In Vivo Behavior of a Custom-Made 3D Synthetic Bone Substitute in Sinus Augmentation Procedures in Sheep

Several synthetic bone substitute materials have been investigated and developed over the last decades to overcome the drawbacks of autologous bone grafting, such as cost, patient disability, and limited availability.¹  Synthetic bone substitutes are often based upon mimicking the properties of natural bone and should serve as a 3D template for initial cell ingrowth and subsequent tissue formation. Geometry, ultrastructure, and mechanical properties of these biomaterials are determinants for successful healing of bone defects, and their aptitude for in vivo resorption is important for allowing simultaneous replacement of the material itself by newly formed bone.

To mimic the structure of cancellous bone, macroporosity with open and highly interconnected pores is of key importance and is often introduced into synthetic bone graft substitutes.²  In fact, it has been suggested that both the chemistry and geometry of the biomaterial in contact with these cells are critical factors to induce cell differentiation into bone-forming cells.^6–8  Pores are necessary for bone tissue formation, as they allow migration and proliferation of osteoblasts and mesenchymal cells and facilitate vascularization. Moreover, they accelerate the biodegradation of the material by increasing the surface area in contact with body fluids. To ensure cell viability and function, ideal scaffolds need to exhibit porosity in different length scales: nano-pores to allow for molecule transport essential for any nutrition, waste removal, and signaling; micropores to ensure cell migration and capillary formation; and millimeter-wide pores to incorporate nerves and blood vessels.^3,4  An optimal ingrowth of osteoblast cells requires pore sizes of 200–400 microns,⁹  while a pore size of even 300 microns is recommended for enhancing vascularization.¹⁰  In general, an open porosity above 50 vol% and pore sizes in the range of 200–800 microns are considered as optimal for bone tissue ingrowth.^10–12 

Moreover, research has suggested that the degree of interconnectivity is more critical than pore size. Furthermore, a porous surface may improve mechanical interlocking between the biomaterial and the surrounding natural bone, providing a good mechanical stability at this critical interface.^5,18,19  However when designing the scaffold architecture, a balance between a high degree of porosity and a reasonable mechanical stability—which could be due to the correct biomaterial chemical composition—must be accomplished to ensure tissue integrity until the bone-healing process begins.¹³ 

Calcium phosphate ceramics such as hydroxyapatite (HA) and tricalcium phosphate (TCP) are considered as the most suitable ceramic materials for bone reconstruction since they are highly biocompatible and osteoconductive, they can be degradable, and their mechanical strength is sufficient for bone reconstruction in non–load-bearing applications.^14–16  Biphasic calcium phosphate (BCP) ceramics characteristics are based on the balance between the less soluble HA and the more soluble TCP. By varying the HA/TCP ratio, the mechanical and biological performance of the ceramic can be tailored.¹⁷ 

In recent years, a number of techniques have been employed to manufacture ceramic bone substitutes meeting all these requirements, including rapid prototyping. This technique for producing synthetic bone substitutes, based on the plotting of 3D scaffold from computer-aided design files, offers a tool for precise control of the overall geometry and the properties of the porous structure.

In this study, a 3D synthetic bone scaffold—manufactured through the rapid prototyping technique following criteria derived for scientific literature both for chemical and geometrical composition—was used as graft for sinus augmentation procedures in an animal model to evaluate the occurrence and progression of bone regeneration process and the scaffold resorption rate during a total healing time of 90 days.

