Onlay-graft of 3D printed Kagome-structure PCL scaffold incorporated with rhBMP-2 based on hyaluronic acid hydrogel

The onlay-graft, one of the most difficult graft conditions, is used for diverse clinical conditions, including plastic and dental surgery. The graft should withstand continuous pressure from overlying tissues and have excellent bone formation capability in a limited bone contact situation. We recently developed a 3D printed Kagome-structured polycaprolactone (PCL) scaffold that has a stronger mechanical property. This study evaluated the clinical feasibility of this scaffold for onlay-graft use. The value of the scaffold containing recombinant human bone morphogenetic protein-2 in a hyaluronate-based hydrogel (rhBMP-2/HA) to enhance bone regeneration was also assessed. 3D-printed Kagome-PCL scaffolds alone (n = 12, group I) or loaded with rhBMP-2/HA (n = 12, group II) were grafted using a rat calvarial onlay-graft model. Following sacrifice at 2, 4, and 8 weeks, all 3D-printed Kagome-PCL scaffolds were accurately positioned and firmly integrated to the recipient bone. Micro-computed tomography and histology analyses revealed a constant height of the scaffolds over time in all animals. New bone grew into the scaffolds in both groups, but with greater volume in group II. These results suggest the promising clinical feasibility of the 3D-printed Kagome-PCL scaffold for onlay-graft use and it could substitute the conventional onlay-graft in the plastic and dental reconstructive surgery in the near future.


Introduction
Autologous bone graft is still considered the gold standard for bone grafts, reflecting its osteogenesis, osteoinduction, and osteoconduction properties [1]. Disadvantages of autografts include extended healing time and need for additional surgeries. To overcome these limitations, bone tissue engineering has sought to temporarily substitute the extracellular matrix.
The influence of several bone morphogenetic proteins (BMPs) to enhance bone formation has been explored. Bone graft substitutes have been fabricated using freeze drying, solvent casting-porogen leaching, and gas foaming. These traditional techniques cannot control the internal architecture, porosity, and geometry of the scaffold [2]. For bone regeneration, the optimal scaffolds should precisely adopt the bone defect shape and intimately match the surrounding tissues [3,4]. Three-dimensional (3D) printing technology has rapidly become a promising alternative to traditional bone graft substitutes, and various alloplastic scaffolds with customized shapes and internal structures have been suggested [5,6].
Synthetic polymers that could be used as 3Dprinted scaffolds include polycaprolactone (PCL), poly(glycolic acid), poly(lactic acid), and poly(lacticcoglycolic acid) [7][8][9]. PCL is a biodegradable polymer that is widely used because of its malleability, excellent mechanical properties, and support of cell proliferation [10]. In recent study of a complex bone defect in rabbit calvaria, the PCL-Kagomestructure scaffold showed better fitting ability and osteoconductivity compared with the conventional grid-structure [4]. The Kagome-structure is typical and features excellent specific rigidity and a porous microstructure with open pores that can be manufactured by 3D printers [11]. A previous study documented the ability of a Kagome-structure scaffold to adopt a precise shape compared to a scaffold having a general lattice structure [3].
The PASS principle (P: primary closure, A: angiogenesis, S: space maintenance, S: stability) may enable successful bone grafting [12]. An excellent blood supply and stable 3D structure can result in successful bone healing that includes adequate angiogenesis. The onlay-graft is a bone graft procedure in which a transplanted bone substitute is placed directly on the surface of the recipient bone. Patients frequently require onlay-grafts, especially for dental implants and facial plastic or reconstructive surgery, which require increased bone volume. In contrast to the inlay-graft, which is performed inside the bony defect, the onlay-graft has a limited contact area for osteogenic cells and factors, and for angiogenesis factors that are mainly from the adjacent native bone with some from the soft tissue above the graft [13]. Moreover, onlay-grafts protrude from the recipient surface. This requires that the mechanical stability of the bone substitutes should be sufficient to withstand the external dynamic force from the overlying soft tissues (figure 1).
For successful onlay-grafts, conditions needed for bone are intimate contact with the recipient surface, proper stiffness to enable sustainable rigid fixation, proper microstructure that includes porosity to enhance osteoconductivity, and no serious inflammatory and immunologic reactions. The conditions for the surgical procedure include maximal blood circulation on the recipient site with little hematoma, proper flap management to ensure tension-free sutures to avoid wound dehiscence and to prevent flap necrosis caused by excessive stress, and proper suture technique to prevent subsequent tearing of the stitches. Thus, the onlay-graft is still challenging, and the risk of complications is high. The success of onlay graft surgery is reliant on the type of procedure and experience of the surgeon.
Recombinant human BMP (rhBMP-2) has been manufactured and refined for medical purposes. The United States Food and Drug Administration (FDA) approved rhBMP-2 incorporated in a collagen sponge in 2007 for bone regeneration [1]. Although collagen is widely used as a carrier of rhBMP-2, the watersoluble protein can be rapidly released from collagen by compression, diffusion, and degradation of the collagen in physiological conditions due to its poor mechanical stability [14]. To overcome these problems, many biomaterials have been researched as rhBMP-2 carriers to achieve stable localized rhBMP-2 concentrations for a sufficient period of time [15][16][17]. Incorporating the rhBMP-2 to the 3D-printed PCL scaffold has been difficult due to hydrophobic property of PCL. Acrylated hyaluronic acid (HA) is a naturally present and high molecular weight polymer with viscoelastic properties. HA has been recently used as a scaffold for rhBMP-2 and human mesenchymal stem cells for bone regeneration [18]. In addition, HA has medicinal value as it is involved in morphogenesis, wound healing, and inflammation [18]. For onlay-grafts, an osteoconductive Kagome-PCL scaffold that incorporates rhBMP-2 into the HA-based hydrogel could act as a dual or sequential delivery system that slowly releases rhBMP-2 and enhances bone formation.
The aims of this study are to explore the onlaygraft use of 3D-printed Kagome-PCL scaffolds for onlay-graft and surgical procedures using a calvarial bone rat model, and to evaluate the results of the onlay-graft with and without rhBMP-2 in the HA-based hydrogel.

