**4. Discussion**

Many studies have been conducted on implant surfaces treated with growth factors like BMP-2, to increase vertical and horizontal bone regeneration without the use of additional bone grafts or barrier membranes [37,38]. Although BMP-2 is one of the effective growth factors, rapid release at an early stage and rapid diffusion into body fluids has limited its clinical applications [39,40]. In order to overcome these problems, the present study was undertaken on a heparin-based release system to provide the sustained and local release of BMP-2 [30,41,42]. Since titanium has good mechanical properties, corrosion resistance, and biocompatibility without biological activity, it is generally used as an implant material [43]. Due to its inertness in vivo, physical adsorption or chemical coatings are

required to enable biomolecules to adhere to titanium surfaces [44]. However, the chemical substances used for this purpose, such as, APTES, EDAC, and NHS, are harmful to the human body [45]. In this study, titanium was coated with heparin using a dopamine surface modification technique, which has been reported to be biocompatible in in vivo toxicity studies [46]. Dopamine is a small molecule substance that possesses catechol and amine functional groups [47]. The catechol component bonds to titanium oxide surfaces and provides an amine group that can bind heparin [48,49].

It has been reported that combinations of different growth factors produce synergistic effects that promote the complex cellular activities that occur during bone regeneration and osseointegration [33,50,51]. However, most studies are limited to cell experiments, the studies reproducing the clinical situation in large animals are rare. Therefore, in the present study, rhBMP-2 and rhPDGF-BB immobilized on Ti surfaces modified with Hepa-DOPA were utilized to establish an effective growth factor delivery system to ensure sustained factor release in appropriate amounts over sufficient time in beagle mandible model. The concentration of rhBMP-2 that is used in this study was set based on our previous studies, which reported 0.75 mg/mL of rhBMP-2 is effective on bone regeneration [34,35]. In PDGF, the previous study of Choo et al. [33] reported that the appropriate concentration of PDGF was from 0.3 mg/mL up to 1 mg/mL, thus we chose 0.75 mg/mL of PDGF concentration within that range. BMP and PDGF combined groups were studied in 1:10 ratio (PDGF:BMP) in previous cell study [6], but we intended to increase the rate of PDGF to get enhanced bone regeneration as ratio of 1:1. The results of our in vitro release experiments indicated that the inclusion of heparin achieved sustained growth factor release, although previous studies have reported a burst release pattern resulting in the release of 70%–90% over the first 6 h [52]. After 24 h, the amounts of PDFG-BB and rhBMP released from PDGF/Hepa/Ti and BMP/Hepa/Ti were approximately 38%, respectively, and the amounts of rhBMP and PDGF-BB released from PDGF/BMP/Hepa/Ti were 45%, respectively. This suggests that PDGF-BB and rhBMP-2 show similar release tendencies in the presence of other growth factors and heparin.

SEM analysis of Ti surface modified by heparin, rhPDGF-BB, or/and rhBMP-2 showed their surface morphologies were similar to that of untreated Ti. XPS showed heparin was successfully grafted on anodized Ti surfaces, as indicated by higher C and N peaks. In a previous study, successful immobilization of rhBMP-2 on surface of carboxymethyl chitosan (CMCS)-grafted Ti was attributed to the covalent bonds formation between the carboxyl groups of CMCS and the amine groups of rhBMP-2 [53]. In the present study, successful anchoring of rhBMP-2 and rhPDGF-BB to Hepa/Ti surfaces was demonstrated by further increases in N content.