The ceramic scaffolds used in this study were fabricated by the direct rapid prototyping technique of dispense-plotting, as previously reported.^20,21  A virtual scaffold model was designed with a cylindrical outer geometry by using 3D computer aided design software. The size of the model was adapted to the shrinkage of the ceramic material in the subsequent sintering process. The inner geometry, that is, the pathway of the material rods, was defined by custom-made software that generates the control commands of the rapid prototyping machine. To build up the green bodies, material rods (consisting of a paste-like aqueous ceramic slurry) were extruded out of a cartridge through a nozzle and deposited using an industrial robot (GLT, Pforzheim, Germany). In this study, hydroxyapatite (HA) and tricalcium phosphate (TCP) powders (Merck, Darmstadt, Germany) were blended to obtain a powder mixture with a HA/TCP weight ratio of 30/70. The characteristic rheological behavior of the aqueous ceramic slurry was achieved by thermal treatment of the raw HA powder at 900°C for 1 hour, and by adding a compatible binder/dispersant system of organic additives of 10.5 wt% relative to the mass of ceramic powder. The rod deposition was controlled in x, y, and z directions to assemble 3D scaffolds layer by layer on a building platform. Rotating the direction of the rod deposition by 60° from layer to layer generated, a 3D network with an interconnecting pore structure. The assemblies made of ceramic slurry were dried at room temperature and subsequently sintered at 1250°C for 1 hour. Finally, the sintered scaffolds were manually reduced to smaller blocks with a volume of about 0.14 cm³ to remove the solid rim that resulted from the turning points at the edge of the printed pathways.

Images of the sintered dispense-plotted scaffolds taken with a light microscope (MZM1, Askania, Mikroskop Technik, Rathenow, Germany) were analyzed using an image analysis system (analySIS, Soft Imaging System, Germany) to determine rod diameters and pore sizes. Density of the sintered scaffolds was measured using weight/volume method in which volume was determined geometrically and by using a helium pycnometer (AccuPyc 1330, Micromeritics, Germany). Semi-quantitative X-ray diffraction measurements (XRD, 3000P, Seifert, Germany) were applied to determine the phase composition of the sintered material, and a Scanning Electron Microscope (SEM; Quanta 200, FEI, The Netherlands) was used to characterize the surface of material rods and fracture faces of sintered samples.

The present study was approved by the Ethical Committee for animal research of the Universities of Teramo and Chieti-Pescara. For the present research, 6 adult sheep, 2 years old and <50 kg of weight, were bred in an authorized farm according to the European community guidelines (E.D. 2010/63/UE). Animals were quarantined for 2 weeks to check the general health status before performing the monolateral sinus augmentation. After surgical procedures, the animals were randomly divided into two different groups and euthanized to explant grafted sinuses after 45 and 90 days, respectively.

The animals were pre-anesthetized with xylazine 0.2 mg/kg i.m. (Rompun, Bayer Healthcare AG, Wuppertal,Germany), followed by diazepam 0.2 mg/kg i.v. (diazepam 0.5%, Intervet Productions S.r.l., Segrate, Italy) and atropine sulfate 6 mg i.m. (atropine sulphate, Fort Dodge) before inducing anesthesia with ketamine 10 mg/kg i.m. (Ketavet 100, Intervet). The animals were intubated, and general anesthesia was maintained by inhalation of 2.5% halothane (Halothane; Merial, Milano, Italy) in a mix of oxygen. For each animal, a monolateral sinus augmentation procedure was carried out with an extra-oral approach, as schematized in Figure 1. The surgical field was prepared to include the main landmarks, namely the angular vein of the eye and the transverse artery of the face. An oblique extra-oral incision was made over the most ventral aspect of the maxillary sinus with a #11 scalpel blade. The subcutaneous tissues were incised, and the masseter muscle and the adherent periosteum were gently mobilized before preparation of the bony window. The lateral wall of the sinus was approached through an oval ostectomy (1 cm superior and 1 cm caudal to the facial tuberosity), carried out using a piezoelectric unit surgery (BioSAF Easy Surgery, Ancona, Italy). Then, the Schneiderian membrane was elevated and two blocks of ceramic biomaterial (of 0.14 cm³) were inserted within the sinus (Figure 1a). Deep and superficial fasciae of the masseter muscle were sutured with a 3–0 multifilament resorbable suture (Vicryl, Ethicon Inc, Somerville, NJ) in a simple continuous pattern. The overall average surgical time for each sheep was <40 min. The animals were given 20 mg/kg of ampicillin intravenously (Vetamplius, Fatro S.p.a., Ozzano dell'Emilia, Italy) every 12 hours for 3 days postoperatively. The wounds were inspected daily for clinical sign of complications. Two groups of three animals each were euthanatized respectively after 45 and 90 days by applying an overdose of thiopental (pentothal sodium, Intervet) and embutramide (Tanax, Intervet).

