Introduction

Natural bone is a hierarchical composite structure1 composed of carbonated apatite that is orderly deposited within a type I collagen matrix2. Although bone has an intrinsic regenerative capacity, large defects or pathological conditions that impair bone integrity can significantly hinder its self-repair ability3. The three-dimensional (3D) architecture of bone presents challenges for in vitro proliferation of injected cells in defect sites. Consequently, tissue engineering has emerged as a promising strategy to mimic bone’s hierarchical structure by developing suitable 3D scaffolds that support new functional bone tissue formation3.

Bone tissue engineering is a rapidly growing field aimed at providing effective substitutes for damaged or diseased bone3,4. Since ceramics and polymers alone cannot fully meet scaffold requirements as bone tissue scaffolds, ceramic-reinforced polymer composites have been developed to integrate favorable properties of both components5,6. Biomimetics is the imitation of natural approaches, systems, and processes for drawing inspiration, driving design idea7, and applying biological concepts to different scientific fields for developing innovative technologies and synthesizing materials and organs8. So, natural strategies and biomimetic approaches have gained considerable attention to fabricate composite bone grafts. Thus, smart biomaterials that actively participate in tissue regeneration and accelerate the early tissue formation stages are crucial for successful outcomes9.

Hydrogels, characterized by deformable three-dimensional (3D) polymer networks10, can be synthesized via physical or chemical crosslinking11, and mimic the extracellular matrix (ECM) at bone injury sites with or without the inclusion of various osteogenic bioactive molecules. Incorporating photoreactive groups into polymers enables in situ hydrogel formation, facilitating uniform cell distribution within scaffolds12. Photocrosslinking with appropriate light sources allows rapid gelation under physiological conditions without pH or thermal damage13, and thus preserving the bioactivity of encapsulated biomolecules and drugs14. Furthermore, the internal structure and mechanical properties of hydrogel can be tailored by polymer selection and reinforcement, thereby optimizing the biochemical and biophysical environment for osteogenic differentiation and creating quasi-in vivo microenvironments9.

Scaffold biocompatibility and functional properties depend on its constituent materials. Chitosan, silk fibroin (SF), and hydroxyapatite (HAp) are widely used in tissue engineering, individually or in combination with other polymers. Among calcium phosphates (CaPs), hydroxyapatite (HAp, Ca10(PO4)6(OH)2) closely resembles the mineral phase of bone15, binds to surrounding hard tissue through specific biological responses16, and enhances tissue compatibility when coated on polymers2. Bombyx mori (B. mori) silk fibroin, a natural protein fiber coated with sericin17,18, exhibits favorable biological properties and its diverse forms have been studied extensively for biomedical applications19,20. Although SF promotes osteogenic gene expression, pure SF scaffolds are insufficient for regenerating large bone defect; however, pre-seeding with undifferentiated mesenchymal stem cells (MSCs) can significantly improve bone formation21. The surface functionalization of SF fiber by HAp deposition combines SF’s mechanical strength with high osteoconductivity of HAp22,23. Chitosan has excellent properties for bone tissue engineering3,24, and it promotes cell adhesion, proliferation3, and the osteo-differentiation of osteoblast ‎precursors seeded on chitosan-based substrates25. Despite early in vivo evidence of chitosan’s bone regenerative capacity reported in 19883, the underlying cellular mechanisms remain under investigation25.

According to the Scopus database, a literature review from 2014 to 2025 reveals numerous composite formulations for bone regeneration, including bioactive molecules, cells, and animal models. Specifically, 42 studies utilized a three-component (Cs/SF/HAp) system, and the Cs/SF mixture were used in nearly all investigations26,27,28,29,30,31,32,33,34,35,36,37,38,39. The final particle-reinforced scaffolds were fabricated using freeze-drying (with27,30,34,36,37,40,41,42 or without26,27,31,32,33,35,36,38,39 chemical crosslinking), electrospinning43,44, plasma splashing39, or thermally induced phase separation45. HAp or other CaPs were used as commercially obtained powders in some studies26,27,28,29,31,32,33,35,36, synthesized via co-precipitation34,37,39,41,42,43,46,47,48 or hydrothermal methods38,49, deposited by alternate soaking44,50, or formed in situ within scaffolds47,48,51,52. Although SF has been widely employed in fibril or nanofibrous forms for the regeneration of bone53,54, nerve55,56, skin57,58, cartilage59, liver60, and vascular tissues61,62, it has been mainly electrospun alone or combined with other biomaterials53,54,55,56,57,58,61,62,63,64,65. However, the native or surface-modified SF fibers have not been extensively applied for bone regeneration. Some representative publications are cited here, and a complete list is presented in the supplementary.

This study aimed to develop and characterize a biomimetic composite scaffold inspired by endochondral ossification combined with bone native structure. The study focused on an engineered photocrosslinkable Cs-SF-HAp hydrogel enriched with mineralized SF fibers that were embedded within a methacrylated chitosan matrix. A photo-activated double crosslinking system to make a soft Cs matrix and a two-step mineralization process, involving modified alternate soaking and heat treatment was introduced to achieve a complete and more crystalline HAp coating on the SF fibers. Expecting superior mechanical properties of fibrous reinforcement than particulate reinforcement66,67, HAp-coated SF fibers promise a structurally-strengthened and osteoconductive composite hydrogel. Furthermore, various hydrogel formulations were prepared, and the structures of fiber-free and fiber-enriched hydrogels were evaluated under different ionic conditions. To investigate the effect of cell density on hydrogel formation, adipose-derived stem cells (ADSCs) were encapsulated within the osteoinductive composite hydrogels, and the resulting constructs were assessed for biological performance and osteogenic differentiation. Also, to reproduce a multicomponent hydrogel and overcome similar problems, we carefully discuss the various fabrication factors and technical and biological aspects considered in our hydrogel scaffold and also outlined the various challenges.

Results

Characterization of extracted and mineralized fibroin fibers

Figure 1a shows the FTIR spectra of the primary fibroin fibers, heat-treated fibroin, and mineralized fibers before and after heat treatment.

Fig. 1
figure 1

(a) FTIR spectra of primary fibroin fibers, heat-treated fibroin, and mineralized fibers before and after heat treatment. (b) XRD spectra of primary and mineralized fibroin fibers before and after heat treatment. (c) Fourier self-deconvoluted (FSD) curve of native (unheated) fibroin fiber. (d) Fourier self-deconvoluted curve of heated fibroin fiber. (e) Curve deconvolution of non-heated mineralized fibroin (f) Curve deconvolution of heated mineralized fibroin. In the deconvoluted curves, the red solid curve is the deconvoluted spectrum, and the dotted curve shows the resultant amide I spectrum. In the FSD curves, the abbreviations stand for side chains (S), (BSS), β-sheets (B), random coil (R), α-helix (A), and turns (T).

