Concentric-mineralized hybrid silk-based scaffolds for bone tissue engineering in vitro models

Abstract

There are many challenges in the development of 3D-tissue models for studying bone physiology and disease. Silk fibroin (SF), a natural fibrous protein used in biomedical applications has been studied for bone tissue engineering (TE) due to its mechanical properties, biocompatibility and biodegradability. However, low osteogenic capacity as well as the necessity to reinforce the protein mechanically for some orthopedic applications prompts the need for further designs for SF-based materials for TE bone. Concentric mineralized porous SF-based scaffolds were developed to improve mechanics and mineralization towards osteoregeneration. Hybrid SF silica microparticles (MP) or calcium carbonate nano-structured microparticles (NMP) were seeded with hMSCs co-cultured under osteogenic and osteoclastic conditions with THP-1 human monocytes up to 10 weeks to simulate and recapitulate bone regeneration. Scaffolds with appropriate pore size for cell infiltration, resulted in improved compressive strength, increased cell attachment and higher levels of expression of osteogenic markers and mineralization after adding the NMPs, compared to systems without these particles. These hybrid SF-based 3D-structures can provide improved scaffold designs for in vitro bone TE.

Introduction

Bone tissue engineering (TE) efforts have focused mostly on developing grafts for large osseous defects using scaffolds, progenitor cells, mechanical stimuli and soluble factors. Aside from bone regeneration, there has also been a focus on three-dimensional (3D) in vitro bone tissue models for drug testing and to study bone physiology and pathology. However, the recreation of a physiological bone structure remains to be accomplished.^[[1]^] Bone regeneration is an active process that recapitulates skeletal development. Regeneration mechanisms depend on many factors including cell–scaffold interactions, which are dependent on biomaterial biocompatibility, architecture, porosity, mechanical properties, degradation rate and the presence of soluble molecules like cytokines, growth factors, hormones, ions, and vitamins to support osteogenic regulatory functions.[2–5] The extracellular matrix (ECM) composition and organization provide mechanical properties for bone, including organic and inorganic matrix components highly organized at multiple levels of structural hierarchy.[6–8] These structures are continuously remodeled by osteoclasts (bone-resorbing cells), osteoblasts (bone-forming cells), and osteocytes (regulating cells).

For 3D in vitro bone tissue models, the engineered ECM should resemble the complex ECM structure of physiological human bone.[1] A fundamental TE design hypothesis is that scaffolds should provide a biomimetic mechanical environment for initial function while also providing sufficient porosity for cell migration. This hypothesis presents conflicting design requirements, since matching bone tissue stiffness requires denser material, in conflict with cell migration/delivery that requires more porous material. While it is clear that other scaffold characteristics, including pore size, pore interconnectivity, permeability, and material surface chemistry influence bone tissue regeneration, pore volume fraction is a primary design variable known to affect tissue regeneration.[9]

To achieve the balance of these properties in bone tissue models, current trends seek scaffold materials that provide mechanical support, slow degradation rates comparable with new bone tissue formation, as well as stimulatory functions to promote osteogenesis. Biomaterials that fulfill these requirements could aid bone therapies. Silk protein is a good example of such as material systems: a mechanically robust yet flexible material, tunable in terms of structure and morphology, and capable of supporting cell attachment in porous sponge-like structures. Silks are fibrous proteins produced by insects and spiders and provide structural and protective functions. Silkworm silk from Bombyx mori is one of the most characterized silks and consists of heavy and light chains that form into fibers with remarkable mechanical strength. Additionally, aqueous processing conditions, biocompatibility and biodegradability of silk fibroin (hereafter termed SF or simply silk), along with facile chemical modifications, are attractive features often sought in biomaterials. These attributes have led to the incorporation of SF into biomaterials in the TE field for a wide range of biomedical applications, including osteogenesis.[4] [10][11][12][13][14][15][16]

Porous three-dimensional silk scaffolds were extensively used in our previous studies for bone tissue engineering and showed excellent biodegradability and biocompatibility. Silk sponges have been used extensively in bone tissue engineering with human mesenchymal stem cells (hMSCs) and shown to facilitate bone formation in vitro and in vivo.[17] Further, silk can be used as a drug delivery vehicle for enzymes, growth factors, and antibiotics, towards biofunctionalization and in combination with inorganic Ca salts the properties of silk can be utilized with chemical compositions and mechanical properties closer to native bone.[14][17][18]

