A Review of In Vivo and Clinical Studies Applying Scaffolds and Cell Sheet Technology for Periodontal Ligament Regeneration

Different approaches to develop engineered scaffolds for periodontal tissues regeneration have been proposed. In this review, innovations in stem cell technology and scaffolds engineering focused primarily on Periodontal Ligament (PDL) regeneration are discussed and analyzed based on results from pre-clinical in vivo studies and clinical trials. Most of those developments include the use of polymeric materials with different patterning and surface nanotopography and printing of complex and sophisticated multiphasic composite scaffolds with different compartments to accomodate for the different periodontal tissues’ architecture. Despite the increased effort in producing these scaffolds and their undoubtable efficiency to guide and support tissue regeneration, appropriate source of cells is also needed to provide new tissue formation and various biological and mechanochemical cues from the Extraccellular Matrix (ECM) to provide biophysical stimuli for cell growth and differentiation. Cell sheet engineering is a novel promising technique that allows obtaining cells in a sheet format while preserving ECM components. The right combination of those factors has not been discovered yet and efforts are still needed to ameliorate regenerative outcomes towards the functional organisation of the developed tissues.


Introduction
Periodontitis is a bacteria-driven infectious oral condition that can lead to severe degeneration of periodontal tissues with high prevalence (42.2%) in adults aged 30 years and older [1]. In severe conditions, bone tissue destruction occurs in forms of craterlike defects around teeth roots. These intrabony defects constitute a major challenge to periodontal disease treatment; they can be treated through conventional root scaling and planning healing but without the formation of new supporting tissue [2], while remaining periodontal pockets can cause aesthetic problems and act as triggers to further destruction. Currently, the outermost goal of periodontal therapy is the simultaneous regeneration of all periodontal tissues, i.e., new alveolar bone, cementum, and periodontal ligament (PDL). Towards this direction, surgical treatment is based on the Guided Tissue Regeneration (GTR) approach, during which appropriate membranes are utilized to protect bone defect from epithelial tissue downgrowth, allowing the healing and regeneration of the underlying tissues [3]. Although GTR provides improvement over conventional open flap surgery, there are many factors that can considerably affect the clinical outcomes and as shown in a recent systematic review and meta-analysis, the improvement may not be statistically significant [4], and large observational studies are still needed to clarify the exact role of these factors to facilitate dental professionals to safely apply this technique according to patient specific conditions and demands.
Many studies have attempted to engineer a suitable environment for periodontal tissue regeneration, by applying the appropriate regulatory signals, progenitor cells, extracellular matrix (ECM) or carrier constructs and adequate blood supply, needed to regenerate all periodontal tissues, bone, cementum, and PDL [5][6][7][8][9]. Although the majority of studies have dealt predominantly with bone regeneration and in clinical practice most surgical approaches are based on guided bone regeneration, recent efforts have focused on the regeneration of PDL, along with bone and cementum [10][11][12][13]. The perpendicular alignment of new highly organized collagen fibers, inserted into the regenerated cementum and bone, is the most fundamental aspect of the whole periodontal tissue complex regeneration and emerging efforts are dedicated to this ultimate goal. Periodontal ligament is a fibrous connective tissue lying between alveolar bone and root cementum, occupying a space of 100 to 400 µm. It originates from neural crest-derived ectomesenchyme and is characterized by large heterogeneity in cell populations [14], extensive blood supply [15], and neural network [16]. Development of PDL starts with root formation, before tooth eruption [17,18]. Root formation starts after the formation of enamel and dentin in the area of the future cementoenamel junction, with the Hertwig's epithelial root sheath (HERS) that is formed by the inner and outer enamel epithelium of the enamel organ. HERS is responsible for the shape and number of tooth roots and induces dentin formation from odontoblasts [18]. After root dentin starts to form, HERS is disintegrated and loses attachment with tooth root. However, its remnants are still present in the form of epithelial cell of Mallasez. By HERS disintegration, dental follicle cells (DFCs) come into contact with newly formed dentin, inducing the formation of cementoblasts that start to secrete cementoid tissue, which is then mineralized to cementum. The initiating factor for cementogenesis is the deposition from the HERS cells of enamel matrix proteins on the root surface [19,20]. Cells of the dental follicle differentiate into fibroblasts, which are responsible for the production of PDL fibers. PDL fibers start to grow from both cementum and alveolar bone, and they gradually elongate during tooth development. In the first steps of PDL formation, collagen fibers are loosely configured near cementum and run in parallel with the root of the teeth [21]. The direction of the fibers changes during toot eruption and appears to be affected by the position of the adjacent teeth. When teeth show up in the oral cavity, dento-gingival, transseptal, and alveolar crest fiber groups appear, while after occlusal contact fibers become apparent and in the apical third of the root [22]. Dense Sharpey's fibers appear to emerge from the alveolar bone in the cervical part of the root and extend towards thin cementum anchored fibers occupying the PDL space. They get thicker, are organized in distinct bundles, and gain their final dimensions and orientation after full occlusal function of teeth [23]. The basic characteristic of these fibers is that they are enclosed within cementum and bone, and this is particularly important for the regeneration of PDL.
The PDL fibrous matrix consists of collagen, reticulin, and oxytalan fibers. The 90% of PDL fibers are collagenous, primarily of type I collagen. They provide the structural strength of PDL, while oxytalan fibers that grow during the development of root and the vascularization network in the PDL, seem to play a role in vascular support [24]. It has been reported that HERS, and cementoblasts have a role in oxytalan fibers development during root development, explaining their closer proximity with cementum [25]. Based on their position and orientation PDL fibers are categorized as alveolar crest, apical, horizontal, oblique, and interradicular fibers which lie between the roots of multirooted teeth.
PDL fibroblasts are responsible for the production and maintenance of the extracellular matrix, which is mainly composed of fibers [49]. These cells synthesize and digest fibrillar collagen and produce various bioactive compounds related to wound healing and remodeling [50]. PDL fibroblasts have the capacity to withstand and dissipate the high occlusal loads exerted upon mastication and thus act as a mechanical load absorbing device [51]. Apart from protection against high mastication forces, together with the gingival tissues, PDL forms an effective shield against oral bacteria [52].