Full size image
Table 1 FTIR characteristic vibrations of silk fibroin and hydroxyapatite.
Full size table

The IR region around 1500–1700 cm−1 including amide II (1500–1600 cm−1) and amide I (1600–1700 cm−1) is usually used to analyze different secondary structures of the silk fibroin. Table 1 includes the positions of different characteristic peaks for silk fibroin and hydroxyapatite. According to FTIR results, the major stretching and bending vibrations of amide I, amide II, amide III, and amide IV were detected for secondary silk I and silk II structures. According to Fig. 1a, heat treatment had no negative effect on the SF fiber, and its chemical structure remained intact. HAp formation causes the stretching and bending vibrations of the \({\text{PO}}_{4}^{{3 - }}\) group (Fig. 1a; Table 1). When the mineral phase is synthesized in the air, partial substitution of the phosphate group with carbonate occurs by dissolution of CO2, resulting in specific peaks of type B carbonate (\({\text{CO}}_{3}^{{2 - }}\))15. Although the Ca/P molar ratio was set at 1.67 for HAp, the initially formed mineral phase exhibited an amorphous structure that gradually became crystalline over time. Heat treatment accelerated the crystallization of the HAp structure, representing distinct phosphate peaks, especially in the range of 900–1100 cm−1 as the characteristic and main peaks of hydroxyapatite15.

Figure 1b shows the X-ray diffraction (XRD) spectra of primary and mineralized fibroin fibers (before and after heat treatment). When the SF fibers are prepared in an aqueous medium, they show the XRD peaks of silk I at the angles of 11.3, 14.4, 18, 19.8, 20.4, 22.2, 25.4, and 29.4°19, and β-sheets in silk II at the angles of 8.5, 20.4, and 24.6°17. Additionally, the water-annealing of the SF fibers (i.e. putting them in water for a long time) changes its crystalline structure from a single-phase state to a mixture of silk I and silk II19. Therefore, the primary SF fibers exhibited a mixed crystalline structure and mineralization intensified the silk I peaks because each stage of mineralization (including 10 cycles of solution exchange) lasted about 3 h. Comparison of XRD results with ICDD 09-0432 confirmed the characteristic diffraction peaks of hydroxyapatite at 25.9, 31.77, 32.2, and 32.9°15 for non-heated mineral coating. The heat treatment intensified the XRD peaks owing to the formation of a higher crystalline structure. The degree of crystallinity (XC) increased from 0.000228 to 0.001717, confirming the positive effect of heat treatment on crystallinity by about 654.35%.

For better evaluation of heating effect on the SF structure, the amide I region of the FTIR spectrum was deconvoluted to resolve secondary structural peaks. Figure 1c and d illustrate the deconvoluted curve fitted on the amide I regions of native (unheated) and heated SF fibers, including all expected secondary structural bands for an ordered structure. Also, the relative area percentage (RA%) of each band relative to the total area of analyzed FTIR region was calculated (Table 2). Although the amide I region is usually considered as 1595–1705 cm−168, the analyzed region was extended to 1750 cm−1 for better curve fitting. The area of Tyr side chains (S band) did not considerably changed by heating. The area of BSS peak (centered at about 1610 cm−1 ) related to intermolecular β-sheet formation and molecular aggregation68 decreased by about 45%. The B1 band (centered at about 1624 cm−1), a well-known band in the most crystallized proteins68 that might be merged with its adjacent BSS band, considerably intensified by heat treatment.

Table 2 The relative area of deconvoluted bands fitted on the FTIR amide I region of original and heated silk fibroin fibers.
Full size table

However, the area of another β-sheet band at the end of the region (B2 band, centered at about 1705 cm−1) decreased because of heat treatment. However, the total area of β-sheet bands (B1 + B2) of heated SF fiber decreased slightly (− 3.918%) compared to unheated fibers. Several bands ascribed to random coil, α-helix, and turns structures are located between major triplet β-sheet bands, whose area changed by heat treatment. Overall, with no considerable change in the area of β-sheets, it could be concluded that heat treatment did not negatively influence on the SF structure.

By performing curve deconvolution on the XRD spectra of non-heated and heated mineralized SF fibers (Fig. 1e, f), the area under the 31.7°-centered peak and total area under deconvoluted curves increased by 52.91 and 65.41%, respectively, confirming the positive effect of heat treatment on the crystallinity improvement.

Fig. 2
figure 2

(i) Optical microscope images (a,b) Silk fibers (c,d). Fibroin fibers (e,f). Mineralized fibroin fibers before heat treatment (g,h). Mineralized fibroin fibers after heat treatment, (j) fibroin fibers after 10 cycles of mineralization, (k) fibroin fibers after 20 cycles of mineralization. (ii) SEM images of fibroin fibers (a) raw fiber before heat treatment, (b) raw fiber after heat treatment, (c) mineralized fiber before heat treatment, (d) mineralized fiber after heat treatment (scale bar = 50 μm).

Full size image

Figure 2i shows the optical microscope images of the SF fibers before (Fig. 2ia–d) and after mineralization (Fig. 2ie–k). Before the heat treatment and because of two separate mineralization steps (Fig. 2if), the HAp layers were formed as two unbonded layers which did not have enough integrity, and could detach from the SF surface. However, heat treatment improved the crystallinity of the apatite phase and integrated two mineral layers, resulting in a complete coating around the SF fibers (Fig. 2ih). The first HAp layer was thin (Fig. 2ij) and acted as a good substrate for the nucleation and growth of a new HAp layer after partial drying, resulting in a thicker coating (Fig. 2ik). Figure 2ii shows the SEM images of primary and mineralized SF fibers before and after heat treatment. The SF fibers exhibited an average diameter of 10–15 μm and there were no visible changes in shapes and surface morphology were observed after heat treatment (Fig. 2iia, b). Mineralization formed a uniform HAp layer (a thickness of about 6–7 μm) on the fibers (Fig. 2iic) and heat treatment did not cause any surface damage in the coating (Fig. 2iid).

Best photocrosslinking condition for optimized hydrogel scaffold

After performing the post-Taguchi analysis, the effect of each factor was determined by the main effect and signal-to-noise ratio (SNR) (Fig. 3), and the condition of maximum SNR was considered to analyze the MTT results and maximizing cell viability.

