Porcine Dermis and Pericardium-Based, Non–Cross-Linked Materials Induce Multinucleated Giant Cells After Their In Vivo Implantation: A Physiological Reaction?

The principles of guided tissue regeneration (GTR) and guided bone regeneration (GBR) are based on the separation of different cells and tissues, competing in the healing process of bone or soft tissue defects.^1,2  In this context, membranes for GTR and GBR should serve as reliable barriers by covering the defect site and acting as stabilizers of underlying augmentation material. Recently, collagen-based membranes have found an increased application in GBR and GTR. In implant dentistry, which often requires bone augmentation to form a sufficient implantation bed, non–cross-linked collagen membranes enable successful integration of alloplastic, xenogenic, or allogenic bone substitute materials while preventing the augmented region from the rapid ingrowth of peri-implant soft tissue.^3–5  Besides their use in augmentation procedures, non–cross-linked collagen materials have also been applied to soft tissue regeneration in the oral cavity. These materials are suitable for treatment of gingival recession or enlargement of peri-implant keratinized gingiva.^6–8 

Collagen, as an integral part of human and animal tissue, plays an important role in wound healing and tissue regeneration, as it possesses angiogenic potential and undergoes enzymatic degradation.^9–11  The turnover of collagen in the organism is mediated by matrix metalloproteases, which are released by recruited neutrophils, monocytes/macrophages, eosinophils, and fibroblasts.^12,13  Due to these favorable properties, collagen is ideal for application in GTR and GBR.^1–5,14,15 

In the past years many different collagen-based membranes from different sources have been developed for application in GTR and GBR. These membranes differ in origin (porcine vs bovine), the physiological compartment from which the collagen is derived (dermis, peritoneum, or pericardium), and physicochemical composition and architecture (bilayered vs multilayered structure). Thus, although they are biological membranes, little is known about how they are integrated within the host tissue.

Collagen-based materials available on the European dental market have been analyzed systematically by the authors. Recently, two types of collagen materials were investigated in the same animal model under identical experimental protocols: Bio-Gide (Geistlich Biomaterials, Wolhusen, Switzerland), a bilayered membrane originating from porcine peritoneum, and Mucograft (Geistlich Biomaterials), a bilayered matrix derived from porcine peritoneum and skin.^6,16  The in vivo analysis of both materials by means of the subcutaneous implantation model in CD-1 mice revealed that they do not have to undergo transmembranous vascularization to be integrated within the host tissue. Histomorphometric investigation further revealed that the materials remained stable over the period of the study and showed no premature dissolution or breakdown. No signs of a multinucleated giant cell–triggered cellular reaction were observed. Furthermore, analysis of the tissue response to both materials showed that mononuclear cells were mainly involved in material degradation. As a main conclusion, it was postulated that implanted membranes become integrated in the host tissue and serve as a physiological extracellular matrix.^6,16 

During the past few years other non–cross-linked collagen materials have been introduced that originate from porcine skin or pericardium. To date, there has been no systematic in vivo study investigating the inflammatory patterns involved in material integration within the host tissue. Accordingly, there is insufficient knowledge about the cellular inflammatory response to these materials. Knowledge about the material-tissue interaction would be of great interest to clinicians and material scientists alike, as the material persistence within the host and its contribution to tissue regeneration could be more predictable.

For this reason the presented in vivo study was performed by initially applying the aforementioned standardized systematic approaches to the porcine-derived non–cross-linked collagen-based materials (BEGO Collagen Membrane, BEGO Collagen Fleece). The two materials were implanted subcutaneously in CD-1 mice, and the material-tissue interactions were assessed histologically and histomorphometrically. The study focused on the influence of the different material sources, pericardium vs dermis; their different physicochemical characteristics; and their effects on material integration, vascularization, and the cellular inflammatory pattern within the peri-implant region.

BEGO Collagen Fleece (BEGO Implant Systems, Bremen, Germany) is a bilayered local hemostatic agent made in a standardized, controlled purification process from porcine dermis. Donor material derives from well-selected veterinary controlled pigs and is purified for potential antigenic properties in a multistep washing and deantigenization protocol. The processed collagen is sterilized by gamma irradiation and lyophilized without further cross-linking treatment.

