure was due to flash-freezing the samples before freeze-drying. Previous studies have investigated the effect of cooling rate on the porous structure of freeze-dried constructs. It has been demonstrated that slower cooling rates produced porous scaffolds with larger pores.⁴⁹ A slow cooling rate would slow down ice crystal formation, thereby enabling the growth of larger ice crystals. When the cooling rate is fast, such as freezing in liquid nitrogen, all the crystallization heat is extracted, and crystallization starts simultaneously in the entire hydrogel. This results in the limited formation of ice crystals, and thus, a non-porous or very small porous structure is formed, which was also observed in this study.

Small pores can effectively reduce cell migration compared to larger pores.^29,50 However, Stenhamre et al. reported that smaller pores, while reducing cell migration, promote cartilage-like formation with larger pores, resulting in chondrocyte differentiation into the osteogenic pathway.⁵⁰ Nonetheless, smaller pores have the benefits of more surface area and cell attachment sites.⁵¹ Additionally, it has been found that larger pores require more cells to fill the free space within the pore, while smaller pores fill the space more easily. Filling these pores can increase cell interactions and provide a more 3D microenvironment to promote tissue formation.⁵⁰ Further studies revealed that regardless of pore size (150–500 µm), cell migration can be attained with good interconnectivity.⁴⁵ Taken together, small to medium pores (80–250 µm) are optimal for cartilage regeneration, i.e., being large enough to promote cell infiltration, but small enough for complete pore filling to facilitate cell-on-cell interaction and cartilage formation. From these findings, it is evident that pre-freezing to −20^oC for 2.5 h before freeze-drying resulted in pores that are suitable for meniscus tissue regeneration, displaying improved interconnectivity across groups and attaining pore sizes of 81–163 µm.

3.7. In vitro assessment of the ECM-infiltrated PCL scaffold

For successful meniscus cartilage tissue engineering, it is critical to control the phenotype of hMSCs and the composition and organization of the ECM they produce. To assess the chondro-inductivity of the scaffolds, hMSCs were statically cultured in chondrogenic media in low oxygen conditions (5% O₂), which was previously demonstrated to support cartilage formation.⁵² The viability of hMSCs on the hybrid meniscal scaffolds at D1 and D21 was evaluated using Live/Dead staining and ImageJ analysis. With regards to the surface of the scaffolds, on D21 (Figure 6A), cells displayed higher viability in all groups, indicating the adequate cytocompatibility of the scaffolds. For example, cell viability in the GelMA group increased from 72.6% to 84.4% between days 1 and 21. The cells on the GelMA hydrogels by D21 varied between stretched and rounded morphologies. Scaffolds containing CS/HA and CS/HAMA displayed a more consistent elongated cell morphology. Previously, it was found that HAMA alone resulted in lower cell viability compared to GelMA.²⁵ However, in this study, the addition of HA or HAMA did not result in any decrease in cell viability. For example, cell viability increased from 83.1% to 90.8% for the GelMA/CS/HAMA group from day 1 to 21. The high cell viability on the surface of the scaffolds is probably due to the biological recognition sites of gelatin, including the arginine-glycineaspartic acid (RGD) sequence,¹⁶ as well as the biological processes involving HA, such as cell attachment, mitosis, and proliferation.¹⁵

Figure 6B depicts the cross-section of the scaffolds on D1, showcasing that cells have infiltrated the scaffold. By D21, the cells were still alive in all groups. On the GelMA scaffolds, the cells grew in random clusters throughout the cross-section and were not evenly dispersed, displaying an area with a high cell density. For scaffolds containing CS/HA and CS/HAMA, the cells mainly clustered around the PCL and grew in the direction of the fiber by D21. These results revealed that the 3D-printed meniscal scaffolds could serve as a micro-pattern to guide cells to secrete an organized fibrocartilaginous matrix, which is crucial for the meniscus due to its complex internal structure and biomechanical function. However, cell expansion into the hydrogel matrix was limited due to the hydrogel entrapping the cells. Each channel of the PCL scaffolds contained a hydrogel material. Once freeze-dried, the hydrogel contracted from the PCL fibers, thereby leaving empty spaces for cells to infiltrate. During the culture period in media, the scaffolds swelled to form a hydrogel again. In turn, this caused the infiltrated cells to get trapped by the PCL fibers, limiting proliferation through the gel and compromising cell viability. For example, cell viability in the GelMA scaffolds at D1 was at 43.5%, while the viability decreased to 38.6% on D21. This limited internal cell proliferation between D1 and D21 may be due to the lack of pores, restricting nutrient diffusion through the swollen hydrogel.