Traditional approaches to fabricate 3D scaffolds[19][20] can limit control over geometry, pore size and distribution, pore interconnectivity, as well as internal channel construction. Random, spontaneously generated, and disconnected pores may decrease nutrient transport, cell migration and survival especially in the center of the scaffold due to transport limitations. To mimic the structure of natural human bone a different system was studied in the present work. Here, the objective was to exploit the versatility with silk materials engineering, to generate composite materials with calcium-based additives, along with a ‘rolling technique’ to generate tubular structures with central channels. The process included microchannel walls[21] on a concentric porous matrix, CaCO₃ nano-structured microparticles (CaCO₃ NMP),[22] and co-cultured adult human stem cells and human leukemia monocytes to form differentiated bone tissue in vitro.[23][24,25] The approach was to mimic physiological tissue properties by including some of the key cells combined with complex bone scaffold structural hierarchically. A concentric porous matrix was fabricated by rolling the SF membranes, allowing for conduits to promote mass transfer through the structures via the channels and a hollow center mimicking the concentric lamellae structure in natural bones. The incorporation of inorganic particles was included to form composite material systems with enhanced mechanical properties. Since Ca²⁺ is an essential component of the mineralization process in bone, adding CaCO₃ NMP may also provide Ca²⁺ ions during cellular differentiation process.

Materials & Methods

Preparation of aqueous silk solutions

A 5-6% (w/v) silk aqueous solution was obtained from B. mori silkworm cocoons using previously described procedures.[26] Briefly, the silkworm cocoons (supplied by Tajima Shoji Co., LTD., Yokohama, Japan) were extracted in 0.02 M sodium carbonate solution and boiling 30 minutes, then rinsed in distilled water and dissolved in 9.3 M lithium bromide. After the complete dissolution with lithium bromide dialysis was carried on against distilled water using a cellulose dialysis tubing (molecular weight cutoff MWCO, 3500, Fisherbrand by Fisher Scientific, Pittsburgh, PA) for 48 h. The silk solution was centrifuged two times for 20 minutes at 11,000 rpm and 5°C (Sorvall centrifuge with SS-34 fixed angle rotor by Thermo Fisher Scientific). Protein concentration was determined by air drying a known volume of the silk solution and weighing the remaining solids. The resulting 5–6% (w/v) silk solution was stored at 4°C until the scaffold preparation.

Synthesis of inorganic microparticles

Calcium carbonate NMP of 2.5μm ± 0.2μm were synthesized by colloidal crystallization between CaCl₂ and Na₂CO₃ using previously described procedures. [22] In a typical experiment, 9.0 ml of 3.2% (w/v) CaCl₂ in 1.25 M glycine–NaCl buffer (pH 10.0) was added and mixed in glass vials with 9.0 ml of a 3.2% (w/v) Na₂CO₃ solution and stirring for 5 min. During preparation, the pH of the solution was adjusted to the desired value with NaOH or HCl. The 3.2% (w/v) Na₂CO₃ (Sigma-Aldrich Saint-Quentin-Fallavier, France) solution was prepared in MilliQTM water and used immediately, while the solution of CaCl₂ (Sigma-Aldrich Saint-Quentin-Fallavier, France) in Gly buffer (Gly, Sigma–Aldrich, Saint-Quentin-Fallavier, France and NaCl VWR International, Fontenay-sous-Bois, France) was prepared and kept frozen until use. The final concentrations in the reaction mixture were 1.6% (w/v) for both Ca²⁺ and CO₃²⁻ions. MilliQTM water was then added to a final volume of 30 ml and the precipitated CaCO₃ NMP were collected by centrifugation. The precipitate was washed with 30 ml of MilliQTM water followed by centrifugation. The pellet was frozen in liquid N₂ before lyophilization and finally stored under vacuum in a desiccator at room temperature until scaffold preparation. Based on previous work,[27][28][29] silica particles have been shown to upregulate osteogenic markers when present on biopolymers scaffolds, and also increased the formation of a collagen/calcium phosphate extracellular matrix, indicating osteoinductive properties. Thus, microparticles of silicon dioxide (silica microparticles) of 2μm ± 0.2μm (Sigma-Aldrich, USA) were used as controls.