Design of Kagome-PCL scaffolds for onlay-graft using micro-computed tomography (micro-CT)
A Kagome-structure model was fabricated as previously described [4,11]. For 3D modeling of the rat calvaria model, micro-CT scans were acquired of calvaria of 6 week-old male Sprague-Dawley rats using a SOMATOM Sensation 16 16-slice multidetector CT scanner (Siemens AG, Forchheim, Germany). The operating conditions were 120 kV, 220 mA, and 0.75 mm thickness. Axial 0.1 mm-thick views were reconstructed using the CT scanner with a H60s medium-smooth kernel. The field of view was 100 × 100 mm on the bregma (perpendicular intersection of the coronal suture and sagittal suture), and data were acquired in a 512 × 512 data matrix.
PCL (Mw = 43 000, Tm = 55 • C-65 • C; Polysciences, Inc., Warrington, PA, USA) was selected to fabricate the scaffold. The Kagome-structure scaffold was designed using a 3D reconstruction model and CATIA V5 R13 software (Dassault systems®; CATIA, Paris, France). The Kagome-structure model was cut along the outline of the calvarial contour at the coronal view of rat bregma (figure 2(A)). The scaffold was composed of a number of Kagome unit cells. The 3D-Kagome model structure was divided into regions of frames and columns. The cylinder-shape scaffold was 3 mm in height and 5 mm in diameter, with four side wings 1.2 mm in height and 3 mm in diameter, with a porosity of 50% ( figure 2(B)).

