Bone formation beyond the skeletal envelope using calcium phosphate granules packed into a collagen pouch—a pilot study

In this proof-of-concept, bone neoformation beyond the skeletal envelope is explored by using a collagen pouch (n = 6) packed with calcium phosphate (CaP) granules placed over the frontal bone in sheep (n = 3). At 13 weeks, macroscopic examination showed specimens covered by an adherent fibrinous envelope with slight vascularization. Histology revealed colonization of the implant by newly formed woven bone and fibrous connective tissue. Surface osteoblasts as well as material-filled macrophages, lymphocytes, polymorphonuclear cells and giant cells were also found in large quantities surrounding the newly formed bone tissue inside the collagen pouch. On the side facing the recipient bone, the collagen membrane had to a large extent been resorbed and bridging bone formation was clearly visible between the test article and recipient bone. On the other side facing soft tissue, the collagen pouch remained intact with a visible fibrous capsule. This study demonstrated that the use of a collagen sleeve as a container for CaP granules allows for good neoformation beyond the skeletal envelope with bridging bone formation clearly visible between the test article and recipient bone. Additionally, in this model, macrophages rather than osteoclasts appear to modulate CaP granule resorption and remodeling into new bone. This construct opens new perspectives for treatment methods that could be used for bone augmentation and restoration of cranio-maxillofacial defects and malformations.


Introduction
Bone reconstruction has always been a challenge.It is indeed a complex process, involving well-defined biochemical mechanisms which are essential for the mobilization of cell populations and crosstalk between them.The intrinsic capacity of bone tissue to regenerate can be challenging in complex clinical cases due to excessive damage caused by for example trauma, tumors, infection, surgery or due to various skeletal diseases or disorders.An improved understanding of bone tissue and the cellular and biochemical mechanisms involved in bone healing has allowed the emergence of numerous bone grafting strategies.These strategies involve, but are not limited to, supplementing the bone healing capacity with bone tissue taken from the patient (autogenous graft) or another patient (allograft), with the use of synthetic bone grafts as well as introducing various growth factors.
Autogenous grafts remain the gold standard for bone grafting as it naturally contains the right amount of growth factors and carries no risks for rejection or transferal of disease as allografts.However, it still carries an increased risk of infection and donor site morbidity [1][2][3].Furthermore, the patient's overall health may prohibit its use, or the autograft may be resorbed too quickly.As a results, a variety of synthetic bone graft substitutes have been developed out of which ceramic-based grafts are the most used today [1,4].
The first ceramic to be used historically is calcium phosphate (CaP).It was first used in 1920 by F H Albee and H Morrison [5] for the treatment of bone fractures in rabbits.Since then, CaP-based cements have developed and are used in a wide range of medical application given its biocompatibility and bioactivity as well as its osteoconductive and osteoinductive properties [6][7][8][9][10].However, its general poor mechanical properties, often unpredictable resorption rate and the biological properties of the new formed bone tissue can limit the use of CaP cements [11][12][13][14].
In bone reconstruction procedures, particularly in oral and maxillofacial dental surgery, bone implants are commonly used in combination with barrier membranes.This practice named guided bone regeneration (GBR) was historically introduced in the late 1990s with the work of Professor Christer Dahlin et al [15], who demonstrated that the addition of a membrane had the effect of maintaining and stabilizing the blood clot [15] which is essential for the incorporation of the implant [16][17][18].
Resorbable or not, these barrier membranes are nowadays used in almost four out of ten dental implant surgeries [19].They mainly play the role of a physical junction that prevents invasion of the bone defect by soft tissue and maintains space for cell migration [15].However, their active role in the bone regeneration process has been highlighted several times in the emerging literature [20].
Among the resorbable membranes, collagen is appreciated for its high biocompatibility, its promotion of wound healing, its functional barrier properties, and its high morphological compliance [21].
The combination of a collagen membrane with CaP granules/construct results in an implant that mimics the natural biochemical composition of the bone tissue.Usually, the collagen membrane is placed between the edges of the defect and the biomaterial [22], or covering it [22][23][24].However, none, to our knowledge, have specifically outlined the use of a collagen pouch filled with CaP granules for the use of either augmenting or restoring craniomaxillofacial defects, malformities or deformities.
In this feasibility study we evaluate then the local tissues effects as well as the performance of CaP granules packed into an acellular collagen pouch without the addition of exogenous growth factors and implanted over the frontal bone of sheep for 13 weeks.We hypothesize that this system would allow for de novo formation of bone tissue limited to the inside space of the pouch, without the need for the direct apposition of the CaP with the bone.If successful, this would be a new way to combine CaP and collagen membranes for guided bone formation.