Sinus explants (woven blocks of <4.5 cm³ in size; Figure 1b) were fixed in 4% paraformaldehyde solution (PBS, pH 7.4) overnight. The fixed explants were firstly analyzed using the micro computerized tomography (micro-CT) technique, as described, to evaluate the biomaterial integration into the host tissue. Following micro-CT analysis, the explants were divided in half and decalcified for at least 30 days in 14% EDTA, as previously reported.^22–24  Decalcified samples were placed overnight in 30% sucrose-phosphate buffered saline (PBS) before freezing. Each explant was completely cryosectioned at 10 μm of thickness. For the morphological and histomorphometric analysis, 4 serial sections were collected every <400 μm of distance. One section was stained with hematoxylin-eosin (HE) to describe the architecture of the newly formed tissues and to evaluate the presence of signs of inflammation. In detail, leucocytes infiltration was scored, as described by Zecha: none (0–5%, leucocytes/total cells), low (5–30%), moderate (30–60%), and high (>60%).²⁵ 

The other serial cryosections were double-immunolabeled using von Willebrand Factor (vWF) and alpha-Smooth Muscle Actin (α-SMA) primary antibodies in order to describe the new blood vessel network organized within the implanted area at 45 and 90 days post-graft. Specifically, vWF was used to quantify the total vascular area (VA), while both vWF and α-SMA indicated the typology and the degree of maturity of the new blood vessels.^26–28  The double immunostaining was performed as previously described.²⁹  In detail, rabbit anti-vWF (diluted 1:400 in PBS/1% BSA, Dako, Glostrup, Denmark) and mouse anti-α-SMA (diluted 1:500 in PBS/1% BSA; Oncogene) were applied together at room temperature overnight, after double exposure to the hotbox oven at 95°C for 5 minutes. The tissue sections were then incubated with a mixture of secondary antibodies. In particular, the secondary goat anti-rabbit vWF antibody (1:100 in PBS; Sigma), and the secondary goat anti-mouse α-SMA antibody (1:100 in PBS; Sigma-Aldrich Biotechnology, St. Louis, Missouri, USA) were used and were labeled with fluorescein isothiocyanate (1:100 in PBS; Sigma), and CY3, respectively. According to the data sheet instructions, bovine aorta tissue collected at the slaughterhouse was used as positive control. Normal goat serum was used as negative control in place of primary antibodies. The specificity of double immunostaining was verified by separately localizing each antigen. Tissue sections were counterstained with DAPI (Vectastain, Burlingame, Calif) to visualize cell nuclei.

The micro-CT two-dimensional (2D) analysis was performed using the X-ray mycrothomography, (Skyscan mod.1072, Kontich, Belgium) operating at 98 μA currents and 100 kV voltages. The samples were analyzed at 20 mm × 20 mm image size, reaching a sample resolution of 19.1 μm pixel size. The data sets were acquired over a rotation range of 180° (with 0.45° rotation step) and reconstructed with a software (NRecon v1.6.3; Skyscan), based on the cone beam algorithm.³⁰  Morphological 3D reconstruction was carried out by using 3D Creator (v2.5; Skyscan).