3D printer fabrication of PCL scaffolds with precision extruding deposition (PED) head
Slic3r version 1.2.9 open source software was used for a tool pathway generation of the STL model (figure 3(A)) [11]. The lab-made precision 3D printer with the PED head is shown in figure 3(B). The PED head comprised a barrel, screw, cartridge body, nozzle, gear set, heating block, and motor ( figure 3(B)). The structural design of the PED head was improved by a parallel-type gear set. The melted material was extruded by rotating the screw at 45 rpm with an air pressure of 250 ± 5 kPa to prevent heating to a temperature exceeding 88 • C, which is the melting point of the PCL pellet. Several Kagomestructure scaffolds were made simultaneously using a 100 µm ceramic nozzle. The onlay-graft scaffold was constructed in three dimensions by flipping the scaffold, as shown in figure 3(C). The scaffold was turned upside down so that the surface quality of the support location degraded during 3D printing. To intimately match the surface of the rat calvaria, supporters were not created on the curved part of the scaffold. However, the wing part floats in the air when upside down for precise printing of the curved part of the scaffold. To overcome this problem, a support was additionally designed to support the wing part during wing printing, as shown in green portions of figure 3(C). Finally, a scaffold was prepared by removing the support from the prepared scaffold.  penicillin streptomycin, trypsin, and low-glucose Dulbecco's modified Eagle's medium were purchased from GIBCO BRL (Grand Island, NY, USA).

Preparation and surgery process for
HA (0.25 mmol, based on the repeating unit MW) was dissolved in 40 ml distilled water and EDC (0.24 g, 1.25 mmol), HOBt (0.17 g, 1.25 mmol) and ADH (2.2 g, 12.5 mmol) were added to the solution. The EDC-mediated coupling reaction between the carboxyl group of HA and the hydrazide group of ADH proceeded with continuous stirring at room temperature for 8 h. HA-ADH was dialyzed against 100 mM NaCl for 2.5 d and against distilled water for 1 d using a dialysis membrane (MWCO14,000; Spec-traPor, Rancho Dominguez, CA, USA). NAS (0.5 g, 3 mmol) was then added to the HA-ADH solution. The reaction continued with stirring at room temperature for 12 h. HA-ADH-NAS was dialyzed extensively against 100 mM NaCl for 2.5 d and against distilled water for 1 d. The product was lyophilized for 3 d to obtain solid acrylated HA.
For gel preparation, the acrylic HA was dissolved in a TEA-buffered solution (0.3 M, pH 8). MMP-sensitive peptide (GCRDGPQGIWGQDRCG) was loaded in the same molar ratio as the acryl and thiol groups [18]. Polyethylene glycol (PEG)-SH4 (MW 10000) was added as a cross-linker with the same molar ratio as the acryl and thiol groups. The HA-based hydrogel was formed by an additional Michaelis Arbuzov reaction [19]. The reaction mixture was incubated at 37.1 • C for gelation. For animal experiments, 1 µg rhBMP-2 was mixed and reacted with the HA gel [20]. Then, 35 µl of the HA gel reaction mixture was injected into the scaffold using a pipette. For the gelation of HA, the scaffold loaded with the gel in the scaffold was incubated at 37 • C for 30 min. The HA loaded scaffold was implanted immediately after gelation in the calvaria of rats in the Kagome-PCL scaffold containing rhBMP-2/HA (group II).

Scaffold characterization
Morphology reproducibility of Kagome-PCL scaffold analyzed via a scanning electron microscope (SEM; SU3800, HITACHI, Japan). The prepared scaffold was coated with platinum for 120 s using a sputtering system to prepare for analysis. Platinum-coated scaffolds were fixed on an SEM-specific jig and analyzed in a vacuum chamber.
In addition, in order to analyze the reproducibility of the pores of the fabricated scaffold, pore size  [11]. © IOP Publishing Ltd. All rights reserved. measurement was carried out, and to check whether the HA gel was well mounted inside the fabricated scaffold, it was observed through a digital microscope (Leica; DMS1000, Germany). Five samples were used for measurement, and five points were measured for each sample.

