*Article* **Effects of Gamma Radiation-Induced Crosslinking of Collagen Type I Coated Dental Titanium Implants on Osseointegration and Bone Regeneration**

**Won-Tak Cho 1,†, So-Yeun Kim 2,†, Sung-In Jung 3, Seong-Soo Kang 4, Se-Eun Kim 4, Su-Hyun Hwang 1, Chang-Mo Jeong <sup>1</sup> and Jung-Bo Huh 1,\***


**Abstract:** This study aimed to compare two methods of crosslinking collagen type I on implanted titanium surfaces, that is, using glutaraldehyde (GA) or gamma-rays (GRs), in a beagle dog model. For in vivo experiments, implants were allocated to three groups and applied to mandibular bone defects in beagle dogs; Group SLA; non-treated Sandblasted, large grit, acid-etched (SLA) implants, Group GA; SLA implants coated with GA crosslinked collagen type I, Group GR; SLA surface implants coated with collagen type I and crosslinked using 25 kGy of 60Co gamma radiation. New bone μCT volumes were obtained, and histologic and histometric analyses were performed in regions of interest. The GR group had significantly better new bone areas (NBAs) and bone to implant contact (BIC) results than the SLA group (*p* < 0.05), but the GA and GR groups were similar in this respect. New bone volumes and inter-thread bone densities (ITBD) were non-significantly different in the three groups (*p* > 0.05). Within the limits of this study, gamma-ray collagen crosslinking on titanium implants can be considered a substitute for glutaraldehyde crosslinking.

**Keywords:** bone regeneration; collagen; gamma radiation; surface modification; titanium implant

#### **1. Introduction**

The interaction between bone and implant interfaces is the key to osseointegration, and various methods of modifying the surfaces of titanium implants have been introduced to improve this process [1–3]. Ti surface modifications influence bone regeneration and biocompatibility and facilitate successful implant fixation without soft tissue intervention [4–6]. Increasing surface roughness and coating implants with biocompatible materials or growth factors are known to increase the osseointegration of Ti implants [7]. In particular, collagen type I is used as a biocompatible polymer because it promotes osteoblast differentiation and provides a suitable environment for bone formation [8–10].

At the molecular level, collagen type I has a tangled, triple-helix structure with two α1 (I) and one α2 (I) polypeptide chains, and many years of clinical use have proven it to be a biocompatible, bioactive, bioresorbable material [11,12]. Implant surfaces coated with crosslinked collagen type I provide a favorable environment for initial osteoblast adhesion and stimulate their proliferation [9]. However, rapid absorption and decomposition

**Citation:** Cho, W.-T.; Kim, S.-Y.; Jung, S.-I.; Kang, S.-S.; Kim, S.-E.; Hwang, S.-H.; Jeong, C.-M.; Huh, J.-B. Effects of Gamma Radiation-Induced Crosslinking of Collagen Type I Coated Dental Titanium Implants on Osseointegration and Bone Regeneration. *Materials* **2021**, *14*, 3268. https://doi.org/10.3390/ma14123268

Academic Editor: Paolo Cappare

Received: 28 April 2021 Accepted: 10 June 2021 Published: 13 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

by enzymes and immune reactions against animal-derived collagen cause type I collagen degradation; therefore, crosslinking is required to improve its in vivo stability [13]. Glutaraldehyde (GA) is commonly used as a crosslinker for collagen-based biomaterials, and GA cross-linking of collagen decreases its antigenicity, makes it resistant to phagocytosis, and invisible to the immune system [9,14,15]. However, like other chemical crosslinking methods, GA has been reported to produce harmful cytotoxic residues and increase proinflammatory cytokine release by macrophages [16–19]. Recently, different types of irradiation-induced crosslinking methods such as gamma-ray and ultraviolet have been used in preference to chemical crosslinkers substances to crosslink polymers like collagen [13,20,21].

Unlike ethylene oxide or GA sterilization, gamma radiation leaves no harmful residues that could potentially harm human health or the environment and is used to sterilize medical devices [6,22]. Moreover, gamma radiation-induced polymer crosslinking enables control of radiation-induced decomposition reactions, e.g., polymer chain scission, which can cause molecular weight reductions, as its effects are not dependent on material compositions [23–25]. Furthermore, when collagen is irradiated with gamma rays, peptide bonds are destroyed due to amino acid deformation, and hydrophilicity is improved by hydrogen bond formation [26]. In addition, enhancements of sandblasted, large grit, acid-etched (SLA) implant surface hydrophilicity have been reported to increase alkaline phosphatase (ALP) by more than 2-fold in cell culture experiments [27].

A previous comparative study concluded that there was no difference between the cytotoxicities of the gamma radiation crosslinked group and a GA-crosslinked group, based on absorbance data. However, gamma crosslinked collagen-coated Ti implants had significantly higher BICs than non-coated controls in a small animal model [28]. Therefore, we compared the effects of GA and gamma-ray crosslinking of collagen type I on the surfaces of SLA Ti implants in a beagle model to determine the effectiveness of gammainduced cross-linking. The null hypothesis was that bone regeneration and osseointegration after GA or gamma crosslinking of collagen type I coated SLA implants are similar.

#### **2. Materials and Methods**

#### *2.1. Experimental Materials*

Collagen type I solution (0.5% (*w*/*v*)) was obtained by dissolving collagen (source: porcine skin, atelocollagen type I, Matrixen-PSP, Sk Bioland Co. Ltd., Cheonan, Korea) in 0.05 M acetic acid (Sigma-Aldrich, St. Louis, MO, USA) at room temperature. The Ti implant fixtures (D 4.0 mm × H 8.0 mm, SLA surface, Cowellmedi Co., Ltd., Pusan, Korea) were placed in a 0.5% (*w*/*v*) collagen type I solution. Bubbles on implant surfaces were removed by sonication (Elmasonic, S 180 H, Elma Schmidbauer, Elma, Germany) for 10 min. Then implants were placed in climate chambers (MIR-253, SANYO, Moriguchi, Japan) to dry for 1 h at 4 °C. Implants in the GA group were crosslinked by placing them in 2.5% (*v*/*v*) GA (DAEMYUNG CHEMICAL, Gyeonggi-do, Korea) for 1 h. Unreacted GA and collagen type I were then removed by washing in distilled water, dried in a vacuum oven (WOV-30, DAIHAN Scientific Co.Ltd., Gangwon-do, Korea) for 3 days [28], and sterilized with ethylene oxide (Manufacturer, City, State, Country). The implants of the GR (gamma-radiation) group were immersed in collagen solution in the same way as in the GA group, followed by ultrasonic cleaning for 10 min, and dried in a climate chamber for 1 h. The GR group implants were then irradiated with 60Co gamma rays (MDS Nordion, Ottawa, ON, Canada) at 25 kGy for1h[28].

#### *2.2. In Vitro Study*

2.2.1. Scanning Electron Microscopy (SEM) Analysis

Surface images of implants were obtained using an SEM unit (Hitachi S3500N, Hitachi, Tokyo, Japan) at magnifications of ×40, ×5000, and ×50,000. For the SEM study, implants were splutter-coated with gold (SCD 005, BAL-TEC, Balzers, Liechtenstein). SEM images were obtained at 15 kV.

#### 2.2.2. X-ray Photoelectron Spectroscopy (XPS)

Implant surfaces were analyzed by XPS (AXIS SUPRA, Kratos Analytical Ltd., Manchester, UK) using a monochromatic Al-Kα (1486.6 eV) X-ray source (1486.6 eV) at 15 kV and 225 W. The binding energy scale was calibrated at the C 1s level (284.5 eV). Implants in each group were subjected to a compositional survey at a pass energy of 160 eV, and core level spectra were obtained at a pass energy of 20 eV. Data analysis was performed using data reduction software (Vision 1.5, Kratos Analytical Ltd., Manchester, UK). Deconvoluted spectra were fitted using a Gaussian−Lorentzian sum function (20% Gaussian and 80% Lorentzian) using XPSPEAK Version 4.1 (Dr. Raymond Kwok, Hong Kong, China).

#### *2.3. In Vivo Experiment*

#### 2.3.1. Experimental Animals

This study was approved by the Ethics Committee on Animal Experimentation of Chonnam National University (CNU IACUC-YB-2018-94). Six beagles (males, three years old, 12 kg) were used in the study.

#### 2.3.2. Surgical Procedure

Beagles were anesthetized with a medetomidine (Tomidin®, Provet, Istanbul, Turkey) 10 μg/kg and tiletamine-zolazepam (Zoletil 50®, Virbac Laboratories, Carros, France) at 5 mg before the procedure and followed by isoflurane inhalation anesthesia (Sevoflurane®, Hana Pharm Co., Seoul, Korea). Anesthesia was maintained using tramadol (Maritrol®, Cheil Pharmaceutical, Uiwang, Korea) 2 mg/kg and carprofen (Rimadyl® inj, Zoetis, Parsippany, NJ, USA) 2.2 mg/kg IV. In addition, infiltration anesthesia at surgical sites was performed using 0.4 mL bupivacaine (Bupivacaine HCl 0.5% Inj., Myungmoon Pharm Co., Seoul, Korea). To prevent infection, 20 mg/kg of cefazolin sodium (Cefazolin®, Chongkundang Pharm Co., Seoul, Korea) was injected subcutaneously.

Mandibular premolars (P1–P4) and M1 molar were extracted after full mouth scaling. Implants were placed after extraction sites had healed for 8 weeks [28,29]. General anesthesia and local infiltration anesthesia were applied as described for extractions. A mid-crestal incision was made at each premolar site, and vertical incisions were made at the mucogingival junction. After mucoperiosteal flap elevation, crestal bone was homogenized by osteoplasty using a bone file and rongeur. Buccal cuboid defects, approximately 5 mm in height from crestal bone, 5 mm deep from the surface of the buccal bone, and 8 mm in width mesiodistally, were created using a straight fissure carbide bur under saline irrigation (JW Pharmaceutical Co. Ltd., Gyeonggi-do, Korea) (Figure 1A). Animals were allocated randomly to the three study groups, which were as follows:


Then, 36 implants (Cowell Medi Co, Ltd., Busan, Korea), 4 mm in diameter and 8 mm high, were implanted in the mandibular defects of 6 animals to expose three threads (Figure 1B). Peri-implant defect sites were grafted with porcine xenografts (Bone-XP, MedPark, Busan, Korea) (Figure 1C), and bone regeneration was guided using resorbable collagen membrane (Bone-D, MedPark, Busan, Korea) (Figure 1D). Surgical sites were sutured with 4-0 Vicryl (Mersilk, Ethicon Co., Livingston, UK). Post-operative care consisted of oral amoxicillin-clavulanate (Amocla®, Kuhnil Pharm Co., Seoul, Korea) 12.5 mg/kg, firocoxib (Previcox, Merial, France) 5 mg/kg, and famotidine (Famotidine®, Nelson, Seoul, Korea) at 0.5 mg/kg for 2 weeks.