Light microscope analysis allowed quantification of the newly deposited bone, scaffold, and soft tissues extensions recorded within the implanted area. Additionally, immunofluorescence analysis was performed to calculate the total VA stimulated by the biomaterial graft. Both analyses were performed utilizing an Axioscop 2 plus (Zeiss, Oberkochen, Germany) and a digital camera (Axiovision Cam, Zeiss), and the data were processed with a KS300 computed image analysis system (Zeiss), as previously described.³¹  Both the analyses were performed on at least 5 different sections/animal and at least 8 different fields/section. Sections were randomly distributed in the central (C) or peripheral (P) portions of the implanted area and were evaluated blindly. HE analysis was carried out at ×50 magnification, and the images obtained were processed by a software, ImageJ 1.44 (National Institutes of Health, Bethesda, USA), to digitally convert scaffold, soft, and newly deposited bone tissues in different colors (Figure 2). On digitally converted images, the relative tissues/scaffold extension was calculated. The quantification of newly deposited bone was performed exclusively by considering the vital bone as the amount of mineralized and vascularized tissues displaying osteocytes within their lacunae.^22,32  The total VA was quantified at ×200 magnification, and the digitized fluorescent vessel signals (vWF positive area) were accomplished using a semi-automated algorithm.³¹  The VA was calculated as the extension expressed in μm² of vWF-positive area/field (15000 μm²).

To compare the effect exerted at the different experimental times, the data were checked for normal distribution by the D'Agostino and Pearson test, and were compared after arctan (x) transformation by two way-ANOVA test. Finally, the posthoc Tukey test (GraphPad Prism 5) was carried out to evaluate the “individual” effect on each examined variable. The data are reported as mean ± Standard Deviation (SD).

The sintered dispense-plotted assemblies had a characteristic mesh-like structure with rod diameters of 300 ± 30 μm, and pore sizes between the rods of about 370 ± 25 μm. By measuring the geometrical density of the sintered scaffolds, a total porosity of about 60 % was calculated. Relative density of the sintered samples determined by helium pycnometry was about 99% th.d., which indicates only a small amount of closed porosity inside the material rods. Two main material phases of the sintered ceramic were detected by semi-quantitative XRD measurements: 30% HA, 60% β-TCP. Also a slight peak of α-TCP (10%) was identified. SEM micrographs of the specimens revealed the anisotropic inner structure of the ceramic scaffolds as result of the layer-by-layer process of dispense-plotting (Figure 3a). The surface of the material rods showed a micro-roughness (Figure 3b and c), which resulted from an open interconnecting microporosity with pores sizes of about 1 μm (Figure 3b and d).

All the animals completed the healing period without any postoperative complications and no clinical symptoms of maxillary sinusitis. All animals displayed the extra-oral surgical windows closed by fibrous tissue. The explants always appeared as uniform blocks of tissue with the scaffolds firmly inserted into the maxillary sinuses.

The 2D micro-CT revealed a complete integration of the scaffolds within the cavities surgically created between the maxillary sinus floors and the Schneiderian membranes both at 45 and 90 days after graft. The micro-CT suggested that the bone integration of the scaffolds was obtained in the points of contact between the biomaterial and the primitive bone (Figures 4 and 5). The stabilization of the scaffolds in the implanted area was then guaranteed by an abundant deposition of non-mineralized connective tissue recorded within all the biomaterial spaces.

The grafted area showed a small inflammatory infiltrate at 45 days: this infiltrate further decreased, and completely disappeared at 90 days (data not shown). As shown in Figure 6, histology confirmed a complete integration of the scaffolds in all the animals, and the persistence for 90 days of the grafted biomaterial. Forty-five days after intervention, the grafted sites showed foci of bone remodeling mainly localized in the periphery of the areas (Figure 6c, dark arrows). In explants, obtained 90 days after sinus augmentation, more mature lamellar bone creating continuous bridges between the peripheral spaces of the scaffolds and the preexistent host bone was recorded (Figure 6e, white arrow). Histological analysis, performed at higher magnifications, revealed that the osteogenic process was mainly stimulated by an endosteal apposition of osteoblastic cells on newly deposited bone next to the scaffolds surfaces (Figure 6c and f). In both groups (45 and 90 days), the majority of scaffold cavities distributed in the central portion of the implanted area were filled by connective fibrous tissue, displaying several fibroblast-like cells (Figure 6a and d) and a widespread blood vessel network. Different typologies of blood vessels were recorded in the scaffold cavities, ranging from capillaries to medium-sized vessels, with the presence or absence of mural peripheral cells (immature and mature vessels, respectively) as indicated by the double immunolabeling with the endothelial (vWF) and mural (α-SMA) markers (Figure 7a).

