Fabrication of bone‐derived decellularized extracellular matrix/ceramic‐based biocomposites and their osteo/odontogenic differentiation ability for dentin regeneration

Abstract The goal of this study was to fabricate bioactive cell‐laden biocomposites supplemented with bone‐derived decellularized extracellular matrix (dECM) with calcium phosphate ceramic, and to assess the effect of the biocomponents on the osteogenic and odontogenic differentiation of human dental pulp stem cells (hDPSCs). By evaluating the rheological properties and selecting printing parameters, mechanically stable cell‐laden 3D biocomposites with high initial cell‐viability (>90%) and reasonable printability (≈0.9) were manufactured. The cytotoxicity of the biocomposites was evaluated via MTT assay and nuclei/F‐actin fluorescent analyses, while the osteo/odontogenic differentiation of the hDPSCs was assessed using histological and immunofluorescent analyses and various gene expressions. Alkaline phosphate activity and alizarin red staining results indicate that the dECM‐based biocomposites exhibit significantly higher osteogenic activities, including calcification, compared to the collagen‐based biocomposites. Furthermore, immunofluorescence images and gene expressions demonstrated upregulation of dentin matrix acidic phosphoprotein‐1 and dentin sialophosphoprotein in the dECM‐based biocomposites, indicating acceleration of the odontogenic differentiation of hDPSCs in the printed biocomposites. The hDPSC‐laden biocomposite was implanted into the subcutaneous region of mice, and the biocomposite afforded clear induction of osteo/odontogenic ectopic hard tissue formation 8 weeks post‐transplantation. From these results, we suggest that the hDPSC‐laden biocomposite is a promising biomaterial for dental tissue engineering.


| INTRODUCTION
Biophysically and biochemically functionalized three-dimensional (3D) constructs have become critical for the successful regeneration of various tissues. 1,2 In particular, tissue engineering strategies, including cell-based regeneration using dental pulp stem cells (DPSCs) and scaffolds combined with bioactive growth factors (i.e., basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), and bone morphogenic protein (BMP)). [3][4][5] have been used to successfully regenerate intricate dental tissues; nevertheless, new bioactive tissue-regenerating materials dedicated to dental tissues are still being investigated. [6][7][8][9] 3D bioprinting has been extensively applied to fabricate cell-laden structures using a mixture of cells and bioactive hydrogels, which can be termed as a bioink. Because this process can enable the stacking of microscale cell-laden struts according to a designed 3D structure, the printing system has been considered as an outstanding tool to attain tissue engineering substitutes. Recently, to successfully regenerate various tissues, such as skeletal muscles 10 and those of the heart, 11 liver, 12 and bone, 13 cell-affordable, biocompatible, and printable hydrogels, including alginate, gelatin, gelatin methacrylate, silk fibroin, collagen, and methacrylated collagen, have been extensively investigated. [14][15][16][17][18][19][20] In general, tissue-specific bioinks to provoke chosen cellular activities can help induce the growth and differentiation of laden cells. Specific microcellular environmental conditions can also be attained with supplementary bioactive components (such as growth factors, RGD ligands, and cytokines) physically or chemically bound in the hydrogels. The developed bioinks provide outstanding regeneration ability of various tissues, but they cannot completely demonstrate the biochemical intricacies of the natural tissue-specific extracellular matrix (ECM).
In particular, decellularized ECMs (dECMs) derived from bovine bone and dentin have been used as a constituting component of bioinks for 3D dentistry constructs, and the fabricated 3D structures represented reasonable osteo/odontogenic differentiation of the laden cells (DPSCs or odontoblast-like cell line). 4,21 However, the regeneration of odontogenic tissues to mimic the organic/inorganic compounds of a native dentin construct is challenging.
Recently, alginate-or Matrigel-based bioinks supplemented with various bioceramics (such as α-tricalcium phosphate [α-TCP], nanohydroxyapatite, and biphasic calcium phosphate) were constructed using a molding and printing process to overcome the weak mechanical nature of hydrogels and improve the osteoinductive properties of the bioinks. [22][23][24] In these studies, viable cells resided well within the structures; however, the potential for in vitro osteogenic activity and in vivo new bone formation and angiogenesis analysis using stem cells have not been fully investigated. Previously, we developed a bioink containing collagen and α-TCP using human adipose stem cells (hASCs) for bone tissue regeneration. 25 We focused on the in vitro osteogenic differentiation of hASCs loaded in the composite bioink with and without osteogenic medium. 25 Our work showed a significant potential for osteogenic differentiation lineage of the hASCs when using the bioink containing bioceramic, but the study was limited in terms of the degree of osteogenic activities of the hASCs loaded in the collagen/ceramic-bioink with and without an osteogenic medium.
Here, we utilized the bioprinting process with a dental-specific bioink containing hDPSCs to manufacture biomimetic 3D dental constructs. To accomplish this, collagen type-I or dECM derived from porcine bone as matrix hydrogels of the cells were physically mixed with an appropriate concentration of β-TCP for the fabrication of a cellladen biocomposite. We assumed that the tissue-specific biochemical cues from the dECM and osteoinductive β-TCP could synergistically effect cell growth and osteo/odontogenic differentiation of the hDPSCs in the printed dental construct. Based on the rheological properties and by assessing printability, we could construct a 3D cellladen mesh scaffold. In vitro biological evaluations (cell viability and growth, calcified tissue matrix, and osteo/odontogenic gene expression) were performed to validate the printed biocomposite with a control, a hDPSC-laden collagen/β-TCP bioink. Subsequently, the hDPSCladen biocomposite was implanted into the subcutaneous tissue of mice to show that the structure can clearly induce osteo/odontogenic ectopic hard tissue formation 8 weeks after transplantation.