ALP activity and calcium deposition as determined by in vitro studies are widely used as markers of early and late differentiation to osteoblasts, respectively [54,55]. We measured ALP activity after culturing MG-63 for 3, 7, or 10 days and calcium deposition after 7 or 21 days. ALP activities of cells that are cultured on PDGF-BB and/or BMP/Hepa/Ti were found to be significantly higher than those of cells cultured on untreated Ti on days 7 and 10 (\* *p* < 0.05 and \*\* *p* < 0.01). As reported in a previous study [6,56], our finding confirmed that rhBMP-2 and rhPDGF-BB both stimulate osteoblast differentiation. Furthermore, the calcium contents of cells cultured on PDGF-BB or/and BMP/Hepa/Ti were significantly higher than those of cells cultured on untreated Ti on days 7 and 21 (\* *p* < 0.05 and \*\* *p* < 0.01). These results indicate that growth factor immobilized Ti substrates can stimulate matrix formation and improve osteoblast cell function. ALP activity and calcium deposition in the PDGF/BMP/Hepa/Ti group were significantly higher than in the PDGF/Hepa/Ti group, but were significantly lower than in the BMP/Hepa/Ti group. In order to investigate the gene expression levels of osteocalcin (OCN) and osteopontin (OPN), which are markers of cell differentiation, the total mRNAs of cells cultured on four Ti surfaces were extracted. On day 21, significant differences were observed between the OCN and OPN expressions of cells cultivated on BMP/Hepa/Ti and PDGF/Hepa/Ti or PDGF/BMP/Hepa/Ti.

In the result of previous cell studies [6], PDGF-BB (5 ng/mL)/BMP-2(50 ng/mL)/Hepa/Ti (a 1:10 ratio of growth factors) was significantly higher than BMP/Hepa/Ti, whereas in this study, PDGF-BB (50 ng/mL)/BMP-2 (50 ng/mL)/Hepa/Ti (a 1:1 ratio) was significantly lower than BMP/Hepa/Ti. Furthermore, ALP, calcium deposition, and gene expression not synergistically increased by the copresence of rhPDGF-BB and rhBMP-2. These results indicate that different concentrations and ratios of PDGF-BB should be selected.

The in vivo animal study was conducted to evaluate the synergic effects of rhBMP-2 (0.75 mg/mL) when combined with rhPDGF-BB (0.75 mg/mL). A total of 40 implants of five groups (control (Ti) group, Hepa/Ti group, PDGF/Hepa/Ti group, BMP/Hepa/Ti group, and PDGF/BMP/Hepa/Ti group) were implanted without using additional implants or barrier membranes. To reproduce clinical situations, a buccal open defect model with a 2.5 mm deep peri-implant bone defect was produced in the mandibles of five beagle dogs. This model was designed with loss of buccal bone and a mesial-lingual-distal 1 mm defect area around the 2.5 mm upper portions of implants. The upper part of implants (microthread) was exposed 2.5 mm above alveolar bone and the lower part (macrothread) was implanted into cortical bone. After a healing period of eight weeks, ISQ values showed the presences of rhBMP-2 and rhPDGF-BB contributed positively to implant stability. Furthermore, ISQ increases were significantly higher in the PDGF/BMP/Hepa/Ti and BMP/Hepa/Ti groups than in the other three groups (\* *p* < 0.05). However, there was no difference between these two experimental groups (*p* > 0.05). μCT analysis showed new bone volumes in the BMP/Hepa/Ti and PDGF/BMP/Hepa/Ti groups were significantly greater than in the other groups (\* *p* < 0.05), but no significant difference was observed between these two groups (*p* > 0.05). Our histomorphometric analysis confirmed vertical new bone formation and osseointegration approaching microthreads around peri-implant bone defects in the BMP/Hepa/Ti and PDGF/BMP/Hepa/Ti groups, whereas osseointegration was unsuccessful in the Ti, Hepa/Ti and PDGF/Hepa/Ti groups, in which a 1 mm gap between implant fixture and bone was observed at lingual portions. microBIC values in the BMP/Hepa/Ti and PDGF/BMP/Hepa/Ti groups were significantly greater than in the other groups (\* *p* = 0.000), but no significant difference was observed between the two (*p* > 0.05). Our in vivo studies were confirmed that PDGF (0.75 mg/mL)/BMP (0.75 mg/mL)/Hepa/Ti (a 1:1 growth factor) had no synergic effect. Based on these results and limitations of present study, further studies on subdivided concentrations and ratios, various implant surfaces and slow releasing systems are required to more comprehensively investigate the effects of combinations of different growth factors on osseointegration and bone regeneration.