Animal procedures and surgical installation of Kagome-PCL scaffolds
The animal procedures were approved by the Institutional Animal Care and Use Committee (No. 2019-12-335) and followed the ethical principles for animal experimentation established by the institute. Sample size was initially calculated for 20 animals using six experimental groups, considering a significance level of 5% and a statistical test power of 95%. To rationalize and distribute the animals, we considered 24 animals to be suitable for the experiment. Twentyfour rats were purchased from Orientbio © (Seongnam, Korea) and divided into six groups. The 3 mm height of scaffolds with and without rhBMP-2/HA was determined following sacrifice of rats at 2, 4, and 8 weeks (table 1).
The surgery to install each onlay-graft is summarized in figure 4(A). In more detail, rats were anesthetized by an intramuscular injection of 20 mg of tiletamline and zolazepam hydrocholoride (Zoletil 50®) and xylazine hydrochloride (Rompun®). The hair from the surgery site was shaved, and the site was disinfected with povidone-iodine. After infiltration of Group II (with rhBMP-2/HA) 1.8 ml of 2% lidocaine HCl (Huons®, Kyeongi-do, Korea) for local anesthesia, a linear incision was made from the distal margin of the scaffold along the sagittal line, and the periosteal was anteriorly dissected to form a pocket beneath (figures 4(B)-(D)). Three to four foramina were made for decortication with a micro-drill (2.0 mm diameter; Neobiotech, Seoul, Korea) at the bregma [21] (figure 4(E)). The scaffold was positioned on the bregma and fixed using miniscrews (4 mm long, 2 mm diameter; Neobiotech) on the four side wings (figure 4(F)). During drilling, the site was irrigated with sterile saline to prevent thermal damage. After flap adaptation, the scaffold was covered with non-damaged soft tissue (figure 4(G)). The surgical site was sutured using a continuous horizontal mattress suture technique and a vertical mattress suture on the anterior third of incision line with 4-0 black silk (figures 4(H) and (I)). Following the surgery, cefazolin (50 mg kg −1 ) and ketorolac tromethamine (1 mg kg −1 ) were administrated intramuscularly twice daily for 3 d to reduce pain and to prevent infection. The site was disinfected with povidone-iodine solution.

Micro-CT analysis
After sacrifice, the scaffold was carefully excised en bloc with calvarial bone using a saw after incision and dissection of skin and subcutaneous tissue ( figure 6). The removed tissues were immediately fixed in 4% (w/v) paraformaldehyde and stored for 5 d. The grafts were obtained by removing the fixation screws after sacrifice. Micro-CT scans were acquired as described earlier in the text. Fixed specimens were wrapped with parafilm to minimize drying. The specimens were scanned by micro-CT using a model 1176 device (SkyScan, Kontich, Belgium) at a high resolution with an aluminum filter (0.5 mm). Scan settings were as follows: resolution of 35.76 µm pixels, energy of 50 kV, and intensity of 500 µA. Projection images of CT scans and reconstructions were saved as 16-bit TIFF files. The image files were transferred to set the measurement software (CTVox; Blue Scientific, Cambridge, UK).
To compare the degree of bone formation in vivo, CT scans were acquired each week. The new bone volume and scaffold shape retention were qualitatively measured by 3D reconstruction using Materialise v 21.0 (MIMICS, Leuven, Belgium). This program integrates CT data composed of tomographic images to reconstruct a 3D model. First, 3D modeling was performed by integrating CT data of pure rat calvaria from the negative control for the reference shape. Then, to measure the volume of the new bone, only the new bone was isolated and reconstructed into a 3D model. Briefly, referring to the reconstructed pure rat calvaria 3D model, the original bone area was removed and only the remaining area was reconstructed into the 3D model. The volume of the isolated new bone was measured as CAD data in the MIMICS program. The bone formation ratio was calculated from the measurement data per total cavity volume (35 µl) of the 3 mm Kagome-PCL scaffold.