| Preparation of decellularized extracellular matrixes (ECMs) from porcine bone and dentin tissues
Porcine bone tissue was extracted from the shins and thighs of the fore and hind limbs of a pig. To remove blood and impurities (fibrous tissue and adipose tissue), the bone tissues were placed in a container with deionized water and washed at 120 rpm for 30 min. This was repeated six times. The bone pieces were then crushed using a grinder to obtain bone powder. For demineralization, 0.5 M HCl (Sigma-Aldrich) was added to the bone powder and stirred for 5 h via magnetic stirring. The stirred solution was sieved using a 100-μm sieve.
The remaining solution was removed via spin-down using a centrifuge and washed several times with distilled water. A 1:1 ratio of a mixed solution of chloroform (Sigma-Aldrich) and methanol (Sigma-Aldrich) was used to remove lipids from the desalted powder, which was washed repeatedly with methanol and then distilled water for 1 h.
The demineralized bone matrix (DBM) was obtained via lyophilization.
A detailed decellularization protocol was used, as in a previous study. 26 The DBMs were briefly washed with distilled water and treated with 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid (EDTA, Sigma-Aldrich) at 37 C for 2 h. Subsequently, the DBMs were treated with 1% w/v penicillin/streptomycin at 4 C for 24 h to remove residual cellular material and then lyophilized to obtain a dECM. 0.01 M HCl (Sigma-Aldrich) with pepsin (Sigma-Aldrich) was added at the density of 1 mg/mL under continuous magnetic stirring for 3 days; 15% w/v of NaCl (Sigma-Aldrich) was also used for the process. Various ECM proteins were centrifuged, and the remnant components were dialyzed using a dialysis sack (molecular weight cutoff: 3.5 kDa; Spectrum Labs). Following completion of the lyophilization procedure, the dECM was obtained.
To obtain the dentin-derived dECM, molar teeth were extracted from a pig's jaw and washed several times with Dulbecco's phosphate buffered saline (Biowest, MO, USA), and the enamel layer was removed completely using a sawing process. Subsequently, demineralization and decellularization procedures were performed, identical to the previous protocol for the bone-derived dECM.

| Characterization of decellularized ECMs (dECMs)
To measure the cellularity of native tissue and the dECMs, doublestranded DNA content was evaluated using the Quant-iT Picogreen ® dsDNA assay kit (Life Technologies). The samples (10 mg/ml) were dissolved in a TE buffer solution (pH 7.5, 10 mM Tris-HCl, and 1 mM EDTA), and then Quant-iT Picogreen ® reagent was added. The mixture was incubated for 5 min at 28 C. A CytoFluor microplate reader (MTX Lab Systems Inc., Vienna, VA) was used to measure absorbance (520 nm).
To quantify the dECM constituents, the amounts of collagen and glycosaminoglycans (GAGs) were assessed. The solubilized collagen, extracted with 0.5 M acetic acid/pepsin at 4 C, was quantified using the Sircol Soluble Collagen Assay (Biocolor Ltd.). The collagen content was determined using a spectrophotometer at 555 nm. Sulfated GAG (sGAG), which was obtained using a papin/pepsin digestion buffer for 3 h at 65 C, was quantified using the Blyscan sGAG Assay (Biocolor Ltd.) at 656 nm.

