**2. Results**

### *2.1. Extraction and Characterization of MC*

One simple method to identify collagen is to scan collagen samples on a UV spectrum (200–400 nm), because the triple helical structure of collagen has absorption at 230 nm. As shown in Figure 1A, commercial atelocollagen and MC have a maximum absorption peak at 230 nm, which is associated with COO<sup>−</sup>, CONH2, and C=O groups in the polypeptides of collagen. Collagen has a few tryptophan, histidine, tyrosine, and phenylalanine components that absorb UV at 250 nm and 280 nm.

**Figure 1.** Characterization of collagen extracted from skins of *P. olivaceus* and of porcine commercial atelocollagen. (**A**) UV-Vis spectra of the two types of collagen; (**B**) Sodium dodecyl sulfatepolyacrylamide electrophoresis evidences the molecular structure and organization of two collagen; (**C**) Fourier transform infrared spectra of collagens exhibits the main vibrations of collagen molecular organization.

Figure 1B represents the SDS-PAGE components of commercial atelocollagen and MC. Commercial atelocollagen and MC both consist of α1 and α2 chains and high molecular weight β and γ components. Thus, the extracted commercial atelocollagen and MC correspond to type I collagen.

The FTIR spectra of commercial atelocollagen and MC from the skins of *P. olivaceus*, which can be used to identify information on the secondary structure of collagen, are shown

in Figure 1C. The changes in the spectral peaks were shown to include changes in the amide A and B bands as well as the amide I–III regions. The amide A band at 3300 cm<sup>−</sup><sup>1</sup> and the amide B band peaks at 2972 cm<sup>−</sup><sup>1</sup> are associated with N–H stretching vibrations and the asymmetrical stretching of CH2. In addition, the amide I peak at 1635 cm<sup>−</sup><sup>1</sup> is mainly associated with the stretching vibrations of C=O, along with the polypeptide backbone or COO<sup>−</sup>. The amide II region of collagen type I appears at approximately 1549 cm<sup>−</sup><sup>1</sup> from N–H bending vibration coupled with C=N stretching vibrations. The amide III region at 1239 cm<sup>−</sup><sup>1</sup> represents the peaks between C=N stretching vibrations and N–H deformation from amide linkages of CH2 groups of the glycine backbone and proline side-chain. As shown in Figure 1C, the intensity and positions of the amide A, B, and I–III bands are similar for the commercial atelocollagen and MC, as well as type I collagen.

### *2.2. Amino Acid Components*

Table 1 shows the amino acid components of commercial atelocollagen and MC from *P. olivaceus*. Commercial atelocollagen and MC were found to have glycine (238.3 and 247.3 per 1000, respectively) and to be low in tyrosine, histidine, methionine, and cysteine. The imino acid content of the extracted collagen was 264.3 and 208.6 per 1000 for porcine atelocollagen and marine atelocollagen, respectively. Based on these differences, the mechanical properties of MC are weaker than commercial collagen. Collagen with a greater imino acid content is more stable in the helix structure, due to the contents of proline and hydroxyproline.


**Table 1.** Amino acid composition of atelocollagen obtained from the skins of *P. olivaceus* and of porcine commercial atelocollagen (per 1000 residues).

### *2.3. Extraction and Characterization of CHA*

The FTIR spectra of raw fishbones and HA and CHA from the bones of *P. olivaceus* are shown in Figure 2A. The FTIR bands of raw fishbones were observed at 1047 cm<sup>−</sup>1, 1644–1740 cm<sup>−</sup>1, 2911 cm<sup>−</sup>1, and 2977 cm<sup>−</sup>1; these bands are both minerals and organic compound of fishbones. HA has bands at 878 cm<sup>−</sup>1, 1000–1100 cm<sup>−</sup>1, 1400–1500 cm<sup>−</sup>1, 3447 cm<sup>−</sup>1, and 3571 cm<sup>−</sup>1. The strongest band from 1000–1100 cm<sup>−</sup><sup>1</sup> is the stretching of PO4<sup>3</sup>− vibrations. The band at approximately 1400–1500 cm<sup>−</sup><sup>1</sup> corresponds to the carbonate group of HA and CHA. The band of OH stretching of HA appears from 1000–1100 cm<sup>−</sup>1. CHA also shows all of the bands of PO4<sup>3</sup><sup>−</sup>, CO3<sup>2</sup><sup>−</sup>, and OH, and because CHA was

extracted by alkaline lysis, the band of CO3<sup>2</sup>− for CHA is more prominent than that for HA (Figure 2A).