Design and preparation of concentric hybrid silk-based scaffolds

UV mask for channels.

Microfluidic semi-circular cross-section microchannel patterns were designed using Layout Editor (Juspertor UG) and printed onto mylar masks using services from Advanced Reproductions (North Andover, MA). For channels with semi-circular profiles, a previous protocol was used.[21] Briefly, the mask pattern was transferred via photolithography to a 100 mm silicon wafer and coated with photoresist (Dow Corning, Midland, MI). A polydimethylsiloxane mold (PDMS) (Dow Corning) was cast onto the photoresist and cured at 60°C for 4 h and then delaminated from the photoresist. The PDMS mold was trimmed to fit the bottom of a 1 cm by side of square Teflon^® container for scaffold assembly. Porous microchannels were fabricated using a composite scaffold prepared from a porous silk film and salt-leached sponge (Figure 1). First, a 1% w/v silk solution was cast onto the patterned PDMS mold, followed by salt-leached sponge assembly.

figure 1.

preparation of concentric-mineralized silk-based scaffolds. a. porous microchannels were prepared using a uv-mask pattern and a composite scaffold made from a porous silk film and salt-leached sponge. caco3 nmp or silica microparticles were added to the silk solution before salt-leaching and masking processes for hybrid or control system fabrication. b. concentric-mineralized silk-based scaffolds were constructed by rolling the grooved mineralized membranes. (scale bars = 20μm and 100μm for membrane and rolled membrane images, respectively) c. sem of cross- section of pure silk (top), silica (middle) and caco3 mineralized (bottom) silk- based concentric scaffolds samples. (scale bars = 20μm for cross- section images)

Salt leaching & Scaffold assembly.

For protein gels, aliquots of 500 μL of silk solution (5-6% w/v) were poured over a patterned PDMS base into cubic Teflon molds 1 cm³. Granular NaCl (1 g of Ø 500–600 μm per mold) (Sigma-Aldrich) was slowly sifted into the silk solution and covered from light as described previously.[30] The molds were cured for 48 h on the bench and then washed for 24 h under magnetic agitation with MilliQTM water to eliminate the dissolved NaCl crystals from the gelled silk sponges. CaCO₃ NMP or silica microparticles (control) were included during the salt-leaching process for hybrid scaffolds construction. The silk was cured for 48 h and then the plates were immersed in distilled water (4 L) for 48 h (water was changed twice) to leach out the NaCl. The remaining silk scaffold was detached, and the thickness was trimmed to 500 μm. Layers of 500 μm with microchannels of 30 μm in diameter were rolled and fixed by rubber O-rings under sterile air to obtain concentric canalized scaffolds.

For sterilization, scaffolds were wrapped by pairs in 20 ml glass vials with 5 ml of MiliQTM water and covered with aluminum foil lids. A saturated steam cycle was employed to autoclave the scaffolds at 121°C for 20 min under a high pressure (Sanyo/Panasonic MLS-3871L autoclave) and drying time was also added to the cycle.

Imaging of silk-based scaffolds

The morphology of the silk-based scaffolds was observed by scanning electron microscopy (SEM) (Zeiss EVO MA10 SEM equipped with an SE1 detector, Carl Zeiss, Oberkochen, Germany) at 3.0 kV and Zeiss EVO 55 Environmental SEM (Carl Zeiss, Oberkochen, Germany) in environmental capture mode. Fractured sections of the silk-based scaffolds were obtained from the hydrated concentric scaffolds using surgical scissors and dried overnight. Samples were mounted on a copper plate and sputter-coated with a 20–30 nm thick Au/Pd layer prior to imaging. The SEM images were used to determine the presence of CaCO₃ NMP and cell distributions.

Pure silk scaffold microchannel morphology was characterized with the same technique. To image the system, porous bulk space and microchannel cross-sections samples were sliced with a stainless-steel scalpel blade. In order to maintain silk structure in the dry state, samples were first frozen at −80°C for 4 h and lyophilized for 12 h. Samples were sputter coated with Pt/Pd (60 s, 40 mA) and imaged with a Supra 55VP field emission SEM (Zeiss, Oberkochen, Germany). The SEM images were analyzed using ImageJ Software to determine pores sizes and porosity.