Study design
The study group consisted of three healthy, skeletally mature female sheep of the breed Blanche du Massif Central with an average age of 3.5 years and an average live weight of 59 kg.Two test articles were surgically implanted over the frontal bone of each animal.A non-implanted article-referred to as the T0 articlewas retained for histological structural characterization as material control.No control group was included as the endpoint was to evaluate the performance of the article, i.e. whether it allows the formation of bone tissue.
After 13 weeks of implantation, the animals were humanely euthanized, the implantation sites carefully removed and submitted for macroscopic, and histopathological analyses.

Materials
The test-article was composed of CaP-granules contained in an absorbable collagen membrane.The CaP ceramic granules (OssDsign AB, Sweden) was prepared from a powder mixture of β-tricalcium phosphate (β-TCP), dicalcium pyrophosphate (β-CPP), and monocalcium phosphate monohydrate (MCPM) as described in previous work [25][26][27] and mixed with glycerol to allow for screw extrusion and spheronization using a Caleva bench-top screen Extruder and Spheronizer 20.The ceramic was extruded until reaching 1 mm size and then transferred into the spheronizer.The resulting granules were separated by size using 0.6-1.0mm sieves and placed in a petri dish to dry at 37 • for 24 h, and ultimately left in sterile water for 48 h to reduce the glycerol content.After 48 h soaking, the granules were once again dried at room temperature under a hood for an additional 24 h before sterilization at 121 • C for 20 min.
The pouch containing the CaP granules was created using a porcine collagen membrane (BioGide®, Geistlich Biomaterials; non-cross-linked porcine types I and III collagen, bilayered).The membrane is supplied sterilized (Gamma Radiation), in individual packages.

Surgery
Sheep were weighed and fasted 12 h before the surgical procedure.These values were used as a reference for comparison at the end of the experiment to ensure that the animals' general condition was maintained.
In an adjacent room to the operating room the sheep were placed under general anesthesia in a prone position.Anesthetic premedication was performed with an intramuscular injection of Diazepam (Diazepam®, TMV, 0.3 mg kg −1 ) and Butorphanol (Torphasol®, Axience, 0.2 mg kg −1 ).Anesthesia was then induced with an intravenous injection of Propofol (Propovet®, Axience, 2-5 mg kg −1 ).An orotracheal tube was inserted fort sustain the anesthesia by inhalation of Isoflurane (IsoFlo, Axience; MAC 1.5%).Each sheep was mechanically ventilated and put on an intravenous infusion with suitable electrolyte solution during the operation.In addition, perioperative analgesia was performed with an intramuscular injection of Flunixine (Meflosyl®, Zoetis, 2 mg kg −1 ).Double antibiotic therapy with Ampicillin (Duphamox LA®, Zoetis, 15 mg kg −1 ) and Enrofloxacin (Baytril 10%®, Bayer, 5 mg kg −1 ) was administered subcutaneously.Anesthesia was monitored with electrocardiogram, peripheral noninvasive arterial blood pressure, oxygen saturation and measurement of rectal temperature.Once anesthetized the craniofacial site was then clipped and scrubbed with povidone-iodine.The site was wiped with 70% isopropyl alcohol prior to painting the surgical site with povidone-iodine solution.
The test article was prepared in the sterile field of the operating theatre.Just prior to implantation, the surgeon created a bag from the porcine collagen membrane (approximately 30 × 40 mm in size) by folding it widthwise and partially sealed it with a single continuous suture using non-absorbable suture (Dafilon® 5-0, B. Braun).Before being completely sealed, the pouch was filled with approximately 2 ml of CaP granules (figures 1(B) and (C)).Subsequently, the test article was soaked in fresh venous blood previously collected from the lateral saphenous vein, until saturation (figure 1(D)).
The surgical procedure was conducted in a surgical theatre under aseptic conditions.A U-shaped skin incision was made to expose the frontal region up to the occipital protuberance (see figure 1(A)).The temporalis muscles with attachments to the frontal bone were elevated and retracted.The frontal bone was then prepared and exposed leaving the periosteum intact.Thereafter, each test articles were placed over the frontal bone (onlay) and fixated with four titanium microscrews (Medicon®), as depicted in figure 1(E).Soft tissue closure was performed with resorbable sutures (Vicryl™ 2-0, Ethicon®), and skin closure was completed with non-resorbable sutures (Prolene 2-0, Ethicon®) and surgical staples (Appose™ ULC Auto Suture™, Covidien™).