Changes in all factors, Eosin Y (EY) concentration, curing time (exposure time to green light), and dithiothreitol (DTT) concentration, influenced cell viability. Despite the least negative effects on the cell viability at the minimum and maximum concentrations, the maximum EY concentration was chosen to ensure gel formation upon light exposure. Similarly, the maximum curing time and DTT were selected to guarantee the formation of the gel network. Because triethanolamine (TEOA) concentration greater than 0.25% (w/v) prevented the formation of a stiff gel, 0.25% (w/v) was considered as the maximum concentration for all samples. Although the extracts after considered time points (Ext-time) showed negative effects on the SNR, the cell viability percentages were usually above 70%, indicating that the negative influence of hydrogel-derived components diminished over time.

Fig. 3
figure 3

(a) Mean effects and (b) signal-to-noise (SN) ratios of the factors involving in the hydrogel preparation.

Full size image

Characterization of the cell-free hydrogel scaffolds with and without SF fibers

Figure 4i presents the FTIR spectra of the hydrogels, before and after immersion in PBS, with and without DTT as a crosslinker, and Table 3 includes the positions of characteristic peaks. Detailed characterization of original and methacrylated chitosan by FTIR and 1HNMR were reported in our previous article11. The deacetylation degree of pure chitosan was about 94%, and the methacrylation degree was calculated about 50% according to 1HNMR result.

Table 3 FTIR characteristic peaks for chitosan, methacrylated chitosan, Eosin Y, triethanolamine, and dithiothreitol.
Full size table

Some important characteristic bands of chitosan were marked in Fig. 4i. Binding methacrylate group to the main chitosan chain causes the specific methacrylate peaks such as C = CH2 in the MECs spectrum. Because chitosan methacrylation usually occurs through NH2 group, a peptide bond is formed, and some specific peaks for C–N bond in the amide I and CONH group are intensified. The intensity of C–N bonds increases by binding the methacrylate group through the amide bond between the carbonyl group of methacrylate and the nitrogen atom of the primary amine group (NH2). Also, methacrylate groups can be connected to chitosan through the OH group and form an ester bond with or without EDC/NHS crosslinker.

Fig. 4
figure 4

(i) FTIR spectra of fiber-free hydrogel with and without DTT as crosslinker in PBS medium. (ii) Dimensional changes of hydrogel scaffolds before and after placing in PBS for 72 h.

Full size image

Eosin Y (EY) and triethanolamine (TEOA) as photoinitiator and co-initiator remain in the hydrogel structure after photo-crosslinking, caused their FTIR characteristic bands that were identified (Fig. 4i; Table 3). For dithiothreitol (DTT), if –SH remains, its stretching vibrations in DTT monomers and dimmers may appear. Overall, the characteristic bands of the remaining DTT molecule (–SH group) and the formation of its –S–S– based structures (ring, dialkyl sulfide or disulfide) were not observed and confirmed.

Figure 4ii shows dimensional changes of hydrogel scaffolds with and without SF fibers after 72 h immersion in PBS, and the calculated shrinkage percentages in diameter, length, and volume are listed in the Table 4. PBS-treated hydrogel scaffolds shrank and the calculated shrinkage percentages for the fiber-free and fiber-enriched (0.1, 0.2, and 0.3% (w/v)) scaffolds confirmed that the SF fibers improved structural resistance to shrinkage, and higher fiber percentages resulted in statistical lower shrinkage in dimensions and volume (P < 0.0001) compared with control (fiber-free) group.

Table 4 Dimensional shrinkage percentages and pore diameters for fiber-free and fiber-enriched hydrogels.
Full size table

Figure 5i shows the SEM images of PBS-treated hydrogels for 72 h to check the effect of the ion-containing medium on the structure. According to Fig. 5i, a more regular structure and more bindings were formed using DTT as a crosslinker. Both scaffolds with and without DTT shrank in several areas (as wrinkles) after placing in PBS. Figure 5ii displays the SEM images of scaffolds containing different percentages of SF fibers before and after 72 h PBS immersion. Table 4 includes the minimum, the maximum, and the average calculated diameters of pores in the fiber-free and fiber-enriched hydrogels. The minimum and maximum pore diameters were calculated from initial data extracted by image analysis before data processing to find normal pore distribution. Although pores with clearly defined openings were selected for diameter calculation, Fig. 5ii shows that many pores that have diameters smaller or larger than minimum and maximum diameters can be found throughout the hydrogel structure. The addition of the SF fibers statistically decreased the pore size compared with fiber-free hydrogel (P < 0.05 for 0.1% (w/v) and P < 0.0001 for 0.2 and 0.3% (w/v)) and prevented considerable contraction of the scaffolds after PBS exposure. Small wrinkles on the pores’ walls and the accumulation of small shrinkages cause the final lower contraction of the fiber-enriched hydrogels.

Cellular assessments of cell-laden photocrosslinkable hydrogels

Determination of suitable cell density for encapsulation

Because cells can physically interfere with hydrogel network formation, different cell densities were encapsulated and evaluated. All hydrogels, cell-free or cell-laden, maintained structural integrity and stiffness post-molding, allowing easy transfer to culture medium. After 24 h of incubation under standard conditions (37 °C, 5% CO2, 80% humidity), the hydrogels containing 1 × 105 cells/mL remained intact, while those containing 5 × 105 and 1 × 106 cells/mL disintegrated. At higher densities, the scaffolds fractured into two pieces, or showed tearing of the outer wall when suspended.

Fig. 5
figure 5

(i) SEM micrographs of fiber-free hydrogel scaffolds with and without DTT as crosslinker and simultaneous investigation of the effect of environmental ions on the structure by putting the scaffolds in PBS. (ii) SEM images of hydrogel scaffold before and after placing in PBS.

Full size image
Fig. 6
figure 6

Hematoxylin-Eosin (H&E) and Periodic acid–Schiff (PAS) staining to evaluate the effect of cell density on hydrogel formation.

Full size image

According to the H&E and PAS staining (Fig. 6), the cell-free hydrogels exhibited a uniform texture, higher crosslink density caused higher staining intensity, and unavoidable air bubbles, formed by mixing the polymer solution, appeared clearly without staining. In contrast, cell-laden hydrogels showed morphological variation in cells due to their spatial orientation, and a general pattern as a bright area (with lower staining intensity) surrounded each cell, indicating a weaker network compared with holes caused by air bubbles. Despite some damage during staining owing to the weakening of the hydrogel network, structural evidence showed that cell presence affects network formation. The integrity of main hydrogel bulk was preserved at low cell density, and different crosslink densities confirmed by the H&E and PAS staining. Increasing cell density and proximity supports the hypothesis that cells act as a physical barrier. Thus, a density of 2.5 × 10⁵ cells/mL was chosen to overcome or minimize this effect. Supplementary Movie S1 shows the distribution of mineralized SF fibers and encapsulated ADSCs post-hydrogel formation.