Pore size, porosity and roughness were determined from representative SEM images. The cross-sectional porosity of scaffolds was calculated using ImageJ software by converting SEM images to binary images (Figure 2a) and thresholding (data not shown). The pore area was then calculated by using the ImageJ ‘Analyze Particles’ feature. The percentage of the colored area relative to the total area in each image was defined as the percent porosity ($\mathrm{\varnothing }$) according to the equation: $\varnothing ={\sum }_{i=1}^n\raisebox{1ex}{$Ai$}\!\left/ \!\raisebox{-1ex}{$A$}\right.100$ where Ai and A are the pore area of each i measured pore and total area of the measured image respectively. Surface roughness statistics based on topographical images were calculated based on ImageJ Software (SurfCharJ plugin) as well with an image as input in which the pixel values represent distance, z, to a surface. Surface roughness was calculated as a correlation to an increase of the standard deviation related to gray values of pixels on processed SEM images (SD +/−). The standard deviation of RA (arithmetical mean deviation), RSK (skewness) and RKU (sharpness) parameters were analyzed. A lower standard deviation correlated to a smoother surface (Table in Figure 2a). The surface roughness calculations were based on 73.14 um x 48.29 um images after converting into 32-bit.

figure 2.

characterization of channel concentric-mineralized hybrid silk- based scaffolds. a. sem images and process images by imagej software of pure silk matrix (left) and caco3 (right) systems were analysed for porosity (pictures, scale bars = 100μm) and surface properties (insets, scale bars = 10μm). data of pore diameter and porosity of pure silk matrix and caco3 systems measured by horizontal line in every 51 and 36 pores, respectively, are shown in the table. roughness was calculated as the standard deviation (+/−sd) of ra (arithmetical mean deviation). rsk (skewness) and rku (sharpness) parameters were also analysed. b. mechanical properties. compression tests were used to evaluate sponge deformation. stress strain curves of hydrated silk-based sponges at 37°c. values of a restricted area close to 25% of deformation generated the linear equation and the trend- line for the calculation of compared e (tensile modulus) and η (compression yield) related to pure silk material. statistically significant differences (n=3, p<0.05) were found for all studies.

The morphology of cells on the scaffolds at 10 weeks of culture were visualized by Zeiss (Germany) Supra 55VP Field Emission Scanning Electron Microscope. After the culture medium was removed, the scaffolds were gently washed with DPBS (pH 7.4) twice and cells were fixed with 2.5% (v/v) glutaraldehyde (Electron Microscopy Science, Hatfield, PA) in 0.1 M PBS (pH 7.4) for 30 min at 4°C. Then, the scaffolds were dehydrated in a series of ethanol (30, 50, 70, 90, and 100% v/v) and treated with hexamethyldisilazane (HMDS; Electron Microscopy Science, Hatfield, PA) before being dried. After the scaffolds were completely dry, samples were coated with a gold-palladium layer prior to SEM analysis.

Mechanical evaluation of silk-based scaffolds

Compression properties of hydrated samples (Ø 5 mm, height 5 mm) were obtained using an Instron 3366 (Norwood, MA) testing frame equipped with a 10 N load cell. The tests were carried out in phosphate buffered saline (0.1 M) (Invitrogen, Carlsbad, CA) at 37°C at a strain rate of 0.5 mm/min. The temperature of the testing room was maintained at 23°C. The specimens to be tested were kept in buffer at room temperature (23°C overnight) prior to testing. The sample modulus and yield strength were determined from the stress-strain curves normalized to the cross-sectional area of the scaffold using a Labview program written in-house as previously described.[31] Briefly, the tensile modulus was defined as the slope of the linear region between 10% and 25% strain. Compression yield strength was defined as the stress at the intersection of the stress-strain curve and a line parallel to the linear region, offset by 0.5% strain.

