Effect of collagen sponge and fibrin glue on bone repair

ABSTRACT The ability of hemostatic agents to promote bone repair has been investigated using in vitro and in vivo models but, up to now, the results are inconclusive. Objective In this context, the aim of this study was to compare the potential of bone repair of collagen sponge with fibrin glue in a rat calvarial defect model. Material and Methods Defects of 5 mm in diameter were created in rat calvariae and treated with either collagen sponge or fibrin glue; untreated defects were used as control. At 4 and 8 weeks, histological analysis and micro-CT-based histomorphometry were carried out and data were compared by two-way ANOVA followed by Student-Newman-Keuls test when appropriated (p≤0.05). Results Three-dimensional reconstructions showed increased bone formation in defects treated with either collagen sponge or fibrin glue compared with untreated defects, which was confirmed by the histological analysis. Morphometric parameters indicated the progression of bone formation from 4 to 8 weeks. Additionally, fibrin glue displayed slightly higher bone formation rate when compared with collagen sponge. Conclusion Our results have shown the benefits of using collagen sponge and fibrin glue to promote new bone formation in rat calvarial bone defects, the latter being discreetly more advantageous.


INTRODUCTION
Bone, in contrast with other tissues, can repair itself without scar formation. However, pathological fractures, great bone defects, insuf cient blood supply, bone or surrounding tissues infections, and systemic diseases may negatively in uence bone healing, resulting in delayed unions or nonunions 7,26 . The common therapeutic approach to treat these clinical situations is the use of bone autografts 5,8 .
Autograft bone is considered the "gold standard" due to the osteoconduction, osteoinduction and osteogenesis promotion as well as angiogenesis without the risk of disease transmission 21 . In oral and maxillofacial surgery, the areas from where autogenous bone can be harvested include extraoral sites as iliac crest, cranial vault and tibia plateau, and intraoral sites as mandibular symphysis, maxillary tuberosity, ramus, tori and exostoses 17,20 . However, the use of bone from these areas has some disadvantages, such as limited availability and increased morbidity associated with a second surgical procedure 6 .
Hemostatic agents have been used to control hemorrhage during surgical procedures for harvesting bone grafts from donor sites. Collagen sponge and brin glue are employed as hemostatic agents and as scaffolds for bone and cartilage tissue engineering 2,16,19,23,29 . Additionally, collagen sponge acts as wound burn dressing material and as intravaginal contraceptive diaphragm 2,14 .
In addition to hemostatic properties, it would be interesting if these materials could favor bone repair 4 . Indeed, the addition of a brin network to collagen sponge increased the osteoblast differentiation in a dose-dependent way, suggesting that these materials may favor bone repair 15 the in uence of  hemostatic agents on bone repair, with positive  and negative results for collagen sponge and brin  glue 4,13,24 . Thus, based on previous promising results, this study was designed to compare collagen sponge and brin glue in terms of bone repair in a rat calvarial defect model.

Co e s o e d r e
Collagen sponges elfoam, er, o aulo, , Bra il , derived from puri ed skin pig, were cut to produce samples with 5 mm in diameter and 2 mm height. This preparation was carried out in a laminar ow cabinet to keep the material sterile. Fibrin glue (Tissucol, Baxter, São Paulo, SP, Brazil), a combination of human brinogen, aprotinin and thrombin, was prepared immediately before using according to the manufacturer's instructions. Solutions of brinogen and aprotinin were heated for 10 min at 37°C, mixed and kept for 1 min at 37°C. In another ask, a solution of thrombin was combined with calcium chloride and kept at 37°C.
ixtures of aprotinin with brinogen and thrombin with calcium chloride were placed in separate syringes connected by a Y-piece that allowed them to blend immediately before implantation in bone defect.