Post-operative procedures
Following anesthesia, animals were transferred to recovery areas and monitored until sternal recumbency was achieved.Post-operative analgesia was provided by injections of Buprenorphine (Buprecare®, Axience, 0.015 mg kg −1 ) at the end of the surgery day and then twice a day for two days, as well as by daily injections of anti-inflammatory medication (Flunixin, Meflosyl®, Zoetis, 2 mg kg −1 ) for seven days.Antibiotics were administered daily for three weeks after implantation of the test article (Amoxicillin, Duphamox LA®, Zoetis, 15 mg kg −1 and Enrofloxacin, Baytril 10%®, Bayer, 5 mg kg −1 ).About two weeks after surgery and complete healing, the surgical sutures and staples were removed.The wounds were then disinfected with an Oxytetracycline solution (Oxytetrin® Spray, MSD).
The general condition of the animals as well as the implantation sites were examined daily until staple removal to detect mortality, morbidity and to prevent any adverse reactions.In addition, neurological/neurobehavioral examinations were performed daily during the first week, weekly during the next three weeks, and then monthly, with particular attention to motion, general behavior or any other relevant abnormal signs like tremor, convulsions, circling, head tilt, abnormal salivation, and lacrimation.In the case of any clinical signs of concern, the animal was immediately examined and treated as needed.
After 13 weeks the sheep were weighed and then humanely euthanized by an intravenous injection of Sodium Pentobarbital (Doléthal®, Vetoquinol, min.182.2 mg kg −1 ).A macroscopic examination was then carried out including examination of the appearance of the skin at the implant site, the skull, and the adjacent tissues.The regions of the implantation were then carefully exposed to allow for an initial macroscopic examination of the implants and the surrounding tissue.The chosen qualitative criteria are indicators of tissue reaction: tissue vascularization, infection markers (rugor, calor, tumor, functio laesa), new tissue formation and if so, their distribution.The parameters considered in this first examination are listed in table 1.
The two implant sites per sheep were removed 'en bloc' (i.e. two implants + underlying bone) and fixed in a 10% neutral buffered formalin (NBF) solution.An identical NBF solution was used to fixate one nonimplanted reference test article (T0).

Histology
Following complete fixation in 10% NBF, a total of six implanted (n = 6) and one non-implanted reference sample were decalcified utilizing a 10% aqueous formic acid solution, dehydrated through successive baths in alcoholic solutions of increasing concentration (50%-99.9%),cleared with xylene and finally embedded in paraffin.Two central longitudinal sections were cut with a microtome (∼4.5 µm thickness) and stained using a safranin-hematoxylineosin stain for qualitative and semi-quantitative analysis, respectively.In order to score the local tissue effects and the inflammatory response at the implantation sites, qualitative and semi-quantitative histopathologic evaluation was performed by a certified pathologist, within ordinal system ranging from minimal, slight, moderate, marked up to severe.Non-implanted samples served as references for the structural characterization of the CaP and collagen pouch.

Statistical evaluation
All values are summarized in tables/figures listing the mean, standard deviation and/or median, minimum, and maximum for continuous data or in tables listing count and percentage for categorical data, where appropriate.

Material characterization
The phase composition of the CaP granules, as determined by XRD, was a mixture of monetite (85%), β-TCP (8%) and β-CPP (7%).No other phases were identified over the limit of detection in the diffraction patterns (figure 2).