**Figure 2.** Characterization of hydroxyapatite isolated from frame of *Paralichthys olivaceus* and sigma. (**A**) Fourier transform infrared spectra of HA and CHA; (**B**) X-ray diffraction spectra of HA and CHA. The energy dispersive spectrometer of (**C**) HA and (**D**) CHA.

XRD analysis is a reliable method for investigating the phase purity and crystallinity of a compound and determining the quantitative and qualitative aspects of a solid compound. Results of XRD are mainly evaluated through comparison with the International Center for Diffraction Data (ICDD) standards. The crystallinity and purity of HA and CHA were defined by XRD analysis. The XRD peaks of the standard and the HA and CHA from the bones of *P. olivaceus* are shown in Figure 2B. The obtained peaks of HA and CHA at 2-theta were identical to 01-086-0740 from the ICDD. The peaks of HA and CHA matched with those of the standard ICDD 01-086-0740 (Hydroxyapatite; Ca5(PO4)3OH) (Figure 2B).

### *2.4. Energy Dispersive Spectrometer (EDS)*

Figure 2C,D displays the EDS data that confirm the content of C, O, Na, Mg, Cl, P, and Ca in the powder. The ratio of Ca/P for the HA powder was 1.578, and that for the CHA powder was 1.96. These results show that CHA has carbonate groups, because it was extracted by alkaline hydrolysis.

### *2.5. Characterization of the CHA-Reinforced Scaffolds*

The mean strut diameter of the CHA-reinforced scaffolds was controlled by adjusting the extruding temperature, nozzle diameter, and speed of the extruder. Based on SEM observations, the strut diameters of PCL, 2.5% CHA/PCL, 5% CHA/PCL, 10% CHA/PCL, 10% HA/PCL, MC/PCL, 2.5% CHA/MC/PCL, 5% CHA/MC/PCL, 10% CHA/MC/PCL, and 10% HA/MC/PCL scaffolds were 518.54 ± 5.72 μm, 607.27 ± 3.34 μm, 571.37 ± 16.43 μm, 561.03 ± 16.5 μm, 484.57 ± 19.68 μm, 508.17 ± 11.62 μm, 599.42 ± 2.94 μm, 526.28 ± 15.34 μm, 539.12 ± 15.34 μm, and 466.67 ± 16.04 μm, respectively. For these results, the strut diameter and pore size of the CHA-reinforced scaffolds were fabricated under the same conditions, but slight differences were observed between the scaffolds (Figure 3A, Table 2).

**Figure 3.** Morphology and characterization of the CHA-reinforced scaffolds. (**A**) SEM image of non-coated group and MC coated group; (**B**) FTIR analysis of non-coated group and MC coated group; (**C**) load-extension curve of non-coated group and MC coated group.

**Table 2.** Strut diameter and elastic modulus of CHA-reinforced scaffolds.


The FTIR spectra of the CHA-reinforced scaffolds were in the spectral range of 4000–650 cm<sup>−</sup>1. The FTIR spectrum of the PCL scaffold was observed at 2860 cm<sup>−</sup><sup>1</sup> (C-H) and 1720 cm<sup>−</sup><sup>1</sup> (C=O) stretching peaks, and the FTIR spectrum of the 10% CHA/PCL scaffold was observed at the same peaks. For the 10% CHA/MC/PCL scaffold, in addition to the peaks of the PCL scaffold, the amide peaks (amide A, B, I–III) of MC were also identified (Figure 3B).

The mechanical properties of the CHA-reinforced scaffolds were evaluated using tensile mechanical testing in a universal testing machine. As shown in Figure 3C, the elastic modulus values of PCL, 2.5% CHA/PCL, 5% CHA/PCL, 10% CHA/PCL, 10% HA/PCL, MC/PCL, 2.5% CHA/MC/PCL, 5% CHA/MC/PCL, 10% CHA/MC/PCL, and

10% HA/MC/PCL scaffolds were 6.29 ± 0.28 MPa, 10.19 ± 0.01 MPa, 9.26 ± 0.33 MPa, 6.85 ± 0.55 MPa, 7.76 ± 0.37 MPa, 6.37 ± 0.16 MPa, 9.38 ± 0.45 MPa, 9.1 ± 0.12 MPa, 7.08 ± 0.52 MPa, and 7.77 ± 0.42 MPa, respectively. The elastic modulus value tended to increase in the scaffolds containing CHA and HA, compared with that of PCL; furthermore, the elastic modulus value decreased as the content of CHA and HA increased. These results may be due to the diameter of the strut and the size of CHA and HA (Figure 3C, Table 2).