BEGO Collagen Membrane (BEGO Implant Systems) is a stratified membrane made in a standardized, controlled purification process from porcine pericardium. Donor animals are well-selected and veterinary-controlled pigs. During the manufacturing process the collagen components undergo lyophilization and sterilization by ethylene oxide gas. Further cross-linking is avoided.

For detailed analysis of microarchitecture and surface texture, scanning electron microscopy (SEM) was performed at the University of Bremen, Advanced Ceramics Group, Bremen, Germany using a Camscan Series 2 scanning electron microscope (Applied Biosystems, Beaverton, Oreg) with a 20 kV accelerating voltage. Prior to SEM imaging all samples were sputtered with gold (K550, Emitech, Fall River, Mass).

The study was approved by the Committee on the Use of Live Animals in Teaching and Research of the State of Rhineland-Palatinate, Germany. Fifty mice from Charles River Laboratories (Sulzfeld, Germany) were kept at the Laboratory Animal Unit of the Institute of Pathology, Johannes Gutenberg University of Mainz, Mainz, Germany. One week before the start of the experiment, animals were allowed to recover and adapt to the conditions of the laboratory. During the experimental period, animals were fed regularly with mouse pellets (Laboratory Rodent Chow, Altromin, Lage, Germany) and water ad libitum. Day and night rhythm was simulated with artificial 12-hour light/12-hour dark cycles.

According to the study protocol, the 50 mice were randomly divided into three groups. The first group consisted of 20 mice treated with the commercially available collagen fleece (BEGO Collagen Fleece). The 20 mice in the second group received a commercially available collagen membrane (BEGO Collagen Membrane). The remaining 10 mice served as a sham-operated control group, meaning they were operated on without implantation of any biomaterial. The observation period was 60 days with the following 5 explantation time points: 3, 10, 15, 30, and 60 days after implantation (n = 4 animals per time point for experimental groups and n = 2 animals per time point for the sham group).

Biomaterial implantation was performed according to previously described methods.^6,16,17  After intraperitoneal anesthesia with 10 mL of ketamine (50 mg mL⁻¹) in combination with 1.6 mL of 2% xylazine, the mice were shaved and an incision was made under sterile conditions in the rostral portion of the interscapular region. A sample of approximately 10 × 10 mm of each of the investigated biomaterials was implanted in the subscapular region in a previously formed subcutaneous pocket under the skin muscle. Wound closure was achieved with 6.0 prolene sutures (Ethicon, Somerville, NJ).

After each observation period, the mice were killed by means of an overdose of ketamine and xylazine. Histologic processing and staining were performed according to well-established and previously described methods.^6,16  After euthanasia, the implanted biomaterial and the surrounding peri-implant tissue were excised and cut in 3 segments representing the margins and the center of the implantation bed. The extracted samples were fixed in 4% buffered formalin for 24 hours, dehydrated in a series of alcohol concentrations and xylene, and finally embedded in paraffin. The prepared samples were cut in paraffin slices 3 to 4 μm thick and afterward deparaffinized and rehydrated. The slices were stained as follows: the first slide was stained with Mayer's hematoxylin and eosin (H&E), and the following 4 slices were stained as previously described with Azan, Sirius red, Movat's pentachrome. and a histochemical detection of tartrate resistance acid phosphatase (TRAP).^18,19  Additional immunohistochemical staining was performed for 4 more slices in order to detect vessel density and biomaterial-associated macrophages. After endogenous peroxidase activity was quenched using 4% H₂O₂ in methanol and blocking was done by means of horse serum, an anti-rabbit polyclonal CD31 antibody (GeneTex, Inc, Irvine, Calif) was used to detect vascular endothelial cells, whereas a second section was stained with a monoclonal F4/80 antibody (eBioscience, San Diego, Calif) for the detection of macrophages. The DAKO REAL EnVision detection system (DAKO, Glostrup, Denmark) was used to visualize these cells. As negative controls, immunohistologic stains in the absence of the primary antibody were performed on 2 control sections. For visualization by light microscopy, the sections used for immunohistochemistry were counterstained with hemalaun.

