**3. Evaluation of implant-soft tissue response**

The soft tissue interface especially the structure of collagen fibre bundles received more attention over the past 10 years with studies that include animal models such as dogs [10–12, 14] and monkeys [23, 24] and human [25, 26] used to explore the structure and dimension of soft tissue-implant interface. Recently, Chai and co-workers [18] have ventured upon the use of three-dimensional oral mucosal models by using the tissue engineering technology to investigate the nature of the peri-implant biological seal.

#### **3.1. Implant-soft tissue interface models**

The advantages and disadvantages of each implant-soft tissue interface study models are described in the next section. These models were developed in order to enhance our under‐ standing of the soft tissue response on various materials with different surface topography and to establish best methods to evaluate the biological seal of peri-implant tissue. Generally, an in vitro study using monolayer cell culture model is conducted to assess the cytotoxicity of the cells and quickly observe cell activities and behaviour towards new dental implant materials. Histomorphometric analysis of en bloc tissue consisting of both soft tissue and implant body is the best method to demonstrate the presence of epithelial and connective tissue attachment at the soft tissue-implant interface. Yet, due to limited opportunity to obtain histological section from human, animal models were developed.

#### *3.1.1. In vitro studies*

Presently, in vitro testing is performed as a prerequisite to in vivo evaluation. However, the in vitro techniques do not reflect the clinical situation and the progress in our understanding of extra- and intracellular processes that occur in connective tissue attachment. Thus, the data cannot be extrapolated into clinical applications. Nevertheless, the study involving monolayer cells is by far the most popular and easy-to-conduct study although more sensitive in vitro evaluations are now available. The cell shape, activities and response can be evaluated morphometrically via immunocytochemical staining [27], or by analysis using scanning electron [28, 29] or fluorescent [30, 31] miscroscopies. Additionally, the gene and protein expressions for cell adhesion and attachment can also be carried out [27, 32, 33]. Most studies used primary human gingival [29, 32, 34] and periodontal [35] fibroblasts as a cell model, which are cultured directly onto the dental implant materials surface. Keratinocytes are also frequently used [27, 36, 37]. Compared to fibroblasts, keratinocytes by far is most difficult to culture. Cochran et al. [35] compared the behaviour of periodontal and gingival fibroblast as well as keratinocytes towards the titanium with different surface textures. They found that human fibroblast and epithelial cell attachment and proliferation are significantly affected by surface characteristics of titanium. Of three cell types, gingival fibroblasts appeared to attach best, followed by periodontal ligament fibroblasts and epithelial cells. Both types of fibroblasts grow and proliferate well on both rough and smooth titanium surfaces compared to epithelial cells once they are attached to the surface [35]. Other study found a significant decrease in the number of gingival fibroblasts on rough titanium (Ti) surfaces compared with smooth polished Ti surfaces [30, 34]. On the other hand, Oates et al. [32] found that the fibroblasts adhesion and attachment are enhanced in rougher surface than smooth surface, in contrast to other findings [33] where focal adhesion kinases were immunogold labelled. In a different study using ceramic, fibroblasts attached more on the milled ceramic and appeared to follow the direction of the fine irregularities on the surface [38]. Nevertheless, most common finding of those studies is that cells were oriented in a parallel order along the grooves of the machined surface but arranged randomly when in contact with a rough surface. Hence, the in vitro models appear to be able to provide an insight and could be used to guide specific cell attachment or specific material with surface characteristics for in vivo models. Animal models are the most common in vivo models carried out compared to human studies.

### *3.1.2. Animal models*

**3.1. Implant-soft tissue interface models**

44 Dental Implantology and Biomaterial

*3.1.1. In vitro studies*

histological section from human, animal models were developed.

The advantages and disadvantages of each implant-soft tissue interface study models are described in the next section. These models were developed in order to enhance our under‐ standing of the soft tissue response on various materials with different surface topography and to establish best methods to evaluate the biological seal of peri-implant tissue. Generally, an in vitro study using monolayer cell culture model is conducted to assess the cytotoxicity of the cells and quickly observe cell activities and behaviour towards new dental implant materials. Histomorphometric analysis of en bloc tissue consisting of both soft tissue and implant body is the best method to demonstrate the presence of epithelial and connective tissue attachment at the soft tissue-implant interface. Yet, due to limited opportunity to obtain

