Non-cross-linked porcine-based collagen I–III membranes do not require high vascularization rates for their integration within the implantation bed: A paradigm shift

Abstract

There are conflicting reports concerning the tissue reaction of small animals to porcine-based, non-cross-linked collagen I–III membranes/matrices for use in guided tissue/bone regeneration. The fast degradation of these membranes/matrices combined with transmembrane vascularization within 4 weeks has been observed in rats compared with the slow vascularization and continuous integration observed in mice. The aim of the present study was to analyze the tissue reaction to a porcine-based non-cross-linked collagen I–III membrane in mice. Using a subcutaneous implantation model, the membrane was implanted subcutaneously in mice for up to 60 days. The extent of scaffold vascularization, tissue integration and scaffold thickness were assessed using general and specialized histological methods, together with a unique histomorphometrical analysis technique. A dense Bombyx mori-derived silk fibroin membrane was used as a positive control, whilst a polytetrafluoroethylene (PTFE) membrane served as a negative control. Within the observation period, the collagen membrane induced a mononuclear cellular tissue response, including anti-inflammatory macrophages and the absence of multinucleated giant cells within its implantation bed. Transmembrane scaffold vascularization was not observed, whereas a mild scaffold vascularization was generated through microvessels located at both scaffold surfaces. However, the silk fibroin induced a mononuclear and multinucleated cell-based tissue response, in which pro-inflammatory macrophages and multinucleated giant cells were associated with an increasing transmembrane scaffold vascularization and a breakdown of the membrane within the experimental period. The PTFE membrane remained as a stable barrier throughout the study, and visible cellular degradation was not observed. However, multinucleated giant cells were located on both interfaces. The present study demonstrated that the tested non-cross-linked collagen membrane remained as a stable barrier membrane throughout the study period. The membrane integrated into the subcutaneous connective tissue and exhibited only a mild peripheral vascularization without experiencing breakdown. The silk fibroin, in contrast, induced granulation tissue formation, which resulted in its high vascularization and the breakdown of the material over time. The presence of multinucleated giant cells at both interfaces of the PFTE membrane is a sign of its slow cellular biodegradation and might lead to adhesions between the membrane and its surrounding tissue. This hypothesis could explain the observed clinical complications associated with the retrieval of these materials after guided tissue regeneration.

Introduction

Recently, a large number of alloplastic, xenogenic or allogenic bone substitute materials have been introduced for the volume enhancement of atrophic human alveolar bone prior to implant-supported reconstruction. The successful integration of these materials into the human body should be performed in accordance with the described requirements of the guided bone regeneration (GBR) method and should not result in the rapid influx of the surrounding peri-implant soft tissue [1]. Protection from this phenomenon can be achieved by barrier membranes, which should be placed above the positioned bone substitute within the reconstructed region and under the periosteal tissue. Barrier membranes have also been used for soft tissue regeneration in the oral cavity. These membranes can be used to treat gingival recessions, where they promote periodontal tissue regeneration after the specific conditioning of the root. This process is termed guided tissue regeneration (GTR). Materials designed for GBR and GTR must possess certain physicochemical characteristics, such as good tissue compatibility and cell occlusivity. In addition, the materials should serve as a barrier membrane, be easily applicable for clinical use and have integrative properties [2]. Considering these requirements, an ideal barrier membrane is a biomaterial that can serve as a barrier membrane for an extended period and then becomes integrated with the surrounding soft tissue after the intended bone regeneration is complete. It has been postulated that it takes 4 weeks to achieve the structural integrity of membranes for periodontal regeneration [3], whereas a longer period, i.e. up to 6 months, has been recommended for bone tissue regeneration [4].

The first generation of materials used for GBR/GTR in biological systems was polytetraflouroethylene (PTFE), which is a highly stable polymer. This non-resorbable material contributed to successful bone and soft tissue regeneration by maintaining its structural integrity while serving as a barrier membrane for as long as it remained in situ [5]. However, a second operation is required for the retrieval of this material, which can potentially damage the newly regenerated tissue [6]. The drawbacks associated with retrieval have led to the development of bioresorbable barrier membranes, such as polyesters, with different biodegradation properties [7], [8]. The application of these biodegradable materials, however, is associated with rapid cellular and enzymatic degradation, and the desired control of the tissue regeneration process is not achieved.

Collagen-based materials have also been used as biodegradable materials in GBR/GTR strategies and have achieved clinical success rates comparable to non-resorbable membranes [9], [10]. Collagen, the most abundant protein in the human and animal body, is physiologically ubiquitous. Collagen’s natural origin, together with its biodegradative characteristics, makes it an attractive candidate for biomaterial-based tissue engineering. During the wound-healing process, collagen undergoes enzymatic degradation mediated by matrix metalloproteases (MMP), which are released by recruited neutrophils, monocytes/macrophages, eosinophils and fibroblasts [11], [12]. Furthermore, type I collagen is known to possess angiogenic potential [13], [14], [15], which is important for successful tissue regeneration.

Native collagen undergoes relatively rapid degradation and therefore does not provide the desired stability needed for a barrier membrane in GBR/GTR applications. Decreasing the collagen degradation rate and corresponding extended stability of the material has been achieved using cross-linking techniques, such as ultraviolet- and gamma irradiation, as well as treatment with hexamethylenediisocyanate, glutaraldehyde, diphenylphosphorilazide and ribose [16], [17]. In contrast, non-cross-linked collagen-based materials and cross-linked collagen membranes have been reported to be partially cytotoxic [18]. In this context, these materials have been shown to inhibit the attachment and proliferation of human periodontal ligament fibroblasts and osteoblasts [19], and they induce a foreign body reaction, i.e. the induction of multinucleated giant cells, in vivo [20]. These findings led to the development of alternative processing techniques in which non-cross-linked native collagen I, which is known to be more resistant to degradation, was combined with collagen III, which is known to be rapidly degraded. The combination of these two collagen types should result in bi- or multilayered membranes that exhibit a regulated biodegradation, i.e. barrier membranes that become integrated within their implantation bed at a later time point.