Histological analyses
After micro-CT scan acquisition, decalcification was performed using 10% ethylenediaminetetraacetic acid for 28 d at room temperature. The prepared samples were embedded in paraffin after dehydration. Coronal sections (4 µm in thickness) were stained with hematoxylin and eosin and Masson's Goldner trichrome to evaluate angiogenesis, new bone formation, and spatial gap between the scaffold and calvarial bone. In addition, the height of the scaffold was measured using histologic images to evaluate the maintenance of the shape of each sample. The height was calculated as the average height of five points on top of the scaffold per sample.

Statistical analysis
All data are expressed as the mean ± standard deviation. Statistical analysis was performed on the micro-CT results using single factor analysis of variance with SPSS version 25.0 software (SPSS, Inc., Chicago, IL, USA). Significance was considered at a level of 0.05.

Results
The fabrication reproducibility of the Kagome-PCL scaffold was confirmed in figure 5. The realization of the pore size of the fabricated scaffold was measured through a digital microscope, which was measured to be an average of 524.34 ± 9.62 µm. In addition, it was confirmed that the interior of the scaffold manufactured by HA was well mounted, as shown in figures 5(D)-(F).
The 3D-printed Kagome-PCL scaffolds were fabricated for an intimate surface fit. No dehiscence, swelling, or hematoma was evident in any rat during the observation period ( figure 6(A)). At sacrifice and removal of the fixation screws, all scaffolds were wellintegrated with the calvarial bone with fibrous tissue coverage ( figure 6(B)).

Micro-CT results
CTVox software was used to analyze bone formation within the Kagome-PCL scaffold in CT images. The differences in bone formation in groups I and II are shown in figures 7 and 8 and table 2.

Histologic results
Close contact was formed between the Kagome-PCL scaffold and the calvarial bone. The close contact did not allow invasion of any soft tissue to the surface. Dense fibrous tissues were integrated from the lateral and coronal sides of the PCL scaffold at the early stage (2 weeks) regardless of the absence or presence of rhBMP-2/HA. New bone formation was observed from the calvarial bone below the scaffold, and active bone remodeling was evident around the decortificated holes. The roofs (third columns) of all the scaffolds were well-maintained without collapse in all samples.
The histological differences in bone formation in groups I and II are shown in figure 9.
The height of the scaffolds was measured (table 3 and figure 10) and were found to be approximately 3 mm (max error rate 2.14%; approximately <65 µm). There was no significant change in the height of the scaffolds.