Cell isolation, expansion and seeding

Cell culture reagents were purchased from Life Technologies Grand Island, NY). Human MSCs were obtained from a commercial source isolated from bone marrow aspirate from a male donor under 25 years old (Lonza, Walkersville, MD) as described previously.[32] The aspirate was combined with hMSCs proliferation medium (α-MEM with 10% FBS, 1% antibiotic/antimycotic, 1% non-essential amino acids (NEAA)) and cultured at 37°C with 5% CO₂ in a humidified environment. Flasks were rocked to allow hMSCs to adhere and media was added every three to four days until the hMSCs reached 80% confluence. Human MSCs at passage two were tested for osteogenic and adipogenic differentiation potential into osteoblasts and adipocytes. THP-1 human acute monocytic leukemia cell line cells (ATCC#TIB-202) were cultured in proliferation medium (RPMI 1640, 10% v/v FBS and 1% v/v antibiotic/antimycotic) at 37°C with 5% CO₂ in a humidified environment and then treated with 200 ng/mL of phorbol-12-myristate-13-acetate (PMA; Sigma, MO) to induce maturation into macrophages. Cell density was not allowed to exceed 1 x 10⁶ cells/mL and the medium was changed every two to three days prior to seeding. All experiments used hMSCs passages between P2-P6. Cell number and viability were determined using trypan blue exclusion. The re-suspended cells were plated at a density of 6×10⁶ cells/mL for seeding.[17][33]

Prior to hMSCs seeding, pre-mineralized or pure silk layers were sterilized by autoclaving and incubated overnight in medium. Cells were seeded at a density of 6x10⁶ cells/mL in a 50 μl drop after trypsinization and hemocytometric counting. Layers were cultured with hMSCs growth medium (MEM α with 10% FBS and 1% antibiotic/antimycotic) to increase cell numbers over four days, after which THP-1 cells were seeded (co-cultured hMSCs osteoblasts & THP-1 osteoclasts systems, or simply the co-culture systems) or monocultures (osteoblast systems). For co-culture systems THP-1 cells were seeded at the same inoculation density that hMSCs were previously seeded (3x10⁵ cells/scaffold) by adding 50 μL suspension onto hMSC seeded scaffolds. Each cell line was allowed to adhere to scaffolds for 3 h. Cell densities were selected to allow cell attachment, spreading and initial proliferation on silk scaffold structures.[17][33] Then, all layers were rolled under sterile atmosphere and secured at the ends by two silicone o-rings and 3 mL differentiation medium was added to each well of 12 well TCPS.

Following seeding, all culture systems were maintained in the same medium, a half and half mixture of RPMI 1640 and α-MEM supplemented with 10% FBS, 1% antibiotic/antimycotic, 1% NEAA, 100 nM dexamethasone (Sigma Aldrich, St. Louis, MO), 10 mM B35 glycerol phosphate (Sigma Aldrich, St. Louis, MO), and 0.05 mM ascorbic acid (Sigma Aldrich, St. Louis, MO) (for osteoblast differentiation, as described previously[17]), and 40 ng/ml phorbol 12-myristate 13-acetate (PMA) (Sigma Aldrich, St. Louis, MO) and 10 ng/ml receptor activator of nuclear factor kappa-B ligand (RANKL) (for osteoclast differentiation, as described previously[33]) with medium changes every 3-4 days.

Metabolic activity and bone markers/osteogenic differentiation

Evaluation of hMSC proliferation.

Proliferation of hMSCs in cultured scaffolds was determined using Alamar Blue assay at, 2, 3 and 5 weeks of culture. At each time point, the scaffolds were rinsed in PBS and transferred to a sterile 12 well plate for each Alamar Blue assay to ensure that only the metabolic activity of cells growing on the films was measured. Then, cell culture media containing 10 vol% Alamar Blue was added to the samples. After 5 h incubation, the fluorescence signal (550 nm excitation 590 nm emission) of a medium aliquot (100 ml) was measured using a plate reader. To estimate the cell number in the cultured scaffolds, a calibration curve was prepared by incubating a specific number of cells in the growth media containing 10% Alamar Blue for 5 h. A no cell control with 10% Alamar Blue in the media was used as a blank and subtracted from the sample. Relative proliferation rate was determined by percentage reduction of Alamar Blue at each time point. A 100% reduced form of Alamar Blue was produced by autoclaving the culture medium with 10% Alamar Blue at 121⁰C for 20 min. Arbitrary units are determined as relative units of fluorescence intensity from the reduction of resazurin found in Alamar Blue to red fluorescent resorufin in the presence of metabolically active cells. Since Alamar Blue is nontoxic to cells, the medium that contained Alamar Blue was replaced with a fresh differentiation medium following each reading. The same samples were measured at each time point. Excel was used to plot metabolic activity as a function of arbitrary units (AU) vs. time. Graphs were normalized to the fluorescence reading of the Alamar Blue solution alone. The sample size per group per time point was n = 3.