Presently, in vitro testing is performed as a prerequisite to in vivo evaluation. However, the in vitro techniques do not reflect the clinical situation and the progress in our understanding of extra- and intracellular processes that occur in connective tissue attachment. Thus, the data cannot be extrapolated into clinical applications. Nevertheless, the study involving monolayer cells is by far the most popular and easy-to-conduct study although more sensitive in vitro evaluations are now available. The cell shape, activities and response can be evaluated morphometrically via immunocytochemical staining [27], or by analysis using scanning electron [28, 29] or fluorescent [30, 31] miscroscopies. Additionally, the gene and protein expressions for cell adhesion and attachment can also be carried out [27, 32, 33]. Most studies used primary human gingival [29, 32, 34] and periodontal [35] fibroblasts as a cell model, which are cultured directly onto the dental implant materials surface. Keratinocytes are also frequently used [27, 36, 37]. Compared to fibroblasts, keratinocytes by far is most difficult to culture. Cochran et al. [35] compared the behaviour of periodontal and gingival fibroblast as well as keratinocytes towards the titanium with different surface textures. They found that human fibroblast and epithelial cell attachment and proliferation are significantly affected by surface characteristics of titanium. Of three cell types, gingival fibroblasts appeared to attach best, followed by periodontal ligament fibroblasts and epithelial cells. Both types of fibroblasts grow and proliferate well on both rough and smooth titanium surfaces compared to epithelial cells once they are attached to the surface [35]. Other study found a significant decrease in the number of gingival fibroblasts on rough titanium (Ti) surfaces compared with smooth polished Ti surfaces [30, 34]. On the other hand, Oates et al. [32] found that the fibroblasts adhesion and attachment are enhanced in rougher surface than smooth surface, in contrast to other findings [33] where focal adhesion kinases were immunogold labelled. In a different study using ceramic, fibroblasts attached more on the milled ceramic and appeared to follow the direction of the fine irregularities on the surface [38]. Nevertheless, most common finding of those studies is that cells were oriented in a parallel order along the grooves of the machined surface but arranged randomly when in contact with a rough surface. Hence, the in vitro models appear to be able to provide an insight and could be used to guide specific cell attachment or

Studies using animals as in vivo models for evaluation of soft tissue response around dental implant have been extensively conducted and are well documented. In animal models, the histological section of peri-implant tissue was made possible, which becomes the gold standard for the implant-soft tissue interface analysis. While dogs models such as the beagle [8, 14] being the most common animal of choice, monkeys [23, 39] and minipigs [40] were also used to demonstrate the presence of epithelial and connective tissue attachment around transmucosal region of dental implants histologically.

The experiments in animals demonstrated that the dimension of the mucosal attachment to implants was similar to the gingival attachment at teeth and was composed of an epithelial portion about 1.5–2 mm long and a cell-rich connective tissue portion close to the implant that was about 1–1.5 mm high [10]. Animal models were also used to evaluate the soft tissue response towards different abutment materials. Abrahamsson et al. [11] investigated the influence of abutment material on the location and the quality of the attachment that occurred between the peri-implant mucosa and the implant. They found no proper attachment formed at the abutment level made of gold alloy and porcelain when compared to those made of titanium and ceramic. In addition, similar finding was noted by Welander et al. [41] when titanium, zirconia and Au/Pt alloy were used. The tissue around abutment made from titanium and zirconia was stable; meanwhile, an apical migration of epithelium was noted on Au/Pt alloy. In another study, Abrahamsson et al. [42] demonstrated that the soft tissue attachment that formed at implants made of commercially pure titanium (c.p. titanium) was not influenced by the roughness of the titanium surface.

Among many, dogs have been the most common animal of choice. This is possibly due to easy access with regard to clinical examinations and oral hygiene procedures of the dogs. It must be noted that non-human primates bear more resemblence to human anatomy and histology than any other animal, thus may offer a higher degree of relevance to human. Nevertheless, the results from animal experiments should always be carefully interpreted since the healing response and immuno-reaction in animals might not be similar to human, so the data might not be comparable. A given sequence of soft tissue integration to implants in a dog may not correspond exactly to an expected outcome in humans. The differences in tissue response during healing between human-human subjects may sometimes become more pronounced between different human to human subjects than between animals and humans. Moreover, the healing response in animals is also less predictable compared to human. In the light of evidence-based dentistry, the result from animal studies should be interpreted cautiously. Additionally, animal studies are also bound to ethical considerations, where study design and calculation of sample size of animals in experiments are to be carried out with caution. Essentially, to have more clinical validity, human randomized control trials should be carried out to obtain more information on the peri-implant tissues.