Cellular morphology on the surface of the scaffold was evaluated by examining the actin cytoskeletal network using a confocal laser scanning microscope on D1 and D21 (Figure 7A). On D21, cells displayed an elongated cytoskeleton compared to D1. Scaffolds containing CS/HA and CS/HAMA cells exhibited an elongated morphology with prominent actin fibers, resembling fibrochondrocyte morphology observed in previous studies using hMSCs in meniscal tissue engineering.^12,53 Low levels of HA were used in this study, as the influence of HA on chondrogenesis is concentration-dependent. Many studies have found that at high concentrations HA has no effect or even inhibits chondrogenesis.^54,55 Nevertheless, using low concentrations has reported some promising results.^27,56 GelMA and HA exhibit intrinsic mixing incompatibility when combined. In a study by Levett et al., this immiscibility led to phase separation between GelMA and HAMA.²⁷ It was reported that this phase separation phenomenon may have rendered cell adhesion sites on the GelMA less available to cells, resulting in rounded cell morphology. In this study, the addition of HA and HAMA resulted in elongated cells, suggesting sufficient mixing without phase separation. Consequently, GelMA was uniformly distributed throughout the construct, facilitating robust cell adhesion and proliferation. Hydrogels containing CS/HA are soft, and the cells display an elongated cell morphology. Hence, our findings suggest that cell morphology is influenced by material composition rather than its stiffness.²⁵

As cell–material interactions are known to influence changes in cellular behavior and phenotype, proliferated hMCSs were tested for the production of ECM components, including collagen and GAGs. To identify cartilaginous tissue formation, H&E staining was performed. After 21 days of in vitro culture, intact hydrogels were still observed in all groups, and evidence of cartilage tissue formation was observed in all samples (Figure 7B). It was observed that elongated fibrochondrocytes aligned in rows embedded in the newly formed matrix composed of collagen fibers stained pink with H&E.

Cells mainly clustered around the PCL and grew in the direction of the fiber. PCL fibers influenced the cell orientation, forming an elongated cell morphology due to its rigid structure.²⁵ Therefore, matrix formation was found to be secreted along the direction of the PCL fibers. Furthermore, it is known that PCL lacks bio-adhesive sites. Hence, while PCL guides the formation of tissue, the hydrogel within the construct promotes the production of ECM, highlighting that cells are influenced by their surrounding environment and the importance of using multi-material constructs.

The solid component of meniscal tissue is primarily composed of a network of collagen fibrils³⁸ and plays a primary role in the biomechanical function of the meniscus.²⁸ As displayed in Figure 8, the formation of collagen was evident in all scaffolds. Interestingly, collagen formation was mostly predominant along the fiber axis, which is corroborated by H&E staining. This is an essential finding in creating a collagen network that replicates the natural collagen orientation, indicating that the 3D-printed PCL meniscal scaffold could serve as a micro-pattern to guide cells to secrete an organized fibrocartilaginous matrix. This is of particular importance in the meniscus, whereby its complex internal structure plays a major role in its biomechanical function.

After 21 days of culture, GAGs were identified in the pores of GelMA and along the PCL fibers in all groups (Figure 8).

GelMA was stained with Alcian Blue and displayed fully pink features on D1, and blue streaks on D21, indicating sGAG production over time. Compared to GelMA, samples containing CS/HA or CS/HAMA stained more positively for sGAGs. CS contains a sulfated polymer, which consequently improves chondrogenesis of the entrapped mesenchymal stem cells.⁵⁷ Samples containing CS at D1 samples display a blue color compared to GelMA alone. The newly produced sGAG within the scaffolds stained more intensely blue and is predominantly found along the PCL fibers, indicating higher sGAG content in these regions than in the original scaffold.