Implantation procedure
All animal procedures were conducted under approval of the Committee of Ethics in Animal Research of the University of São Paulo. Thirty male Wistar rats weighting 250-300 g were anesthetized by a combination of ketamine (7 mg/100 g body weight) (Agener União, Embu-Guaçu, SP, Brazil) and xylazine (0.6 mg/100 g body weight) (Calier, Juatuba, MG, Brazil). The cranium of each rat was shaved and disinfected with povidone-iodine alcoholic solution (Rioquímica, São José do Rio Preto, SP, Brazil) and an incision was made along the sagittal suture to expose the parietal bones. Unilateral calvarial bone defects with 5 mm in diameter were created with a trephine under saline irrigation to remove both external and internal cortical and to preserving the dura mater. The bone defects were randomly treated (ten for each material) with either collagen sponge or brin, or kept empty (control). The skin was sutured with nylon 3.0 (Ethicon Ltd., São Paulo, SP, Brazil). After surgery, the animals received a single dose of antibiotics solution containing benzathine benzylpenicillin (156.000 IU/100 g body weight), benzylpenicillin procaine (78.000 IU/100 g body weight), benzylpenicillin potassium (78.000 UI/ 100 g body weight), dihydrostreptomycin base (65 mg/ 100 g body weight), and streptomycin base (65 mg/100 g body weight) (Fort Dodge, Campinas, SP, Brazil) and unixin meglumine analgesic (10 mg/100 g body weight) (Schering-Plough, Vila Olímpia, SP, Brazil). At the end of 4 and 8 weeks, key time points to determine the progression of bone formation in this animal model, the rats were euthanized and the calvariae harvested and processed for histological and histomorphometric analyses.

Histological analysis
Harvested calvariae were xed in 10 buffered formalin for 36 h and decalcified in buffered EDTA (Merck, Darmstadt, Germany) for 10 days. Decalci ed tissues were then embedded in paraf n and tissue sections with 6 μm thickness -from the central portion of samples -were stained with hematoxylin and eosin. Histological images were acquired under light microscopy (Axioskop 40, Carl Zeiss, Oberkochen, Germany) using a digital camera Axiocam ICc3 (Carl Zeiss) attached to the microscope and processed with the software Axion Vision (Carl Zeiss).

Histomorphometric analysis
Before the decalcification for histological processing, samples xed in 10 buffered formalin for 36 h, were submitted to micro-computed tomography (micro-CT) for morphometric analysis using the SkyScan 1172 system (Bruker-SkyScan, Kontich, Belgium). Images were acquired at 60 kVp and 200 mA and reconstructed using the software NRecon (Bruker-Skyscan) with smoothing 1, ring artifact correction 5 and beam hardening correction 20%. The micro-CT analyses were carried out using the 3D Ctan software (BruKer-Skyscan) to evaluate bone volume, percentage of bone volume, bone surface and trabecular number, thickness and separation.

Statistical analysis
The results were expressed as mean±standard deviation (n=5) and compared by two-way ANOVA (treatment x time point) followed by Student-Newman-Keuls test when appropriated. The level of signi cance was set at p 0.05.