#### *3.1.3. Human studies*

The composition of the connective tissue interface towards implants was studied in both animal experiments and human biopsy materials. While human studies are very limited due to ethical issues, the evidence of epithelial and connective tissue attachment around periimplant regions are obtained mostly from failed implant [43], autopsy[44, 45] or clinical studies [1, 46], where the presence of connective tissue attachment on these studies is still difficult to demonstrate. Most of the human studies that have been carried out were clinical studies in which the traditional periodontal parameters were used for monitoring the soft tissue responses around dental implants intra-orally. According to clinical studies that involve the marginal bone levels, we can conclude that bone level is stable as it implies that the soft tissue integration has not migrated apically [1, 46, 47]. Liljenberg et al. [48] in their study of soft tissue biopsies of edentulous ridge mucosa and peri-implant mucosa revealed that the composition of both tissues were nearly identical in terms of collagen, cells and vascular structures. The peri-implant mucosa harboured a junctional epithelium that contained significantly enhanced numbers of different inflammatory cells infiltration. On the other hand, Piatelli et al. [44] found that there was no inflammatory infiltrate in epithelium or connective tissue in human autopsy biopsies of titanium dental implants. It is also interesting to note that the collagen fibres in the coronal part were parallel to implant surface while in the apical region the fibres were in a perpendicular fashion was found. Additionally, Glauser et al. [49] used both hard and soft tissue biopsies of mini titanium implants with different surface characteristics to demonstrate the establishment of junctional epithelium attachment to the implant surfaces. They noted that collagen fibres and the fibroblasts were oriented parallel to the implant surface. The oxidized and acid-etched implants revealed less epithelial downgrowth and longer connective tissue than machined implants [49]. As for different types of materials, Vigolo et al. [46] assessed the peri-implant mucosa around abutments made of gold alloy and titanium and found no difference between the two types of abutments with regard to peri-implant marginal bone level and soft tissue parameters. Meanwhile, Nevins et al. [26] using en bloc biopsy demonstrated intimate contact of junctional epithelium cells to implant surface and connective tissue with functionally oriented collagen fibres running towards the implant surface designed with Laser-Lok microchannels. Nonetheless, it is unethical to remove implant in order to attain en bloc tissue for histological analyses in human, and data from autopsy did not necessarily represent the ultrastructural nature of the peri-implant interface. In addition, not all animal experiments can be replicated in human samples due to cost and ethical considerations. For this reason, the investigation of peri-implant interface for improvement of connective tissue attachment is rather difficult to conduct in human. Thus, the need of development of different models for histological analyses may be essential.

#### *3.1.4. Three-dimensional oral tissue engineering*

As the opportunity to undertake human studies is limited, many studies that evaluated the peri-implant interface were carried out using animal models. With advances in knowledge on tissue regeneration, tissue-engineered oral mucosal equivalents (three-dimensional oral mucosal model, 3D OMM) have been developed for clinical applications and also for con‐ ducting in vitro studies on biocompatibility, mucosal irritation, disease and other basic oral biological phenomena such as for grafting of oral mucosal defects [50, 51]. The 3D OMM consists of both epithelium and connective tissue layers, grown in the laboratory using collagen membrane as the scaffold. Therefore, evaluation of cell-cell interaction between epithelium, connective tissue and implant surface using 3D OMM is possible and could become an alternative method to study the nature of peri-implant interface. The use of 3D OMM will permit histological preparation and histomorphometric analysis of the interface. With the modification of culture technique, Chai et al. [18] have constructed 3D OMM and have demonstrated the presence of peri-implant tissue with features that mimicked those seen in vivo when tested with titanium. Chai and co-workers [19] further developed the 3D OMM and succeeded in obtaining formed peri-implant-like-epithelium (PILE) on the polished, ma‐ chined, sand-blasted and TiUnite titanium surfaces. Using the 3D OMM, ultrastructural investigation of the soft tissue-implant interface with transmission electron microscopy (TEM) is also possible. It is also interesting to note that the presence of hemidesmosome-like structure as an epithelial attachment to the material surface is shown using this model (**Figure 4**). Moreover, the biological seal of peri-implant tissue can also be demonstrated quantitatively with 3D OMM [52, 53]. This can be carried out via assessment of penetrative behaviour of radioisotope material through the 3D OMM model [52]. Alternatively, the biological seal of peri-implant can also be assessed through the measurement of degree formed by pocket or non-pocket epithelial attachment at the oral mucosal model-material interface [18, 53]. Although only limited study is available on the use of 3D OMM for evaluating the peri-implant interface, this model appears to have a more promising prospect than the monolayer cell culture model. This model is a useful method to evaluate the soft tissue response prior to investigation with an animal model.