Figure 9A and B presents the sGAG content in scaffolds between D1 and D21, as well as the GAG secretion profile over time. As observed, CS is highly soluble as it is readily leached from the scaffolds containing CS (Figure 9C), leading to an overall reduction in GAG content between D1 and D21. The secretion of CS at D1 is high; however, over time, the quantity of secreted sGAG steadily decreases. Figure 9D presents more details on the secretion of sGAG between D14 and D21. In scaffolds containing CS, there is a significant drop in sGAG secretion between D14 and D17; however, after D17, there is an increase in the secreted sGAG. This increase in GAGs in the media indicates the production and secretion of newly synthesized sGAG from the seeded cells within the scaffold. Overall, at D21, the amount of sGAG retained and secreted (both normalized to wet weight) was significantly higher in the constructs containing CS/HA and CS/HAMA (Figure 9A). This finding is consistent with previous reports, whereby GelMA alone may result in reduced GAG production.²⁵ The total sGAG retained in the construct (relative to secreted sGAG) was also higher in the samples containing CS/HA and CS/HAMA, i.e., 72%, 92%, and 94% for GelMA, GelMA/CS/HA, and GelMA/CS/HAMA, respectively (Figure 9B). It was previously reported that HA interacts with ECM components in matrix retention, particularly when loading is applied.^16,58 This may explain the smaller percentage of sGAGs lost to the media in samples containing HA and HAMA compared to GelMA alone. This result suggests that the use of GAGs to produce a biomimetic hydrogel provides cues that may be important for tissue homeostasis. After 21 days, the net GAG retained in each construct was 0.02%, 0.21%, and 0.23% of its wet weight for GelMA, GelMA/CS/HA, and GelMA/CS/HAMA, respectively, i.e., approximately 15–23% of the native meniscus.

Immunohistochemistry between D1 and D21 revealed that the different hydrogel compositions could support the synthesis of either collagen type I or collagen type II (Figure 9E). In hMSC-laden GelMA hydrogels, both collagen types I and II were observed. However, hydrogels containing GAGs displayed a higher intensity of collagen type I. Furthermore, cells on the surface of all constructs produced more collagen than internally in the scaffolds, which is consistent with previous studies.^25,59 This is most likely due to the lack of cell infiltration and lower nutrient concentrations caused by diffusion gradients within the hydrogel. Cells on the surface of the constructs were able to proliferate faster and produce more ECM, as they had better access to nutrients and oxygen.²⁵ Different collagen types are found in each region with varying quantities, with collagen type I predominating within the meniscus.⁶⁰ Collagen type I is found throughout the meniscus, whereas collagen type II is exclusively found in the inner two-thirds region. Collagen type I is most abundant in the peripheral region of the meniscus and is responsible for 90% of the composition by dry weight, while other collagens (type II, III, IV, and XVIII) are present in quantities less than 1%.⁶¹ In the inner region, collagen constitutes approximately 70% of the dry weight, of which 60% is collagen type II and the remaining 40% is collagen type I.⁶² Our results indicate that a scaffold with regions containing different ECM-like material compositions may be used to promote regiondependent collagen synthesis similar to the native tissue.

3.8. Mechanical properties of cell-seeded hybrid scaffolds

Mechanical properties are crucial for the success of biomaterials in tissue engineering. A decrease in mechanical properties indicates the dissolution of the hydrogel, suggesting a loss of structural integrity. However, an increase in mechanical properties over time may be an indication of ECM production.²⁵ In this study, no significant degradation was observed, and all constructs were easy to handle at D21. The addition of CS/HA to the GelMA matrix on D1 caused a decrease in the equilibrium and compressive modulus (Figure 10A and B). This decrease is likely due to a reduction in crosslink density caused by GAG molecules hindering the crosslinking of GelMA. Conversely, the addition of HAMA increased the equilibrium and compressive modulus. Moreover, the addition of HAMA enhanced the hydrogel’s mechanical properties compared to GelMA alone or with the addition of unmethacrylated HA. This improvement could be attributed to the increased crosslink density of the matrix, with HAMA concurrently crosslinked along with the GelMA. The different degradation characteristics of the hydrogels containing HA and HAMA provide insight into their crosslinking mechanisms. For instance, the GelMA/CS/HA hydrogel completely digests in papain. However, for GelMA/CS/HAMA, when the gelatin component of a gel is removed by papain, an intact and stable network of crosslinked HAMA remains. The existence of this stable HAMA network indicates that an interpenetrating HAMA network exists within the bulk GelMA network, which explains the significantly higher equilibrium modulus in constructs containing HAMA compared to the HA at D1. Additionally, after one day in culture media, the wet weight of constructs containing HA was greater than GelMA alone and constructs containing HAMA (Figure 10C). This is attributed to a reduction in crosslink density, thereby enabling more water uptake. During culture, the hydrogel scaffolds were subjected to enzymatic degradation. As the hydrogels are GelMA-based, the presence of cells may result in rapid degradation of the hydrogel via matrix metalloproteinases (MMP) activity.²⁵ After 21 days of culture, both the compressive and equilibrium modulus increased in all groups. This change in mechanical properties is a result of a balance between the production of new ECM and hydrolytic degradation of the hydrogel. Therefore, the enhanced mechanical properties are a result of the cell-laden matrix. Additionally, there was no significant change in wet weight within each group over time (Figure 10C), suggesting that the rate of degradation of the gel matched the rate of new matrix production.