Eight weeks after implant placements, animals were sacrificed by potassium chloride intravenous injection (JW Pharmaceutical Co. Ltd., Gyeonggi-do, Korea) under general anesthesia, and mandibular bones were harvested and fixed in neutral buffered formalin (Duksan Pure Chemical. Co. Ltd, Gyeonggi-do, Korea) for 2 weeks.

**Figure 1.** Surgical procedures used to place surface-treated implants in beagle mandibles. (**A**) Creation of buccal cubic defects, (**B**) Implant placement, (**C**) Distribution of bone graft material, (**D**) Collagen membrane placement.

#### 2.3.3. Micro-Computed Tomography (μCT) Analysis

Mandibles were wrapped with Parafilm M® (Heathrow Scientific, Vernon Hills, IL, USA) and scanned by μCT (Skyscan-1173, ver. 1.6, Bruker-CT Co., Kontich, Belgium) at 130 kV and an intensity of 60 μA to obtain the μCT images of regions of interest (ROIs). We used a pixel resolution of 24.15 μm to determine new bone volumes (NBVs) in defect areas around implants. μCT image reconstructions were performed using Nrecon reconstruction software ver. 1.7.0.4 (Bruker-CT Co., Kotich, Belgium). The study used 1 mm diameter ROIs around implants (Figure 2).

**Figure 2.** The μCT images of regions of interest (ROIs) which included 1 mm around each implant. (**A**) buccal view, (**B**) occlusal view.

#### 2.3.4. Histologic Analysis

After μCT analysis, mandibular bone specimens were dehydrated in an ethanol series (Duksan Pure Chemical. Co. Ltd, Gyeonggi-do, Korea) 70, 80, 90, and 100%, infiltrated with resin (Technovit 7200, Heraeus KULZER, Hanau, Germany) for a week, fixed to an embedding frame, and embedded using a UV curing system (KULZER EXAKT 520, Heraeus Kulzer, Norderstedt, Germany). Polymerized specimens were sectioned at 400 μm at implant centers using a diamond cutter (KULZER EXAKT 300 CP Band System, Exakt Apparatebau, Norderstedt, Germany). Then, they were polished to a thickness of 30 μm using an EXAKT grinding machine (KULZER EXAKT 400CS, Exakt Apparatebau, Norderstedt, Germany), mounted on slides, and stained with hematoxylin and eosin (H&E). Images of stained specimens were obtained using a light microscope (Olympus BX, Olympus, Tokyo, Japan). BIC and ITBD values and new bone areas (NBAs) were measured using an image analysis program (ver. 7.5, i-solution, IMT i-solution. Inc., Vancouver, BC, Canada) by a trained investigator (Figure 3). ROIs were set at exposed three upper threads and 1 mm around fixtures, as shown in Figure 2.

NBAs (%) = New bone area (mm2)/Total ROI area (mm2) <sup>×</sup> 100 (1)

BIC (%) = Length of the new bone to implant contact (mm2)/Total ROI length of implant (mm2) <sup>×</sup> 100 (2)

ITBDs (%) = New bone area of inter thread (mm2)/Total area of inter thread (mm2) <sup>×</sup> 100 (3)

**Figure 3.** Histometric measurements in regions of interest (ROIs). ROIs were fixed from implant platforms to the third thread and at 1 mm around implants in occlusal view. (**A**) NBA: New bone area, (**B**) BIC: Bone-to-implant contact, (**C**) ITBD: Inter-thread bone density.

#### 2.3.5. Statistical Analysis

Results are presented as means ± standard deviations (SDs), and the analysis was performed using SPSS Ver. 25 (SPSS Inc., Chicago, IL, USA). Since NBAs, ITBDs, NBV, and BIC values were not normally distributed by the normality test, the Kruskal-Wallis one-way analysis was used to determine the significances of intergroup differences. The Mann-Whitney U test was applied as a post hoc test. Statistical significance was accepted for *p* values < 0.05.

#### **3. Results**

#### *3.1. In-Vitro Study*

#### 3.1.1. Collagen Crosslinked Ti Implant Surface Morphologies

When collagen was crosslinked using GA or 25 kGy Gamma rays on SLA implant surfaces, surface morphologies were similar due to their rough SLA surfaces (Figure 4).

**Figure 4.** The scanning electron microscopy (SEM) images. (**A**) The SLA (Sandblasted, large grit, acid-etched implant surface) group, (**B**) The GA (glutaraldehyde) group, and (**C**) the GR (gamma-radiation) group. [Original magnifications: ×40, ×5000, and ×50,000].

#### 3.1.2. XPS Findings

Surface elemental compositions were determined by XPS (Figure 5). The SLA group had the lowest nitrogen content (0.33%), followed by the GA group (6.22%) and the GR group (17.64%). Since the major component of collagen is gelatin (a protein), a large amount of nitrogen indicates good crosslinking [30] (Table 1).

h h h

**Figure 5.** Surface XPS spectra of the three study groups. (**A**) The SLA group, (**B**) the GA group, and (**C**) the GR group.


**Table 1.** Atomic concentrations (at. %) on implant surfaces as determined by XPS.

#### *3.2. In Vivo Study*

#### 3.2.1. Clinical Findings

All beagles survived the surgical procedures without complications, such as inflammation or infection. Mandibular jaw segments were harvested after sacrifice.

#### 3.2.2. Micro-Computed Tomography (μCT) Findings

In regions of interest, NBV was 64.78 ± 3.24% in the GR group, 61.42 ± 7.07% in the GA group, and 56.06 ± 7.31% in the SLA group. Thus, although NBV was relatively high in the GR group, differences were not significant (Figure 6).

**Figure 6.** The volumetric ROI analysis of new bone.

#### 3.2.3. Histological Findings

The histological results of the SLA, GA, and GR groups are shown in Figure 7. No abnormal inflammatory cells or singularities were found in any group. However, new bone formation was observed between the third and second threads in the SLA group but distributed evenly in all the GA and GR groups. The crosslinked groups exhibited more new bone formation than the SLA group, but the new bone formation was similar in the GA and GR groups.

**Figure 7.** H&E stained sections at 8 weeks post-implantation. (**A**) The SLA group, (**B**) the GA group, (**C**) the GR group, (**a**) ×12.5, (**b**) ×40, (**c**) ×100. Note: NB = New bone, GM = Bone graft material, CT = Connective tissue, M = Membrane.

3.2.4. Histometric Findings

Histometric results are summarized in Table 2 and Figure 8. NBA values of the SLA, GA, and GR groups were 38.27 ± 9.34%, 52.37 ± 7.93%, and 43.77 ± 8.81%, respectively, and were significantly higher in the GR group than in the SLA group (*p* < 0.05).


**Table 2.** Mean values of new bone areas (NBAs), inter-thread bone densities (ITBDs), and bone to implant contacts (BICs) as determined by histometric analysis.

\* Indicates statistical significance (*p* < 0.05).

**Figure 8.** Histometric analysis within regions of interest (ROI). (**A**) New bone area (%), (**B**) Inter thread bone density (%), (**C**) Bone-implant contact (%). \* Indicates statistical significance (*p* < 0.05).

ITBD results of the SLA, GA, and GR groups were 49.52 ± 5.11%, 58.10 ± 12.33%, and 64.10 ± 5.65%, respectively, and no significant intergroup difference was found (*p* > 0.05). Corresponding BIC results were 47.3 ± 6.58%, 54.61 ± 9.4%, and 60.19 ± 11.23%, and BIC was significantly greater in the GR group than in the SLA group (*p* < 0.05). On the other hand, the results of the GR group were similar to the values of the GA group in NBA, ITBD, and BIC (*p* > 0.05).

#### **4. Discussion**

Commercially available dental implants are generally considered to have high biocompatibility and surfaces suitable for bone regeneration [4,31], and this is supported by the results of prospective and retrospective clinical studies, which reported implant 10-year survival rates exceeding 90% [32–35]. Nevertheless, dental implant failure due to osseointegration failure often occurs unexpectedly and remains an important clinical problem [36,37]. Therefore, studies on implant surface modification methods have also been conducted to improve osseointegration using surface treatments and collagen as bioactive material [38,39]. However, extracted collagen's mechanical properties and stabilities are inferior; thus, its potential is limited [40,41]. GA has been used as a collagen crosslinking agent for several decades, but some GA probably remains in situ after crosslinking. Protocols for removing unreacted GA have been proposed to solve this problem, but unfortunately, these methods have also been reported to have cytotoxic side effects [42,43]. On the other hand, gamma-ray-based crosslinking does not leave harmful residues and has recently been used to crosslink polymers, including collagen [6]. Therefore, this study was conducted to evaluate and compare the merits of crosslinking collagen type I on the surfaces of SLA implants with gamma-ray radiation or GA in a large animal model.

Collagen type I is a useful biopolymer and widely used clinically due to its low immunogenicity, biocompatibility, and biomedical potential [42]. In addition, collagen is known to promote osteoblast adhesion when coated on implant surfaces [9]. Previous in vivo studies have confirmed that collagen treatment promotes bone regeneration following implantation of crosslinked collagen-coated Ti implants and that collagen treatment enhances bone to implant adhesion to bone and accelerates bone formation [44,45]. Likewise, in the present study, NBAs and BIC values were higher in the GR group than in the SLA group, similar to the GA and SLA groups, which suggests 25 kGy gamma-ray exposure provides better crosslinking than GA. Furthermore, XPS analysis showed surface nitrogen levels (17.64%) were higher in the GR group than in the GA group (6.22%). However, the GR group did not significantly differ compared to the GA group (*p* > 0.05).