Discussion
Successful onlay-graft outcome requires that the graft material be placed at the recipient site as tightly as possible. In addition, considering the unique nature of the onlay-graft, the shape and size of the graft material should be maintained over time.
Maxillofacial bony defects have complex curvatures, which makes it difficult to achieve tight contact between the recipient site and the conventional graft. The result can be malunion or displacement of the graft. The current 3D printing technology can fabricate bone graft materials with a customized size and shape for such complex conditions. This is a marked improvement over the classical types of bone grafts. Tight contact between the graft and the recipient site enhances the speed of bone engraftment due to faster vascularization and cellular migration. In addition, the intimate contact could also improve graft stability at the recipient site, which could enhance the graft success rate [13,22]. In our previous study [4], we showed that in the inlay-graft condition, the positions of the 3D-printed Kagome-PCL grafts were far more stably maintained with time than that of the conventional 3D printed grid-type PCL graft because of the tighter contact of the Kagome type at the recipient site.
Studies have sought to engineer scaffolds by considering solvent casting, fiber bonding, phase separation, gas-induced foaming, and salt leaching to mimic the complex geometries of the human body [23]. Among them, PED with a mechanical screw dispensing system was developed to fabricate more delicate PCL scaffolds [24]. This involved the extrusion of a highly viscous material through a smalldiameter nozzle (<100 µm diameter), and desired mechanical properties, structural integrity, and controlled pore size and interconnectivity [3,11,24]. In this study, using this technique, 3D-printed Kagome-PCL scaffolds were accurately manufactured to meet the needs for mechanical strength and delicate structure (figures 7(A) and 8(A)).
The scaffold height was designed to be 3 mm since a graft height of at least 2 mm is required for cosmetic surgery [13,25]. Continuous pressure over the graft from overlying soft tissues, such as muscle, skin, or mucosa, might cause graft resorption and displacement, with an unsatisfactory result. Therefore, the absorbable graft should endure the overlying soft tissue pressure without change of the graft height and shape until newly regenerated bone replaces the graft materials as the graft gradually degrades with time [1]. In this study, similar with our previous studies [4,11], the Kagome graft had sufficient mechanical strength to resist overlying pressure in the critical period. Collapse was not observed in any sample, and new bone grew well from the surface of the recipient native bony surface to the graft (table 2).
Wound dehiscence at the graft site is the most frequent cause of graft failure of bone grafts [26]. In particular, an onlay-graft increases the possibility of wound rupture because the excess volume of the graft could extend to overlying soft tissues [22]. Once wound dehiscence occurs, the graft material is exposed and is infected. Thus, a careful surgical design is important to prevent wound dehiscence. In this study, the skin incision was performed to cover the graft by non-damaged soft tissue. No wound dehiscence or graft failure occurred in any of the rats in this study.
PCL is a popular material for 3D-printed bone graft materials [7]. It is degraded by the hydrolysis of its ester linkages in the human body and has therefore received great attention for use as an implantable biomaterial. PCL has been approved by the US FDA for specific human applications that include drug delivery, suture, or as an adhesion barrier [27]. However, PCL suffers from some tissue engineering pitfalls that include slow degradation rate, poor mechanical properties, and poor cell adhesion [26].
To overcome the limitations of PCL for bone regeneration, rhBMP-2 was incorporated into the grafts. rhBMP-2 is a potent growth factor that enhances the osteoblastic differentiation of stem cells, angiogenesis, and new bone formation. Maintaining an adequate concentration of rhBMP-2 at the graft site for a required period of time is essential for graft success because of its short biological half-life, watersoluble property, and rapid clearance [14][15][16]. Therefore, a proper carrier of rhBMP-2 may be critical for its successful use. Recently, to regulate the expression and concentration of rhBMP-2, several hydrogels have been developed as a delivery scaffold. They include HA [18], crosslinking of aqueous gelatin with glutaraldehyde [28], and functional nanoparticlehydrogel complexes [29]. In particular, hydrogels are hydrophilic hydrated polymers that easily deliver oxygen and nutrients to tissues [30]. The hydrophilic properties allow incorporation of rhBMP-2 and other water-soluble growth factors into hydrophobic polymer scaffolds like PCL. HA, a biologically active molecule that regulates the tissue repair process  on multiple levels, also showed a positive effect on wound healing [31]. Benefits of HA in both early and late wound healing have been described [32,33]. A recent study documented that HA hydrogels promoted comparable skin-wound healing outcomes with basic fibroblast growth factor [34]. In addition, HA-incorporated composite scaffolds showed great potential for enhancing osteogenesis and mineralization, and as local delivery carriers for osteoinductive components [35]. Concerning our results, an HA hydrogel could enhance bone regeneration and prevent wound dehiscence, which is the most common complication of the onlay-graft [13]. Our histological and radiological results also indicate a synergistic effect of the Kagome-PCL scaffold and HA hydrogel. This dual scaffold could deliver rhBMP-2 slowly and consistently without degeneration, which is supported by the observations of gradual angiogenesis and retention of HA through 8 weeks ( figure 9(B)). In this study, the 3 mm Kagome-PCL scaffold could endure external forces from the overlying soft tissue for 8 weeks in vivo. New bone formed in the first column cavity of the scaffolds in group I at 8 weeks. However, the amount of new bone at 8 weeks in group I (24.57%) was significantly less than that in group II at 2 weeks. The new bone occupied approximately 48.4% and 69.5% of the total cavity volume of the scaffolds containing rhBMP-2/HA at 2 and 8 weeks, respectively.
In this study, HA remained during the resorption that occurred for up to 8 weeks. In the in vivo environment, as the MMP-sensitive peptide cross-linker in the HA-based hydrogel is degraded by the surrounding cells, the stiffness of the hydrogel is gradually reduced. Thus, the exposure rate of the hydrogel to the surrounding tissue is important in determining the degradation speed of HA. With this in mind, the design of the Kagome-PCL scaffold could be considered more advantageous as a container to induce the delayed degradation of the HA-based hydrogel than the conventional grid-structure scaffold. In our  preliminary study, the HA-based hydrogel degraded at a much slower rate in the Kagome-PCL scaffold than in the conventional grid scaffold, even though the porosity ratio was the same in both scaffolds. This might be because the total surface area of the hydrogel that is exposed to surrounding tissue is only about 26% in the Kagome-PCL scaffold but about 90% in the conventional grid type scaffold (data not shown). The effective delivery of rhBMP-2 and synergistic effect by the Kagome-PCL scaffold/HA hydrogel system could be crucial in the onlay-graft procedure, which has high risks of wound dehiscence and lack of osteogenic cells from recipient bone. Also, as excessive dose of rhBMP-2 has often adverse effect, this slow release feature of rhBMP-2 in the Kagome-PCL scaffold with HA hydrogel could have additional benefit of reduction the therapeutic dose and side effects of rhBMP-2 while maintaining bone healing potency [14].
Further studies are needed to reveal the duration and release profile of rhBMP-2 with this dual scaffold. These studies should include assessments of the postulated release profile of bigger rhBMP-2/HA containing Kagome-PCL scaffolds in animal experiments of larger sample size. After verifying the release profile of this dual scaffold, the optimal condition for clinical application should be determined, including the concentration of rhBMP-2, porosity of the Kagomestructure, and maximum size of the Kagome-PCL scaffold.