Cementum
PDL regeneration requires deep understanding of the hierarchical complexity of dental cementum, as cementum is a basic component of periodontal attachment apparatus that provides anchoring of the principal collagen fibers of the PDL to the root surface. In addition, cementum has an important role in PDL regeneration, as its components and specific microtopography tailor the responses of PDL cells [53]. Cementum is a natural composite containing inorganic hydroxyapatite (HA) nanocrystals and organic matrix rich in collagen fibers (predominantly type I collagen), while non-collagenous matrix proteins like proteoglycans, acidic glycoproteins, growth factors, and attachment proteins occupy the interfibrillar spaces. There are three types of cementum [54]. Acellular cementum covers the cervical two-thirds of the root surface and has a thickness ranging from 50 to 200 µm [55,56]. Its main function is to anchor tooth through periodontal ligament fibers (Sharpey's). It contains cell-free mineralized matrix, densely packed and radially oriented collagen fibers. The apical portion of the root and the furcation areas are generally layered by cellular mixed stratified cementum. Cellular cementum is characterized by a stratified structure with intrinsic and extrinsic collagen fibers. Extrinsic collagen fibers are derived from PDL, while intrinsic contain entrapped cementocytes. Acellular afibrillar cementum is a type of acellular cementum usually found along the cementoenamel junction [57]. It has a thickness of~15 µm and is composed of a matrix with mineralized glycosaminoglycans, but without either cementoblasts or collagenous fibers. Ideally, in periodontal engineering approaches, efforts should be made to regenerate cement-like tissue as close as possible to acellular extrinsic fiber cementum, because that type of cementum is the most appropriate to ensure attachment [55,57].

Cell-Guided PDL Regeneration
Periodontal ligament (PDL) regeneration is a challenging and ambitious task, since it demands a highly coordinated spatiotemporal healing procedure, which includes bone formation within the periodontal defect, along with cementogenesis and PDL fiber formation and attachment on to the root surface [58]. Additionally, challenges arise from the avascular nature of the tooth surface, by the bacterial accumulation along with the techni-cally challenging operating environment due to limited access [58]. Tissue engineering has emerged recently, targeting to potentially regenerate various tissues and organs, including the periodontium [59]. A tissue engineering approach encompasses the use of 3D scaffold, combined with bioactive molecules and cells, and has the potential to regulate the healing process and bypass the abovementioned challenges [59].
As cell-based PDL regeneration attempts have increased over the last years, the application of post-natal progenitor cells has risen, making them an attractive choice for tissue engineering applications. MSCs present the most extensively applied cell type for cell-based regeneration, due to their multi-differentiation capacity, immunomodulation, anti-apoptosis, angiogenesis, and cell recruitment [60]. Different cell types have been applied in previous attempts of cell-based periodontal regeneration, such as bone marrow MSCs (BMMSCs), PDLSCs, gingival fiborblasts (GFs), and dental pulp stem cells (DPSCs).
PDL stem cells (PDLSCs) are the cells that have been isolated from the PDL, and possess characteristics similar to those of MSCs, and a unique potential to regenerate complex PDL tissues [61]. Gingival fibroblasts (GF), which are the most common cell type present in the gingival tissue, are known to modify their behavior and translocate into periodontal defects [62]. It has been found that GF have the ability to form mineralized tissue and express bone-related proteins, while they have been used in regenerative applications, reinforcing the hypothesis that GF possess stem cell characteristics [62]. BMMSCs have demonstrated the ability to proliferate extensively and to differentiate into multiple cell lines; however, their application in periodontal defects has provided some contradictory results [63]. Their effectiveness has been directly related to the morphology of the defect, where increased bone formation has been documented in fenestration and grade III furcation defects, whereas BMSCs application in three-wall intrabony defects had limited effects on new bone formation [63]. Dental pulp stem cells (DPSCs) have also emerged as a potential cell source for tissue engineering applications, as they are easily accessible, can be obtained in large numbers in a non-invasive way, and have multilineage differentiation potential [64]. DPSCs administration has been considered as a possible treatment strategy and DPSCs have been applied in vivo targeting periodontal regeneration. Although their application might be beneficial in terms of bone regeneration, their effectiveness regarding cementum or PDL regeneration is still questionable [65].
Nonetheless, tissue engineering applications do not always render the desired results, due to the immune response triggered by degradation of the scaffolds [58]. Additional problems that arise are low survival of expanded and grown cells in vitro before implantation into the living body, inability of injected cells to attach to the site of implantation, and lack of vascularization or difficulty in revascularization in the site of interest [66,67]. Techniques that utilize extracellular matrix (ECM) production in vitro prior to cell transplantation, such as cell sheets or cell pellets, have gained attention in attempts to overcome the problems encountered by the tissue engineering strategies.
Cell sheet engineering is a novel technique that allows the acquirement of cells in a sheet format, without the application of proteolytic enzymes or other disruptive method, thus enabling the preservation of ECM components [68] (Figure 1). Different methods have been employed to harvest cell sheets, such as the use the temperature responsive culture dishes, the use of polymerized human fibrin-coated dishes, and the use of Vitamin C (Vc) treatment [69]. The use of temperature-responsive culture dishes was the first method applied to obtain cell sheets, and has been the most extensively implemented, with the utilization of Poly(N-isopropylacrylamide) (PIPAAm), which is a temperature responsive polymer [70]. A smart biointerface from PIPAAm was developed, which allowed the control of cells attachment through the manipulation of temperature [71]. In normal cell culture conditions of 37 • C, the surface remains hydrophobic, allowing cells to attach and proliferate, and changes into hydrophilicity below the critical tempeature of 32 • C, leading to cells detachment from culture surface without the use of proteolytic enzymes [70]. This technique exhibits numerous advantages over conventional methods, as the cells preserve the integrity of adhesion proteins, such as E-cadherin and laminin 5, retain its ECM components secreted by the cells, and have minimal cell loss [70]. Cell sheets can be directly applied into a defect area either as a coating, as numerous cell sheets can overlap each other, creating a three-dimensional structure, or even shrink and create a cell pellet, that can be applied as a graft into the area of interest. Different methods have been employed to harvest cell sheets, such as the use the temperature responsive culture dishes, the use of polymerized human fibrin-coated dishes, and the use of Vitamin C (Vc) treatment [69]. The use of temperature-responsive culture dishes was the first method applied to obtain cell sheets, and has been the most extensively implemented, with the utilization of Poly(N-isopropylacrylamide) (PIPAAm), which is a temperature responsive polymer [70]. A smart biointerface from PIPAAm was developed, which allowed the control of cells attachment through the manipulation of temperature [71]. In normal cell culture conditions of 37 °C, the surface remains hydrophobic, allowing cells to attach and proliferate, and changes into hydrophilicity below the critical tempeature of 32 °C, leading to cells detachment from culture surface without the use of proteolytic enzymes [70]. This technique exhibits numerous advantages over conventional methods, as the cells preserve the integrity of adhesion proteins, such as E-cadherin and laminin 5, retain its ECM components secreted by the cells, and have minimal cell loss [70]. Cell sheets can be directly applied into a defect area either as a coating, as numerous cell sheets can overlap each other, creating a three-dimensional structure, or even shrink and create a cell pellet, that can be applied as a graft into the area of interest.