The average extension of the newly deposited bone tissue slightly increased over the time considered, as showed in the Table. However, the new bone formation was not uniformly present within the grafted area. In fact, the points of significant greater deposition were always recorded at the periphery of the cavities (Figure 8). In addition, the process of bone formation remained active during the 90 days after sinus augmentation, increasing significantly either in the central or in the peripheral portions of grafted areas (Figure 8). On the contrary, the scaffold surfaces significantly decreased, showing an inverse behavior with respect to the fibrous tissue filling the scaffold interspaces (Table). The average analyses of the total VA indicated a significant increase in the extension of blood vessels at day 90 (Table). Moreover, topographic analysis of VA indicated that blood vessel remodeling was more active in the peripheral fields, and that the greatest increase in VA was recorded in the peripheral fields of the grafted areas 90 days after sinus augmentation with the calcium phosphate scaffolds (Figure 7b).

Scaffolds for bone regeneration are generally required to have biological, biomechanical, and biomaterial functions, including the following: (1) appropriate porosity for the diffusion of nutrients and the invasion of vascularity from a surrounding tissues, (2) appropriate material surface chemistry with biocompatibility to allow cells to adhere and express their normal phenotypes, (3) sufficient mechanical properties as a load-bearing construct during the regeneration process, and (4) adequate biodegradability after sufficient new bone formation. Recent development in design and fabrication technologies has made it possible to design a 3D scaffold with a controlled architecture by computational modelling and simulation.^20,33–36 

The design of the scaffold microstructure affects the regeneration process, as well as initial biological and biomechanical properties, that is, the temporal change in the function of the bone/scaffold system is also important for the successful bone regeneration.³⁷  During the regeneration process, the degradation of the scaffold and formation of the new bone occur in the same time frame as confirmed by the present results. Calcium phosphate ceramic scaffolds for bone tissue engineering can be manufactured with variable chemical and geometrical characteristics.³⁸  The pore sizes are well known to be crucial for tissue ingrowth and vascularization, as high porosity and large pores may enhance bone ingrowth and integration of the biomaterial.^6,39,40  The minimum recommended pore size for a scaffold is 100 μm based on the early work of Tsuruga,⁴¹  but subsequent studies have shown a better osteogenic response in scaffolds with pores of 300 μm.^39,42  Relatively larger pores favor direct osteogenesis since they allow vascularization and high oxygenation, while smaller pores result in osteochondral ossification; however, the type of bone ingrowth depends on the biomaterial and on the geometry of the pores. The present in vivo results clearly confirmed this hypothesis.

The custom-made scaffold with rod diameters of 300 ± 30 μm and pore sizes between the rods of about 370 ± 25 μm displayed a prompt and complete tissue integration already detectable after 45 days from the sinus augmentation procedure. In addition, the in vivo approach allowed demonstration of the good immune-tolerability of the bone substitute that did not induce any signs of inflammatory reaction during the first 90 days of healing. In fact, the 3D scaffold was well tolerated by the host tissue that underwent a progressive process of tissue regeneration. The scaffold was, in fact, completely colonized by a highly vascularized fibrous tissue enriched by several foci of newly deposited bone, mostly located at the periphery of the grafted area.