Fig. 7
figure 7

(a) Light microscopic images of encapsulated cells in the photocrosslinked hydrogels with and without fibroin fibers (b) changes in the amount of released alkaline phosphatase (ALP) after 3, 7, and 14 days (c) the percentages of lactate dehydrogenase (LDH)-based cytotoxicity and viability (d) relative gene expressions by real time-PCR assay. MHydro, FHydro and mFHydro stand for MEC-based hydrogel scaffold, fibroin-containing MHydro, and MHydro containing mineralized fibroin fibers, respectively.

Full size image

Cellular morphology, viability, and osteogenic gene expression

The light microscope images of encapsulated cells on days 7 and 14 (Fig. 7a) showed many round viable cells distributed throughout the fiber-free hydrogel (MHydro sample). Fibroin-enriched hydrogel (FHydro sample) also showed cells with round morphology on day 14. Although cells did not show typical adherent morphology of 2D cell culture or tissue, their appearance suggested viability. The mineralized SF fiber hydrogel (mFHydro sample) exhibited fewer visible cells because of mineral particles, though viable round cells were identifiable on days 7 and 14. To investigate viability and activity of encapsulated ADSCs, alkaline phosphatase (ALP), a positive osteogenic marker, and lactate dehydrogenase (LDH), as a cytotoxicity indicator, were quantified in the culture medium. Figure 7b shows ALP quantity after 3, 7, and 14 days released from different hydrogel scaffolds. On day 3, MHydro exhibited the highest ALP activity (25.00% higher than other groups, P < 0.01). ALP levels in MHydro and mFHydro increased on day 7 compared to their ALPs on day 3 by 4.00% and 25.00% (P < 0.01), respectively, while FHydro remained constant. On day 14, the ALP activity significantly increased in MHydro, showing 33.00% and 38.30% higher levels than same values on days 3 and 7 (P < 0.0001). SF fibers enhanced ALP expression in FHydro by 33.00% on day 14 compared with earlier time points (P < 0.001). Mineralized SF fibers (mFHydro) showed ALP levels on day 14 higher than on days 3 and 7 by 66.25% and 33.00%, respectively (P < 0.0001), and 25.00% higher than FHydro (P < 0.001). Since LDH cytotoxicity (Fig. 7c) was below 10%, the ALP activity and LDH cytotoxicity proved an acceptable cytocompatibility for ADSCs encapsulated into all hydrogel scaffolds. Osteogenic differentiation was further evaluated by analyzing Runx2, Col1A1, and Osteocalcin (OCN) gene expression (Fig. 7d) after 7 and 14 days. On day 7, Runx2 was significantly (P < 0.0001) upregulated in all hydrogels compared with control (ADSCs) (196.40%, 206.70%, 395.80% for MHydro, FHydro, mFHydro). Despite nearly similar Runx2 expressions for MHydro and FHydro, mineralization significantly enhanced Runx2 compared to MHydro (P < 0.0001) and FHydro (P < 0.001) by 67.27% and 61.66%, respectively. On day 14, Runx2 expression remained elevated in all hydrogels but the trend of Runx2 expression changed (160.50%, 363.50%, 207.40% for MHydro (P < 0.01), FHydro (P < 0.0001), and mFHydro (P < 0.0001), respectively), and FHydro showed the highest Runx2 expression compared with MHydro (P < 0.0001) and mFHydro (P < 0.01) on day 14 (77.93 and 50.78%, respectively) and FHydro on day 7 (51.12%, P < 0.01). SF fibers increased Col1A1 expression on day 7 compared with control and MHydro (P < 0.05) by 28.00% and 25.17%, respectively, while mineral phase in mFHydro significantly (P < 0.0001) upregulated Col1A1 compared with the control group (407.40%) and other hydrogels (661.86 and 296.41% for MHydro and FHydro, respectively). On day 14, Col1A1 expression was significantly downregulated in all hydrogels compared with control (P < 0.01) and values of day 7 (-25.08, -97.50, and − 95.70% for MHydro, FHydro (P < 0.01), and mFHydro (P < 0.0001), respectively). Downregulation in FHydro (P < 0.01) and mFHydro (P < 0.05) was also significant compared with MHydro (-59.93%, -56.91%, respectively). None of the hydrogels induced osteocalcin (OCN) expression after 14 days, although SF fibers with or without mineralization slightly enhanced OCN expression.

Discussion

In this study, SF fibers were successfully coated with a HAp microlayer, the HAp coating was characterized using various methods, and heat treatment improved its crystallinity. MECs was synthesized through an optimized route, the best photocrosslinking conditions were identified, and incorporation of mineralized SF fibers into the hydrogel resulted in a porous and shrinkage-resistant scaffold. ADSCs encapsulated in the MECs-based hydrogel (with and without SF fibers) remained viable for 14 days, and the mineralized SF fibers upregulated osteogenic gene expression.

Bone matrix consists of type I collagen, which serves as a template for mineralization, and can be replaced with suitable polymers that mimic bone architecture. Silk fibroin is an appropriate natural polymer due to its biocompatibility, good cell adhesion, and low immunogenicity83. Since fibers are expected to provide greater mechanical reinforcement than particle-reinforced materials66,67, SF fibers were chosen for their mechanical support and inherent osteogenic potential. HAp-coated SF fibers can serve as substitutes for collagen fibers in natural bone mineralization. Moreover, SF fibers are easy to obtain and prepare, making the composite more time- and cost-efficient compared to other forms (particles, nanofibers) that require additional processing.

The alternate (cyclic) soaking process is a rapid and efficient coating method in which the target substrate is repeatedly immersed in Ca2+- and \({\text{PO}}_{4}^{{3 - }}\)-containing solutions2. Surface hydroxyl and carboxyl groups enhance hydrophilicity84, while carboxylate groups serve as nucleation sites for CaP deposition2,21 through ion exchange with Ca2+ to form complexes such as –(COO−)2Ca2+ or –[(COO−)Ca2+]84. Initial HAp nucleation also involves weaker ion–polar interactions between Ca2+ and peptide carbonyl groups (C=O) in the silk fibroin, as well as hydroxyl groups on its surface22. Upon immersion in the \({\text{PO}}_{4}^{{3 - }}\)-containing solution, \({\text{PO}}_{4}^{{3 - }}\) ions bind to surface-bound Ca2+84, indicating that Ca2+-surface interactions govern deposition process2. Optical microscopy shows partial HAp coverage after the first step (Fig. 2ij) and complete coverage following the second (Fig. 2ik).