After machining Ti, its surface is contaminated by adsorbed organic entities such as atmospheric hydrocarbons, water, or cleaning fluids [46,47]. Previous studies that analyzed the chemical compositions of different implant surfaces by XPS have reported carbon deposition percentages ranging from 17.9 to 76.5% [48]. Therefore, gamma irradiation at 25 to 35 kGy has been recommended for the rapid disinfection and sterilization of medical devices. Ueno et al. [49] found that deposited hydrocarbons can be removed by high-energy UV or gamma radiation and that the removal of hydrocarbons improves Ti biocompatibility and induces osseointegration. Our XPS results returned surface carbon figures in the GR, GA, and SLA groups of 0.93, 6.22, and 20.96%, respectively, suggesting that surface carbon was removed by gamma irradiation [30]. This observed reduction in surface carbon levels by gamma irradiation is consistent with the results of previous studies [6].

Accordingly, the present study suggests that gamma irradiation-induced collagen crosslinking enhances Ti implant biocompatibility and bone adhesion in beagle mandible models. Collagen cross-linked implants using gamma irradiation may improve the osseointegration in adverse circumstances requiring transcrestal sinus lift procedures [50]. Besides, in patients with a history of systemic disease, increased implant-bone osseointegration may be an important factor for long-term implant survival [51]. Meanwhile, Misch [52] recommended that the occlusal implant area be made small. Since the increased osseointegration increases the mechanical strength of the bone tissue, the occlusion of the implant prosthesis can be properly distributed [53].

Furthermore, if a substance that induces a stem cell response, such as rhBMP-2, is attached to the collagen-crosslinked implant with gamma rays, better osteoinductivity can be expected. However, the study was limited by the model used, the number of beagles involved, and its short duration. Furthermore, there was no difference in the histological aspect compared to the GA group. In addition, it is considered necessary to compare it with other biocompatible materials other than collagen. Accordingly, we recommend additional experiments be performed to establish a scientific basis for the clinical effectiveness of crosslinking collagen on Ti implants using gamma radiation.

#### **5. Conclusions**

This study was conducted to assess the effects of gamma radiation-induced collagen crosslinking on osseointegration and bone regeneration in defect areas around SLA implants. Within the limitations of this study, gamma-ray collagen crosslinking was found to be at least as effective as GA crosslinking in terms of bone regeneration efficacy. According to our results, gamma-radiation can be used to effectively crosslink collagen on implant surfaces and not raise concerns about toxic residues. Additional animal studies are required to determine optimum gamma-radiation dose criteria and to more comprehensively evaluate the effect of irradiation on osseointegration.

**Author Contributions:** Conceptualization, J.-B.H.; methodology, J.-B.H., C.-M.J., S.-E.K., and S.-S.K.; formal analysis, W.-T.C. and S.-H.H.; investigation, W.-T.C. and S.-H.H.; data curation, W.-T.C., S.-H.H., S.-I.J., and S.-Y.K.; writing—original draft preparation, J.-B.H., S.-Y.K., and W.-T.C.; writing review and editing, C.-M.J., W.-T.C., and J.-B.H.; supervision, J.-B.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Research Foundation of Korea (NRF) funded by the Korean government (MSIT)(Grant no. NRF-2020R1A2C1004927).

**Institutional Review Board Statement:** The study was approved by the Institutional Animal Care and Use Committee (IACUC) of Chonnam National University (CNU IACUC-YB-2018-94, 07-01-2019).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data sharing not applicable.

**Conflicts of Interest:** The authors have no conflict of interest to declare.

#### **References**


## *Article* **Settable Polymeric Autograft Extenders in a Rabbit Radius Model of Bone Formation**

**Lauren A. Boller 1, Madison A.P. McGough 1, Stefanie M. Shiels 2, Craig L. Duvall 1, Joseph C. Wenke <sup>2</sup> and Scott A. Guelcher 1,3,4,\***


**Abstract:** Autograft (AG) is the gold standard for bone grafts, but limited quantities and patient morbidity are associated with its use. AG extenders have been proposed to minimize the volume of AG while maintaining the osteoinductive properties of the implant. In this study, poly(ester urethane) (PEUR) and poly(thioketal urethane) (PTKUR) AG extenders were implanted in a 20-mm rabbit radius defect model to evaluate new bone formation and graft remodeling. Outcomes including μCT and histomorphometry were measured at 12 weeks and compared to an AG (no polymer) control. AG control examples exhibited new bone formation, but inconsistent healing was observed. The implanted AG control was resorbed by 12 weeks, while AG extenders maintained implanted AG throughout the study. Bone growth from the defect interfaces was observed in both AG extenders, but residual polymer inhibited cellular infiltration and subsequent bone formation within the center of the implant. PEUR-AG extenders degraded more rapidly than PTKUR-AG extenders. These observations demonstrated that AG extenders supported new bone formation and that polymer composition did not have an effect on overall bone formation. Furthermore, the results indicated that early cellular infiltration is necessary for harnessing the osteoinductive capabilities of AG.

**Keywords:** autograft extender; bone; polyurethane

#### **1. Introduction**

Autograft (AG) bone is considered the gold standard in bone grafting. It is osteoinductive, osteoconductive, and osteogenic, and it does not pose a risk for disease transmission [1–3]. AG comes in various forms including both cancellous and cortical [3]. Cancellous AG is most often harvested from the iliac crest (IC); however, other donor sites such as the posterior superior iliac spine, femur, proximal tibia, and distal radius are utilized [4–7]. Cancellous AG contains mesenchymal stem cells (MSCs), osteoblasts, and growth factors including bone morphogenetic proteins (BMPs), which contribute to its osteoinductivity [3,8]. The trabeculae present within cancellous AG allow for enhanced cellular infiltration and vascularization in comparison to cortical AG [8]. Cortical AG is ideal for defects that require structural support as it offers superior mechanical properties compared with cancellous AG. However, cortical AG is less osteoinductive than cancellous AG, and its density results in slower revascularization and inhibits cellular infiltration [8,9]. Despite its osteogenic properties, AG is a scarce resource with multiple drawbacks including donor site morbidity (10–39% of patients), limited availability, the need for a second

**Citation:** Boller, L.A.; McGough, M.A.P.; Shiels, S.M.; Duvall, C.L.; Wenke, J.C.; Guelcher, S.A. Settable Polymeric Autograft Extenders in a Rabbit Radius Model of Bone Formation. *Materials* **2021**, *14*, 3960. https://doi.org/10.3390/ma14143960

Academic Editor: Jung-Bo Huh

Received: 31 May 2021 Accepted: 8 July 2021 Published: 15 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

surgical site [6], and rapid resorption dependent on the bone density and embryologic origin of the AG [10].

The use of allograft from donors is an alternative to AG. Allograft is more readily available than AG and provides structural support, but it does not possess the same osteoinductive capacity as AG due to its processing [2]. Furthermore, allograft faces potential immune rejection and slow osseointegration with host bone [11]. Synthetic materials such as recombinant human bone morphogenic proteins (rhBMPs) have emerged as substitutes for AG [12–14], but none of these alternatives has been shown to match all of the benefits provided by AG. Furthermore, the use of the FDA-approved rhBMP-2 treatment (INFUSE® bone graft, Medtronic) is limited to a few clinical indications [15–17].

To overcome the limitations in AG including availability and rapid resorption, various approaches to increase the overall volume of AG while maintaining its osteogenic and osteoinductive properties have been employed. Clinically, AG is typically blended with an 'extender' to reduce the volume of AG needed for implantation [18,19]. An early study demonstrated the utility of demineralized bone matrix as an AG extender [20] More recently, tissue engineered approaches to incorporate synthetic bone substitutes with AG have been investigated. Calcium phosphates (CaPs) such as β-tricalcium phosphate (β-TCP) and hydroxyapatite were evaluated as AG extenders for spinal applications [21–24]. Similarly, poly(propylene fumarate)- and poly(lactide-co-glycolide) (PLGA)-based polymer AG extenders have also been evaluated for spinal applications [25–28], while AG extenders utilizing bioactive glass particles have been investigated in the femur [29].

Lysine-based poly(ester urethanes) (PEURs) and poly(thioketal urethanes) (PTKURs) have been previously investigated in bone regeneration applications [30–33]. The mechanical properties of these materials can be easily altered, and the addition of ceramic particles, AG, and allograft supports new bone formation at various anatomic sites [34–37]. Previous work has demonstrated selective, cell-mediated, first-order degradation of PTKUR in vivo [38]. Furthermore, low-porosity PTKURs utilized in rabbit intertransverse processes [39] and femoral plugs [33] exhibited new bone formation, but minimal PTKUR degradation was observed. Slow degradation is advantageous in applications in which mechanical stability is required; however, in applications utilizing biologics, faster graft resorption is necessary to harness the osteoinductivity. In a previous study, PEUR was used to deliver rhBMP-2 and demonstrated balanced polymer resorption and new bone formation [34,40]. Therefore, we compared PEUR [41,42] with PTKUR [38] as an AG extender to test the hypothesis that faster degrading PEUR would support increased cellular infiltration and bone formation in a rabbit radius model.

Herein, settable and resorbable PTKUR-AG and PEUR-AG extenders were implanted into a 20 mm critical-sized segmental defect in the rabbit radius to investigate the effects of polymer composition on cellular infiltration, new bone formation, and polymer resorption. In this study, PTKUR or PEUR was blended with fresh IC AG and the resulting material subsequently molded to size and implanted in the defect. In vivo outcomes assessed post-operatively with X-ray, μCT, histology, and histomorphometry were compared to an AG control.

#### **2. Materials and Methods**

#### *2.1. Materials*

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA) with the exception of anhydrous diethyl ether purchased from Fisher Scientific. Lysine triisocyanatepolyethylene glycol (LTI-PEG) prepolymer (NCO = 21.7%) was obtained from Ricerca Biosciences LLC (Concord, OH, USA).

#### *2.2. Polyester Triol and Thioketal Diol Synthesis*

The polyester triol (molecular weight 450 g mol−1) was synthesized utilizing a previously published method [43]. Briefly, glycerol, 70% ε-caprolactone, 20% glycolide, and 10% DL-lactide monomers were mixed for 40 h under argon at 140 ◦C. The resulting fluid was vacuum dried at 80 ◦C for 48 h. Thioketal (TK) diol was synthesized utilizing a previously published method [32]. Briefly, 2,2-dimethoxypropane and thioglycolic acid were reacted in the presence of bismuth (III) chloride at room temperature for 24 h. The resulting solution was filtered, dissolved in tetrahydrofuran, and added dropwise to LiAlH4 under anhydrous conditions. The reaction was refluxed at 52 ◦C for 18 h and the product filtered and vacuum dried for 48 h.