| Rheological properties of bioinks
A rotational rheometer (Bohlin Gemini HR Nano; Malvern Instruments) equipped with cone-and-plate geometry (cone angle of 4 , diameter of 40 mm, and gap of 150 μm) was used to measure the rheological properties of the prepared bioinks. The storage modulus (G') of the various bioinks was measured in terms of shear stress (0.1-1000 Pa, temperature: 25 C, frequency: 1 Hz) and temperature sweeps (10-50 C, frequency: 1 Hz, strain: 1%). The shear stress value at the limit of the linear viscoelastic region for the G' vs. shear stress curves is marked as the yield stress (τ y ). All tests were performed in triplicate.

| Fabrication of cell-laden biocomposites
Cell-laden structures (10 Â 10 Â 1.3 mm 3 ) were fabricated using a three-axis robot system (DTR3-2210 T-SG; DASA Robot) equipped with a dispensing system (AD-3000C, Ugin-tech) and a 25G dispensing needle (inner diameter: 250 μm). The printing conditions, including pneumatic pressure and processing temperature, were appropriately selected, except for the nozzle moving speed of 10 mm/s. After fabricating the cell-laden constructs, crosslinking was performed using 1 mM genipin solution (Challenge Bioproducts) in a medium for 30 min at 37 C with 5% CO 2 .

| Characterization of cell-laden composite scaffolds
To visualize the surface morphologies, optical microscope (BX FM-32; Olympus) and field emission scanning electron microscopy (FESEM; JSM-7500f; JEOL Ltd.) were used. In addition, the distribution of elemental phosphate (P) and calcium (Ca) was evaluated using energy-dispersive spectroscopy (EDS).
To characterize the crystal peaks of β-TCP, Wide-angle X-ray diffraction (X'Pert PRO MRD; PANalytical, UK) with CuKα radiation, under the beam conditions of 40 kV and 20 mA with spectrum collection at 2θ = 20-40 and the step size of 0.1 was used.
The compressive properties of the cell-laden structures (6 Â 6 Â 4 mm) in a wet state were measured using a universal testing instrument (SurTA; Chemilab). Briefly, the compression speed was set at 0.5 mm/s and the compressive modulus was calculated using 5%-10% strain of the stress-strain curves. The cell-laden constructs were cultured in a 6-well culture plates supplemented with GM and incubated at 37 C in 5% CO 2 .
The medium was replaced every 2 days. To induce osteogenic differentiation of the hDPSCs, 100 μM dexamethasone (Sigma-Aldrich), 10 mM β-glycerophosphate (Sigma-Aldrich), and 50 μM ascorbate-2-phosphate (Sigma-Aldrich) were mixed with the GM. The fabricated cell-laden structures were cultured in osteogenic differentiation medium (DM) after 7 days of culture. The DM was changed every 2 days.

| Osteogenic activities
The alkaline phosphatase (ALP) activity was assessed via measuring the release of p-nitrophenol (pNP) released from pNP phosphate (pNPP). Briefly, the cell-laden constructs were rinsed using PBS, followed by incubation using Tris buffer (10 mM, pH 7.5) containing 0.1% Triton X-100 for 10 min. Afterwards, 100 μl of the lysate was added into 96-well culture plate containing equal volume of pNPP solution. Then, the ALP activity was quantified using aa microplate spectrophotometer using a wavelength of 405 nm. Furthermore, the cell-laden biocomposite were stained with ALP by twice rinsing the composite using PBS. Then, ALP buffer (100 mM Tris-Cl, pH 9.5, 100 mM NaCl, and 10 mM MgCl 2 ) was used for equilibration. Then, the biocomposites were immersed in BCIP/NBT (Sigma-Aldrich) for 30 min, followed by enzymatic activity termination via rinsing the samples using PBSS containing 20 mM EDTA. The ALP stained biocomposite were visualized using an optical microscope.
To evaluate the degree of calcium mineralization, the cell-laden biocomposites were stained with Alizarin red S staining. Briefly, the biocomposites were rinsed twice using PBS, followed by fixation in 3.8% formaldehyde solution (Sigma-Aldrich). Then the biocomposites were stained with 40 mM Alizarin red S with pH of 4.2 for 60 min.
The stained biocomposites were visualized using an optical microscope. To quantify the degree of mineralization, the Alizarin red S stained biocomposites were rinsed three times with distilled water and destained using 10% cetylpyridinium chloride in 10 mM sodium phosphate buffer (pH 7.0) for 15 min. The degree of calcium mineralization was measured using a microplate spectrophotometer at wavelength of 562 nm. To assess the mineralization by the cells, the optical density value of the cell-free composites was taken away from the value of the cell-laden structures. All data values were defined as ± SD (n = 5).