The histologic and histopathological evaluation was performed by members of the author team (S.G. and M.B.) at the REPAIR-lab in vivo laboratory (Institute of Pathology, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany) with an ECLIPSE 80i microscope (Nikon, Tokyo, Japan). Microphotographs were recorded using the Nikon DS-Fi1/digital camera together with a Nikon digital sight control unit to systematically outline the tissue responses on the biomaterial surfaces and in the center of the biomaterials. The qualitative histologic analysis was carried out to describe the degradation and integration processes of the implanted biomaterials in the animal organism and the define cells that are involved in the tissue responses to the biomaterials. A further aim of the histologic analysis was to examine the process of biomaterial vascularization and possible unfavorable biomaterial interactions, such as encapsulation, hemorrhage, or necrosis of the peri-implant tissue.

The thickness measurements of both investigated biomaterials were made at each of the experimental time periods in order to obtain information about the volume stability of the biomaterials at the implantation site. Analysis was carried out according to previously published methods^6,18,20  by using a specialized scanning microscope (Eclipse 80i histological microscope, Nikon) connected to a DS-Fi1/digital camera (Nikon) and an automatic scanning table (Prior, Rockland, Mass). Furthermore, the dedicated NIS-Elements software (Nikon) allowed generation of total scans (ie, 1 large image, assembled from 100–120 single images) at a magnification of ×100 and a resolution of 2500 × 1200 pixels, which were then used for histomorphometric analysis by means of the annotations and measurements tool. Thickness measurements were carried out by measuring the thickness of the biomaterials within each biopsy at up to 15 points along the margins. With the obtained values, the mean membrane thickness and the standard deviation for each biomaterial group at the time of explantation was calculated.

To conduct the histomorphometric measurements of the material-associated multinucleated giant cells within the implantation beds of both collagen-based biomaterials, the corresponding TRAP-stained slides were digitized according to the previously described method following established and published methods.^6,16  In brief, the counting tools of the NIS Elements software were used to determine the amounts of the multinucleated giant cells adhering to the biomaterials. Furthermore, the area of the implantation bed of the materials was also measured using the area tool of the software. For the further statistical comparison the number of multinucleated giant cells was related to the implant area (giant cells/mm²).

For histomorphometric measurements of the 2 different vascularization parameters, the aforementioned system composed of the microscope in combination with the digital camera, the scanning table, and the computer system running the Nikon NIS Elements software was used according to previously described methods.^6,16  Briefly, the CD31 slides for both material groups and the control group stained via immunohistochemistry were first digitized. After that, the area of the implantation beds of the biomaterials (or the area of the subcutaneous pocket in the case of the control group) was measured by means of the area tool of the Annotations and Measurements section of the NIS Elements software. In the next step, the single vessels within the implantation beds were also manually marked using the area tool. For further comparison of the vascularization pattern of the implantation beds, the measured data were used to determine both the vessel density, which reveals the number of vessels per square millimeter (vessels/mm²), as well as the percent vascularization that showed the percentile of vessels of the relative total implantation beds.

Afterward, the obtained values were used for statistical intraindividual analysis of materials thickness using a variance analysis followed by a least significant difference post hoc assessment using the SPSS 16.0.1 software (SPSS Inc, Chicago, Ill). Statistical differences were considered significant if P values were <.05 and highly significant if P values were <.01. The GraphPad Prism 5.0 software (GraphPad Software Inc, La Jolla, Calif) was used for plotting graphs.

The SEM imaging of the BEGO Collagen Fleece revealed a bilayered character (Figure 1a). The fleece had a spongy structure with pore sizes of smaller diameters near the more compact surface (Figure 1a, upper part; Figure 1b and c) and a larger diameter near the rough and open porous surface (Figure 1a, lower part; Figure 1d and e).

The collagen membrane possessed a stratified architecture (Figure 2a) with one smooth surface and one rather rough surface (Figure 2b through e). The smoother side had a dense and a more plane structure (Figure 2b and c), while the open porous structure had a rougher side (Figure 2d and e). Between both surfaces, the membrane was composed of a multilayered structure of collagen layers in a comb-like arrangement with connecting collagen filaments (Figure 2a).

The mice in all groups survived material implantation without complications and did not show any macroscopically notable events until the end of the study. Within the control group, no adverse macroscopic and histologic (tissue) reactions were observed, which indicating a lack of contamination during the surgical procedures. Thus, the skin wound healed completely until day 30 of the study.