Silk fibroin (SF) fibers contain hydrophobic and hydrophilic domains, with the hydrophilic regions composed of less conserved (GAGAGX) and highly conserved (GAGAGS) motifs (G = glycine, A = alanine, S = serine, X = valine/tyrosine)23. Crystalline domains are rich in glycine, alanine, and serine, while amorphous domains contain amino acids with bulkier side chains85 that serve as HAp nucleation sites and mimic type I collagen in natural bone21. Sericin, a secondary silk protein, is rich in polar amino acids (consisting of strong polar groups such as carboxyl, amino, and hydroxyl groups)86 that, in order of abundance, include serine, aspartate, glycine, threonine85,86,87, glutamate85,86,87,88, alanine, tyrosine, and arginine86. Although sericin removal (degumming) enhances the biocompatibility of fibroin-based scaffolds by reducing immune reactions18,21, complete elimination of sericin can reduce the number of surface functional groups available for HAp nucleation. Consequently, non-uniform degumming may explain the sparse HAp layer formed during the first mineralization step (Fig. 2ij), followed by complete surface coverage after the second (Fig. 2ik). Washing after each soaking cycle prevents HAp nucleation in solution, and heterogeneous nucleation is favored both thermodynamically and kinetically due to lower interfacial energy between the substrate surface and a growing nucleus than a small nucleus in the solution2.

During alternate soaking process, initial CaP deposition consists of loose pre-nucleated aggregates that densify over time89, forming metastable phases such as octacalcium phosphate, dicalcium phosphate dihydrate, and amorphous calcium phosphate (ACP)90,91. XRD analysis (Fig. 1b) and peak deconvolution (Fig. 1c, d) confirmed that heat treatment enhanced HAp crystallinity. Broad peaks corresponded to low-crystalline and non-stoichiometric HAp, indicating a mixture of crystalline HAp and ACP92. As ACP transforms into crystalline HAp over time93, even non-heat-treated samples showed partial crystallinity due to the delay before XRD measurement. Although XRD resolution limited crystallite size estimation, SEM images (Fig. 2iic, d) revealed ultrafine and uniformly distributed HAp particles. The low crystallinity and nanoscale size of carbonated HAp coating can resemble natural bone minerals92,94. Silk fiber remains structurally stable up to 150°C95, enters a glassy state at ~ 178 °C96, and degrades near 250°C95. In addition, silk I and silk II can be degraded at approximately 271 °C and 276 °C, respectively96. Therefore, both FTIR (Fig. 1a) and XRD confirmed that heat treatment at 220 °C did not negatively influence SF fiber integrity.

The effects of various key factors, influencing the formation of the photocrosslinkable hydrogel, on ADSCs were evaluated. Although some factors negatively affected cell viability (Fig. 3), these effects could be ignored by considering the necessary conditions for the hydrogel scaffold and its intended application. Despite a slight decrease in viability, a higher DTT concentration was selected because DTT contributes to the hydrogel network and any residual molecules can be inactivated by pH and temperature. Additionally, the adverse effect of prolonged light exposure was considered insignificant, since light must penetrate the entire depth of the large target scaffold (an 8–10 mm diameter, 15 mm length cylinder) to ensure proper hydrogel network formation. Removing potentially harmful substances can improve viability over time, and cell behavior analysis confirmed cell survival up to 14 days along with their participation in osteogenic differentiation. Overall, biocompatibility of the hydrogel fabrication system can be effectively controlled. Chitosan, a linear amino polysaccharide, is derived by deacetylation of chitin10, and mimics the structures of proteins and amino acids97. Since inter- and intra-chain hydrogen bonds between hydroxyl groups and the carbonyl of acetyl groups stabilize chitin’s conformation97,98, methacrylated chitosan may resemble chitin due to similarities between acetyl and methacrylate groups. This similarity could promote hydrogen bonding and partial secondary structure formation under suitable conditions11. Despite Hofmeister’s ion ranking describing to ion effect on protein solubility in 188899,100, subsequent studies confirmed the agreement of other macromolecules and properties100. Various aqueous ions can induce intramolecular hydrogen bonding between methacrylate carbonyls and adjacent hydroxyl groups11. This bonding may explain the shrinkage and size reduction observed in hydrogel scaffolds in PBS (Fig. 4ii), resulting in localized contractions and wrinkled pore walls (Fig. 5).

In the human body, structural gradients are mainly found at the interface between tissues. Bone exhibits structural gradients, with density and porosity varying from spongy to cortical bone101. Its integrity depends on porosity and pore size102, including Haversian/Volkmann canals (~ 50 μm for healthy bone), quasi-ellipsoidal osteocytic lacunae (few µm), and canaliculi (channel networks between the lacunae, < 1 μm)103. From the point of view of bone tissue type, cortical bone porosity ranges 20–100 μm in diameter102, and porosity with smaller diameter than 20 μm are attributed to smaller Haversian canals or lacunae104. Spongy bone usually has an irregular porous structure with pore size of 300–500 μm, while its pore size varies depending on its location within the body102,105. Adding SF fibers with DTT double-crosslinking produced more regular pores, reducing size from ~ 92 μm (MHydro) to ~ 64 μm (mFHydro). Although ~ 100 μm pores are preferred for cell penetration and bone ingrowth106, some studies show 50–125 μm can suffice under non-load-bearing conditions, and 75–100 μm pores support the ingrowth of unmineralized osteoid tissue107.

In this study, relatively large-dimension scaffolds (8–10 mm in diameter and 15 mm in length) were fabricated without cell encapsulation. However, incorporation of a high density of encapsulated cells compromised the structural integrity of the hydrogel matrices. Mesenchymal stem cells, with an average diameter of 10–15 μm, likely act as physical barriers that interfered with effective crosslinking between polymer chains via functional groups. Histological analyses using H&E and PAS staining (Fig. 6) revealed that encapsulated ADSCs formed spherical aggregates, adversely impacting the surrounding microenvironment and inhibiting the formation of a robust hydrogel network. This disruption resulted in localized weakening and subsequent hydrogel disintegration. Upon immersion in culture medium, contraction induced by hydrogen bonding caused condensation of the entrapped fluid within the scaffold, and the weakened outer walls were unable to resist the resulting internal pressures. Therefore, a minimum ADSC number was determined after 24 h to maintain the structural integrity of fiber-free hydrogel. This optimized cell density was used for fabricating fiber-enriched hydrogels. If long-term assessment of cell density effects was done, fiber-enriched hydrogels could not be intact and the validity of further analyses was undermined. Therefore, all experimental evaluations were confined to a 14-day period, during which multiple properties of the composite hydrogel scaffolds were examined. Obviously, if appropriate conditions can be provided so that a higher cell density can be used without worrying about the hydrogel disintegrating in the long term (e.g. in vivo or clinical applications with enclosed bone defects), better results are expected.