#### *2.3. AG Extender Fabrication*

PTKUR- and PEUR-AG extenders were fabricated by adapted two-component reactiveliquid molding methods as previously described [39]. Briefly, polyisocyanate comprised of either TK diol or polyester triol, 10 pphp iron acetylacetonate (FeAA) catalyst in caprolactone 0.5% (*w*/*w*), and LTI-PEG prepolymer were mixed together. Morselized AG (70 wt%) was added to the mixture and stirred by hand until homogeneous. The resulting mixture was injected as a viscous paste that was cured to form a solid implant in situ. The targeted index (NCO:OH) was 200.

#### *2.4. AG Extenders in a Rabbit Radius Defect*

Adult New Zealand White rabbits were used in this study (*n* = 12). The protocol was approved by the Ethics Committee of the U.S. Army Institute of Surgical Research (A-18-035). Animals were randomly assigned to PEUR-AG, PTKUR-AG, or AG control treatment groups (*n* = 4 per group). Assuming an effect size of 0.999 (determined from a previous study [41]) and alpha of 0.05, an a priori power analysis determined that a sample size of *n* = 3 would provide a power of 0.95. Thus, 4 animals per group were considered to provide sufficient power for this study. Animals were premedicated with slow-release buprenorphine (0.1 mg kg<sup>−</sup>1) and anesthetized with isoflurane (1–3%). For all groups, the animal's left hindlimb and right forelimb were shaved and prepared for sterile surgery using alternating washes of alcohol and povidone-iodine. The left IC was exposed, and AG (0.6–0.7 g) was harvested using an oscillating saw. Excess soft tissue was removed and a bone mill (R. Quétin) was used to morselize the harvested bone. The IC harvest site was closed and the right radius exposed. An oscillating saw was used to create a 20 mm segmental defect in the radius. AG extenders were prepared as explained above and shaped to size (5 mm × 20 mm). AG control (morselized AG without PTKUR or PEUR) was molded to shape and carefully placed within the defect. A surgical elevator was used to place the AG extenders in the defects to ensure correct placement. AG extenders were allowed to cure in situ (Figure 1) after which the radial site was closed. Post-operative X-ray images (Faxitron X20) were taken throughout recovery and Calcein green and Xylenol orange fluorochromes were injected at 4 and 8 weeks post-operatively, respectively, to evaluate bone remodeling temporally. Animals were anesthetized and euthanized at 12 weeks. The radii were harvested and placed into formalin for further analysis.

#### *2.5. μCT Analysis*

μCT analysis was performed using a μCT50 (SCANCO, Brüttisellen, Switzerland). Radii were scanned at 70 kVp energy, 200 μA source current, 1000 projections per rotation, 800 ms integration time, and 17.2 μm voxel size. In order to spatially evaluate bone growth throughout the defect, bone area was calculated for each axial section (17.2 μm) totaling 20 mm. The area of interest (AOI) included the proximal onset of the defect and extended the length of the defect. It is not possible to distinguish AG from old or new bone utilizing μCT; thus, the ulna was included in analysis due to bone formation observed within the interosseous syndesmosis interfacing the ulna in some of the samples. The bone area was plotted as a function of defect length where 0 mm and 20 mm represented the proximal and

distal ends of the defect, respectively. The bone volume (BV) and total volume (TV) within the AOI were measured to calculate the bone volume fraction (BV/TV). Additionally, trabecular thickening (Tb. Th.), trabecular spacing (Tb. Sp.), and trabecular number (Tb. N.) were evaluated.

**Figure 1.** Surgical images. AG control, PTKUR-AG extender, and PEUR-AG extender in the 20 mm defect prior to closure.

#### *2.6. Histological Evaluation*

Non-decalcified histology was utilized to evaluate cellular infiltration, new bone formation, and residual polymer (*n* = 4 per treatment group) [38,42]. After formalin fixation, radii were dehydrated and embedded in poly (methyl methacrylate). Serial coronal sections were cut from the center of each defect with an Exakt band saw. Sections were polished and stained with Sanderson's Rapid Bone Stain to assess osteogenesis and remodeling. Safranin O staining was also performed to assess endochondral ossification. An unstained section was utilized to analyze fluorochrome binding. High-magnification histological images were obtained via bright-field and fluorescent microscopy (Olympus BX41, Tokyo, Japan).

For quantitative histomorphometry, images were taken at 4× via-bright field and fluorescent microscopy (Biotek Cytation). The AOI was defined as a 20 × 5 mm rectangular region that encompassed the entirety of the graft and defect. The ulna was excluded from the AOI. The same AOI was used for both Sanderson's Rapid stained and fluorescent sections. Quantification of new bone, infiltrating cells and tissue, and residual polymer was performed using Metamorph (Version 7.0.1). Bone was thresholded either as red (Sanderson's rapid) or green/orange (fluorochromes). Residual material was thresholded as black stain, and infiltrating cells were thresholded as blue/teal. The thresholded area was reported as an area percentage of the total AOI.

#### *2.7. Statistical Analysis*

Data were analyzed utilizing GraphPad Prism (Version 8.4.1) and reported as mean ± standard deviation. Treatment group outcomes at 12 weeks were evaluated using an ANOVA with a Tukey's multiple comparison test. Treatment group outcomes compared at 4 and 8 weeks were evaluated using a two-way ANOVA with Tukey's multiple comparison test. Statistical significance was set at *p* < 0.05.

#### **3. Results**

#### *3.1. Surgical Outcomes*

The surgical procedures and subsequent healing were uneventful. No fractures of the radii occurred. As shown in Figure 2, X-rays displayed healing progression from 0 to 12 weeks in all of the groups. The AG control presented challenges in implantation and shape maintenance during the surgical procedures due to the lack of a settable polymeric extender. However, AG control remained in place throughout the study and displayed at least partial bridging of the defect along the radial side of the defect within three of the four samples (Figure 2A). Both AG extenders were coherent throughout surgical placement and remained stable throughout the entirety of the study (Figure 2B,C). The AG extenders displayed new bone growth at the host bone/graft interfaces, and graft remodeling was observed in both AG extenders, specifically near the proximal and distal ends of the defect where new bone and decreasing residual graft were observed. Bridging of the defect was not observed in any of the AG extender samples. Both AG extenders exhibited increasing opacity within the grafts over the 12-week time course, and no qualitative differences in new bone formation within the graft were observed between PTKUR- and PEUR-AG groups.

**Figure 2.** Representative X-ray images of (**A**) AG control, (**B**) PTKUR-AG, and (**C**) PEUR-AG acquired immediately after the surgical procedures and after 4, 8, and 12 weeks of healing. Areas of bone remodeling and formation are noted by yellow arrows.

#### *3.2. In Vivo Bone Analysis*

Representative μCT images revealed no significant difference in total bone (including new bone, residual AG, and host bone) between groups at 12 weeks (Figure 3A). New bone formation was observed in the interosseous membrane in the space between the radius and ulna. Bone area and volume were quantified by μCT analysis (Figure 3B,C). All groups displayed similar trends of increased bone area at the proximal and distal ends of the defect with a gradual decrease in bone area as the center of the defect was approached (Figure 3B). BV/TV in PEUR-AG extenders trended higher compared with PTKUR-AG (*p* = 0.070) and AG control (*p* = 0.337), but the differences were not significant (Figure 3C). Additional bone morphometric parameters including trabecular thickness (Tb.Th.), trabecular separation (Tb.Sp.), and trabecular number (Tb.N.) did not show significant differences between groups (Supplementary Figure S2).

**Figure 3.** μCT analysis of bone remodeling. Representative μCT images of (**A**) AG control, PTKUR-AG, and PEUR-AG 12 weeks post-operatively. (**B**) Total bone area at 12 weeks measured as a function of defect length by μCT from the proximal to distal interfaces of the defect. Corresponding dotted lines representative standard deviation. (**C**) Bone volume/total volume (BV/TV) at 12 weeks for each treatment group.

New bone formation measured via histological analysis was observed within all groups (Figure 4A and Figure S1). Quantitative histomorphometric analysis at 12 weeks showed no significant difference in new bone formation between PTKUR- and PEUR-AG extenders (Figure 4B). While μCT analysis showed no significant difference in BV/TV between the AG control and extender groups (Figure 3C), histomorphometric analysis showed significantly higher new bone formation compared with both AG extenders (Figure 4B). This discrepancy can be explained in part by the different regions of interest used for μCT (entire defect including the ulna) and histomorphometry (center of the defect excluding the ulna). Although the AG control displayed greater new bone formation, healing within the samples appeared to be inconsistent (Figure S1).

**Figure 4.** New bone formation in AG extenders. (**A**) Representative images of Sanderson's Rapid stained AG control, PTKUR-AG, and PEUR-AG histological sections. The AOI (20 mm × 5 mm) used for analysis is indicated by the yellow box. (**B**) Histomorphometric analysis of area percentage of new bone (red) at 12 weeks within the defect. Statistical significance determined using one-way ANOVA, \*\* *p* < 0.01.

Representative histological sections show the ingrowth of new bone at the graft interface indicating osseointegration in all groups (Figure 5A). While some specimens in the AG group showed increased adipogenesis in the marrow cavity compared with the extender groups (Figure 5A), images of histological sections from all AG specimens show variable adipogenesis (Supplementary Figure S1). Osteoblasts were observed around the perimeter of bone ingrowth, suggesting active ongoing remodeling (Figure 5B).

**Figure 5.** Osseointegration of AG extenders. (**A**) Histological images demonstrate osseointegration of the AG extenders at the host bone/material interface. H represents new bone, S represents scaffold, and \* represents new bone growth. (Scale bar, 1 mm) (**B**) New bone growth occurring within the graft. Yellow arrows point to osteoblasts. (Scale bar, 100 μm).

Additionally, Safranin O staining revealed faint orange staining (cartilage) indicating that previous endochondral ossification had occurred within the AG control group (Figure 6A), while ongoing endochondral ossification was observed at 12 weeks in both AG extenders (Figure 6B,C).

**Figure 6.** Endochondral ossification within AG extenders. Histological images demonstrate endochondral ossification with (**A**) AG control, (**B**) PTKUR-AG extenders, and (**C**) PEUR-AG extenders at 12 weeks. (Scale bar, 100 μm).