ALP activity analysis.

Alkaline phosphatase (ALP) activity was measured by biochemical assay from Stanbio Laboratory (Boeme, TX), based on conversion of p-nitrophenyl phosphate to p-nitrophenol, which was measured spectrophotometrically at 405 nm. Samples for alkaline phosphatase (ALP) activity analysis were rinsed in PBS and stored at −20°C prior to testing.

qPCR.

At the third and fifth week of culture, samples for quantitative polymerase chain reaction (qPCR) analysis were rinsed in DPBS (pH 7.4) and stored in Trizol (Life Technologies, Grand Island, NY) at −80°C prior to analysis. RNA was isolated by single step acid-phenol guanidinium method. For RNA purification, Qiagen RNEasy kit (Qiagen, Valencia, CA) was used. Reverse transcription was performed on RNA using the High-Capacity cDNA Reverse Transcription kit (Life Technologies, Grand Island, NY). Assays on demand (Life Technologies, Grand Island, NY) were used for the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), alkaline phosphatase (ALP), collagen type 1 (COL-1), bone sialoprotein (BSP), osteocalcin (BGALP), osteonectin (SPPL) and osteopontin (SPOCK). Gene expression was quantified with a Stratagene Mx3000P QPCR System (Stratagene, La Jolla, CA) based on fluorescence intensity after normalization with an internal reference dye and baseline correction. Differences of gene expression were generated by using comparative Ct method (Ct [delta] [delta] Ct comparison) as previously used in our lab[34] and recommended by the manufacturer (Perkin Elmer User Bulletin #2, Applied Biosystems, Foster City, CA). The data were normalized to the expression of the housekeeping gene, glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) within the linear range of amplification and differences. The threshold cycle (Ct) was selected in the linear range of fluorescence for all genes.[35]

Statistical analysis

All the experiments were independently replicated with a minimum sample size of N = 3. Values were expressed as average plus or minus the standard deviation (±SD). Student’s t-test was performed for paired observations. A value of p<0.05 was considered statistically significant. ANOVA and a post-hoc analysis (Tukey & Fisher) were carried out to give pairwise statistical comparison and a value of p<0.05 was considered statistically significant.

Results & Discussion

Scaffolds design

The incorporation of CaCO₃ NMP into silk solutions supported the formation of a homogeneous material with interconnected pores, without loss of gelation capacity (Figure 1). Using this method, microchannel cross-sectional shape and dimensions were controlled and 1cm² films around 500 μm thick were rolled. Given that microvasculature in bone has diameters that range between 10-50μm[36] a channel diameter of 25μm was considered appropriate for the structures formed here. This circular structure should also provide improved transport to the deeper regions of the scaffolds. Interconnected pores along the structures, combined with the presence of longitudinal channels, can promote osteo-differentiation and active bone formation in a more homogeneous manner. The designed structures should also be relevant for vascular ingrowth along with the mass transfer needs.[13]

Silk and silk-mineralized scaffold characterization

When larger pores were observed in transverse cuts of the hybrid scaffolds (average pore diameters around 410 vs. 250 μm in the nonhybrid systems), lower porosity (about 17%) was found when CaCO₃ NMPs were included in the silk matrix before gelation (Figures 1, 2). This change could be advantageous to stimulate osteogenesis in vitro by reducing cell proliferation and enhancing cell aggregation. [37–39]

Higher standard deviations of RA (around 20%), RSK (around 14%) and RKU (around 40%) were observed for the hybrid scaffolds in comparison to the silk only systems. Roughness is a desirable property to enhance a cellular attachment, and the microparticles can also act as nuclei for mineralization.