| Immunofluorescence
The hDPSCs on the biocomposites were stained with (OPN) and dentin sialophosphoprotein (DSPP) immunofluorescence to assess degree of osteogenesis and odontogenesis. Briefly, the biocomposites were rinsed twice using PBS and hDPSCs were fixated using 3.7 formaldehyde for 60 min. Then, the biocomposites were permeabilized using 2% Triton X-100 for 1 h and treated using 2% bovine serum albumin

| Real-time polymerase chain reaction (RT-PCR)
To evaluate the various osteogenic and odontogenic markers, including osteopontin (OPN), osteocalcin (OCN), biglycan (BGN), dentin sialophosphoprotein (DSPP), and dentin matrix acidic phosphoprotein-1 (DMP-1), RT-PCR (Applied Biosystems) was performed after 21 and 28 days of culture. Briefly, the total RNA was isolated from the cultured hDPSCs on the biocomposites using TRI reagent (Sigma-Aldrich), followed by purity and concentration evaluation using spectrophotometry (FLX800T; Biotek). Then, reverse transcription system using ReverTra Ace qPCR RT Master Mix (Toyobo, Japan) was utilized to synthesize complementary DNA (cDNA) from the RNase-free DNase-treated total RNA. Finally, StepOnePlus Real-Time PCR System (THUNDERBIRD ® SYBER ® qPCR Mix (Toyobo, Japan) was used to quantify the gene expressions. The gene-specific primers sequences are listed in Table 1.

| In vivo implantation of the hDPSCs-laden composite structure
The bioprinted collagen/β-TCP and dECM/β-TCP composite structures  2.14 | Statistical analyses SPSS software (SPSS, Inc.) was used to conduct statistical analyses. A single factor analysis of variance was employed, and a value of p* < 0.05, was considered statistically significant.