A recently published manuscript has shown that the subcutaneous implantation of a non-cross-linked collagen I–III matrix (Mucograft®) in wild-type mice led to integration of the material within the subcutaneous connective tissue, which was a result of the mononuclear-cell-based tissue reaction to the material in combination with a mild vascularization of the implantation bed. Similar histological findings were observed after the implantation of this matrix in human oral tissue [21]. Interestingly, similar results were also obtained on histological examination of transplanted oral connective tissue in the human oral cavity (unpublished data). Clinically, the use of this porcine-derived collagen matrix, and the use of another bovine-origin non-cross-linked matrix (Copios®), contributed to an increase in the soft tissue volume that was comparable to that achieved using the patient’s own connective tissue for transplantation [21], [22]. Based on these findings, the hypothesis was made that xenograft- or allograft-derived non-cross-linked collagen membranes/matrices would demonstrate similar biodegradation characteristics to natural collagen I/III within the human connective tissue. In contrast to undergoing rapid biodegradation and transmembrane vascularization shortly after transplantation, these materials become integrated within their implantation bed and exhibit mild vascularization [21], [22]. These data, however, contradict recently published data showing that the subcutaneous implantation of bilayered non-cross-linked collagen-I/III membranes (BioGide®) in rats resulted in a homogeneous and transmembrane vascularization within 2 weeks of implantation and the persistence of the material within the subcutaneous tissue for as little as 4 weeks [20]. Another aspect of the report that could not be observed in the studies with Mucograft® was a decrease in the thickness of the continuous membrane and membrane breakdown within 4 weeks [23]. It is important to note that Mucograft® is produced by the same manufacturer and using the same processing techniques as Bio-Gide®. The two materials differ mainly in the thickness of their two layers. This difference could be the reason why Bio-Gide® is referred to as a “membrane”, whereas Mucograft® is considered a “matrix”.

The previously reported rapid transmembrane vascularization and the relatively fast decrease in the membrane thickness of the non-cross-linked collagen membranes [20], [23] are not in accordance with the above-mentioned requirements for GBR/GTR, which suggest that a membrane should serve as a barrier membrane, protecting the ongoing regenerative process under its surface. Furthermore, these findings also contradict the logic behind the use of non-cross-linked collagen membranes I/III in GBR and GTR, because non-cross-linked collagen I/III membranes would have no integrity and would thus undergo rapid biodegradation.

Another in vivo study, which analyzed the tissue reaction to Bio-Gide® in a rat calvarian model, documented a mononuclear cell-based tissue reaction in combination with a small number of capillaries within the inter-fibrillar spaces 4 weeks after implantation [24]. In this period, a clear identification of the membrane was observed. Histologically, a loose connective tissue was detected on the upper layer of the membrane surface, which was located below the skin, whereas the underlying membrane surface, which was located on the dura mater, appeared more vascularized [24]. Furthermore, the material was detectable after 9 weeks. The authors did not report a significant increase in vascularization of the material between 4 and 9 weeks [24]. The findings suggest that, as opposed to rapidly degradable characteristics, these membranes exhibit the integrative characteristics of non-cross-linked collagen I/III membranes [20]. Interestingly, the authors stress the necessity of the transmembrane vascularization of collagen membranes for GBR/GTR [20], [23] whilst unknowingly contradicting their own histological results [24].

The misdiagnosis of the histological analysis of the tissue–biomaterial interaction by non-experts is a possibility, especially if a series of single images of the implantation bed was evaluated without concentrating on the findings within the materials entire implantation bed. This scenario may have contributed to the findings and conclusions of these reports [20], [23], [24].

The aim of the present study was to assess the tissue reaction to Bio-Gide® in the same model and protocol that was used for Mucograft®. As mentioned above, the Bio-Gide® membrane is macroscopically a thinner variant of the Mucograft® matrix, which is a bilayered membrane. One side of the membrane is smooth, with low porosity, whereas the other side exhibits a more porous, three-dimensional network. In this histological study, special emphasis was placed on the cellular tissue reaction of the material, its vascularization and its integration into the tissue. These observations were made after the generation of a total scan of the complete implantation bed. Total scans allow the assessment of all aspects of the tissue’s reaction to the material without the possibility of misdiagnosis, which would arise if the observer focused only on particular regions of the slides. Emphasis was also placed on the potential changes in the thickness of the material, which was measured histomorphometrically throughout the study. A silk fibroin-based dense micronet from pure race silkworms was used as a positive control. Previously published data have shown that silk fibroin from a polyhybrid silkworm pretreated with formic acid can result in different characteristics. Formic acid treatment for 60 min produced a dense micronet, whereas treatment with formic acid for 30 min resulted in a porous micronet [25]. Another reason for choosing a 60 min treatment for the silk fibroin from pure race silkworms was to evaluate whether the observed tissue reaction of Wistar rats to polyhybrid fibroin treated with formic acid for 6 min could be reproduced in mice.

A commercially available PTFE membrane was used a negative control. As previously mentioned, PTFE membranes are the golden standard for GBR/GTR, and the tissue reaction to these so-called inert materials should be reproducible and independent of the species used. This in vivo study was performed over a period of 60 days, with observations taken at 3, 10, 15, 30 and 60 days.