### *2.6. Cell Viability on the CHA-Reinforced Scaffolds*

Cytotoxicity to the CHA-reinforced scaffolds was evaluated using a cell live/dead assay 7 days after cell seeding on the scaffolds. At 7 days after cell seeding, the contents of the CHA and HA did not affect cytotoxicity and cell distribution. In addition, on the 7th day, the presence of an MC coating on the CHA-reinforced scaffold did not affect the cell distribution. However, the detection amount of PI in the MC-coated scaffolds was lower than that of the MC-uncoated scaffolds (Figure 4A). As a result of the cell viability for 3, 5, and 7 days, it was possible to confirm a larger number of living cells in the MC-coated scaffold group.

**Figure 4.** Cell viability of non-coated and MC-coated scaffolds on MC3T3-E1. ( **A**) cell fluorescence image stained FDA/PI; and (**B**) cell viability for 3, 5, and 7 days.

*2.7. Alkaline Phosphatase Activity of the CHA-Reinforced Scaffolds*

The ALP activities of PCL, 2.5% CHA/PCL, 5% CHA/PCL, 10% CHA/PCL, 10% HA/PCL, MC/PCL, 2.5% CHA/MC/PCL, 5% CHA/MC/PCL, 10% CHA/MC/PCL, and

10% HA/MC/PCL scaffolds were 100 ± 5.08, 104.63 ± 4.91, 110.78 ± 3.28, 100.59 ± 1.27, 111.31 ± 1.81, 152.20 ± 5.09, 154.74 ± 2.59, 153.75 ± 3.26, 152.50 ± 2.62, and 153.05 ± 2.38, respectively. These results confirm that the ALP activity was increased in the MC-coated group on the scaffold (Figure 5A).

**Figure 5.** In vitro effect of scaffolds on (**A**) ALP activity and (**B**) mineralization activity during osteogenic differentiation of MC3T3-E1 cells. \* *p* < 0.05 was considered to indicate a statistically significant difference compared with non-coated scaffolds.

### *2.8. Mineralization of the CHA-Reinforced Scaffolds*

The PCL, MC/PCL, 2.5% CHA/MC/PCL, 5% CHA/MC/PCL, and 10% CHA/MC/ PCL scaffolds were stained with Alizarin Red S stain, and calcium deposition on the scaffolds was observed. Along with the increasing concentration of CHA, the scaffolds created calcium in a dose-dependent manner. At 21 days, the 10% CHA/MC/PCL scaffold showed eight times more mineralization than the PCL scaffold (Figure 5B).

### *2.9. In Vivo Experiments*

To confirm the bone regeneration ability of the scaffolds, the PCL, 10% CHA/MC/PCL, and 10% HA/MC/PCL scaffolds were implanted into a defect in the mouse calvarial defect model.

In the defect site, a difference in bone regeneration was confirmed by micro-CT between the non-treatment, PCL, CHA, and HA reinforced scaffolds. Twenty weeks after surgery, the area of the bone defect was determined by micro-CT. The 10% CHA/MC/PCL and 10% HA/MC/PCL scaffolds showed better regeneration, due to the synergistic effects of their 3D structure with CHA and MC for promoting bone regeneration, than the non-treatment group. Further investigation of the 3D reconstruction was conducted by analyzing the bone defect areas. This analysis showed that the bone volume of the 10% CHA/MC/PCL and 10% HA/MC/PCL scaffold groups have a greater quantity of bone volume regenerated than those of the non-treatment group (Figure 6A).

Histological analysis was performed with hematoxylin and eosin (HE), picrosirius red, and Masson's trichrome (MT) staining of the bone defect area. At low magnification, the new bone and host bone were separated to define a 3-mm bone defect. Although the bone defect did not completely regenerate, the CHA- and HA-reinforced scaffold groups showed better regeneration than the non-treatment and PCL scaffold groups. As shown in the HE and MT staining images (Figure 6B), both CHA- and HA-reinforced scaffold groups formed regenerative tissue around the scaffolds. In addition, picrosirius red staining confirmed that the collagen formation around the scaffold was improved.

**Figure 6.** In vivo performance evaluation. **(A)** 3D reconstruction images of bone defect areas and (**B**) histological analysis of the effect of scaffolds on bone regeneration in vivo. Note: "S" represents scaffold; "black arrow" indicates new bone; "white arrow" indicates new type I collage. \* *p* < 0.05 was considered to indicate a statistically significant difference compared with PCL group.