#### *3.3. PTKUR and PEUR Graft Remodeling*

Histological analysis revealed residual polymer (black) in AG extenders (Figure 7A and Figure S1). High-magnification images demonstrated that PTKUR-AG underwent slower resorption as evidenced by the higher amount of dense residual polymer in the PTKUR-AG sections compared with the extensive resorption evident in the PEUR-AG sections (Figure 7A). These findings were confirmed via histomorphometric analysis (AOI represented in Figure 4A) in which PTKUR-AG exhibited significantly more residual polymer compared with the PEUR-AG group (Figure 7B). All groups supported cellular and tissue infiltration (teal/blue), but significantly greater cellular and tissue infiltration was observed in the PEUR-AG group compared with PTKUR-AG and AG control (Figure 7B).

**Figure 7.** AG extender remodeling. (**A**) Representative images of residual polymer in AG control, PTKUR-AG, and PEUR-AG extenders. P denotes residual polymer and \* denotes implanted AG. (Scale bar, 1 mm) (**B**) Histomorphometric analysis of area percentage of infiltrating cells and tissue and residual polymer within the defect after 12 weeks post implantation. Statistical significance determined using Two-way ANOVA, \* *p* < 0.05, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001.

Bone remodeling throughout the healing process was observed in all groups, especially at the proximal and distal host bone/graft interfaces (Figure S1). Remodeling was observed within the PTKUR- and PEUR-AG grafts around the periphery of implanted AG at 4 and 8 weeks, indicating mineralization nucleating from implanted AG particles within the extenders (Figure 8A). PTKUR- and PEUR-AG extenders exhibited increased bone remodeling at 4 weeks (green) compared with 8 weeks (orange/red); however, these differences were not significant (Figure 8B). Additionally, increased bone remodeling was observed at the graft/host bone interface, indicating the osseointegration of both AG extenders (Figure 8C). The AG control demonstrated significantly greater bone remodeling compared with the AG extenders at both 4 and 8 weeks (Figure 8B); however, inconsistent healing was observed as only two of the four controls exhibited complete bridging along the lateral edge of the defect (Figure S1).

**Figure 8.** Dynamic histomorphometric analysis at 4 and 8 weeks. (**A**) Representative fluorescent images of AG, PTKUR-AG, and PEUR-AG groups. The AOI is indicated by the yellow box. (**B**) Histomorphometric analysis of area percentage of active bone remodeling at 4 (green) and 8 (orange/red) weeks within the defect. (**C**) Representative images of bone remodeling at the host bone–graft interface, demonstrating osseointegration in PTKUR-AG extenders and PEUR-AG extenders. HB indicates host bone and G indicates grafts. Statistical significance determined using two-way ANOVA, \* *p* < 0.05, \*\*\* *p* < 0.005 \*\*\*\* *p* < 0.001.

#### **4. Discussion**

In this work, we implanted PTKUR-AG and PEUR-AG extenders in a 20 mm critical sized rabbit radial defect to evaluate the effects of polymer composition on both bone formation and graft remodeling in vivo. Both PTKUR- and PEUR-AG extenders supported new bone formation and utilized less AG than the AG control. Furthermore, the polymeric component of the AG extenders degraded and simultaneously maintained AG within the defect for 12 weeks. PEUR-AG extenders degraded more rapidly compared with PTKUR-AG extenders. However, new bone formation in both AG extenders was delayed compared with the AG control.

To understand the effect of polymer composition on bone formation and graft remodeling, AG extenders were implanted in a 20 mm critical-size radial defect in rabbits [44,45]. This model was selected as no external fixation was required [2]. No graft failure was observed in any of the groups throughout the 12 weeks, suggesting that AG extenders exhibited sufficient compression-resistant properties. Previous studies in the spine and mandible demonstrated that an elastic modulus >1 MPa provided compression-resistant

properties [46,47]. We previously reported PTKUR-AG and PEUR-allograft moduli of 6.08 MPa and 4.38–9.47 MPa, respectively [39,48].

Previous studies performed in the rabbit radius have reported bone growth from the proximal end of the defect, the distal end of the defect, and the interosseus membrane [44,48–51]. Similarly, we observed bone formation in these directions. Due to the inability of μCT to distinguish between new bone, residual AG, host bone, and ossification extending from the ulna within the interosseous membrane to the radius, the ulna was included in μCT analysis. Interestingly, BV/TV trended higher in PEUR-AG compared with PTKUR-AG. These differences were not significant, but they were likely due in part to increased degradation of the PEUR, allowing for increased bone formation throughout the defect and within the interosseous membrane. Furthermore, μCT bone area quantification indicated increased bone at the proximal and distal end of the defect, indicating bone formation at the interfaces. Consistent with previous studies utilizing AG [52–54], new bone formation via creeping substitution at the host bone/graft interface was observed.

Histomorphometric analysis was performed to evaluate new bone formation specifically in the 5 mm × 20 mm defect space; thus, bone present in the interosseus membrane was excluded from analysis. Transverse sections were obtained from the center of the defect to evaluate bone formation at its most stringent point. Ultimately, no significant difference in bone between PTKUR- and PEUR-AG extender was observed via histomorphometry. The AG control demonstrated significantly increased new bone within the defect compared with AG extenders at 12 weeks via histomorphometric analysis, but new bone formation appeared to be variable throughout the defect. These differences were not observed in overall BV/TV between groups, suggesting that PTKUR- and PEUR-AG promoted bone formation, particularly in the interosseus membrane surrounding the defect while AG control promoted greater bone formation within the defect site itself.

In agreement with an earlier PTKUR-AG study in a biologically stringent intertransverse process defect [39], residual polymer was observed in PTKUR-AG extenders at 12 weeks. PEUR degradation occurred more rapidly than PTKUR degradation, as evidenced by significantly less residual polymer within the defect at 12 weeks. PTKUR degrades in response to specific cell types including osteoclasts, macrophages, and other ROS-secreting cells [38], while PEUR degrades via autocatalytic hydrolytic degradation [55]. Furthermore, AG control exhibited the least amount of cellular infiltration at 12 weeks, suggesting that cells recruited to the AG were osteoprogenitor cells that underwent direct differentiation. PTKUR-AG extenders exhibited less cellular infiltration than PEUR-AG extenders, demonstrating that cells were able to more readily infiltrate the graft as the polymer degraded. Early vascularization of cancellous AG begins at 2 days and is followed by the recruitment of MSCs in response to the osteoinductive signals of AG within the first weeks after implantation [56]. In contrast, AG particles were encapsulated in residual polymer in the AG extenders (Figure 5A), which delayed the rate of new bone formation. These findings are further confirmed by Safranin O staining in which faint positive Safranin O staining for cartilage was demonstrated within the AG control group, suggesting that new bone formation occurred via endochondral ossification and was near completion by 12 weeks. However, more intense positive Safranin O staining was observed in the AG extenders, suggesting that ongoing endochondral ossification was still occurring at 12 weeks. Thus, AG particles encapsulated in residual polymer retained their osteoinductivity beyond the first few weeks after implantation. Although the AG extenders showed delayed endochondral ossification suggesting a longer total healing time, the slower healing coupled with the observation that AG remains stable throughout the study suggests that PEUR and PTKUR extenders can reduce the risk for rapid resorption.

Dynamic bone histomorphometry is a widely utilized method for evaluating bone remodeling [57,58]. As mentioned above, ossification within the interosseus membrane was excluded from dynamic histomorphometric analysis. In agreement with our static histomorphometric findings, bone remodeling at 4 and 8 weeks was greater in AG controls compared with PTKUR- and PEUR-AG extenders within the defect. It is likely that the lack of polymer in AG controls allowed for more extensive cellular infiltration and new bone formation. Despite a smaller 10 mm defect size, a previous study utilizing highly porous cell-seeded hydroxyapatite scaffolds did not observe fluorochrome binding within the scaffold until six weeks post implantation [51]. Herein, fluorescent staining beginning at 4 weeks was apparent within the grafts in the AG extender groups, suggesting that embedded AG maintained bioactivity. Additionally, abundant osseointegration at the host bone/graft interface in both AG extenders was observed, further confirming bioactivity. AG is resorbed by osteoclasts and new bone is deposited by osteoblasts in a process known as creeping substitution [56].

The complete degradation of synthetic polymers requires from 4 to 24 months in vivo [59], which can delay the creeping substitution of AG particles encapsulated in polymer. However, while the encapsulation of AG particles within residual polymer delays new bone formation, it also protects AG from rapid resorption that can result in unpredictable healing [10]. While osteobiologics such as recombinant human bone morphogenetic protein-2 (rhBMP-2) have been shown to promote predictable bone healing at 12 weeks in the rabbit radius defect model [60], rhBMP-2 is approved by the FDA for only a limited number of indications, including lumbar fusion, ridge augmentation, and fresh fractures of the tibial shaft. Furthermore, rhBMP-2 delivered on a collagen sponge has weak mechanical properties, similar to AG. Thus, PEUR- and PTKUR-AG extenders may be most beneficial in clinical scenarios where long-term mechanical stability is required, such as posterolateral spine fusion and fractures of the mid-diaphysis. This study is limited by a single intermediate time point for endpoint outcomes (12 weeks), at which time bone healing and complete resorption of residual polymer and AG were not observed. Future studies should focus on optimizing the rate of polymer degradation to increase the rate of new bone formation while protecting the AG from excessive resorption.

#### **5. Conclusions**

In this work, PTKUR- and PEUR-AG extenders were compared with an AG control in a rabbit radius model of bone regeneration. PTKUR- and PEUR-AG extenders both maintained AG in the defect throughout the study and demonstrated bone formation along the host bone/graft interface comparable to AG control. Polymer resorption and subsequent cellular infiltration were observed within the defect space in both AG extenders but did not have an effect on overall bone formation. These results suggest that early polymer degradation and cellular infiltration are necessary for harnessing and maximizing the osteoinductive capabilities of AG.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/ma14143960/s1, Figure S1: Fluorescent and Sanderson's Rapid stained histological sections. Figure S2: Bone morphometric parameters.