At day 3 after implantation, the material was detectable within the subcutaneous tissue (Figures 3a1 and 4A). At this early time point, mononuclear cells were predominantly detectable on both material surfaces, and only a few single cells were found within the central regions of the collagen fleece (Figure 4a). At day 10, the material was also detectable within the subcutaneous tissue, and the amount of the mononuclear cells on both surfaces had increased (Figures 3a2 and 4b). Only a few cells were able to penetrate toward the center into the oval parts of the bulk material; thus, the core of the collagen fleece was free of invaded cells (Figure 4b). At day 15, the membrane was located within the subcutaneous tissue, and a thin layer of connective tissue had formed on both material surfaces, which contained mainly mononuclear cells (Figures 3a3 and 4c). Additionally, single multinucleated giant cells were found at the surfaces of the biomaterial (Figure 4c). At day 30 after implantation, the material was still detectable within its implantation bed and the amount of the cellular wall on both material surfaces had increased (Figures 3a4 and 4d). The number of multinucleated giant cells also increased, and loss of material integrity was initiated (Figure 4d). Until this time point, the peri-implant tissue of the material showed a mild vascularization, and microvessels were located within the peripheral regions of both surfaces (Figure 4e). At day 60, only remnants of the implanted material were found within a relatively well-vascularized connective tissue and associated only with mononuclear cells (Figure 4f). The collagen fleece seemed to have undergone nearly complete biodegradation. The application of TRAP staining revealed only a few signs of TRAP activity within mononuclear cells or multinucleated giant cells (Figure 5a).

The tissue reaction to the BEGO Collagen Membrane could be divided into 2 different reactions, which seemed to be specifically related to each of the membrane surface characteristics. Starting on day 3 after implantation, the collagen membrane was detectable within the connective tissue and a mononuclear cell wall was observed on both membrane surfaces (Figures 6a1, 7a1, and 7b1). At this time point, very few single cells were observed invading the periphery of the compact side of the membrane, and no cellular invasion into the spongy membrane side was detectable (Figure 7a1 and 7b1). At day 10, the material was also easily located within its implantation bed, while the thickness and density of the mononuclear layers on both membrane surfaces were increased (Figures 6a2, 7a2, and 7b2). Furthermore, penetration of few mononuclear cells was observed mainly within the compact surface at this time point (Figures 6a2 and 7a2). The peri-implant tissue facing the compact surface showed few vessels, and no vessels were observable within the membrane body (data not shown). The peri-implant tissue at the spongy membrane surface showed no signs of vascularization.

At day 15 after implantation, the membrane was still detectable within the subcutaneous tissue (Figure 6a3). At this time point, a moderate number of multinucleated giant cells was detectable on the compact material surface along with some vessels (Figure 7a3), and few of these cells could be found on the spongy surface (Figure 7b3). At day 30 after implantation, the collagen-based material was still detectable, and the amount of the multinucleated giant cells had increased on the compact surface (Figure 7a4). The compact surface showed evidence of disintegration (Figures 6a4 and 7a4). At this time point, the spongy layer demonstrated mononuclear cells penetrating into the membrane (Figure 7b4). However, still very few multinucleated cells were found within the peri-implant regions of the spongy layer (Figure 7b4). At day 60, the membrane was still observed within the subcutaneous soft tissue and the compact surface layer seemed to have been diminished (Figures 6a5 and 7a5). Only a small dense layer remained as evidence of its previous existence (Figure 7a5). The previous tissue reaction toward this membrane layer that was associated with the occurrence of multinucleated giant cells was changed, and only adherent mononuclear cells were detectable (Figure 7a5). At this time point, still no expressed cell penetration into the membrane was detectable (Figures 6a5, 7a5, and 7b5). Peri-implant mononuclear cells penetrated the spongy surface layer on the other side (Figure 7b5). No transmembranous vascularization was found at any of the study time points. Furthermore, only a few TRAP expressions in mononuclear or multinucleated cells were observable at anytime point (Figure 5b). The collagen membrane was detectable within the time period between day 3 and day 60 (Figure 6A1 through A5). At day 60, the membrane was well integrated within its implantation bed and showed no breakdown of material integrity (Figures 6a5, 7a5, and 7b5).