The aim of this review was to discuss recent advancements and strategies for PDL regeneration in terms of clinical outcomes derived from in vivo models and clinical studies, by applying cell sheet technology and scaffold constructs. A search strategy was applied to include most of the available literature in Web of Science, Pubmed and Scopus databases. Search terms included "periodontal", "periodontal ligament", "regeneration", "in vivo", "scaffolds", "clinical trial", "clinical study", "clinical", "stem cells", "progenitor cells", "precursor cells", "pluripotent stem cells", "multipotent stem cells", "embryonic stem cells", "ips cells", "somatic cells", "mesenchymal stem cells". Hand searching from selected review articles and other included articles was also performed.

In Vivo Studies
The included in vivo studies employed the intrabony/furcation periodontal defect model or the fenestration periodontal defect mode ( Figure 2) as an orthotopic model and different ectopic models to test the regenerative capacity of tissue engineering constructs consisting of scaffolds, matrices, membranes, hydrogels etc., either loaded with growth factors or seeded with different cells/cell sheets. The aim of this review was to discuss recent advancements and strategies for PDL regeneration in terms of clinical outcomes derived from in vivo models and clinical studies, by applying cell sheet technology and scaffold constructs. A search strategy was applied to include most of the available literature in Web of Science, Pubmed and Scopus databases. Search terms included "periodontal", "periodontal ligament", "regeneration", "in vivo", "scaffolds", "clinical trial", "clinical study", "clinical", "stem cells", "progenitor cells", "precursor cells", "pluripotent stem cells", "multipotent stem cells", "embryonic stem cells", "ips cells", "somatic cells", "mesenchymal stem cells". Hand searching from selected review articles and other included articles was also performed.

In Vivo Studies
The included in vivo studies employed the intrabony/furcation periodontal defect model or the fenestration periodontal defect mode ( Figure 2) as an orthotopic model and different ectopic models to test the regenerative capacity of tissue engineering constructs consisting of scaffolds, matrices, membranes, hydrogels etc., either loaded with growth factors or seeded with different cells/cell sheets. Twenty-six studies used ectopic models to assess the periodontal regenerative capacity of cell sheet transplantation (Supplementary Table S1). Twenty-five of these studies used nude mice, while one study used rats. The included studies used a variety of biomaterials in an attempt to simulate the orthotopic conditions and ectopically assess the potential for periodontal regeneration of each cell sheet, such as ceramic bovine bone (CBB), chemical conditioned root dentin (CCRD), dentin block, polyglycolic acid (PGA) film, gelfoam scaffold, treated dentin matrix (TDM), hydroxyapatite/tricalcium phos-

Cell Sheet Engineering Ectopic Models
Twenty-six studies used ectopic models to assess the periodontal regenerative capacity of cell sheet transplantation (Supplementary Table S1). Twenty-five of these studies used nude mice, while one study used rats. The included studies used a variety of biomaterials in an attempt to simulate the orthotopic conditions and ectopically assess the potential for periodontal regeneration of each cell sheet, such as ceramic bovine bone (CBB), chemical conditioned root dentin (CCRD), dentin block, polyglycolic acid (PGA) film, gelfoam scaffold, treated dentin matrix (TDM), hydroxyapatite/tricalcium phosphate (HA/TCP), titanium (Ti), teeth roots, platelet-rich fibrin (PRF) fabricated into bioabsorbable fibrin scaffolds, decalcified dentin matrix (DDM), polycaprolactone (PCL) scaffold, Matrigel, and micro/macro-porous biphasic calcium phosphate (MBCP) blocks. Two studies implanted the cell sheet/material complex into jawbone implant sockets, using bioengineered tooth root (bio-root) structure from HA/TCP, wrapped with the cell sheet [72,73]. Furthermore, two studies assessed the regenerative potential of cell sheets combined with titanium samples [66,74]. In the study by Washio et al. [66], hPDLCs sheets/titanium implant complexes were transplanted into mandibular bone defects, were histological observation demonstrated the formation of cementum and PDL-like tissue on titanium surface. Those findings support the prospect of future efforts towards the formation of a stable periodontal complex around dental implants.
Several different pretreatments were used in the included studies with intendence to enhance the regenerative potential of cell sheets. Five studies used the Vc pretreatment as the method of choice for the cell sheet fabrication [72,73,75,82,98]. Li et al. [91] assessed the effect of low-intensity pulsed ultrasound (LIPUS) stimulus on PDLSC sheet formation and periodontal tissue regeneration in vivo. Their results highlighted the positive effect of LIPUS-treated PDLSC sheets on ECM synthesis and PDL-like tissue regeneration compared with the untreated PDLSC sheet group. The effect of pretreatment of human PDLSC (hPDLSC) sheets with recombinant human bone morphogenetic protein-2 (rhBMP-2) targeting the regeneration of dental cementum and the periodontal complex was evaluated in an ectopic model of nude mice [81]. Pretreated hPDLSC sheets exhibited significantly more mineralization and collagen ligament accumulation as compared with the control group, thus enabling the formation of PDL cementum-like complex [81]. Yang et al. [76] assessed the effect of conditioned medium (CM) from developing apical tooth germs on hPDLSC sheets, that were then transformed into cell pellets to be used for periodontal tissue engineering. Cementum-like mineralized tissues and PDL-like fibrous tissues were identified in the CM treated group, whereas control group cultured without CM rarely formed cementum/PDL-like tissue [76]. Platelet-rich derivatives were used in two studies, one used platelet-rich plasma (PRP) as pretreatment, while the study by Wang et al. used platelet-rich fibrin (PRF) as a bioabsorbable scaffold [78,85]. PRP pretreatment resulted in significantly enhanced osteogenic differentiation of PDLSCs and increased bone and collagen formation in vivo compared with untreated control [85]. Moreover, the use of PRF as a bioabsorbable scaffold was more beneficial in terms of PDL and bone tissue formation when combined with jaw BMMSC instead of PDLSC sheets [78].