RESULTS
Three-dimensional micro-CT reconstructions showed the lack of bone formation inside control defects at 4 and 8 weeks ( Figure 1A and J) while some bone formation was observed in defects treated with either collagen sponge ( Figure 1B and K) or brin glue ( Figure 1C and L) at the same periods. In histological sections, we observed a connective tissue with no signs of bone formation inside control defects at both time points (Figure 1D G, M and P). A histologically similar new bone tissue with homogeneously distributed osteocytes and partially lined by active osteoblasts was observed in defects treated with either collagen sponge ( Figure  1E, H, N and ) or brin glue ( Figure 1F, I, O and R) at 4 and 8 weeks. Also, a cement line in the interface between this new woven bone and the lamellar one ( Figure 1Q and R) and newly formed blood vessels were noticed in defects treated with collagen sponge or brin glue ( Figure 1H and R). In addition, none of the defects exhibited foreign body reaction, chronic in ammation or infection.
Bone volume (Figure 2A) and its percentage ( Figure 2B) were affected by treatment (p=0.0001), time point (p=0.0001) and interaction treatment x time point (p=0.013). At 4 weeks, bone volume and its percentage had no statistically signi cant difference among treatments (p>0.05). At 8 weeks, bone volume and its percentage were higher (p<0.05) in defects treated with fibrin glue compared with collagen sponge, which were higher (p<0.05) than control defects. Additionally bone volume and its percentage were higher (p<0.05) at 8 weeks compared with 4 weeks. Bone surface ( Figure 2C) was affected by treatment (p=0.0001) and time point (p=0.006), but not by interaction treatment x time point (p=0.295). By taking 4 and 8 weeks together, bone surface was larger (p<0.05) in defects treated with brin glue compared with collagen sponge, which were larger (p<0.05) than control defects. Additionally, bone surface was larger (p<0.05) at 8 weeks compared with 4 weeks. Trabecular number ( Figure 2D) was affected by treatment (p=0.005) and time point (p=0.0001), but not by interaction treatment x time point (p=0.116). By taking 4 and 8 weeks together, trabecular number was higher (p<0.05) in defects treated with brin glue compared with collagen sponge and control defects, which had no statistically signi cant difference (p>0.05). In addition, trabecular number was higher (p<0.05) at 8 weeks compared with 4 weeks. Trabecular thickness ( Figure 2E) was affected by time point (p=0.0001), being larger at 8 weeks compared with 4 weeks, but it was not affected by treatment (p=0.120) and interaction treatment x time point (p=0.894). Trabecular separation ( Figure 2F) was affected by treatment (p=0.0001), but not by time point (p=0.443) and interaction treatment x time point (p=0.118). By taking 4 and 8 weeks together, trabecular separation was larger (p<0.05) in control defects compared with defects treated with brin glue and collagen sponge, which had no statistically signi cant difference (p>0.05).

DISCUSSION
This study was designed to evaluate if treatment with either collagen sponge or brin glue is able to favor bone formation in rat calvarial defects compared with non-treated defects. The newly formed bone observed was histologically similar irrespective of treatment, in contrast with the lack of detectable bone repair inside control defects. In general, morphometric parameters indicate more bone formation in treated defects when compared with the non-treated ones, with a tendency to be larger in defects treated with brin glue.
Histological sections are useful to observe and describe morphological features of tissues; however, they do not allow the reconstruction of three-dimensional images from serial sections, as a tissue segment is removed during the sample preparation 28 . Thus, in this study, the morphological description was based on decalci ed histological sections and the morphometric analysis was performed on three-dimensional images obtained from micro-CT following established parameters.
In addition to the hemostatic effect, collagen sponge and brin glue may affect tissue healing including bone tissue. The use of collagen sponge on perforated cortical model promotes bone augmentation possibly by acting as a scaffold for cells and keeping the space for bone growth 27 . Also, collagen sponge has been successfully used as a BMP-2 carrier to enhance bone formation 3,9,10 . Despite the association with BMP-2 resulted in more bone formation, the effect of collagen sponge by itself could not be neglected 3 . Agreeing with these observations, our results clearly indicate that treatment with collagen sponge elicited better bone response compared with untreated defects. Regarding brin glue, the results are controversial since some studies have reported that brin glue either alone or associated with biomaterials may promote bone formation, whereas others did not show improvement in bone repair 1,13,18 . Our ndings indicated that, as well as collagen sponge, brin glue favors bone formation compared with untreated defects. Despite the previous studies showing the positive effect of both collagen sponge and brin glue on bone repair, the comparison between them regarding potential of bone formation has been poorly explored. In this context, our results showed that brin glue promotes slightly more bone formation compared with collagen sponge. In keeping with our study, brin glue exhibits better osteogenic ability than collagen sponge as a BMP-2 delivery vehicle 30 .
In conclusion, we have shown, by histological and morphometric analyses, the bene ts of using collagen sponge and brin glue to promote new bone formation in rat calvarial bone defects, with slight advantage to brin glue. Thus, these materials could be combined with cells and growth factors as either cell therapy or tissue engineering approaches aiming at enhancing and/or accelerating bone repair.