The histomorphmetric analysis of the collagen fleece revealed that the thickness significantly decreased between day 3 and day 30 after implantation (P < .05) before material defragmentation was initiated between day 30 and day 60 (Figure 3b). The most obvious change in the fleece thickness was observed between day 15 and day 30, and no statistically significant difference was measured (Figure 3b).

The histomorphometric analysis of membrane thickness revealed a continuous decrease of membrane thickness over time, but only at day 60 was a significant membrane thickness decrease measured when compared to values at day 3 (P < .05) (Figure 6B).

The histomorphometric analyses revealed that no multinucleated giant cells were found within the implantation beds of either collagen-based material at day 3 and day 10 after implantation (Figure 8a). Starting with day 15, a comparable low amount of multinucleated giant cells was found within the implantation beds of both collagen materials (BEGO Collagen Fleece: 0.47 ± 0.14 giant cells/mm²; BEGO Collagen Membrane: 0.26 ± 0.12 giant cells/mm²). No significant differences were found between these groups (Figure 8a). Furthermore, no intraindividual significant differences were found when comparing the values of day 10 and day 15 after implantation (Figure 8a).

At day 30 after implantation, significant increases in the amounts of multinucleated giant cells were measured in both study groups (P < .01 or P < .001) (Figure 8a). Thus, the number of giant cells within the BEGO Collagen Fleece study group (4.38 ± 1.40 giant cells/mm²) was significantly higher compared with the values found within the implantation beds of the BEGO Collagen Membrane group (1.74 ± 0.62 giant cells/mm²) (P < .001) (Figure 8a).

At day 60 after implantation, significant decreases in the amounts of the material-associated multinucleated giant cells were found in both collagen-based biomaterials (P < .01 or P < .001) (Figure 8a). Thus, no multinucleated giant cells were found at this time point within the implantation beds of the BEGO Collagen Fleece as this material was nearly completely degraded. In contrast, low multinucleated giant cell amounts were found within the implantation beds of the BEGO Collagen Membrane (0.51 ± 0.09 giant cells/mm²) (Figure 8a). Also at this time point, no significant differences were found in the amounts of multinucleated giant cells between the both study groups (Figure 8a).

Analysis of the vessel density showed that no vessel ingrowth was detected in either collagen-based material at day 3 and day 10 after implantation (Figure 8b). However, a mild vascularization was found in the control group (0.69 ± 0.22 vessels/mm²), that was significantly higher compared with the values of both other study groups (P < .01) at day 3 after implantation (Figure 8b). At day 10 after implantation, the vessel density values of the control group were similar to that of day 3 (1.19 ± 0.54 vessels/mm²) without intraindividual differences, while these values were again significantly higher compared with that of both material study groups (P < .001) (Figure 8b).

At day 15 after implantation, a mild vascularization was found in the BEGO Collagen Fleece group (0.68 ± 0.19 vessels/mm²), which was significantly higher compared with the values of this group at day 10 (P < .001) (Figure 8b). Furthermore, no vessels were found within the implantation beds of the BEGO Collagen Membranes at this time point (Figure 8b). Thus, a significant difference was measured between the values of both material study groups (P < .01) (Figure 8B). At this time point, the values of the vessel density of the control group were still on a similar level (1.29 ± 0.32 vessels/mm²) compared with the former study time point without significant differences, and these values were again significantly higher compared with the values of the material study groups (P < .05 or P < .001) (Figure 8B).

At day 30 after implantation, a further significant increase in vessel number was measured for the BEGO Collagen Fleece group (1.39 ± 0.33 vessels/mm²) compared with the values of day 15 (P < .001) (Figure 8b). In contrast, still no vessel ingrowth was detected in the BEGO Collagen Membrane group at this time point (Figure 8b). Thus, a significant difference was found between the values of these 2 groups (P < .001) (Figure 8b). Furthermore, the value of the control group (1.66 ± 0.45 vessels/mm²), which was still on a similar level compared with day 15 after implantation without significant changes, showed no significant differences compared with values of the BEGO Collagen Fleece study group, while significant differences were measured compared with the values of the BEGO Collagen Membrane group (P < .001) (Figure 8b).

At day 60 after implantation, no vessels were found within the implantation beds of either collagen-based biomaterial, so that a significant decrease in the BEGO Collagen Fleece group was measured (P < .001) (Figure 8b). The values of the control group at this time point (1.66 ± 0.45 vessels/mm²) were also decreased compared with day 30 after implantation, but without significant difference; however, they were significantly different compared with the other study groups (P < .001) (Figure 8b).