Five studies assessed the effect of coculture of different cells on the properties and effectiveness of cell sheets, as well as their regenerative abilities [77,84,[87][88][89]. Coculture of PDLSCs with a different cell line seems to be beneficial on the properties of the cell sheet that results from the coculture system. More specifically, hPDLSCs were cocultured with hBMMSCs, and the mixed cell sheet was used to create a cell pellet which was applied in vivo for ectopic transplantation, showing enhanced cementum/PDL-like tissue regeneration with neovascularization when compared to the non-mixed cell pellet [77]. Furthermore, the in vivo application of cell sheets from the coculture of PDLSCs with either urine-derived stem cells (USCs) or jaw BMMSCs resulted in increased expression levels of bone-and ECM-related genes and proteins and led to the formation of a complex tissue like the native periodontal tissue [87][88][89]. Panduwawala et al. fabricated triple-cell sheets from PDLSCs and human umbilical vein endothelial cells (HUVECs), either from combination of the different cell sheets (PDLSCs-HUVECs-PDLSCs) or cell sheets from the coculture of these cells and found that both conditions resulted in periodontal fiber formation similar to PDL, as well as vascular lumen formation [89]. Liu et al. assessed the regenerative capacity of PDLSCs from healthy subjects (HPDLSCs) and patients diagnosed with periodontal disease (PPDLSCs) when cocultured with DFCs [84]. DFCs seem to enhance the stemness of both HPDLSCs and PPDLSCs, and the cocultured HPDLSC sheet managed to regenerate the PDL complex, whereas in the case of PPDLSC sheet, fibers did not adhere well while inflammatory cells were also present in the regenerated tissue [84].

Orthotopic Models
Twenty of the studies used orthotopic models to assess the regenerative capacity of cell sheet transplantation in periodontal defect models, with or without biomaterials (Table 1).    The studies investigated the potential of cell sheets towards periodontal tissue regeneration in vivo through various animal models and experimental strategies. Different periodontal defect models were used, where two studies used one-wall bone defect model [94,101], two studies used three-wall intrabony defects [104,105], and each of the following defect models were used in one study, a dehiscence defect model [110], a class III furcation defect model [92], a two-wall intrabony defect model [107], a horizontal defect model [106], and a fenestration defect model [99], while the rest of the studies did not specify the morphology of the defect. Six studies assessed cell sheet application in a rat model [86,94,99,103,109,112], six studies used dogs as the animal model of choice [101,102,[106][107][108]111], four studies used miniature pigs [75,100,104,105], two studies used rats [90,97], one study used sheep [110], while there was also 1 clinical study [8].
To produce the cell sheets, the different orthotopic studies used a variety of cells. Most of the included studies (eight studies) formed cell sheets from PDLSCs [75,86,90,[100][101][102][103]106], followed by PDLCs (six studies) [8,97,107,109,110,112], BMMSCs (three studies) [101,108,110], and dental follicle stem cells (DFCs) (three studies) [94,107,111], gingival fibroblasts (two studies) [99,110], DPSCs (two studies) [104,105], while the following cells, osteoblastic cells [97], alveolar periosteal cells (APCs) [101], and stem cells from human exfoliated deciduous teeth (SHEDs) [95], were used in one study, each. Tsumanuma et al. [101] assessed the effect of three-layered cell sheets from different cell lines in one-wall surgically created defects in dogs and found that the application of PDLC sheets resulted in more newly formed thick acellular/cellular cementum, denser collagen fibers and enhanced PDL formation compared to the BMMSC and APC sheets groups. In the study by Guo et al. [107], the effectiveness of DFC sheets and PDLC sheets towards periodontal regeneration was assessed in a two-wall intrabony defect in dogs. While new periodontal attachment was observed in both groups, complete periodontal regeneration involving PDL and cementum was detected only in the DFC sheet group, which also exhibited enhanced bone formation when compared to the PDLC sheets. Whereas the study by Yang et al. showed similar periodontal regeneration potential between DFC sheets and SHED sheets [94]. More specifically, the regenerated tissues observed in both experimental groups were all consisting of fibroblasts and collagen fibers, which were perpendicularly arranged and well organized, similar to that of native PDL [94]. Another study showed the superiority of the application of a complex cell sheet containing two cell lines, PDLCs and osteoblastic cells, against the application of each single cell sheet containing either cell line [97]. In detail, complex cell sheet application resulted in new bone formation and complete PDL regeneration, restoring the functional connection between the alveolar bone and tooth root, whereas control groups exhibited incomplete recovery in both mineralized tissue and soft tissue formation [97]. When comparing three different cell sheets in a surgically created dehiscence periodontal defects in sheep, Vaquette et al. found that BMMSC and PDLC sheets demonstrated similar results in terms of new bone formation, PDL and cementum regeneration after 10 weeks, whereas both groups exhibited superior regenerative potential when compared to GF sheets [110].
Wei et al. assessed different methods for the obtainment of cell sheets, the use of temperature responsive culture dishes and the application of Vc, and its effect on periodontal regeneration potential of PDLSC sheet [75]. Vc-induced PDLSC sheets application into the defect area resulted in increased bone/cementum-like matrix formation, which was significantly higher compared to the PDLSC sheets from the temperature responsive culture dishes [75]. Two studies assessed the effect of different pretreatments, such as inflammatory stimulation or hypoxia, on the regenerative potential of cell sheets [90,107]. The study by Yu et al. showed that 24-hour hypoxic pretreatment of PDLSCs enhanced their regenerative potential in vivo in terms mineralized tissue and cementum formation, and PDL regeneration [90].
The cell sheets in the different studies were combined with a variety of biomaterials, such as HA/TCP, CBB, TDM, Matrigel, gel foam scaffold, platelet-rich fibrin granules, polycaprolactone scaffold, polyglycolic acid (PGA), and porous β-TCP.
The effect of platelet-rich fibrin granules on PDLSCs sheet fragments targeting periodontal regeneration was assessed in the study by Zhao et al. [102]. In the tooth reimplantation model used in this study, the combined application of PDLSC/PRF was more effective in regenerating PDL-like tissues and avoiding ankylosis and inflammation, compared to the other groups [102].