| Appropriate concentration of bioceramic in 3D-printed scaffold laden with hDPSCs
In a previous study, we demonstrated an optimal printing method using bioinks (collagen/hASC/β-TCP) for obtaining multilayered cellladen scaffolds. 25 Under selected printing parameters, we were able  In biomedical scaffolds, pore geometries, including pore size and strut size, are important design parameters. To circumvent the effect of pore geometry on the cellular responses of the laden cells, we set the pore (500 μm) and strut (500 μm) sizes by controlling the printing parameters, as shown in Table 2. Figure 2c shows the optical and live (green)/dead (red) images of the 3D-printed biocomposite structures for various bioink formations. After the crosslinking process was complete, the structures exhibited slightly different pore geometries than the initially printed structure, because they were gradually squeezed owing to the weight of the embedded ceramic fraction. This squeezing was more accelerated in the structures with relatively higher bioceramic concentrations than in the pure collagen bioink ( Figure S1a,b).
According to this result, the strut size and thickness of the printed composites differed slightly from those of the originally designed geometry (Figure 2d-e). In addition, the initial cell viability is an important criterion for determining printing stability. Figure 2f shows the initial cell viability, determined using the live/dead images of the hDPSCs in Figure 2c, after the crosslinking procedure. As indicated by the results, the cell viability at higher ceramic concentrations (bioceramic >30 wt%) was significantly reduced due to the higher wall shear stresses in the nozzle, signifying that the concentration of the bioceramic (below 20 wt%) was safe for the laden cells. Furthermore, as expected from the strut size and thickness of the printed compos-      Abbreviations: BDNF, Brain-derived neurotrophic factor; bFGF, Basic fibroblast growth factor; BMP, Bone morphogenetic protein; EGF, Epidermal growth factor; EG-VEGF, Endocrine gland-derived vascular endothelial growth factor; FGF, Fibroblast growth factor; GDF-15, Growth/differentiation factor 15; GDNF, Glial cell line-derived neurotrophic factor; GH1, Growth hormone 1; HB-EGF, Heparin-binding epidermal growth factor-like growth factor; HGF, Hepatocyte growth factor; IGF-1, Insulin growth factor 1; IGFBP, Insulin-like growth factor-binding protein; LOD, limit of detection; MCSF R, Macrophage colony-stimulating factor 1 receptor; NGF R, Nerve growth factor receptor; NT-3, Neurotrophin-3; NT-4, Neurotrophin-4; PDGF-AA, Platelet-derived growth factor A chain; PIGF, Placenta growth factor; SCF R, Stem cell factor receptor; SCF, Stem cell factor; TGF α, Transforming growth factor alpha; TGF β1, Transforming growth factor beta-1; TGF β3, Transforming growth factor beta-3; VEGF, Vascular endothelial growth factor; β-NGF, Beta-nerve growth factor. components derived from native tissues 31 For regenerative dentistry, dentin-derived dECM (d-dECM) has been used as a potential bioink mixed with a 1:1 ratio [alginate hydrogel (3% w/v) and d-dECM] to regenerate dentin. 21 The cell-laden scaffold using the d-dECM-based bioink showed good cytocompatibility and odontogenic activity. 21 However, this approach of using the dentin-derived dECM-based bioink can be highly inefficient because the yield rate for obtaining dentin-derived dECM from native dentin tissue is considerably lower compared to that of bone-derived dECM ( Figure S2).
To overcome this low yield rate, we used dECM derived from bovine bone tissues. To successfully obtain dECM biomaterials, complete elimination of the cellular components is required, without a significant loss of the bioactive components such as collagen, fibronectin, elastin, GAG, and proteoglycans. The images in Figure 4a show the decellularization procedure from native porcine bone tissue, and

| Osteogenic and odontogenic differentiation of biocomposites
To observe the degree of osteogenic differentiation of the biocomposites, we conducted ALP and alizarin red S staining at 21 and 28 days (Figure 6a

| Ectopic hard tissue formation
In this work, a comparative study of the hDPSCs-laden structures (CTS-20 and dECM-20) and dECM structure without the cell was performed to assess ectopic bone-formation after subcutaneous implantation in the dorsal area with the procedure, shown in Figure 8a. Experimental animals were healthy during the experimental period, indicating that any evidence for toxicity or side effects was not observed.
In general, the blood vessel formation in the implanted scaffold is a critical key component because the formation can specify efficient graft/host interactions and even signifies the stable cell-survival for long term period. 33 To evaluate the blood vessel formation of the implanted dECM without cells, CTS-20, and dECM-20, the implanted structures were histologically evaluated using hematoxylin and eosin (H&E) staining (Figure 8b). Figure 8c showed the vascularization of the implanted structures which were assessed using the H&E staining.
As expected, vessel formation was plentiful in the dECM-20 structure compared to that of the CTS-20, which could be due to the synergis-

| CONCLUSION
In this study, we fabricated a cell-laden bone-derived dECM/β-TCP/ hDPSC biocomposite, which was appropriately printed using various printing parameters for dental tissue engineering. To achieve the hDPSC-laden composite, we formulated a bioink using various weight fractions of the bioceramic, with the goals of reasonable initial cell viability after printing and a mechanically stable 3D mesh structure. In addition, to achieve a biochemical cue to provide appropriate dentalspecific microcellular environmental conditions, we accommodated bone-derived dECM, which has biological components similar to those of the dentin-derived dECM. According to the results of various in vitro cellular activities and in vivo work, the dECM-based biocomposite demonstrated meaningful cell viability and cell growth and even effectively accelerated the osteo/odontogenic differentiation of hDPSCs. Based on these results, we can conclude that the proposed 3D-bioprinted biocomposite, supported with the biochemical cues derived from the dECM and β-TCP, can serve as a potential dental tissue engineering material.