Strain modulus E and compressive strain η were calculated from stress-strain curves for pure silk, control (silica hybrid system) and the CaCO₃ hybrid system. The values of E and η for pure silk structures significantly increased for the CaCO₃ hybrid systems and the control hybrid material. Values of E for the CaCO₃ hybrid systems were 2.1-fold higher than pure silk and 1.9-fold higher than the silica hybrid scaffolds, while η values were 3-fold higher than pure silk and 2.2-fold higher than the control. In similar findings, internal channels in collagen scaffolds offered improved mechanical performance when compared to pure porous collagen or hybrid CaP collagen-based scaffolds, and similar performance when collagen was mixed in high proportion with hydroxyapatite nanoparticles.[40–42]

Osteoblast and co-culture concentric pure silk and silk-mineralized scaffolds

In terms of cell growth and differentiation, the three systems showed similar profiles during 5 weeks of culture when osteoblasts or co-cultures were seeded into the scaffolds. DNA content of the silk-based scaffolds was quantified over 5 weeks in osteogenic medium (Figure 3a). After 2 weeks of culture, the CaCO₃ hybrid systems supported a higher number of cells (represented by DNA content) than the pure silk or silica systems, with a significant difference until the 3rd week. Alkaline phosphatase (ALP) levels were higher in the presence of CaCO₃ NMP in the matrices after 3 weeks for osteoblasts and the co-culture systems. A decrease in ALP (2-fold) was observed for the co-cultures and could be related to the presence of osteoclasts.[43] On other hand, the presence of CaCO₃ NMP in the scaffolds enhanced ALP levels both for the osteoblast and co-culture systems (about 12% and 40%, respectively). Control silica hybrid systems did not demonstrate this performance until 3 weeks. The cell proliferation and ALP results suggested improved outcomes for cell attachment on the CaCO₃ hybrid scaffolds, correlated with gene expression. Compared with silica systems, CaCO₃ hybrid matrices showed an increase of ALP as well as higher values compared with the pure silk material, both for the osteoblasts and co-culture systems at 3 weeks. Compared with silk, silica systems seeded with osteoblasts or co-cultures did not show significant differences at 3 weeks or 1 week respectively.

figure 3.

(a) Cell proliferation studies for hMSC-derived osteoblasts (top) or co-cultured hMSC-derived osteoblasts & THP-1-derived osteoclasts systems (bottom) maintained on pure silk, silica or CaCO₃ scaffolds carried out for 5 weeks. (b) ALP activity of hMSC-derived osteoblasts or co-cultured hMSC-derived osteoblasts & THP-1-derived osteoclasts systems on pure silk and hybrid systems (silica control scaffolds and CaCO₃ scaffolds) measured after 1 and 3 weeks of incubation. (c) Osteogenic expression relative to pure silk on hybrid (□ and ■) and control ( and ) co-cultured systems studied after 3 and 5 weeks respectively. More relevant statistically significant differences are showed after an ANOVA and a post-hoc anlaysis (Tukey & Fisher) with n = 3 or larger for each group of samples. * p < 0.05).

For the different matrices, the analysis of ALP, COL-1, BGALP, SPPL and SPOCK showed that the hybrid systems reached higher levels of gene expression than on the pure silk matrices over 5 weeks (Figure 3c and ESI 2).

For the cell culture systems, the osteoblast cultures had higher gene expression levels than the co-cultures for some of the genes towards the end of the study, but were not significantly different for COL-1, BSP and SPPL at week 3 and COL-1 and SPPL at week 5. Except for BSP expressed from CaCO₃ systems (statistical differences of the 3 w and 5 w silica and CaCO₃ systems relative to the normalizing pure silk matrices are shown on the ESI 2 figure), the genes tested for co-cultures were more highly expressed with the hybrid matrices versus the pure silk systems (1.1 to 5.9 times higher depending on the gene, and up to more than 30 times for SPPL from CaCO₃ hybrid matrices). During the study, ALP, COL-1 and BSP early expression genes and middle to late expression genes BSP and BGALP demonstrated higher expression on the silica than the CaCO₃ hybrid matrices, while SPPL and SPOCK late expression genes showed higher expression on the CaCO₃ hybrid matrices (Figure 3c). At the 5th week, the late gene SPOCK was expressed with significant difference between the silica and CaCO₃ hybrid systems, and SPPL for the CaCO₃ co-cultures were expressed around 2-fold higher than the silica hybrid co-culture scaffolds.