Iwasaki et al. used a decellularized amniotic membrane (amnion), instead of an engineered cell sheet, with or without PDLSCs in a surgically created periodontal defect in rat maxillary molars; and found that the presence of PDLSCs enhanced periodontal tissue regeneration four-weeks post-transplantation, as indicated by the radiological and histological analysis [103]. In the study by Jiang et al., decellularized sheets from human PDL cells were combined with 15-deoxy-∆12,14-prostaglandin J2 (15d-PGJ2) nanoparticles along with or without polycaprolactone/gelatin (PCL/GE) nanofibers as potential candidates for periodontal regeneration in rat periodontal defect model [112]. The application of decellularized hPDLCs sheets resulted in successful bone tissue ingrowth, as well as cementum-like and PDL-like tissue formation on the root the mandibular first molar, despite the presence or absence of PCL/GE nanofibers [112]. Farag et al. [109] assessed the effect of decellularized PDLSCs sheet combined with PCL scaffold on periodontal regenera-tion in a rat periodontal defect model, where the beneficial role of the decellularized matrix was demonstrated. More specifically, the decellularized sheets were infiltrated with cells, and exhibited significantly higher new attachment of periodontal fibers when compared with the PCL scaffolds alone, while the regenerated PDL fibers were more organized and inserted with a perpendicular allignemnt into the root surface [109]. In the study by Yang et al. [111], the use of TDM particles or HA/TCP combined with DFCs resulted in increased bone formation when compared to the groups without materials. Furthermore, the presence of DFCs had a positive effect on the density of bone formation and the extent of PDL-like tissue formation compared to the control group without cells [111].

In Vivo Studies with Scaffolds for PDL Regeneration
A large variety of biomaterials in the form of simple, biphasic, or multiphasic scaffolds have been proposed for the regeneration of damaged periodontal tisses. These 2D and 3D constructs have been developed based on concepts of complete regeneration of the periodontal apparatus (bone, PDL, and cementum) or partial regeneration of specific compartments, such as PDL or bone/cementum tissues, with most of them investigating the osteogenic capacity of their materials focusing on bone regeneration. As GTR still constitutes the "gold standard" in periodontal surgical interventions, various degradable or non-degradable membranes (2D structures) have been utilized to prevent epithelium downgrowth to allow a smooth healing of damaged periodontal connective tissue and ligament. The use of scaffolds aims to develop biocompatible and bioactive platforms that, with the help of other attached or loaded molecules and growth factors, can lead to timely and guided cell migration, proliferation, and differentiation to promote tissue regeneration. Towards this direction, different animal species (dogs, miniature pigs, rats, etc.) and in vivo models have been evaluated, including scaffolds placement in surgically created periodontal defects or ectopic tissue formation by subcutaneous implantation in animals.

Periodontal Defect Model
Most in vivo studies evaluating scaffolds employed the intrabony/furcation periodontal defect model or the fenestration periodontal defect model ( Table 2). Surgical creation of one-or two-wall bone defects near the roots of molars and premolars, PDL and cementum removal and scaffold placement in close proximity to root dentin, are the major steps for the furcation model, while defects with standardized height (usually 5 mm in apico-coronal direction at the furcation region) are created around premolars or molars in the furcation model.    Beagle dogs have been widely used in studies evaluating various therapeutic strategies for periodontitis and the regeneration of periodontal tissues with scaffolds [137]. The rational for their use lies in the similarities of their periodontal tissues' architecture and oral microflora with humans [138]. In addition, proper hygiene can be achieved without sedating the animals, which ensures animals convenience and low risk of complications to proper healing and regeneration [139,140]. Significant limitations of dog models are that dogs do not exert lateral movements during mastication and that they present greater bone remodeling rate that could yield high regenerative potential and subsequently to optimized results of in vivo studies [141,142]. Two different surgical protocols are used; one comprising the creation of supraalveolar critical-size furcation defects and the other creating intrabony defects [143]. Hydrogels [13,125] composite collagen-hydroxyapatite scaffolds [123], bioceramic [116,144] and bioactive glass scaffolds [122], nanoparticlesloaded polymeric or collagen scaffolds [124], and combined micropatterned polymeric scaffolds [130], either cell loaded or not, have been tested in the beagle or other dog models. The whole periodontal complex was regenerated, with obliquely inserted ligament-like fibers when a biphasic scaffold consisting of gelatin and β-TCP/HA particles (BH) and biphasic cryogel scaffold (BCS) was implanted in beagle dogs, loaded with BMP-2 and protected by a functionally graded membrane [13]. The use of FGF-2 was advantageous in combination with a nano-β-TCP collagen scaffold for the development of accelular cementum on the surface of exposed roots and the formation of PDL-like tissue [124]. Similar were the findings of Momose et al [125], with the use of a collagen hydrogel scaffold loaded with FGF-2; however, both studies although verified the presence of PDL-like fibrous tissue, they did not observe tissue attachment and functional Sharpey's fibers formation. Bioceramic diopside ceramics proved more efficient in producing large quantities of bone, cementum and well-oriented collagen fibers compared to β-TCP, that presented only limited new bone or osteoid deposition [144]. Cell-loaded chitosan/anorganic bovine bone composite scaffolds and collagen sponges presented greater volumes of new bone and cementum, with dense PDL fibers [115], in an oblique or perpendicular orientation [127]. On the other hand, limited positive effect was reported from Liu et al, that used collagenhydroxyapatite scaffolds loaded with BMSCs, possibly explained from the limited survival of cells within the scaffold due to the poor blood supply of labial alveolar bone [123].
Miniature pigs have been used as a more convenient and reliable animal model in many studies in dentistry, but surprisingly very few studies exist on testing scaffolds for PDL regeneration in miniature pigs [113,118]. Pigs' bone anatomy and morphology, healing, and rate of remodeling are considered to be close to those of humans, and therefore is a suitable animal species, as evidenced from a lot of studies in recent years [137]. In addition, pigs have anatomically and functionally temporomandibular articulation close to that in humans and as omnivores, they masticate with lateral jaw movements, representing a more suitable model for mimicking the mastication cycle [145]. However, they have inherent limitations, such as larger teeth surrounded of large bone volume, long junctional epithelium, and different oral microflora. Only a few studies have used the periodontal defect model in the miniature pig in combination with scaffolds for PDL regeneration [113,118]. Hybrid tooth constructs from PGA/PLLA and PLGA scaffolds for tooth and bone parts respectively were seeded with DSCs and remained for 12 and 20 weeks [113]. Despite the new cementum formation, periodontal ligament fibrous tissue resembling natural Sharpley's fibres was found scarcely. A hyaluronic hydrogel scaffold releasing IL-1-resceptor antagonist was used in an effort to optimize regeneration by restricting the inflammatory stage of periodontal wound healing [118]. Although bone and cementum-like tissue were formed and PDL-like fibers were anchored to cementum, no distinct effect was observed in the group with the IL-1-resceptor antagonist.