The histomorphometric analysis of the percent vascularization also showed that the values of the control group were significantly higher at day 3 (0.05 ± 0.01 %) and day 10 (0.09 ± 0.01 %) after implantation compared with the values of the collagen-based material study groups (P < .01 or P < .001), as no vessels were found within the implantation beds of the latter 2 at these time points (Figure 8C).

At day 15 after implantation, a mild vascularization was measured in the BEGO Collagen Fleece group (0.09 ± 0.01 %), that was significantly higher compared with the values of the former study time point (P < .001) as well as to the values of the BEGO Collagen Membrane group (P < .001), which showed no signs of vessel ingrowth (Figure 8c). Furthermore, the values of the control group (0.09 ± 0.03 %), which were on a similar level compared with day 10 after implantation, were similar to the values of the BEGO Collagen Fleece group, and only a significant difference was measured compared with the BEGO Collagen Membrane group (P < .001) (Figure 8c).

At day 30 after implantation, a further significant increase in the percent of vascularization was found in the BEGO Collagen Fleece group (0.18 ± 0.04 %) P < .001); these values were again significantly higher compared with the values of the BEGO Collagen Membrane group, which showed still no signs of vessel ingrowth in its implantation beds (P < .001) (Figure 8c). At this time point, the values of the BEGO Collagen Fleece group were also significantly higher compared with the values of the control group (0.12 ± 0.03 %) (P < .01), in which values were on a similar level compared with day 15 after implantation (Figure 8c). Additionally, the values of the control group were significantly higher compared with the values of the BEGO Collagen Membrane group (P < .001) (Figure 8c).

At day 60 after implantation, no vessels were found in either collagen-based material group, so that no significant differences in percent vascularization were found between either study group (Figure 8c). A significant decrease in the percent vascularization was calculated for the BEGO Collagen Fleece group compared with day 30 (P < .001) (Figure 8c). Furthermore, at this time point the values of the control group (0.16 ± 0.02 %) were on the same level as at day 30 after implantation, and significant differences were measured in comparison to the both material study groups (P < .001) (Figure 8c).

In recent years, non–cross-linked collagen-based materials began to replace other material classes, such as nonresorbable or fast biodegradable materials, for applications in GTR and GBR.²¹  Currently, several animal-based tissue sources are attractive for processing to membranes or matrices for GBR and GTR.

Recently, our group investigated the tissue integration of 2 porcine-based non–cross-linked materials.^6,16  The implantation of a bilayered membrane (Mucograft) of porcine peritoneum combined with porcine skin revealed that the material used was detectable within the entire study period of 60 days.⁶  The dense surface disappeared after 30 days, but the spongy component was well integrated within its implantation bed without any breakdown or signs of transmembranous vascularization. This material induced a tissue reaction that was associated only with mononuclear cells, such as macrophages and fibroblasts, and no multinucleated giant cells were detectable during the observation period. Similar observations were made relating to a porcine, peritoneum-derived non–cross-linked collagen I-III (Bio-Gide) after application in the same animal model.¹⁶  Furthermore, the results of these 2 independent, in vivo studies revealed that non–cross-linked collagen materials do not have to undergo transmembranous vascularization in order to contribute to a sufficient integration within their peri-implant region. These results led to the conclusion that collagen-based materials can be supplied with nutritive substances from their peri-implant tissue through diffusion processes and do not need a foreign-body–triggered granulation tissue for the tissue integration and cellular degradation process.^6,16  Accordingly, these materials possess the desired stability for tissue regeneration according to GTR and GBR.

Based on these observations, the present in vivo study compared the tissue reaction of a porcine dermis-derived collagen fleece to that of a porcine pericardium-derived collagen membrane. The goal of the study was to apply the aforementioned established protocols in order to analyze whether the previously observed tissue and cell reactions to porcine-based collagen materials are animal specific or are related to the physiological compartment from which the material is derived or its physicochemical characteristics.

