Silk Fibroin-Based Hydrogels Supplemented with Decellularized Extracellular Matrix and Gelatin Facilitate 3D Bioprinting for Meniscus Tissue Engineering

Abstract

The human meniscus transmits high axial loads through the knee joint. This function is compromised upon meniscus injury or treatment by meniscectomy. 3D printing of meniscus implants has emerged as a promising alternative treatment, as it allows for precise mimicry of the meniscus architecture. In this study, silk fibroin (SF) known for its excellent mechanical properties is used to fabricate hydrogels for 3D bioprinting with infrapatellar fat pad-derived mesenchymal stem cells (IFP-MSCs). Extracellular matrix (ECM) derived from bovine menisci and gelatin are added to 10% SF to promote cell adhesion and printability. To examine the mutual influence of cells and biomaterial, experiments are conducted with and without IFP-MSCs. The cells are found to influence crosslinking, β-sheet formation, and mechanical strength. Variations between printed and casted hydrogels are identified for cell number, metabolic activity, secondary structure, and mechanical strength. Remarkably, the printed hydrogels with IFP-MSCs exhibited a compressive Young's modulus of 0.16 MPa, which closely resembled that of human osteoarthritic menisci. After initial low viability, IFP-MSCs in the casted hydrogels are able to proliferate within the biomaterial. The chondrogenic differentiation medium upregulated the expression of chondrogenic markers in the casted hydrogels, indicating promising prospects for future meniscus tissue engineering (TE).

Keywords: 3D bioprinting; meniscus tissue engineering; silk fibroin.

Figure 1

Characterization of bovine meniscus and ECM powder by biochemical assays. A) DNA content reduction by decellularization (n = 3). B) Loss of sGAGs (n = 3). C) Preservation of collagen represented by hydroxyproline content (n = 3).

Figure 2

Crosslinking kinetics of SF‐based or ECM‐based hydrogels with 10 u mL⁻¹ HRP and 0.01% H₂O₂ at 37 °C. The means of three technical replicates are illustrated. A–C) Crosslinking kinetics of hydrogels with varying SF and ECM concentrations (n = 1). D) Combined fluorescence by addition of the individual fluorescence values of SF10 and ECM2.5‐ECM10 for comparison with measured combinations (n = 1). E) Crosslinking times of SF and SF‐ECM gels. Data were compared using one‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a–j) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 1).

Figure 3

The influence of IFP‐MSCs on SF‐based hydrogel crosslinking using 10 u mL⁻¹ HRP and 0.01% H₂O₂ at 20 °C. A) Crosslinking times of SF‐based hydrogels with and without IFP‐MSCs. Data were compared using two‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a–g) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 3). B) Fluorescence maxima of SF‐based hydrogels (n = 3). C) Crosslinking kinetics of SF‐based hydrogels without cells (n = 3). D) Crosslinking kinetics of SF‐based hydrogels with 10⁶ IFP‐MSCs/mL (n = 3).

Figure 4

3D (bio)printing SF10‐ECM5‐G3 in comparison to casted gels. A) Printability test of SF10‐ECM5‐G3 at 20 °C and SF10‐ECM10 at 4 °C without crosslinking. B) Metabolic activity of IFP‐MSCs in SF10‐ECM5‐G3 gels (n = 1, with three technical replicates). C) Casted and printed cell‐free controls after crosslinking with HRP and H₂O₂.

Figure 5

Mechanical properties of hydrogels and OA human menisci. The means of three technical replicates are illustrated. A) Compressive Young's modulus of hydrogels on day 14 was calculated from unconfined compression tests at 10% strain. Data were compared using two‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a‐g) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 1). B) Poisson ratio of OA human menisci (n = 3). C) Young's modulus of OA human menisci at 10% strain (n = 3). D) Transparency change caused by β‐sheet formation. (scale bar = 10 mm).

Figure 6

Secondary structure changes caused by 3D (bio)printing. The means of three gels as technical replicates were illustrated for the cell‐free controls. A) β‐sheet content of casted and printed cell‐free controls in 1X PBS. Data were compared using two‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a–d) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 1). B) RC content of casted and printed cell‐free controls in 1X PBS (n = 1). C) β‐sheet content of casted and printed gels with IFP‐MSCs in comparison to cell‐free controls (n = 3). D) RC content of casted and printed gels with IFP‐MSCs in comparison to cell‐free controls (n = 3). E) Amide I region in casted cell‐free controls in 1X PBS (n = 1). F) Amide I region in printed cell‐free controls in 1X PBS (n = 1). G) Amide I region in casted gels with IFP‐MSCs (n = 3). H) Amide I region in printed gels with IFP‐MSCs (n = 3).

Figure 7

The effect of 3D bioprinting on weight and stability. The IFP‐MSCs were cultivated in casted or printed hydrogels exposed to growth or chondrogenic differentiation medium, while the cell‐free controls were cultivated in growth medium. A) Hydrogel wet weight (n = 3). B) Water content calculated with wet and dry weight (n = 3). C) Compressive Young's modulus at 10% strain tested by unconfined compression. The only significant difference was found between casted (growth) day 14 and printed (growth) day 14 (n = 3).

Figure 8

The effect of 3D bioprinting on cell viability. The IFP‐MSCs were cultivated in casted or printed hydrogels exposed to growth or chondrogenic differentiation medium. A) Confocal microscopy images after live and dead staining of hydrogel cross sections (central area) and SEM images of cell‐free controls. The scale indicates 300 µm in the confocal microscopy images and 500 µm in the SEM images (n = 1). B) Living cells counted in ImageJ. Data were compared using two‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a–c) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 3). C) Calculated viability (n = 3).

Figure 9

Gene expression of IFP‐MSCs. The IFP‐MSCs were cultivated in 2D, casted, or printed hydrogels exposed to growth or chondrogenic differentiation medium. The Ct values were normalized to GAPDH and to the 2D control in growth medium of day 14 by the 2^−ΔΔCt method. Data were compared using two‐way ANOVA with Tukey post‐hoc test. Different letters above the bars (a–j) indicate statistically significant differences between the groups (p < 0.05). Groups sharing a letter are not significantly different (p ≥ 0.05) (n = 3).