Rat species is the most used animal model implementing the periodontal defect model in recent articles. Despite rats presenting continuous teeth eruption and periodontal remodeling with cementum and bone apposition, that can yield optimized results in respect to potential regenerative materials and approaches, their ease of handling and low maintenance cost, along with low ethical or social concern, have made them prevail in periodontal tissue regeneration studies (Table 2). In general, the rat periodontal fenestration defect model with an extraoral (buccal) surgical approach is used and is widely adapted as a valid model before proceeding to larger animal testing. The advantage of this approach is that the chance of gingival tissue ingrowth is eliminated, however it is technically more demanding [143]. Athymic nude, Sprague-Dawley, Fischer 344, and Wistar rats have been used in different studies in combination with varying scaffolding materials for PDL regeneration. Hydrogel scaffolds [128,131,133] and various fibrous polymeric constructs like PCL [12,109,114,132], PLGA/PCL [122,130], PCL/PEG [120], and PLGA [136] scaffolds or membranes have been implanted in periodontal defects in rats, fabricated mostly by electrospinning. A biomimetic F/CaP coating process was applied on PCL scaffolds and was more effective in creating new alveolar bone, PDL, and cementum compared to uncoated scaffolds. Cell seeding with primary HPDLs cell sheets and PDLSCs on PCL scaffolds resulted in well-organized periodontal tissue complex with PDL fiber angulation similar to native tissue. On the contrary, when simple PCL scaffolds were used [12,109], PDL-like tissue was aligned parallel to the root surface, with few fibers inserting the cementum layer. A biphasic scaffold was developed, with micropatterned PLGA/PCL compartment for PDL regeneration and amorphous PCL for bone formation [130]. Each compartment was seeded with different cells, i.e., the PDL compartment with hPDLs and bone with hGFs and modified to incorporate vectors encoding BMP-7 (bone compartment) and PFGD (PDL compartment). Different combinations were evaluated in terms of either PDL compartment micropatterning or not, and gene delivery, and the optimum outcomes regarding PDL formation aligned obliquely were received for micropatterned PDL irrespectively of the single or dual gene delivery.
Although rabbits have been used in evaluating therapeutic factors for treatment of periodontitis [146,147], their use in studies evaluating periodontal or peri-implant tissues regeneration is very limited [148,149]. This is due to their bone composition and remodeling rate which is different to human. Sowmya et al. [10] created defects in the maxilla of New Zealand White Rabbits and implanted tri-layered nanocomposite hydrogel scaffolds loaded with FGF 2 (PDL compartment) and platelet-rich plasma (PRP)-derived growth factors. They concluded that although PDL, cementum and bone was developed in the tri-layered scaffolds, more organized bone tissue was developed when the growth factors were used.
Other animals for PDL regeneration include sheep and mice, with very limited available data. In a recent study, Vaquette et al [110] used an ovine periodontal defect model to evaluate if different cell types for seeding biphasic electrospun PCL/β-TCP scaffolds could exert a different effect on PDL regeneration and concluded that although robust cementogenesis and PDL regeneration was evidenced in cases where PDLCs and Bm-MSCs were seeded, no cementum formation was observed when GCs were used. In the study of Zheng et al. [129], β-TCP scaffolds were seeded with gene-transfected BMSCs and implanted in nude BALB/c mice periodontal defects. Cell-seeded scaffolds were able to regenerate PDL and cementum, but defects filled with neat β-TCP scaffolds presented only fibrous tissue formation without new cementum or oriented fibers.
To summarize the findings, scaffolds alone cannot promote PDL regeneration and anchoring into new bone and cementum, irrespective of their composition or structure. In most of the studies, the use of cell seeding or loading scaffolds with growth factors was more effective in providing not only higher bone volume but also obliquely or perpendicular attachment of newly formed PDL fibers. Multiphasic scaffolds, or patterned scaffolds that mimic the structural compartments of periodontal tissues, provided the topographical cues necessary for cells to promote regeneration of PDL and the whole periodontal tissue complex. Another additional property of scaffolds towards functional PDL fibers orientation is the presence of a calcium-based component, although there is no clear evidence whether its chemical similarity to bone and cementum or its topographical orientation is the prevailing factor that guides PDL regeneration.

Subcutaneous Placement Model
Another commonly applied model for evaluating PDL regenration involves the subcutaneous placement of scaffolds in the dorsum of athymic or nude mice and rats. Pockets of certain dimensions are surgically created on the back of rats and the materials are implanted subcutaneously. This model has the disadvantage of not resembling the actual clinical conditions in periodontal area, especially in terms of oral microflora. Subcutaneous placement of scaffolds yields results regarding the induction of any inflammatory reactions and vascularization [150]. This model is commonly applied as tissue ingrowth occurs and angiogenesis can be validated, and although timing is not the same as with human histological findings, safe predictions can be made validating the clinical translation of the model [151].