Histomorphometrical evaluation directly confirmed an adequate trophic support of the newly generated tissues exerted by a widespread blood vessel network. Scaffold placement was accompanied by an active process of angiogenesis, as demonstrated by the increasing total VA occurring during the first 90 days. In addition, the analysis showed the presence of both capillaries and large vessels in the grafted area. On the basis of vWF and α-SMA co-expression, several of these newly organized blood vessels reached a high degree of structural organization, as indicated by the recruitment of mural cells.⁴³  Few mural cells (pericytes) were recorded in capillaries between the endothelial cells (vWF positive), while they were frequently localized at the periphery of large new blood vessels where they differentiated into vascular smooth muscle cells. The presence of smooth muscle in the vasculature of grafted compartment suggests that the nervous system or local mediators may contribute to the control of regenerated tissue by modulating wall blood flow. The persistence of an active process of vascular remodeling localized at the periphery of the grafted area could be interpreted as a favorable condition to sustain the active process of new extracellular matrix/bone deposition evident at the edge of the graft, thus confirming a high osteoconductivity of the scaffold itself.

Moreover, the rate of scaffold degradation should be taken into account. Scaffolds with a high degradation rate should not have high porosities since rapid resorption of the biomaterial will compromise the mechanical and structural integrity before their substitution by the newly formed bone. In contrast, scaffolds with low degradation rates and strong mechanical properties can be highly porous, because the higher pore surface area interacting with the host tissue can accelerate degradation due to macrophages via oxidation and/or hydrolysis. The great surface of interaction between scaffolds and host tissues was essential to stimulate a prompt and persistent process of osteogenesis that, as demonstrated by the morphological analysis, was obtained by an endosteal apposition of osteoblastic cells on newly deposited bone matrix next to the scaffold surface or bone tissues. For this reason, as first demonstrated by the micro-CT and then confirmed by the histological analysis, bone deposition involved mainly the periphery of the grafted area, thus creating bridges of bone integration between scaffold and the native bone after 90 days from the sinus augmentation procedure. Newly formed bone deposition was accompanied by a progressive degradation of scaffold that became significantly relevant after 90 days. Unfortunately, the amount of resorbed scaffold had not been completely replaced by mineralized tissues. In contrast, in some areas of scaffold degradation it was possible to find fibrous tissue.

There is, however, an upper limit in porosity and pore size set by constraints associated with mechanical properties. An increase in the void volume results in a reduction in mechanical strength of the scaffold, which can be critical for regeneration in load-bearing bones. The extent to which pore size can be increased while maintaining mechanical requirements is dependent on many factors including the nature of the biomaterial and the processing conditions.

These considerations indicate that the presented direct rapid prototyping technique is well suited for the production of calcium phosphate bone substitutes for bone tissue engineering, as it allows control of porosity, pore design, and external shape of an implant of any desired HA/TCP phase ratio. The analyzed scaffold showed good osteoconductive properties required for clinical purposes. Since bone regeneration proceeds within the scaffold from the periphery near to the points of contact between the biomaterial and native bone, the complete reconstruction of the alveolar ridge may need a long period of time. However, the process could be accelerated by creating different points of bone nucleation through engineering it with stem/osteoprogenitor cells. The preclinical trials proposed in this study made it possible to document the good osteoconductive properties and the high biocompatibility of this innovative scaffold, certifying the high translational value of its therapeutic efficacy and safety in oral and maxillofacial reconstructive medicine.

Abbreviations

BCP

biphasic calcium phosphate

HA

hydroxyapatite

HE

hematoxylin and eosin

micro-CT

micro computerized tomography

PBS

phosphate buffered saline

TCP

tricalcium phosphate

VA

vascular area

vWF

von Willebrand Factor

α-SMA

alpha-Smooth Muscle Actin

This work was supported by Tercas Foundation and, partially, by the Ministry of Education, University and Research (M.I.U.R.), Rome, Italy. The authors gratefully acknowledge the help of Sabine Hamisch, UweHamhaber, BioCerEntwicklungs-GmbH, Ludwig-Thoma-Str. 36c, 95447 Bayreuth, Germany, in providing the materials and of Maura Turriani, Delia Nardinocchi, and Oriana Di Giacinto for their precious technical support in cell cultures and histological analysis.