At D21, samples containing HAMA had the greatest mechanical properties compared to the other groups, with the highest compressive modulus at 0.36 MPa and equilibrium modulus at 0.26 MPa. It was reported that sGAGs attract water within the meniscus tissue. This water uptake and collagen fiber orientation are known to have a symbiotic relationship concerning the compressive properties of the tissue.²⁸ In this study, all groups exhibited an increase in compressive properties, accompanied by an overall reduction in sGAGs from D1 in both GelMA/CS/HA and GelMA/CS/HAMA groups. Therefore, it is hypothesized that the sGAG level does not fully account for the changes in compressive properties, but rather the location-specific distribution of GAGs plays a significant role. From histological analysis, it is evident that newly secreted ECM is concentrated along the PCL fiber, potentially resulting in narrower spaces between the fibers. This may restrict water flow and mimic the natural synergetic relationship between the collagen fibers and GAGs in natural tissue. Furthermore, GelMA/CS/HA exhibited the most significant change in mechanical properties compared to GelMA/CS/HAMA during cell culture, with the compressive modulus increasing by 41% between D1 and D21. The initial mechanical properties are dependent on the crosslink density of the polymer network, and a lower crosslink density facilitates diffusion of the newly secreted matrix.⁶³ Therefore, hydrogels that are initially stiff may hinder the formation and distribution of ECM, while softer materials facilitate new ECM infiltration, consequently becoming stiffer after in vitro culture.^63,64 The same trend is observed here, whereby the HA group features enhanced mechanical properties; however, interestingly, both HA and HAMA groups have similar collagen and GAG content. Therefore, this change in mechanical properties may be explained not only by the production of the matrix, but also by the interaction of molecules within the ECM. It is expected that a higher interaction between the unmethacrylated HA with the newly produced ECM compared to HAMA enables the retention of molecules within the matrix when a compressive load is applied.⁵⁸ With longer in vitro culture, more distinct changes in mechanical properties would likely be observed between the groups.²⁷ Both the equilibrium and compressive moduli of hydrogels alone were found to be lower than the native meniscal tissue. However, this is mitigated by the reinforcement of hydrogels with PCL, which provides structural integrity to the scaffolds. The observed changes in mechanical properties of hydrogels over time in the in vitro cell culture study are indicative of cell proliferation and ECM production.²⁷

From these results, GelMA/CS/HAMA is suggested as a possible candidate for the peripheral region of the full meniscus scaffold to enhance the production of collagen type I. Conversely, GelMA/CS/HA is suggested for the inner region due to its ability to increase its compressive properties faster over time and exploit the interaction of HA and CS molecules. However, further analysis of in vivo biological responses to these materials is warranted to fully substantiate their effectiveness in meniscus tissue engineering strategies.

4. Conclusion

In this study, we successfully fabricated scaffolds from PCL fibers using 3D printing technology that mimics the structure of a human meniscus. We demonstrated that by varying the fiber spacing within layers and introducing offsets, scaffolds with different pore sizes and mechanical properties could be produced. The 3D-printed PCL framework was infiltrated with hydrogel combinations of GelMA, GelMA/CS/HA, and GelMA/CS/HAMA, and the scaffolds were subsequently freeze-dried. Additionally, the pre-freeze process of freeze-drying had a significant impact on the size and distribution of pores within the scaffold, with higher pre-freezing temperatures favoring ice crystal formation and subsequently resulting in significantly larger pores. Overall, pre-freezing to −20°C for 2.5 h resulted in pore sizes of 81–163 µm; pre-freezing to −80°C resulted in pore sizes of 68–99 µm; and flash freezing caused a significant decrease in pore size to 17–50 µm. Hybrid ECM/PCL scaffolds seeded with hMSCs successfully infiltrated the scaffold, with the PCL fibers serving as a micro-pattern to direct the proliferation of cells and newly secreted ECM. The hMSCs displayed a predominantly elongated morphology and successfully secreted ECM. By D21, the collagen content within the scaffolds was found to be 22–25% of the native meniscus tissue, higher than previous reports for meniscus regeneration. The matrix produced along the fiber orientation increased the compressive modulus over time, possibly due to the restriction of water flow between the fibers. This could also be attributed to the synergetic relationship between collagen fibers and GAGs in the native meniscus. Taken together, the enhanced cellular proliferation, ability to synthesize ECM similar to the native tissue, capability to provide a micro-pattern for directing the orientation of the newly synthesized ECM, and improvement in mechanical properties over time make this hybrid scaffold a promising candidate for meniscus tissue engineering applications.