Regarding PDL regeneration, a few studies have used this model for pre-clinical testing of scaffolds (Supplemenatry Table S2). Biphasic or multiphasic scaffolds with PCL and β-TCP [152], or HA [153] have been tested in combination with osteoblasts, PDLCs sheets, or DPSCs. PCL/HA micropatterned 3D printed scaffolds with spatiotemporal delivery of recombinant human amelogenin, CTGF and BMP 2 from PLGA microspheres, seeded with DPSCs developed CEMP1 + mineralized tissue and aligned collagenous fibers resembling PDL-like tissue, while in the absence of biological cues (scaffolds without microspheres) similar tissue characteristics were received, although suboptimal [153]. Strong attachment of PDL and higher bone apposition was corelated with the CaP coating of a PCL/β-TCP scaffold seeded with PDLCs [152]. To better mimic periodontal tissue architecture in an ectopic rat model, dentin matrix or slices have been used in association to biphasic composite scaffolds [92,[154][155][156]. Dentin is either treated with 37% orthophosphoric acid to expose dentin tubules [155] or treated with EDTA [92]. A patterned PDL-like layer was designed on top of dentin surfaces, with multiple perpendicularly oriented channels to guide fibroblasts alignment and increase vascularization [155]. Although this patterning resulted in vascular structures and fibers development aligned along the PDL-like layer, cementum was formed only in the case of cell seeded scaffolds. A fiber-guiding microchannel pattern from chitosan of low molecular weight with pores of 450µm and high elastic modulus was successful in guiding fibroblasts and promoting PDL regeneration [157]; however, in agreement with the study of Vaquette et al [154], PDL lacked functional orientation. Biomimetic fabrication of scaffold microarchitecture is crucial to functional orientation of new PDL tissue formation, as evidenced also by Yu et al [11], who used a bilayer scaffold of mineralized collagen and concentrated growth factor in comparison with a deproteinized bovine bone mineral. They concluded that apart from the hierarchical PDL-like microenvironment of the scaffold, its stiffness, degradation rate similar to natural bone and the good interfacial stability of the two scaffold components, allowed the smooth healing and regenerative process, leading eventually to functionally oriented PDL fibers inserted in the newly formed cementum tissue.
Electrospun membranes have been applied to mimic the PDL and are used as an intermediate layer between dentin and the bone compartment of scaffolds [154] based on their use as suitable materials for GTR in periodontal tissues engineering. Other desirable properties of electrospun membranes are their efficacy in loading nanoparticles, antibiotics and/or growth factors. Vaquette et al [154] utilized a PCL membrane to stabilize PDL cell sheets and verified that the heat press-fitting treatment they employed improved the membrane adhesion to the scaffold and facilitated new PDL formation and attachment, although its orientation was not perpendicular to new cementum. New fibrous tissue along the dentin surfaces with large areas of no attachment were observed in mesoporous HA/chitosan scaffolds, while new cementum and continuous soft tissue formation were observed when these scaffolds were loaded with recombinant human amelogenin.
Based on the results of the included studies a calcium phosphate mineral [92,152] or complex PDL-like patterned structures [153,155] should be present to induce the perpendicular orientation of new PDL fibers.

Other Models
Other in vivo models include the extraction of teeth and implantation in jawbone sockets, implantation in calvaria defects and regeneration after experimental periodontitis model in maxillary molars. HA/TCP scaffolds, electrospun scaffolds or membranes from either PCL or PLGA and PEG-DA based hydrogels have been evaluated (Supplementary  Table S3). Smart antibacterial hydrogels with capacity to control inflammation were used after loading with SDF-1 in an experimental periodontitis model and resulted in complete in situ periodontal regeneration with arrangement similar to normal periodontium [157]. Despite the materials used as scaffolds after implantation in jawbone defects after teeth extraction, functional regeneration of PDL was observed, along with complete regeneration of cementum and bone [72,158]. Electrospun sheets, with parallel-aligned fibers similar to ECM were efficient in allowing topographical alignment of cells to guide the development of organized PDL tissues [158]. In the model of Kim et al [159], in the case of PDL/bone removal after teeth extraction and reimplantation with aligned PCL/gelatin membranes, the authors concluded that when cells were cultured on aligned membranes under cycling mechanical loading, the combination of alignment and load was efficient in regenerating all tissues, however PDL fibers alignment deviated from normal. Nevertheless, when PDL remained in the sockets, the regenerated bone was in close contact to intact PDL without interfering connective tissues. In a rat calvaria model, Chen et al [160] evaluated electrospun multiphasic scaffolds of PCL, type I COL, and PEG-stabilized ACP nanoparticles loaded with rhCEMP. They reported cementum-like tissue formation, limited bone regeneration, and thick connective tissue formation with parallel oriented fibers.

Clinical Studies Involving Scaffolds and Growth Factors
The regenerative therapies in treatment of periodontitis implicate various bone grafts and GTR aimed to promote de novo formation of periodontal complex (Table 3). Although much effort was made in development of new scaffolds for periodontal tissues regeneration at preclinical level, there are limited data on their clinical application.    Safety and efficacy of natural and synthetic scaffold materials currently used in clinical periodontology are well documented. Nevertheless, a new synthetic zinc-substituted nanostructured material based on monetite (Sil-Oss ® ) for the treatment of intra-bony defects was developed and tested in 30 patients [176]. The authors did not find significant differences with synthetic HA in terms of clinical findings and bone mineralization; however, the new material showed a significant increase in bone fill percentage as compared with HA at 3 and 6 months of observation.
With development of CBCT and 3D printing technologies it became possible to construct patient specific scaffolds. 3D printing provides better control over the scaffold microarchitecture and allows fabrication of complex multistructures, recreating features of bone, cementum, and PDL. In 2015, for the first time Rasperini et al. [169] fabricated custom-made PCL scaffold using selective laser sintering and implanted it into a large periodontal defect of a 53-year-old male. The scaffold contained an internal compartment for rhPDGF-BB delivery and extended pegs for PL regeneration and guidance. After one year of follow there were no signs of chronical inflammation and the scaffold remained covered, favoring partial root coverage, and 3 mm of clinical attachment gain. However, it later became exposed to the intraoral environment, contaminated by microbes and, consequently, lost. Apparently, the slow degradation time of PCL polymer and mismatch of its mechanical properties with the surrounding tissues was the reason for its failure. Another clinical case of horizontal alveolar bone augmentation by 3D printed bioceramic (30% HA-70% β-TCP) scaffold was recently reported by Mangano et al. [181]. The authors demonstrated histological and histomorphological assessment of this retrieved scaffold after seven years of implantation. Interestingly, despite complete integration of biomaterial, which remained unloaded for so many years, it was not fully resorbed and even preserved its initial microarchitecture.
In recent years, the biological growth factors have been extensively applied for periodontal regeneration as components of biomimetic scaffolds or locally in a form of solution. Local delivery of bioactive molecules by scaffolds can create favorable microenvironment for differentiation of stem cells in the surrounding periodontal tissues. The most frequently used delivery system for growth factors is tricalcium phosphate (β-TCP). Due to its porous microstructrure it entraps biological factors, helps in stabilization of blood clot, and serves as a scaffold for new bone formation.