**Author Contributions:** Conceptualization, L.A.B., M.A.P.M., C.L.D., J.C.W. and S.A.G.; methodology, L.A.B., M.A.P.M., S.A.G., S.M.S. and J.C.W.; formal analysis, L.A.B. and M.A.P.M.; investigation, L.A.B., M.A.P.M., S.M.S.; resources, L.A.B., M.A.P.M., S.M.S., J.C.W. and S.A.G.; data curation, L.A.B. and M.A.P.M.; writing—original draft preparation, L.A.B. and M.A.P.M.; writing—review and editing, L.A.B., S.M.S., C.L.D., J.C.W. and S.A.G.; visualization, L.A.B. and M.A.P.M.; supervision, J.C.W. and S.A.G.; project administration, S.M.S., J.C.W. and S.A.G.; funding acquisition, J.C.W. and S.A.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the National Institutes of Health (NIH) (R01AR064772 and T32DK101003) and the United States Army Institute of Surgical Research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors' and do not necessarily reflect the views of the National Institutes of Health or the Department of Defense.

**Institutional Review Board Statement:** All surgical and care procedures were approved by the Institutional Animal Care and Use Committee of the US Army Institute of Surgical Research, Fort Sam Houston, TX. Procedures were performed in compliance with the Animal Welfare Act, Animal Welfare Regulations, and the Guide for the Care and Use of Laboratory Animals.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


## *Article* **Promotion of Bone Regeneration Using Bioinspired PLGA/MH/ECM Scaffold Combined with Bioactive PDRN**

**Da-Seul Kim 1,2,†, Jun-Kyu Lee 1,†, Ji-Won Jung 1, Seung-Woon Baek 1,3,4, Jun Hyuk Kim 1, Yun Heo 1, Tae-Hyung Kim <sup>2</sup> and Dong Keun Han 1,\***


**Abstract:** Current approaches of biomaterials for the repair of critical-sized bone defects still require immense effort to overcome numerous obstacles. The biodegradable polymer-based scaffolds have been required to expand further function for bone tissue engineering. Poly(lactic-co-glycolic) acid (PLGA) is one of the most common biopolymers owing to its biodegradability for tissue regenerations. However, there are major clinical challenges that the byproducts of the PLGA cause an acidic environment of implanting site. The critical processes in bone repair are osteogenesis, angiogenesis, and inhibition of excessive osteoclastogenesis. In this study, the porous PLGA (P) scaffold was combined with magnesium hydroxide (MH, M) and bone-extracellular matrix (bECM, E) to improve anti-inflammatory ability and osteoconductivity. Additionally, the bioactive polydeoxyribonucleotide (PDRN, P) was additionally incorporated in the existing PME scaffold. The prepared PMEP scaffold has pro-osteogenic and pro-angiogenic effects and inhibition of osteoclast due to the PDRN, which interacts with the adenosine A2A receptor agonist that up-regulates expression of vascular endothelial growth factor (VEGF) and down-regulates inflammatory cytokines. The PMEP scaffold has superior biological properties for human bone-marrow mesenchymal stem cells (hBMSCs) adhesion, proliferation, and osteogenic differentiation in vitro. Moreover, the gene expressions related to osteogenesis and angiogenesis of hBMSCs increased and the inflammatory factors decreased on the PMEP scaffold. In conclusion, it provides a promising strategy and clinical potential candidate for bone tissue regeneration and repairing bone defects.

**Keywords:** bone regeneration; poly(lactide-co-glycolide); magnesium hydroxide; extracellular matrix; polydeoxyribonucleotide; porous scaffold

#### **1. Introduction**

Bone fracture is the most common injury, which has high healing efficiency by oneself, but critical-sized bone fraction indispensably requires orthopedic surgery. To enhance the bone repair rate, various methods has been used, including autograft, allograft, xenograft, and artificial bone graft materials (e.g., tricalcium phosphate, hydroxyapatite, and bioglass). Among these treatments, autograft, obtained from the patient's other position, is regarded as a 'gold standard' because of its high regeneration rate and superior osteoconductivity and osteoinductivity without any immune response [1–3]. Although autograft has numerous advantages including no risk of disease transfer, there are some limitations, such as

**Citation:** Kim, D.-S.; Lee, J.-K.; Jung, J.-W.; Baek, S.-W.; Kim, J.H.; Heo, Y.; Kim, T.-H.; Han, D.K. Promotion of Bone Regeneration Using Bioinspired PLGA/MH/ECM Scaffold Combined with Bioactive PDRN. *Materials* **2021**, *14*, 4149. https://doi.org/10.3390/ ma14154149

Academic Editor: Pedro de Sousa Gomes

Received: 28 May 2021 Accepted: 22 July 2021 Published: 26 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

restricted bone supply, donor site morbidity, and poor capability to accommodate defects. To conquer these hurdles, scaffold implantation is considered as an ideal way for bone tissue regeneration. Current bone tissue engineering (BTE) approaches still have numerous limitations such as low biocompatibility, mechanical property, osteoinductivity, and osteoconductivity. The polymer-based scaffold has been studied because of its biodegradability and biocompatibility. Among them, Poly(lactic-co-glycolic) acid (PLGA) was approved by the Food and Drug Administration (FDA) for diverse types of bone implants. However, it has been reported that its degradation byproducts, lactic acid and glycolic acid cause an acidic microenvironment at implanting site [4,5]. In previous studies, magnesium hydroxide (MH) performed outstanding pH neutralization ability for diverse tissue regeneration [6–11], in particular, bone repair [12–14]. However, MH as a hydrophilic inorganic molecule is difficult to disperse evenly in the hydrophobic polymer-based scaffold. To disperse metal ion molecule, in the prior study, the MH modified with a ricinoleic acid (mMH) was attempted to PLGA porous scaffold [9]. Plus, the incorporation of mMH into PLGA implant would be used to attenuate acid-induced inflammation triggered by the degradation products from the polymer and to improve the hydrophilicity of the scaffold.

As noted above, to overcome the limitation of the BTE scaffold, the extracellular matrix (ECM) isolated from mammalian tissues has been attempted in the scaffold for enhancing biocompatibility [15–17] and mimicking the natural composition of bone tissue. Especially, bovine-derived decellularized bone extracellular matrix (bECM), comprising mostly of calcium and phosphate, can improve not only biocompatibility also osteoconductivity of the scaffold.

Aside from these improvements, because bone repairing takes a longer time than other tissue in general, the ideal BTE scaffold should have an osteoinductive property. In this respect, in order to enhance osteoinductivity and bioactive function of the scaffold, polydeoxyribonucleotide (PDRN) was applied in bone regeneration. PDRN is a natural bioactive molecule normally extracted from salmon trout (*Oncorhynchus mykiss*) gonads, which is a short DNA form (50 to 2000 base pairs). Recently, some studies reported that PDRN has great effects on improving tissue regeneration since it plays as an adenosine A2A receptor agonist. Adenosine A2A receptor is a member of the G protein-coupled receptor (GPCR) family that has been proven as effective in improving angiogenesis and reducing inflammation [18–20]. Additionally, PDRN provides building blocks, nucleotides, and nucleosides to produce nucleic acids using less energy via the salvage pathway [21].

In this study, we designed a bioinspired scaffold by integrating mMH (M), bECM (E), and PDRN (P) into a porous PLGA (P) scaffold. We hypothesized that mMH could suppress the detrimental effect caused by PLGA degradation, reduce osteoclastogenesis; bECM could mimic the natural bone tissue microenvironment, improve osteoconductivity; PDRN could promote angiogenesis during bone repair. Therefore, the functionalized biodegradable PMEP scaffold would be applicable for effective bone regeneration with synergistic effects from these bioactive molecules.

#### **2. Materials and Methods**

#### *2.1. Materials*

Poly(D,L-lactide-co-glycolide) (PLGA, lactide:glycolide = 75:25, I.V. = 0.8–1.2) was purchased from Evonik Ind. (Essen, Germany). Magnesium hydroxide (MH), L-ascorbic acid, dexamethasone, and β-glycerophosphate were purchased by Sigma Aldrich (St. Louis, MO, USA). Ricinoleic acid was purchased from TCI product (Tokyo, Japan). The bovine bone-derived extracellular matrix powder (bECM; InduCera) was supplied by Oscotec Inc. (Seongnam, Korea). Polydeoxyribonucleotide (PDRN) was obtained from Goldbio (St. Louis, MO, USA). D-Plus™ cell counting kit 8 (CCK-8) cell viability assay kit was obtained from Dongin LS (Seoul, Korea).

#### *2.2. Scaffold Preperation*

The modified Mg(OH)2 was synthesized with ricinoleic acid (mMH) following the process with the previous study [9]. All scaffolds (PLGA, PME, and PMEP) were prepared by the freeze-drying method. In brief, the ice particles (200–300 μm) were prepared by spraying deionized water into liquid nitrogen as a porogen for the porous scaffold. A 20 wt% mMH and PDRN, and 50 wt% bECM (compared to PLGA) were mixed with 0.5 g of PLGA in 0.3 M dichloromethane solution. The mixtures and ice particles were stuffed into round PTFE mold (ø5 × 2 mm2). The filled molds were freeze-dried for 2 days to remove the ice and remaining organic solvent, then the porous scaffolds were obtained.

#### *2.3. Scaffold Characterization*

The cross-section morphology of the scaffolds was observed using scanning electron microscopy (SEM; GENESIS-1000, Emcraft, Gwangju, Korea). The thermal property of the scaffolds was analyzed by a thermal gravimetric analyzer (TGA 4000, PerkinElmer, Waltham, MA, USA). To assess the neutralization capacity of the scaffolds, the mass and pH changes were measured in 500 μL phosphate-buffered saline (PBS) solution (pH 7.4) with 20 μg/mL protease K (Bioneer, Daejeon, Korea) for 14 days. The inorganic compositions of the scaffold were measured using inductively coupled plasma-optical emission spectroscopy (ICP-OES, Optima 8000, PerkinElmer, Waltham, MA, USA). The water contact angle (WCA) was analyzed using the sessile drop method at room temperature to evaluate the hydrophilicity and hydrophobicity of the scaffolds.