Follow-up and clinical observations
All animals showed a good general condition during the study.Weigh-ins did not reveal any significant weight loss compared to before the surgery.On the contrary, the animals tended to gain some The results of the clinical observations are summarized in table 2. Superficial wounds were noted in the groin and legs, requiring regular local treatments (Vétédine® Solution, Vétoquinol).As the animal's general condition were not affected, this had no impact on the study.

Macroscopic observations
No abnormalities were observed before implantation in any sheep.All the implants were successfully implanted over the frontal bone without any difficulty.
The test articles showed a color variation after being soaked in venous blood, where 50% (3/6) of the test articles were bright red, while the remaining 50% (3/6) had a pale pink color.
Macroscopic examination of the skin and adjacent tissues at the implantation site after three weeks implantation showed no signs of abnormalities in two out of the three (2/3) animals.No migration of the test articles compared to the implantation day could be observed.After dissection of the implanted sites in the two normal animals, the implants were visible, intact (no signs of membrane rupture or granules leakage) and covered by an adherent envelope with slight vascularization (figure 3).Furthermore, the implants felt

Histological observations
The test article consists of a combination of two materials, a collagen membrane and CaP granules.The collagen membrane is composed of an amorphous material and an eosinophilic fibrillar component (Geistlich; Bio-Gide absorbable membrane) (figures 4(A) and (B)).In the reference article, the CaP granules were identifiable as grey amorphous material in the center along the membrane.
At 13 weeks, the five implanted articles still had similar architecture to the non-implanted article (T0).In all sites, the implant was colonized by abundant heterogeneous tissue.Newly formed vascularized bone and fibrous connective tissue filled the cavity of the pouch (figures 4(C)-(E)).Small multifocal aggregates of CaP granules were still identifiable within the implant.
The side of the collagen pouch facing soft tissue was still discernible and covered by a fibrous capsule (see figure 4(C)) colonized by fibroblasts.Mild and variable signs of resorption of the collagen membrane were observable.However, the membrane in contact with the bone recipient was almost completely resorbed, with the formation of a bridging bone tissue (figure 3(D)).
The implanted sites showed biodegradation reflected histologically by a moderate level of cellular activity (figure 5).A large number of material-filled macrophages were observed around the CaP granules.In addition to the material-filled macrophages, the most represented cell populations were giant cells, polymorphonuclear cells, and lymphocytes and were found in the soft connective tissue that filled the pocket.A significant amount of osteoblast was found on the surface of the newly formed bone tissue.
No signs of inflammatory reaction were noted, except for the excluded site.Histological examination revealed the presence of neutrophilic granulocytes and macrophages.For the excluded test article, the tissue response and the bone remodeling were not observed (figure 5).