Conclusion
All 3D-printed Kagome-PCL scaffolds were accurately placed and firmly integrated to the recipient bone. Micro-CT and histology analyses revealed the maintenance of the height of the scaffolds in all animals, confirming the sufficient mechanical strength of the scaffold. In addition, 3D-printed Kagome-PCL scaffolds maintained rhBMP-2 mixed in HA during the observation period. Newly formed bone in the scaffold was observed in the absence (group I) and presence (group II) of rhBMP-2 mixed in HA. New bone volume was greater in group II. These results suggest the promising clinical feasibility of the 3Dprinted Kagome-PCL scaffold for use in onlay-graft procedures.

Data availability statement
The data generated and/or analyzed during the current study are not publicly available for legal/ethical reasons but are available from the corresponding author on reasonable request.
Part of the figure Created with BioRender.com (Agreement number:DI22LU9CFV).

Author contributions
Jeong-Kui Ku performed surgery, analyzed the data, and wrote the manuscript. Kang Gon Lee manufactured the HA hydrogel with rhBMP-2, performed data analysis, and wrote the manuscript. Min-Soo Ghim manufactured the 3D printed Kagomestructure PCL scaffold, analyzed the data, and wrote the manuscript. Young-Kyun Kim provided advice concerning the scaffold application. Sang-Hyug Park provided advice concerning the scaffold fabrications. Yongdoo Park supervised the study and provided critical discussion for the HA hydrogel with rhBMP-2. Young-Sam Cho supervised the study and provided critical discussion of the 3D printed Kagome-structure PCL scaffold. Bu-Kyu Lee conceived the study determined the shape of the scaffold, administered and supervised the study, and wrote the manuscript.