To encapsulate ADSCs in hydrogel scaffolds, the cells were suspended in FBS instead of culture medium, as FBS supplies essential nutrients and reduces mechanical and chemical stress during mixing with the polymer solution. Microscopic analysis (Fig. 7a) showed that encapsulated ADSCs maintained a rounded morphology after 14 days, unlike the typical spread shape of adherent cells reported in various studies. Since no 3D SEM images of cells in hydrogels exist, the spatial arrangement of cells within pores or their entanglement in the gel mass remain unclear. Although sample shrunk and SEM imaging was not feasible for assessing cell morphology, H&E and PAS staining (Fig. 6) revealed that the intensity of the dye around the cells was much lower than in other parts of the hydrogel, meaning that the cell was surrounded by a weak network and a soft matrix. Different views have been presented regarding the round morphology of cells encapsulated within hydrogels. Some studies have shown that although substrate stiffness generally affects cell morphology in 2D scaffolds, MSCs in 3D constructs might maintain their round morphology regardless of hydrogel stiffness, indicating natural adaptation to the 3D microenvironment108,109. Other studies have shown that the round morphology of cells encapsulated in hydrogels depends on the degradation of the matrix by the cells, such that the cell remains round if it cannot degrade the surrounding matrix110,111,112.

Viability analysis of encapsulated ADSCs (Fig. 7c) revealed reduced cell viability in fiber-enriched hydrogels (FHydro and mFHydro) compared to the fiber-free sample (MHydro). Although cells were gently and manually mixed with the polymer solution to minimize mechanical stress, interactions with the fibers, particularly the rough mineral coating, and cell–fiber collisions likely contributed to reduced viability. This limitation appears inherent to the current hydrogel preparation method. Certainly, cell viability and ALP activity may not always change in perfect parallel, and ALP upregulation can non‑synchronously occur in MSCs where cell proliferation or viability does not increase correspondingly. For example, resveratrol treatment in human periosteum‑derived MSCs led to a decrease in proliferation/viability at higher doses, while significantly increasing ALP activity during osteogenic induction113.

The non-fiber-enriched chitosan hydrogel (MHydro) significantly improved Runx2 expression at 7 and 14 days, but not Col1A1 (Fig. 7d). Chitosan resembles glycosaminoglycans in connective tissue ECM, and promotes MSC attachment, proliferation, and osteo-differentiation3. Although amine group (NH2) promotes cell attachment25 and some functional groups such as amine, hydroxyl (OH), and methyl (CH3) can enhance Runx2 and COl1A1 expression114, the direct osteogenic effect of chitosan is unconfirmed and it may act via non-specific cell adhesion25. In the silk fibroin-enriched hydrogel (FHydro), Runx2 expressions at day 3 and 7 were nearly similar but increased after 14 days. Reports on the osteogenic potential of silk fibroin are inconsistent, some describe weak osteogenic differentiation, while others confirm upregulation of osteogenesis-related genes. Its β-sheet and amide groups provide a stiff matrix that promotes osteogenesis21. Although Notch signaling maintains osteoprogenitors and inhibits their osteo-differentiation115, silk protein suppresses Notch pathway, raising some osteogenic markers such as ALP, Runx2, and osterix21. Moreover, B. mori silk stimulates mineral deposition by bone marrow stromal cells in vitro23. However, plain SF scaffolds fail to fully repair large bone defects in vivo116, and pre-seeding with stem cells or adding bioactive inorganic materials improves bone formation21. Surface-modified SF fibers in mFHydro enhanced osteogenic properties. Compared with FHydro containing non-mineralized SF fibers, ALP activity of encapsulated ADSCs increased over time and osteo-differentiation was better regulated (Fig. 7b). The HAp coating on SF fibers likely improves cell-matrix interactions and osteo-differentiation, although its mechanism is unclear23. HAp upregulates ALP and COl1A1 gene expressions117, and stimulates osteoblast to synthesize more bone matrix118. Because OCN is expressed in highly differentiated cells119, it is obvious that differentiated ADSCs have not yet entered the mineralization phase after 14 days. Matrix stiffness can activate transcription factors that regulate genes for specific cell-lineage differentiation. Higher stiffness enhances actomyosin contractility120 and promotes MSCs to differentiate into osteoblasts121. Compressive moduli below 25 kPa do not provide enough cytoskeletal tension in cells cultured on or encapsulated in the stiff matrices to induce differentiation, requiring osteogenic ligands for Runx2 expression120.

Conclusions

In this study, a fiber-enriched photocrosslinkable MECs-based hydrogel, incorporated with ADSCs, was developed for bone regeneration. The SF fibers were extracted and fully coated by a HAp microlayer to integrate the inherent osteogenicity of the silk fibroin with the bioactivity of HAp. Heat treatment improved the crystalline structure of HAp while it did not negatively affect SF structure. Adding the various amounts of the SF fibers to the basic chitosan hydrogel (MHydro) and using an extra DTT crosslinker decreased the volume shrinkage of composite hydrogels under ionic conditions. Incorporated ADSCs into MHydro and SF-enriched hydrogels were alive for 14 days. Also, the SF fibers in the SF-enriched hydrogels upregulated bone-specific genes, and HAp mineral enhanced gene expression more. The composite MEC-based hydrogel, enriched with the mineralized SF fibers and incorporated with ADSCs, can be a promising 3D in-situ forming and injectable structure to fill bone defect sites without surgical intervention.

Materials and methods

All methods were performed in accordance with the relevant guidelines and regulations.