The histologic results of the collagen fleece revealed that this material serves as a stable material during the first 30 days after implantation and then undergoes a nearly complete biodegradation and disintegration between day 30 and day 60 after implantation. Within the first 30 days, mononuclear cells started to penetrate into the central region of the material from both surfaces followed by material breakdown after this time point until the end of the study. These data suggest that the material is able to induce its own biodegradation by changing the metabolic and cellular conditions of the surrounding peri-implant tissue. Accordingly, the material-specific physicochemical characteristics, such as its pore system and its source (dermis), might make this material suitable for clinical applications in which a wound bed has to be stabilized, before tissue formation. Based on the present data, this material does not seem to have properties to serve as stable barrier membrane for applications in GBR, as it might not be able to shield the regeneration process as long as it might be needed,²¹  but can serve as a spongy matrix to support the initial steps of wound healing. In such cases, the used collagen fleece has to be nonpermeable for the surrounding peri-implant cells at early time points. One interesting finding was the presence of TRAP-negative multinucleated giant cells and nearly no TRAP-positive multinucleated giant cells within the implantation bed of this material. The 5a subform of the TRAP5 enzyme, which is described as being expressed by mono- and multinucleated cells that are involved in inflammatory tissue reactions in order to generate reactive oxygen species.^22–27  The latter enables cells, such as macrophages and multinucleated foreign body giant cells (the result of fusion of macrophages) to perform the degradation of foreign bodies, such as a wide range of biomaterials.²³  On the other hand, the expression of the TRAP5b molecule, which showed a high similarity to TRAP5a, within bone tissue is suggested to be involved in matrix and collagen degradation during bone resorption.²⁵  No further knowledge exists about the involvement of these molecules in inflammatory and physiological milieus.

Thus, the question arises whether the choice of the material source and/or material preparations might be able to induce TRAP-positive multinucleated giant cell formation with an inflammatory character. Interestingly, previous studies on collagen-based biomaterials showed the absence of TRAP-negative and TRAP-positive multinucleated giant cells.^6,16  The present study, however, is still not able to clarify the role of the multinucleated giant cells within the implantation bed of the analyzed materials. The observed results call for reflection on the role and the meaning of these cells, even in the peri-implant tissue of biologically derived materials, such as the collagen fleece.

Multinucleated giant cells are also known to be able to produce the vascular endothelial growth factor, and this seems to contribute to an increased vascularization in the implantation bed.^19,20  In relation to the present data, the occurrence of these cells might explain the increased vascularization of the implantation bed of the BEGO Collagen Fleece between day 15 and day 30, which could be accompanied by material breakdown. It must be noted that the appearance of the multinucleated cells also increased between day 15 and day 30. We have previously shown that the number of these cells decreases after material degradation.²⁸  It is likely that these cells were induced in relation to specific physicochemical characteristics of the material and are involved in material degradation. Thus, their presence within the implantation bed was not accompanied by any observable tissue damage, as no signs of hemorrhage, necrosis, or material encapsulation were observed at any time point within the experimental groups.

In the present study, the tissue response to a non–cross-linked collagen-based membrane was also investigated. The results showed that the collagen membrane was able to persist as a barrier membrane within its implantation bed for up to 60 days. Furthermore, the histologic data revealed that the 2 material surfaces induced 2 different cellular reactions. These observed tissue reactions were mainly confined to each of the material surfaces and seemed to occur independently. Both membrane sites initially induced a tissue reaction based on mononuclear cells. TRAP-negative multinucleated giant cells were also observed on the dense material surface as soon as 15 days after implantation. At this time, point the peri-implant tissue adjacent to the compact surface layer showed few vessels. On that note, the combination of mononuclear and multinucleated cells on the compact material site contributed to its nearly complete biodegradation within the first 30 days. Interestingly, very few multinucleated giant cells were observed at the spongy material surface between days 10 and 60. The spongy surface did not allow an expansive cell penetration into its core for the first 30 days after implantation. From this time point until the end of the study, however, a controlled cell and tissue penetration pattern was observed that allowed the cells to fill the gap between the material fibers. Thus, the spongy material surface served as a stable barrier membrane until the end of the study while undergoing a stepwise integration. No transmembranous vascularization could be observed at any of the study time points. These findings support the hypothesis of our research group that collagen-based membranes for GBR application are permeable for nutrients by means of the diffusion processes. In this way, the bony defect region can be supplied with nutrition.^6,16  Based on the observed histologic results in this study, the investigated collagen membrane can serve as a barrier membrane as it does not undergo a premature breakdown within the observed early and late study time points. With the different tissues of a bony defect side and the goals of GBR and GTR in mind, the 2-component material seems to fulfill the requirements that a material should possess in this context. It can be assumed that the compact membrane surface, which should be aligned in the direction of the soft tissue, allows integration within this tissue without allowing a cell and tissue ingrowth into the membrane bulk toward the spongy surface and the underlying bony defect area. Furthermore, the spongy side allows a delayed cell and tissue ingrowth into its interstices. This side is recommended as coverage for the bone defect, in which bone regeneration becomes orchestrated. Keeping in mind that bone tissue has a considerably slower growth potential, this delayed ingrowth could lead to stepwise integration of the material within the newly formed bone tissue, when regeneration reaches the upper part of the defect side.^2,29,30 