#### *2.4. Cell and Cytotoxicity Assay*

Human bone-marrow mesenchymal stem cells (hBMSCs) were cultured in DMEM/low glucose media supplemented with 10% FBS (Hyclone, Logan, UT, USA) and 1% antibiotic– antimycotic solution (Gibco, Thermo Scientific Inc., Waltham, MS, USA). The cells were maintained under a humidified atmosphere with 5% CO2 at 37 ◦C. The viability and proliferation of the cells were determined using a Live-dead viability/cytotoxicity kit (Invitrogen, Thermo Scientific Inc., Waltham, MS, USA) and the fluorescence images were obtained using LSM880 (Zeiss, Jena, Germany) at 1, 3, and 7 days. The CCK-8 assay was conducted on the 3D scaffold at the same days.

#### *2.5. Wound Healing Assay and Tubule Formation*

The scratch wound healing assay was conducted to assess the migratory capacity of hBMSCs by PDRN. The cells were seeded into a 6-well culture plate at the density of <sup>3</sup> × <sup>10</sup><sup>5</sup> cells/well and cultured for 1 day. The confluent wells were scratched, then washed with PBS solution. Cells were cultured with DMEM/low glucose containing 1% (*v*/*v*) FBS and added 100 μg/mL of PDRN. After 12 and 24 h, the plates were photographed and quantified the healed area using Image J software. To assess the angiogenic effects of the PDRN, 250 μL of matrigel matrix (Corning, Brooklyn, NY, USA) was added to pre-cooled 24-well plate and then incubated at 37 ◦C for 1 h. The human umbilical vein endothelial cells (HUVECs) were seeded onto coated well at the density of 1.2 × 105 cells/well with EBM-2 (Lonza, Basel, Switzerland) containing 1% FBS, then added 100 μg/mL of PDRN. After 18 h, the cells were stained with calcein AM (C1430, Thermo Scientific Inc., Waltham, MS, USA), then photographed with a fluorescence microscope (U-RFL-T, Olympus, Tokyo, Japan). The tube length and branch point were quantified using Image J software.

#### *2.6. RNA Extraction and Quantitative Real-Time PCR (qRT-PCR)*

The RNA from scaffolds was extracted using Trizol reagent (15596018, Ambion, Invitrogen, Thermo Scientific Inc., Waltham, MS, USA) following the manufacturer's instructions. The RNA concentration and quality were measured by spectrophotometer (ND-1000; Thermo Scientific, Waltham, MA, USA). The cDNA was synthesized using PrimeScript RT Reagent Kit (Perfect Real Time, Takara, Tokyo, Japan). The qRT-PCR was performed using each primer and SYBR Green PCR Master Mix (Applied Biosystems, Thermo Scientific

Inc., Waltham, MS, USA). The expression of osteogenic, angiogenic, and inflammationrelated genes was calculated with the 18S rRNA as a reference using the 2−ΔΔCt method. The primers used were as follows: 18S rRNA: forward, 5 -gcaattattccccatgaacg-3 and reverse, 5 -gggacttaatcaacgcaagc-3 ; IL6: forward, 5 -gatgagtacaaaagtcctgatcca-3 and reverse, 5 -ctgcagccactggttctgt-3 ; IL-1β: forward, 5 -tacctgtcctgcgtgttgaa-3 and reverse, 5 -tctttgggtaatttttgggatct-3 ; VEGF: forward, 5 -actggaccctggctttactg-3 and reverse, 5 tctgctccccttctgtcgt-3 ; MMP2: forward, 5 -caccaccgaggattatgacc-3 and reverse, 5 - cacccacagt ggacatagca-3 ; ALP: forward, 5 -atgaaggaaaagccaagcag-3 and reverse, 5 -ccaccaaatgt gaagacgtg-3 ; RUNX2: forward, ggtcagatgcaggcggccc-3 and reverse, 5 -tacgtgtggtagcgcgtgg c-3 ; OCN: forward, 5 -cagcgaggtagtgaagagacc-3 and reverse, 5 -tctggagtttatttgggagcag-3 .

#### *2.7. Osteogenic Differentiation In Vitro*

To assess the capacity of osteogenic differentiation on the 3D scaffold, hBMSCs were seeded onto the scaffold at the density of 5 × 105 cells/scaffold. After 1 day, the medium was replaced with an osteogenic differentiation medium, DMEM/low glucose, containing 50 μM L-ascorbic acid, 0.1 μM Dexamethasone, and 10 mM β-glycerophosphate. After 7 days of osteogenic differentiation, the scaffolds were fixed with 10% formalin for 20 min, rinsed with deionized water, and stained with an alkaline phosphatase (ALP) staining kit (MK300, Takara, Japan). The stained samples were incubated in 15 and 30% sucrose solution in order, and embedded with frozen section media (FSC 22, Leica Biosystems, Wetzlar, Germany). The frozen samples were sectioned with a cryostat microtome (CM3050S, Leica Biosystems, Wetzlar, Germany). For the quantification of ALP activity, the scaffolds were lysed using an ALP assay kit (MK301, Takara, Tokyo, Japan). The assay was conducted according to the produced protocol.

#### *2.8. Tartrate-Resistant Acid Phosphatase Staining and Activity*

The osteoclastogenesis was identified by Tartrate-resistant acid phosphatase (TRAP) staining (MK-300, Takara, Tokyo, Japan). The RAW264.7 cells, mouse macrophage cell line, were seeded into a 24-well culture plate at the density of 2 × 104 cells/well. After 1 day, 100 ng/mL of receptor activator of the NF-κB ligand (RANKL) was treated with RAW264.7 cells to induce differentiation into osteoclast. After 3 days, the scaffolds were put into trans-well inserts for co-culture with osteoclast. The TRAP staining was executed after 3 days.

#### *2.9. Statistical Analysis*

All experimental results were obtained through more than three independent experiments, and the values were described as mean ± standard deviation (SD). The statistical significance was analyzed by one-way ANOVA using Tukey's post hoc method in Graph-Pad Prism 7.0 software [12] (GraphPad Software, Inc., San Diego, CA, USA). The statistically significant difference was defined as the *p* value being less than 0.05. The differences were considered significant when \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, and # *p* < 0.0001.

#### **3. Results and Discussion**

#### *3.1. Scaffold Characterization*

The biodegradable porous scaffolds containing PLGA, mMH, dECM, and PDRN were fabricated using the etching method with ice particles. In Figure 1A, the SEM images represent cross-section morphology that the pores of the scaffold were well-distributed and interconnected, so that the cells could easily attach and migrate in the scaffold during bone regeneration. Moreover, the 200–300 μm of porogens were used, that it is known as appropriate size for osteogenic differentiation in many other studies [22,23]. This size of porogen could be beneficial to cell ingrowth into the pore structures. The proportion of inorganic molecules in the scaffolds was analyzed by TGA (Figure 1B) and induced coupled plasma-optical emission spectroscopy (ICP-OES, Table 1). The PMEP scaffold consists of 195.44 ppm of magnesium, 265.74 ppm of calcium, and 151.20 ppm of phosphorus, respec-

tively. The PME scaffold consists of 201.46 ppm of magnesium, 270.44 ppm of calcium, and 136.43 ppm of phosphorus, respectively. Interestingly, the amount of phosphorous in the PMEP was slightly higher than the PME because of the phosphate backbone in PDRN. The porous scaffolds with dissimilar surface roughness can cause different wettability and thus affect the permeability. The WCA was conducted to evaluate the wettability of the scaffold. The angles on PLGA, PME, and PMEP scaffold were 104.59, 93.99, and 77.12◦, respectively. As bioactive molecules were added, the contact angles decreased. In other words, the PMEP scaffold has more hydrophilic property than the PLGA and PME ones.

**Figure 1.** Scaffold characterization. (**A**) representative scanning electron microscopy (SEM) images of PLGA, PLGA/MH/ECM (PME), and PLGA/MH/ECM/PDRN (PMEP) scaffold. (Scale bars = 200 μm). (**B**) thermal gravimetric analysis (TGA) thermograms of each scaffold. Change of (**C**) pH and (**D**) mass during in vitro degradation in PBS solution with protease K at 37 ◦C for 14 days.



The degradation of porous PLGA, PME, and PMEP scaffolds was observed in the presence of 20 μg/mL protease K for 14 days at 37 ◦C. The accelerative condition was conducted using protease K due to relatively high molecular weight of PLGA. In Figure 1C, pH value of the PLGA in PBS solution drastically decreased to 4.3 after 14 days of degradation. However, pH value of the solution containing the PME and PMEP specimens initially reached 8.3 and 8.5 and dropped slowly for 14 days to reach 5.5 and 5.4, respectively due to neutralization ability of mMH. The PME and PMEP scaffolds showed fast degradation performance than the PLGA only scaffold since those were containing numerous soluble bioactive molecules (Figure 1D).

#### *3.2. Biocompatibility of the Scaffold*

To confirm cytotoxicity of the scaffolds in vitro, in Figure 2A, calcein AM and ethidium homodimer 1 (EthD-1) stainings were conducted with hBMSCs at 1, 3, and 7 days. Because of its well-known biocompatibility of PLGA, the EthD-1 positive cells indicating dead cells were observed rarely in all the scaffolds even the PLGA only group. What is more, the

cells were observed evenly along with the pores of the scaffold. However, the population of calcein AM positive cells, the live cells, was getting increased in the PME and the PMEP than the PLGA at 1, 3, and 7 days, respectively. In Figure 2B, the cell viability was quantified using CCK-8 in 1, 3, and 7 days. The initial adhesion rate of hBMSCs on the PMEP significantly increased due to its surface hydrophilicity for cell recruitment (*p* < 0.01). Because a hydrophilic surface promotes the adhesion of the cells [24], the initial cell adhesion rate significantly increased on the PMEP scaffold.

**Figure 2.** Biocompatibility of the scaffolds. (**A**) live-dead assay images on each scaffold at 1, 3, and 7 days (scale bar = 100 μm). (**B**) cell viability of the hBMSCs onto each scaffold at 1, 3, and 7 days in vitro. The differences were considered significant when ns = not significant (*p* ≥ 0.05), \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001 (*n* ≥ 3).

As mentioned previously, the PLGA produced acidic byproducts during hydrolytic degradation, so that the cell slightly proliferated for 7 days. However, the cell viability on the PME and particularly, the PMEP scaffold was remarkably enhanced for 7 days, *p* < 0.05 and *p* < 0.001, respectively. Consequently, the incorporation of mMH, bECM, and PDRN could constrict the adverse effect on cell cytotoxicity caused by hydrolytic degradation of PLGA.