Representative SEM images are shown in Figure 4. The surface of hybrid scaffolds presented significant differences after 10 weeks of culture under osteogenic differentiation (Table ESI 1). SEM-based surface analysis by ImageJ was used to evaluate roughness throughout 10 weeks of culture (Figure 4b), during which the roughness of the hybrid scaffolds increased 60% while no significant difference was observed for pure silk co-cultures. Higher roughness generated by the CaCO₃ NMP can be an advantageous property to enhance cellular attachment (Figures 2a, 4a) and provide nuclei for mineralization (Figure 4b).[17,46]

figure 4.

sem analysis of pure silk and hybrid (left & right respectively) scaffolds: a. before cells seeding. surfaces roughness differences are indicated by a circle. b. after 10 weeks of co-culture with hmsc-derived osteoblasts & thp-1-derived osteoclasts. matrices porosity differences are indicated by a triangle. caco3 microparticles aging evolution is indicated by arrows for hybrid systems in figures a and b. continuous cell attachment covering a zone of the hybrid scaffold is indicated by a rectangle. scale bars = 10 μm)

The matrices did not show a loss in mechanical integrity based on visual observation, increase porosity, or the presence of cracks or ruptures under SEM observations after 10 weeks, for either the pure silk or hybrid systems (Figure 4b). Cell differentiation related to lower porosity is known to stimulate osteogenesis in vitro by suppressing cell proliferation and fostering cell aggregation, thus a lower porosity of the hybrid versus pure silk systems with Ca²⁺ could support improved differentiation.[47]

During 10 weeks of study, many of the differentiated cells remained on the surface of the co-cultured hybrid scaffolds. The particles had a different morphology after 10 weeks, probably due to partial cell degradation of the matrices along with exposure to culture media (Figure 4b). Surface variations may account for some of the differences between co-cultures systems at 10 weeks when comparing the osteoblasts in both the pure silk and hybrid systems (data not shown). The presence of active osteoclasts and macrophages in co-cultures has been shown to increase osteoblast activity through secretion of molecules such as platelet-derived and angiogenic factors by osteoclast progenitors as residual undifferentiated THP-1 cells and THP-1 macrophages.[48][49] The hybrid co-cultures resulted in increased ECM deposition (Figure 4b) and changes on the surfaces of the matrices. These changes features to support bone remodeling. Further, a range of studies with different cell types and biomaterials have demonstrated increased osteoblast differentiation and activity on rougher surfaces.[50–52]

The attachment and morphology of cells cultured on the pure silk and hybrid scaffolds were observed by SEM. After 10 weeks the cells attached to the surfaces and exhibited elongated or star-shape morphologies. (Figure 4b) Cell−cell contacts formed a continuous sheet covering parts of the scaffold surface when CaCO₃ NMP were present (Figure 4b). The SEM images demonstrated that the hybrid scaffolds were favorable for attachment, spreading and ECM deposition.

Conclusions

Previous studies have demonstrated that silk biomaterials support tissue engineering of bone and can be used to study bone remodeling.[17,30,53,54] However, silk by itself is not osteogenic and the mechanical properties of silk porous scaffolds are generally lower than those of native bone (Young’s modulus of 100 kPa vs. 10 MPa for bone). In the present study, hybrid silk-CaCO₃ 3D structures were generated and then seeded with cultures of hMSCs and THP-1 cells for osteoblasts and osteoclasts, respectively, for 10 weeks. Co-cultures with the hybrid systems vs. pure silk showed better performance towards cell attachment, differentiation and mineral deposition over time. Hybrid scaffolds supported increased proliferation and expression of osteogenic marker genes and mineralization compared to the pure silk matrices. The presence of CaCO₃ nano-structured microparticles promoted osteogenic differentiation, with better attachment of THP-1 osteoclast differentiated cells to provide a platform for bone-like structure formation. The CaCO₃ nano-structured microparticles also offer the possibility of loading other compounds as delivery vehicles for drugs, such as different factors during tissue-specific formation or antibiotics to prevent surgical infections.[55] The systems can be further improved by incorporating microtubes in the center of the structures to add a central vascularization path.[17]