However, the histologic analysis also revealed the presence of mainly TRAP-negative multinucleated giant cells within the implantation bed of this biological membrane. As mentioned previously, the question arises as to whether the occurrence of these cells is a physiological process involved in biodegradation of some collagen-based materials or is associated with collagens of some particular regions. Multinucleated cells, which are also called osteoclast-like cells, are present within the implantation bed of several investigated material classes ranging from bone substitute granules^18,31,32  to silk-based polymers^20,28  and nonresorbable expanded polytetrafluororethylene-based materials.¹⁶  The presence of these cells has also been shown to contribute to enhanced material vascularization.²⁰  Interestingly, some reports note the occurrence of multinucleated giant cells in relation to porcine collagen-based biomaterials.^33–35  However, the pathobiological significance of the presence of different types of multinucleated giant cells still remains to be elucidated. This could prove to be of importance for the evaluation of materials with respect to their suitability for clinical application. On a final note, it must be mentioned that both materials are prepared using different sterilization methods, that is, gamma irradiation for the BEGO Collage Fleece and ethylene oxide for the BEGO Collagen Membrane. It has been shown that the sterilization of polymers by gamma irradiation can cause changes in their physical or mechanical properties, while the usage of ethylene oxide seems to be much more stable.³⁶  Thus, the observed higher inflammatory tissue reaction with the involvement of multinucleated giant cells could also be related to the gamma irradiation for the BEGO Collage Fleece, as this process could have destroyed parts of the collagen fibers, making them seem to be foreign bodies as they had lost their natural structure. However, little is known about the exact sterilization procedures of both materials (regarding exposure times and concentrations of both sterilization agents), and no knowledge exists about the impact of both agents on the structure of the collagen molecule; thus, no correlation is possible between the presented in vivo results and the respective sterilization procedure of both collagen-based materials.

In this study, the inflammatory response to 2 different collagen-based materials—porcine pericardium and porcine dermis—wa assessed after their in vivo implantation in CD-1 mice for up to 60 days. The cellular pattern involved in material integration was assessed by means of histologic analysis and histomorphometric techniques. The results of the present study showed the following: the implantation of the 2 materials resulted in multinucleated giant cell induction in the case of the dermis-based collagen fleece and the compact layer of the pericardium-based collagen membrane, while the spongy layer of the collagen membrane induced a mainly mononuclear-triggered tissue reaction. The present data underline that although both biomaterials are biologically derived, different tissue origins within the same animal species can induce different cellular tissue responses, specifically mononuclear vs multinucleated giant cells. The presence of the latter are considered a foreign body response. The present study data emphasize that more investigations about the role of multinucleated giant cells within the implantation bed of collagen-based materials are necessary in order to be able to classify their potential negative or positive role in biomaterial integration and implantation bed vascularization. It has to be clarified whether the clinical use of such collagen-based materials is superior compared with synthetic-based polymers.

Abbreviations

GBR

guided bone regeneration

GTR

guided tissue regeneration

H&E

hematoxylin and eosin

SEM

scanning electron microscopy

TRAP

tartrate resistance acid phosphatase

The authors would like to thank Ms Ulrike Hilbig and Mr Mykhaylo Reshetnykov for their technical assistance and the University of Bremen, Advanced Ceramics Group for SEM imaging.

This research was partially funded by the International Bone Research Association Foundation, Basel, Switzerland and partially by the research funds of the University of Mainz, Mainz, Germany.