Synthesis of methacrylated chitosan

Previously, we presented an optimized route for synthesizing methacrylated chitosan and discussed the structural characterizations and synthesis efficacy in detail11. Chitosan powder (Cs, Sigma-Aldrich, USA) was dissolved in diluted methacrylic acid and the pH was adjusted to about 5.8–6. After warming the solution up to 60 °C, the required amount of methacrylic anhydride (MAH) (equivalent to the number of functional groups of Cs) and the calculated amount of EDC/NHS as crosslinker was added at once, the solution was stirred for 6 h and kept for extra 18 h at room temperature. Finally, the solution was dialyzed against distilled water for 48 h at room temperature, freeze-dried, and stored at − 20 °C until further use. By considering the deacetylation degree (DD) and calculating the mole of functional group (nf) for the known amount (m) of Cs, the theoretical required volume of MAH and the amount of EDC/NHS can be calculated by the following equations11. :

$${{\text{n}}_{\text{f}}}=\left[ {\left( {m \times {{\text{D}}_{\text{D}}}} \right)/\left( {161 \times {{\text{D}}_{\text{D}}}+365 \times (1 - {{\text{D}}_{\text{D}}})} \right)} \right] \times 100$$
(1)
$${{\text{V}}_{{\text{MAH}}}}=330.802 \times {{\text{n}}_{\text{f}}}$$
(2)
$${{\text{m}}_{{\text{EDC}}}} \approx 220 \times {{\text{n}}_{\text{f}}} \times {\text{p}}$$
(3)
$${{\text{m}}_{{\text{NHS}}}} \approx 115 \times {{\text{n}}_{\text{f}}} \times {\text{p}}$$
(4)

Where p stands for the percentage of functional groups that are modified by crosslinking. The DD value of native Cs was determined by 1HNMR.

Extraction, mineralization, and characterization of silk fibroin (SF) fibers

The B. mori silk fibers were cut into sections (length < 1 mm) and stirred in boiling 0.02 M sodium carbonate solution for 1 h to remove sericin. Then, the fibers were drawn out, washed with cold distilled water (25 °C) twice, stirred in distilled water for 20 min, and air-dried18.

The SF fibers were coated with a layer of calcium phosphate (CaP) using an alternating soaking method. By keeping the Ca/P molar ratio at ~ 1.67, CaCl2 and (NH4)2HPO4 solutions were prepared at concentrations of 600 and 360 mM, respectively. The SF fibers were soaked in each solution for 6 min and washed with deionized water twice after each soaking step to remove remaining Ca2+ and \({\text{HPO}}_{4}^{{2 - }}\) ions and prevent CaP precipitation among fibers. To form a complete CaP coverage on the SF fibers, two coating sets were performed including 10 cycles of changing the solutions in each set and 24 h rest time between each set to dry the new CaP layer well. Finally, a heat treatment regime was done to increase crystallinity: (1) heating slowly up to 220 °C at a heating rate of 2 °C/min (2) maintaining for 5 h isothermally (3) cooling down slowly in the furnace to reach the room temperature. Despite the suitable thermal stability of SF up to 250°C122, uncoated SF fibers were similarly heated to ensure SF thermal stability. The morphology and structures of coated and uncoated SF fibers were analyzed by optical microscope (BEL INV2, BEL Engineering, Italy), scanning electron microscope (SEM, TESCAN VEGA-II, Czech Republic), X-ray diffraction (XRD, CuKα1 = 1.54059 Å, Bruker D8 advance, Germany), and Fourier transform infrared spectroscopy (FTIR, KBr method, PerkinElmer Spectrum One, USA). To evaluate the effect of heat treatment on the SF structure, the amide I regions in the FTIR spectra of original (unheated) and heated SF fibers were analyzed by Fourier self-deconvolution (FSD) method to separate secondary structural peaks and identify their relative contribution to the overall sample structure123. The absorption spectra of amide I were derived from the FTIR spectra, Gaussian peaks (ascribed to Tyr side chains (1605–1615 cm−1), aggregate β-strand/β-sheets (1616–1621 cm−1), β-sheets (1622–1627, 1628–1637, and 1697–1703 cm−1), random coils (1638–1646 and 1647–1655 cm−1), α-helices (1656–1662 cm−1), and turns (1663–1670, 1671–1685, and 1686–1696 cm−1)) were fitted on them68, and normalized to the total area of analyzed FTIR range to calculate the fraction of each band. Additionally, peak analysis was performed on the peak at about 2θ = 25.6° and the degree of crystallinity (XC) was calculated from the following formula15 to evaluate the effect of heat treatment on structural crystallinity of the CaP coating:

$${{\text{X}}_{\text{C}}}={\left( {\frac{{{{\text{K}}_{\text{A}}}}}{{{{\text{B}}_{1/2}}}}} \right)^3}$$
(5)

where KA = 0.24 and B1/2 is the FWHM (full width at half maximum) of the corresponding peak (in degrees).

Because XRD characteristic peaks of deposited CaP and fibroin were merged, curve deconvolution was also used to calculate crystallinity degree to more reliably evaluate the effect of heat treatment on structural crystallinity. Because the strongest XRD peak of hydroxyapatite can be centered at about 31.7°, the XRD data in the range of 31–32° was smoothed partly and deconvolution curves were found. Then, the area under the 31.7°-centered peak and the total area under all deconvoluted curves were measured to evaluate the effect of heat treatment on crystallinity enhancement.

Determining the best photocrosslinking condition based on cell viability

The concentrations of the polymer, photoinitiator, and crosslinker as well as duration of photoexposure are important factors that might especially affect cellular viability and their suitable amounts have to be determined. Therefore, various parameters based on their effects on cellular viability were evaluated using the MTT assay. By considering the biocompatibility of MECs24 and triethanolamine (TEOA, as co-initiator) up to 5%124, their concentrations were fixed at 0.7% (w/v) and 0.25% (v/v), respectively. Different levels were chosen for curing time, concentrations of Eosin Y (EY, as photoinitiator, Sigma-Aldrich, USA), and dithiothreitol (DTT, as crosslinker, Sigma-Aldrich, USA). Also, four extraction times (1, 3, 5, and 7 days) were considered for preparing sample extraction according to ISO 10993-5 and 12 (Table 5). Finally, the Taguchi method including a standard L16 array (Table 6) was used to design the experiment by Minitab software.

According to ISO 10993-5 and 12, each hydrogel sample was placed in FBS (fetal bovine serum)-free DMEM cell culture medium (Biowest, USA) for considered extraction time to get their extracts.