#### *3.3. Confirmation of Angigenic Ability*

Angiogenesis is a physiological process by which new blood vessels form from the pre-existing vascular network, allowing the delivery of oxygen and nutrients to the body's tissues. Angiogenesis has been studied as a therapeutic target in regenerative medicine. Bone is also richly vascularized tissue, so that new blood vessels play a critical role in maintaining the bone cells survival and stimulating their activity. However, in situ vascularized bone regeneration still remains in the extreme challenge [25–28].

As mentioned previously, PDRN has a pro-drug activity carried out through two different mechanisms. First, PDRN supplies purines and pyrimidines, promoting DNA synthesis or repair through the 'salvage pathway' [18,21]. Next, PDRN stimulates adenosine A2A receptor, as suggested by Thellung et al. studied the effect of PDRN using 3,7-dimethyl-1 propargylxanthine (DMPX), a selective adenosine A2A receptor antagonist [29]. Adenosine and adenosine A2A receptor were considered clinically important to enhance angiogenesis. Wang et al. studied that adenosine enhances cell growth and induces tube formation in HUVECs in vitro [30]. In Figure 3, the biological ability of PDRN was investigated in angiogenesis and wound healing for effective bone tissue repair. When PDRN treated, HUVECs had formed a significant number of branch points and longer lengths of tubes. On the same side of Figure 2, because PDRN could enhance the growth and migratory ability of hBMSCs, the wound closure rates also highly increased to 34.8 and 31.9% in PDRN treated groups compared to control at 24 and 48 h, respectively. To conclude, these

outstanding biological abilities of PDRN give a synergistic effect to achieving a novel strategy for bone regeneration.

**Figure 3.** Biological effects of PDRN. (**A**) tubule-forming assay; images of HUVECs stained with calcein AM (scale bar = 200 μm) (**a**) and quantification of branch point (**b**) and tube length (**c**). (**B**) wound healing assay; optical images (scale bar = 200 μm) (**a**) and quantification of closed area at 24 and 48 h (**b**). The differences were considered significant when # *p* < 0.0001 (*n* ≥ 3).

#### *3.4. Biological Abilities of the PMEP Scaffold with hBMSCs: Anti-Inflammation and Angiogenesis*

Since mineralization is affected by numerous mechanisms, biomaterials should have a variety of functions, such as vascularization, inhibition of inflammation, as well as osteogenesis to reach effective bone regeneration. The quantitative real-time PCR (qRT-PCR) was conducted to determine the expression of inflammation and angiogenesisrelated genes on 3D scaffolds with hBMSCs. The effect of the scaffolds was assessed in osteogenic media at 7 and 21 days. As shown in Figure 4A, the PME scaffold restricted the expression of inflammatory genes, interleukin-6 (IL-6) and interleukin-1β (IL-1β), compared to the PLGA scaffold. Plus, the PMEP scaffold effectively suppressed the abovementioned gene expression even in comparison with the PME one. In recent studies, the researchers demonstrated that the PDRN affects to increase expression of vascular endothelial growth factor (VEGF) and to suppress the production of pro-inflammatory cytokines by stimulating the A2A receptor [18,21,31]. As a result, the PDRN could promote angiogenesis and inhibit inflammation during bone repair. In Figure 4B, the PME scaffold exhibited a negligible difference in the expression of angiogenesis-related genes, including VEGF and matrix metalloproteinase-2 (MMP2). Likewise, the PMEP scaffold promoted the highest gene expression of VEGF and MMP2 on both days. It is notable that the addition of PDRN on the scaffold has effects on not only reducing the inflammatory response but also significantly enhancing vascularization. In prior analysis (Figure 3), we confirmed the effectiveness of PDRN, treated directly in cells in the 2D environment. Further, incorporation of mMH, bECM, and PDRN in the biodegradable porous 3D scaffold displayed attenuating inflammatory response and enhancing angiogenesis, simultaneously. These results suggest that the PDRN effect is not only for angiogenesis but also may influence several factors containing the healing process.

**Figure 4.** Anti-inflammatory and angiogenic effects on the scaffolds using hBMSCs. Gene expressions of hBMSCs onto the scaffolds related to (**A**) anti-inflammation: IL-6 and IL-1β, and (**B**) angiogenesis: VEGF and MMP2 at 7 and 21 days. The differences were considered significant when ns = not significant (*p* ≥ 0.05), \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, and # *p* < 0.0001 (*n* ≥ 3).

#### *3.5. Induction of Osteogenesis in 3D Scaffold*

To identify the osteogenic capacity of the scaffolds, hBMSCs were seeded onto the scaffold. ALP is known as an early marker of osteogenesis. After 7 days of osteogenic differentiation, the ALP staining was conducted on each scaffold. Figure 5A showed that the PMEP scaffold formed more degrees of staining with less collapsing of internal structure compared to other scaffolds. Moreover, the PMEP scaffold enhanced ALP activity, which was even significantly higher than the PME one. These results implied that the PME scaffold could induce osteogenic differentiation of hBMSCs effectively, and by adding PDRN, the osteogenesis was more enhanced. Further investigation of cell differentiation was verified through gene expression analysis of the osteogenic markers. The expression of osteogenesis-related genes including ALP, runt-related transcription factor 2 (RUNX2), and osteocalcin (OCN) was also evaluated at 7 (Figure 5B) and 21 days (Figure 5C). In general, the RUNX2 and OCN are, respectively, used as mid- and late-responsive genes for bone formation. The results exhibited that the PMEP scaffold significantly up-regulated ALP, RUNX2, and OCN at all days. The mRNA expression of genes in the PMEP scaffold respectively increased by 2.48-, 2.05-, and 3.07-fold higher than the PLGA group at 7 days. The expressions on 21 days were also up-regulated by 2.08-, 1.75-, and 1.94-fold higher in the PMEP scaffold, respectively. These results indicated that the MH provides biocompatibility, bECM has osteoconductivity, and the PDRN promotes angiogenesis and osteogenesis by stimulating the A2A receptor. In conclusion, the PMEP scaffold has the potentials that not only effectively induce early-stage of osteogenesis, but also affect the maturation of hBMSCs for bone regeneration.

**Figure 5.** Osteogenic differentiation onto scaffolds using hBMSCs. (**A**) optical images of the scaffolds stained with ALP for 7 days in osteogenic medium (**a**). Scale bars indicate 400 μm. What is more, the quantification of ALP activity onto each scaffold (**b**). (**B**,**C**) gene expressions of hBMSCs related to osteogenesis onto the scaffolds; ALP, RUNX2, and OCN at 7 and 21 days of osteogenic differentiation. The differences were considered significant when ns = not significant (*p* ≥ 0.05), \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, and # *p* < 0.0001 (*n* ≥ 3).

#### *3.6. Attenuation of Osteoclastogenesis*

Recently, to develop the bone repairing materials, one of the most important issues is bone homeostasis between osteoclasts and osteoblasts, because excessive differentiation of osteoclasts affects bone tissue resorption, which would occur metabolic bone-related diseases such as osteoporosis [32–34]. In general, when bone fracture occurs, both osteoblasts and osteoclasts are activated. Immoderate osteoclastogenesis and osteoblastogenesis cause eventually the delay of bone formation or nonunion. Thus, the BTE scaffold should control initial immoderate osteoclastogenesis, which is critical for promoting osteoblast activity and enhancing bone mineral density [35–37]. To evaluate the control ability of the 3D scaffold, we designed an indirect co-culture system (Figure 6) using a trans-well insert. In Figure 6A, optical images represent TRAP positive cells after 3 days of RANKL treatment. The stained cells (purple) significantly decreased in the PMEP group compared to the control, the PLGA and the PME. To quantify the osteoclast activity, TRAP activity was analyzed with the 3D scaffolds. In the PLGA group, the activity slightly increased in comparison to RANKL treated control group. However, the secreted bioactive molecules from the PME and PMEP scaffold attenuated RANKL-induced differentiation into osteoclast of RAW264.7 cells for 31.7 and 74.4%, respectively, than control. Overall, the PMEP scaffold has multifunctional abilities in inhibition of local inflammation, promotion of angiogenesis, and attenuation of osteoclastogenesis. The bioinspired PMEP scaffold would be clinically utilized as a bone grafting material for tissue regeneration of various sizes and shapes.

**Figure 6.** RANKL-induced osteoclastogenesis of RAW264.7 cells for 3 days. Experimental design of osteoclastogenesis using porous scaffold (right above). (**A**) optical images of TRAP+ cells (scale bar = 100 μm). (**B**) quantification of TRAP activity. The differences were considered significant when ns = not significant (*p* ≥ 0.05), \*\*\* *p* < 0.001, and # *p* < 0.0001 (*n* ≥ 3).

#### **4. Conclusions**

Because of their biodegradability and biocompatibility in physiological environments, biodegradable synthetic polymers are commonly used for a wide range of biomedical applications, especially bone repairing. Among them, PLGA has been clinically used as a bone grafting material. However, PLGA-based bioimplants often occur clinical failure due to low mechanical property and local acidification. Our findings proposed that mMH could enhance the mechanical property and neutralize acidification. The bECM was introduced to improve osteoconductivity by providing natural calcium and phosphate rich environment. Additionally, the DNA-derived bioactive molecule, PDRN facilitated biocompatibility and in situ vascularization during osteogenesis.

Taken together, we investigated that the synergistic interaction of mMH, bECM, and PDRN in the PMEP scaffold. The bioactive PMEP scaffold could inhibit osteoclastogenesis and promote adequate cell proliferation, angiogenesis, and osteogenesis in vitro. In the future study, we expect that the PMEP scaffold can regenerate the new bone in vivo by multifunctional abilities. This versatile biodegradable scaffold would apply to a novel bone tissue engineering as an advanced biomedical device.

**Author Contributions:** D.K.H. conceived and supervised the project. D.-S.K. and J.-K.L. contributed equally to this work. D.-S.K., J.-K.L., J.-W.J., S.-W.B. and J.H.K. performed the experiments and analyzed the data. The manuscript was written by D.-S.K., J.-K.L., Y.H. and T.-H.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by Basic Science Research Program (2020R1A2B5B03002344) and Bio & Medical Technology Development Program (2018M3A9E2024579) through the National Research Foundation of Korea funded by the Ministry of Science and ICT (MSIT), and the Korea Medical Device Development Fund grant funded by the Korea government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, Republic of Korea, the Ministry of Food and Drug Safety (202011A05-05), and Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HR16C0002).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**