Discussion
This study investigated the performance of CaP granules packed a collagen pouch placed and fixed over the frontal bone in sheep.At 13 weeks, the CaP component of the construct had to a large extent been resorbed and replaced with newly formed bone and connective tissue.Moreover, the degradation process of the collagen membrane seems related to the tissue it faces.On the side facing soft tissue, the collagen pouch still remained intact with visible fibrous encapsulation.The presence of macrophages, polymorphonuclear cells, and fibroblasts near the surface of the membrane apposed to the soft tissues is compatible with a process of membrane degradation by phagocytosis.This mode of degradation is consistent with the biodegradation process of collagen in vivo, but also of implanted exogenous collagen currencies [35][36][37].On the side facing native bone, the collagen membrane had to a large extent resorbed with bridging bone formation clearly visible between the test article and native bone.
Kozlovsky et al noticed only moderate biodegradation of an identical membrane with central intramembrane neo-ossification on the side of the collagen membrane facing a rat calvarial bony defect at an earlier implantation timepoint (nine weeks) [38].The observed ossification appeared to be independent of the newly formed bone tissue in the recipient bone [38].Additional immuno-fluorescence studies at various timepoints with and without CaP granules would be needed to further investigate this biodegradation process and to identify the role of membrane ossification and vascularization on the bridging bone formation observed in the current study.
No bone formation was observed in either the surrounding soft tissue or on adjacent bone.This indicates the potential of the construct to form bone beyond the skeletal envelope that is incorporated into the recipient's bony structure over time.
The newly formed bone tissue was identifiable inside the pouch by its 'woven' arrangement with many trabeculae.It also showed cell populations involved in bone formation and bone remodeling.Indeed, numerous surface osteoblasts, according to the definition given by Shapiro et al [39], are first identifiable in the immature bone tissue by their cuboidal shape [40].These were mainly localized at the edge of the trabeculae.The presence of osteoblasts is consistent with the results of Franchi's research [41], in which these cells appeared in the peri-implant area of the femur of sheep after two and four weeks implantation with titanium-and zirconium screws, respectively.
The association of macrophages with osteoblasts reflects active tissue remodeling within the test article during which crosstalk between these two cell populations is crucial.It has indeed been known since the publication of Chang et al's [42] work that both pro-inflammatory (M1) and anti-inflammatory (M2) macrophages are directly involved in the bone remodeling process.However, the emerging literature specifies their involvement in the initiation of bone regeneration [43,44], but also their influence on the anabolic function [42,45], the modulation, and differentiation of osteoblasts [46][47][48].Depending on their polarization [49] (M1 or M2), macrophages are in fact involved in each phase of the process, i.e.: (1) by cleaning up damaged tissue at the peri-implant area immediately after clot formation [16,50] (2) by initiating tissue reorganization through large enzyme production (3) by inducing [28] bone formation through the secretion of growth hormones such as transforming growth factors, fibroblast growth factors, epithelial growth factors, or also bone morphogenic proteins [48], all involved in the recruitment and differentiation of osteogenic cell lines [42,48,51,52] (4) by supporting tissue neovascularization via the secretion of angiogenic molecules [49,50,53,54] and finally (5) by playing a prominent role in the biodegradation of materials [55].Furthermore, given the large number of CaP-filled macrophages and very low numbers of osteoclasts observed in this current study, it seems that CaP resorption is mainly macrophage-driven.This is consistent with the findings of recent studies which was reported the presence of CaP particles in macrophages after 12, 13, and 52 weeks following implantation in calvarial sheep models of similar CaP composition [28,56,57].
Based on prior studies and the ISO 10993-6 standard, a period of 13 weeks was considered adequate as an initial observation point in this pilot study to evaluate the test article ability to induce bone formation and how it affects surrounding tissue.The newly formed bone tissue in the pouch did however not show lamellar bone deposition after this period, in contrast to various similar works after a similar or even shorter implantation time [41,58].We speculate that this difference in osteogenesis is explained by (1) the different mechanisms observed during membranous (flat bones) and endochondral (long bones) ossification, especially during bone healing [59,60], (2) the lack of decortication or bone damage [61], (3) the different biochemical and biomechanical nature of the implants and finally (4) different material characteristics such as surface topography, wettability and porosities [62][63][64][65][66].
The difference in maturation between the present study and other recently published research papers could also be explained by the difference in vascularization between the long bones, and the frontal bone surface.A proper vascularization of the implantation site is indeed crucial for osteogenesis and implant survival [67].On the other hand, too limited blood supply to the implantation site or a lack of vascularization is described as one of the main factors in implant failures [68].The angiogenic-osteogenic coupling ensures the supply of oxygen and nutrients, precursor cells, but also platelets and immune cells, secreting growth factors and cytokines [17,69,70].The implantation of the test articles in our study was performed on a bone surface with limited vascularization, and without creating tissue damage, and therefore without the direct creation of an adequate hematoma.This hematoma is however essential to initiate the creation of a vascular network around and specific to the implant [7,16,46,50].Many studies have indeed shown the role of the blood clot in the early healing processes [71,72].It acts as natural scaffold having a concentration of growth factors and blood cells which recruits other cells essential in angiogenesis and bone healing [17].The blood treatment of the test articles prior to implantation was therefore intended to recreate the optimal conditions for a clot building to initiate the implant incorporation and remodeling into new bone.
The collagen membrane is the most used resorbable membrane for GBR because of its high biocompatibility, its promotion of wound healing, its functional barrier properties, and its high morphological compliance [21].Usually, this membrane is placed between the edges of the exposed bone and the biomaterial [22], or placed overlaying it [22][23][24].In this study, the CaP granules are contained in a closed space formed by the collagen membrane, thus optimizing the benefits of this barrier membrane to the system while avoiding migration of the granules into the surrounding soft tissue.This physical barrier helps to form a completely delimited local environment in which the progress of the implant incorporation into the host bone is favored and stabilized.
In the present study, the incorporation of the test article results in the biodegradation of the collagen membrane in contact with the recipient bone which allows for bridging formation between the newly formed bone and the native bone.The de novo formation of bone tissue observed in the present study appears otherwise to have been limited to the interior of the pouch.No ossification or CaP granules were observed in the soft tissue outside the pouch.
The proper management of the bone graft also depends on the fixation of the membrane.The implant designed for our study combines the CaP and the barrier membrane in a single unit, which allows for simultaneous handling of both components and easier fixation.This is relevant as obtaining adequate fixation tends to be a challenge for surgeons, due to the small size of the surgical sites in oral and maxillofacial surgeries.Additionally, the work of An [73] and Wessing [74] underlines the positive influence of the fixation membrane in GBR on bone regeneration by increasing the expression of osteogenic factors and leading to a much higher formation of new bone [73,74].Moreover, a too mobile combination of bone graft and membrane tends to decrease the rate of bone regeneration and seems to increase the risk of graft failure [75].Given that the titanium screws used for fixation in this study is the only non-resorbable component a resorbable fixation system should be introduced to allow for a completely resorbable implant.
A limitation of this study is the absence of control to compare the formation and characteristics of the newly formed tissue.However, this was a feasibility study to determine whether the collagen membrane and CaP combination constructed in a pouch were able to induce and drive the formation of new bone tissue beyond the skeletal envelope.Future comparative studies should be carried out to objectify the respective involvement of the collagen and of the CaPpreparation used and their importance in the bone regeneration process.Repeating this study with different compositions of CaP could also help to objectify the biological effect of the components, particularly on cell recruitment and the biomineralization process.Moreover the combination with other cellular and molecular techniques, such as immunohistochemistry, would enable the mechanisms of action to be identified with greater precision, particularly macrophage phenotype.
This could lead to different opportunities for use in maxillofacial surgery (for example, during alveolar ridge preservation, where this device could allow easier handling and fixation) or in orthopedic surgery, in the context of large bone defects.