Table 5 Selected factors and their levels.
Full size table
Table 6 Tagouchi orthogonal array of designed experiment.
Full size table

Human ADSCs (isolated from a surgically removed human abdominal adipose tissue according to the ethical approval (IR.PII.REC.1399.005) from the ethics committee of the Pasteur Institute of Iran and informed consent from the patient) were seeded in a 96-well plate (1 × 104 cells/well) and cultured for 24 h under standard culture conditions (37 °C, 5% CO2, 80% humidity) before treating with hydrogel extracts. After discarding the supernatant, 100 µl of each extract (without adding FBS) was added to each well (5 wells for each sample), cultured for 24 h under standard culture conditions, the MTT assay was performed, and an ELx808 ELISA reader (BioTek, USA) read the absorbance at 570 and 630 nm to calculate cell viability percentage (Viab%).

$${\text{Viab}}\% =\left( {{\text{O}}{{\text{D}}_{{\text{sample}},570\;{\text{nm}}}} - {\text{O}}{{\text{D}}_{{\text{sample}},630\;{\text{nm}}}}} \right)/\left( {{\text{O}}{{\text{D}}_{{\text{control}},570\;{\text{nm}}}} - {\text{O}}{{\text{D}}_{{\text{control}},630\;{\text{nm}}}}} \right) \times 100$$
(6)

Preparation and characterization of cell-free photocrosslinked hydrogel scaffolds with and without SF fibers

After dissolving the MECs in ultrapure deionized water at a concentration of 0.7% (w/v), Eosin Y (EY, 0.1 mM), triethanolamine (TEOA, 0.25% (w/v)), and dithiothreitol (DTT, 10 mM) were added. After enough stirring, the mixture was poured into a cylindrical Teflon mold (8 mm in diameter and 15 mm in length), exposed to the green light at a wavelength of 525 nm for 8 min, and the chemical structure was analyzed by FTIR. Because the ionic nature of aqueous mediums causes contraction behavior in MECs hydrogel11, non-mineralized SF fibers at the concentrations of 0, 0.1, 0.2, and 0.3% (w/v) were mixed with the best hydrogel composition, exposed to green light for 8 min, and placed in phosphate-buffered saline (PBS, as an ionic model medium) for 72 h. Finally, the microstructures of the hydrogel scaffolds were studied by SEM, and the percentages of diameter, length, and volume contractions were calculated. Additionally, different SEM images of each scaffold type were analyzed with ImageJ software to evaluate the effect of crosslinking and fiber percentage on pore size. Pores with clearly defined openings were selected, their cross-sectional areas were measured, at least 40 data points were selected after data processing and outlier removal, and the diameter of a circular pore with the average measured area was calculated.

Preparation and cellular assessments of cell-laden photocrosslinkable hydrogel scaffolds

Investigating the effect of cell density on the formation of hydrogel network

Despite many reports of encapsulating millions of cells in hydrogel scaffolds, our observations confirmed that a large number of cells prevented the successful formation of a hydrogel network. Thus, four cell densities (zero, 1 × 105, 5 × 105, and 1 × 106 cells/mL) of ADSCs were encapsulated in the fiber-free optimized hydrogel scaffolds and the persistence of the physical shape of the scaffolds was checked after 24 h. Moreover, the scaffolds were fixed with 15% formalin, cross-sectioned with microtome, and stained by Hematoxylin-Eosin (H&E) and Periodic acid–Schiff (PAS) to evaluate the effect of physical presence of encapsulated cells on hydrogel network.

Cell encapsulation in hydrogel scaffolds

To investigate different biological features, ADSCs were encapsulated in the three hydrogel scaffolds: (1) MEC-based hydrogel scaffold (MHydro group) to evaluate the main component only, (2) SF-containing MHydro (FHydro group) to evaluate the osteogenic effect of SF fibers, and (3) mineralized SF-containing MHydro sample (mFHydro group) to evaluate the osteogenic effect of mineral phase. For each group, MEC, EY, TEOA, and DTT were dissolved in ultrapure sterilized water and mixed at their best concentrations. After suspending ADSCs in FBS, an adequate volume of cell suspension was added to the 0.3% (w/v)-containing polymer mixtures to provide a cell density of 2.5 × 105 cells/mL. The mixtures were manually and gently homogenized, placed in a sterilized Teflon mold, and photocrosslinked by exposing them to the green light for 8 min. Then, the photocrosslinked hydrogels were removed from the mold, placed in the 3%FBS-supplemented DMEM, and examined at proper time points.

Evaluation of cell morphology, viability, and osteogenic gene expression

To evaluate the morphology of the encapsulated ADSCs, each scaffold was suspended in the culture medium and examined with an optical microscope (BEL INV2, BEL Engineering, Italy) at the appropriate magnification after 7 and 14 days.

Evaluating the viability of encapsulated ASDCs in a hydrogel using MTT assay is challenging because dye penetration into the hydrogel, MTT reduction by the cells, dissolution of formazan, and its removal from the hydrogel must be completely done. Also, preventing the reduction effect of the scaffold or removing its negative effect should not be ignored. So, a lactate dehydrogenase (LDH)-based cytotoxicity assay was used as a fast and robust alternative method125. After 3, 7, and 14 days, 100 µL of supernatants were transferred to a 96-well plate (3 replicates for each time point), cytotoxicity was assessed using the Lactate Dehydrogenase Cytotoxicity Detection KitPLUS (Roche Diagnostics Gmbh, Mannheim, Germany) according to the manufacturer’s instruction, and cytotoxicity percentage was calculated using the following equation126:

$${\text{Cytotoxicity}}\% =\left[ {\left( {{\text{rLDH}} - {\text{LC}}} \right)/\left( {{\text{HC}} - {\text{LC}}} \right)} \right] \times 100$$
(7)

Where rLDH, LC, and HC stand for released LDH, low control (supernatant of tissue culture plastic (TCP)), and high control (lysate of cultured cell in TCP), respectively. Because FBS and phenol interfere with LDH absorption, the phenol-free DMEM containing as little FBS as possible (3% (v/v)) was used. To measure the alkaline phosphatase (ALP) activity, an Abcam Colorimetric ALP assay kit (USA) was used according to the manufacturer’s instructions.

To analyze the osteogenic capacities of the hydrogel groups, quantitative real time-PCR (qPCR) was conducted after 7 and 14 days to measure the expressions of Runx2, osteocalcin (OCN), and type I collagen (Col1A1) genes. Total RNAs were extracted using FavorPrep Blood/Cultured Cell Total RNA Mini Kit (FAVORGEN, Taiwan) according to the manufacturer’s protocol and their integrities and purities were assessed by detecting the ratios of 260/230 and 260/280 about 2 using a NanoDrop ND-1000 Spectrophotometer (Thermo Scientific, USA). After synthesis of cDNA using a BioFACT 2 Step 2X RT-PCR Pre-Mix (Taq) (BIOFACT, Korea), a qPCR test was done using a qPCRBIO SyGreenMix Lo-ROX (PCRBIOSYSTEMS, UK), and gene expressions were evaluated by REST2009 software by considering 3 technical replicates of samples in each group at the significance of 0.05. Table 7 includes the forward and reverse sequences of qPCR primers.

Table 7 The forward and reverse sequences of qPCR primers.
Full size table

Statistical analysis

Obtained data were presented as mean ± standard deviation, the homogeneity of variance was checked using the Kolmogorov–Smirnov and Levene’s tests, and statistical comparisons were performed through one-way ANOVA at a significance of 0.05.