*3.1.3. Human studies*

46 Dental Implantology and Biomaterial

histological analyses may be essential.

*3.1.4. Three-dimensional oral tissue engineering*

The composition of the connective tissue interface towards implants was studied in both animal experiments and human biopsy materials. While human studies are very limited due to ethical issues, the evidence of epithelial and connective tissue attachment around periimplant regions are obtained mostly from failed implant [43], autopsy[44, 45] or clinical studies [1, 46], where the presence of connective tissue attachment on these studies is still difficult to demonstrate. Most of the human studies that have been carried out were clinical studies in which the traditional periodontal parameters were used for monitoring the soft tissue responses around dental implants intra-orally. According to clinical studies that involve the marginal bone levels, we can conclude that bone level is stable as it implies that the soft tissue integration has not migrated apically [1, 46, 47]. Liljenberg et al. [48] in their study of soft tissue biopsies of edentulous ridge mucosa and peri-implant mucosa revealed that the composition of both tissues were nearly identical in terms of collagen, cells and vascular structures. The peri-implant mucosa harboured a junctional epithelium that contained significantly enhanced numbers of different inflammatory cells infiltration. On the other hand, Piatelli et al. [44] found that there was no inflammatory infiltrate in epithelium or connective tissue in human autopsy biopsies of titanium dental implants. It is also interesting to note that the collagen fibres in the coronal part were parallel to implant surface while in the apical region the fibres were in a perpendicular fashion was found. Additionally, Glauser et al. [49] used both hard and soft tissue biopsies of mini titanium implants with different surface characteristics to demonstrate the establishment of junctional epithelium attachment to the implant surfaces. They noted that collagen fibres and the fibroblasts were oriented parallel to the implant surface. The oxidized and acid-etched implants revealed less epithelial downgrowth and longer connective tissue than machined implants [49]. As for different types of materials, Vigolo et al. [46] assessed the peri-implant mucosa around abutments made of gold alloy and titanium and found no difference between the two types of abutments with regard to peri-implant marginal bone level and soft tissue parameters. Meanwhile, Nevins et al. [26] using en bloc biopsy demonstrated intimate contact of junctional epithelium cells to implant surface and connective tissue with functionally oriented collagen fibres running towards the implant surface designed with Laser-Lok microchannels. Nonetheless, it is unethical to remove implant in order to attain en bloc tissue for histological analyses in human, and data from autopsy did not necessarily represent the ultrastructural nature of the peri-implant interface. In addition, not all animal experiments can be replicated in human samples due to cost and ethical considerations. For this reason, the investigation of peri-implant interface for improvement of connective tissue attachment is rather difficult to conduct in human. Thus, the need of development of different models for

As the opportunity to undertake human studies is limited, many studies that evaluated the peri-implant interface were carried out using animal models. With advances in knowledge on tissue regeneration, tissue-engineered oral mucosal equivalents (three-dimensional oral mucosal model, 3D OMM) have been developed for clinical applications and also for con‐

**Figure 4.** Hemidesmosome-like structures (black arrows) formed from 3D OMM and specimens (Ti). P = polished and M = machined surfaces. (Reproduced with permission from [19]).

#### **3.2. Analyses of the soft tissue-implant interface**

The soft-tissue implant interface can be investigated through histomorphometric and histo‐ logic analyses. Of both, the preparation for latter analysis is very difficult to carry out especially if the implant is attached to the tissue. The histological studies also allow identification of specific protein markers expressed by any of the tissue or cells in response to dental implants. The histological sections can then be analysed under different types of microscopies. Among the known microscopic analyses for assessing the peri-implant interface are light microscopy (LM), scanning electron microscopy (SEM), confocal laser scanning microscopy (CLSM), focus ion beam (FIB) and transmission electron microscopy (TEM). Similar to SEM, CLSM allows assessment of peri-implant interface and cell-cell interaction without the need of histological processing as for light microscopy. With these two, direct visualization of implant-soft tissue interface is possible with appropriate preparation for each microscopy. The tissue or specimens can also be fluorescently labelled for the identification of adhesion molecules or cells and examined under CLSM [43]. While the use of 3D oral mucosal model with implant material intact may allow direct examination of the connective tissue attachment, the method to prepare the specimen still remains challenging and technically demanding. More studies in term of optimization of certain promising technique such as FIB for TEM analysis must be explored in order to obtain the ultrastructural nature of the implant-soft tissue interface.