Conclusion
In this proof-of-concept study, we demonstrate that the use of a collagen pouch as a container for CaP granules allows for neoformation beyond the skeletal envelope after 13 weeks with bridging bone formation clearly visible between the test article and native bone without the addition of growth factors.This new approach in GBR may allow for new treatment options beyond cranio-maxillofacial defects and should be studied further in both orthopedic and dental applications.

Figure 1 .
Figure 1.(A) Schematic overview of the surgical approach and implantation site location (top view).(B) Calcium phosphate (CaP) granules.(C) Native test article with CaP granules contained in a collagen membrane.(D) Test article after blood impregnation.(E) Implants positioned on the frontal bone at the time of implantation.

Figure 2 .Table 2 .
Figure 2. XRD-pattern of the CaP granules, including phase identification for each peak.

Figure 3 .
Figure 3. Photograph of an implant site exposed 13 weeks after implantation.Intact pouches (outlined in dotted lines) covered by an adherent envelope with slight vascularization.

Figure 4 .
Figure 4. Histological assessment of the non-implanted article and after 13 weeks of implantation in cranial ovine model-representative microphotographs (SHE staining).(A) and (B) Gross picture of the non-implanted article (T0) consisting of collagen a collagen membrane (CM) forming the pouch of CaP granules.(C) Overview of the test article on the recipient bone (RB) with the collagen pouch filled with CaP granules.(D) Biodegradation of the CM with bridging newly formed bone (NB); BV; blood vessel; (E) magnification on the different newly formed tissue and of cell population: surface osteoblasts (Ob) surround the newly formed tissue in which CaP is degraded by macrophages; CT: connective tissue, FC: fibrous capsule, P: membrane collagen pouch (i.e.test article).

Figure 5 .
Figure 5. Histopathological semiquantitative evaluation after 13 weeks implantation in a cranial ovine model.•: median; error bars represent the rang of minimum and maximum values; ∆: values of the infected site.

Table 1 .
Criteria for macroscopic examination of the implant region after termination.