PDGF was discovered in 1989 by Lynch and co-workers and since then is thoroughly investigated as a treatment option for regeneration of the whole periodontal complex. PDGF can bind the superficial receptors of periodontal ligament cells and bone cells and enhance their chemotaxis and proliferation [184,185]. It can also promote angiogenesis and wound healing by stimulation of VEGF release [186]. Sarment et al. [162] found that locally delivered rhPDGF stimulates release of pyridinoline cross-linked carboxyterminal telopeptide of type I collagen and enhancing bone turnover.
Nevins et al. [167] demonstrated encouraging long-term results of rhPDGF application for treatment of large osseous defects. During the observation period of 36 months, they noticed significant clinical and radiographic improvements in patients that were treated by rhPDGF-BB at 0.3 mg/mL with TCP as compared to other treatment modalities. The authors defined the positive treatment outcome when clinical attachment level inceased by 2.7 mm and linear bone growth was higher than 1.1 mm.
Safety and efficacy of rhPDGF was documented by Jayacumar et al [166]. They found significantly higher linear bone growth (3.7 mm vs. 2.8 mm) and bone fill (65.6% vs. 47.5%) in the experimental group as compared to the TCP control after 6 months over baseline, while no pronounced adverse effects such as pain, fever or swelling were noticed. Similar results were reported by Maroo and Murphy in 2014 [168], who have found even greater gain of the amount (4.05 mm vs. 2.50 mm) and percentage of defect fill (94.3% vs. 68%) at the rhPDGF treated sites compared to TCP after 9 months of follow up.
Another application of rhPDGF for management of gum recession has been reported by McGuire and Scheyer [165]. In this case series, it was suggested that application of TCP, rhPDGF, and collagen membrane is similarly effective for the reduction of gingival recession as subepithelial connective tissue graft. Three years later, a randomized clinical trial of this group demonstrated the regeneration of the periodontal complex when using rhPDGF-mediated therapy through histological analysis. Particularly, in 9 months post surgery, the rhPDGF-treated sites showed oblique orientation of the Sharpey's fibers and their insertion into newly formed cementum [165].
Several clinical trials utilize FGF-2 in combination with resorbable scaffolds for periodontal defect treatment [171,172,178]. FGF-2 stimulates angiogenic and mitogenic activity of periodontal ligament MSCs and plays an important role in wound healing. Randomized clinical trials showed that administration of 0.3% of rhFGF-2 significantly improved the bone fill percentage, and its superior efficacy compared to EMD treatments [171,172]. Periodontal therapy with FGF-2 was found to be efficient in smokers.
Platelet-rich plasma (PRP) and platelet-rich fibrin (PRF) contain a cocktail of growth factors such as PDGF, VEGF, insulin-like growth factor, and transforming growth factor which accelerate wound healing and new bone formation. Recent meta-analysis pointed out favorable clinical outcomes for treatment of periodontal intarabony defects using PRF in combination with bone graft materials [187]. However, yet there is no histological evidence in clinical studies that blood plasma coagulation can promote true periodontal regeneration.

Clinical Studies Involving Caffolds Combined with Cells
Another promising approach for periodontal tissue regeneration is stem cell-based therapies There are several sources of stem cells that were described in literature such as bone marrow, PDL and dental pulp, exfoliated deciduous teeth, gingival and human umbilical cord. Isolation of allogenic MSCs from human umbilical cord is not invasive and is less expensive than the convetional cell isolation procedures. Besides, these cells are pluripotent, as they can differentiate into osteoblasts, cemetoblasts and PL fibroblasts and are capable for self-renewing. In this direction, Dhote et al. [188] applied umbical cord MSCs in combination with β-TCP scaffold and platelet-derived growth factor-BB (PDGF-BB) and found significant gain of clinical attachment level and radiographic defect fill as compared to open flap debridement (OFD). Ferrarotti et al [189] used autologous DPSC micrografts collected and directly seeded on collagen sponges for treatment of chronic advanced periodontitis. The authors reported improved clinical and radiographic parameters as compared to control sites treated with collagen sponges alone. Safety and favorable clinical results were reported also by Baba et al. [190], who treated inrtabony defects in 10 patients with autologous bone marrow stem cells in complex with PRP in a woven-fabric composite poly-L-lactic acid scaffold. However, there was no control group for further comparisons. By contrast, Chen et al. [191] demonstrated no significant differences in clinical and radiographic findings between demineralized bovine bone scaffolds with PDL-derived MSCs compared to the scaffold alone. During 12 months of follow up, the clinicians did not observe considerable adverse effects and changes in blood formula related to MSCs and scaffolds implantation and considered that it was safe. Sanchez et al. [192] performed a pilot clinical study on 20 patients and did not find additional clinical benefits of using PDL MSC-based cell therapy as compared to xenogenic bone substitute alone after 12 months.
In a more recent study by Apatzidou et al. [193], three options of periodontal bone defects treatment were compared: a complex of alveolar bone marrow MSCs with collagen scaffolds and autologous fibrin/platelet lysate (aFPL), a collagen scaffold with aFPL and no scaffold as a control. Although there were no inter-group differences after 12 months of follow up, all treatment approaches led to significant clinical improvements in terms of radiographical bone fill and soft tissue healing. The authors suggested that application of MSCs based therapy might have potential in defects with a complicated, non-contained morphology over the extended period.
Quite a different approach for severe bone defect treatment was reported by Iwata et al. [8], who applied PDL-derived cell sheets in combination with TCP ( Figure 3). Clinical and radiographic findings after 6 months revealed clinical attachment gain (2.5 ± 2.6 mm) and increase of bone height of 2.3 ± 1.8 mm. The results maintained over the follow up period, while no serious adverse effects were recorded.  Several clinical studies utilize autologous GFs seeded on bioresorbable scaffolds such as collagen [194][195][196], plasma mesh rich in growth factors [197] and acellular dermal matrix allograft (ADMA) [198] for management of gingival recession. This cell tissue engineering strategy does not require extensive grafting, as a very small piece of gingiva is collected, so it decreases patients' morbidity. Favorable results of GF-based therapy on collagenous matrix in terms of recession coverage and increase of keratinized tissue width were reported by Mohammadi et al. [195] and Dominiak et al. [194] at 3 and 6 months, respectively. Histological evaluation revealed complete resorption of collagenous membrane in 3 months, formation of new keratinized tissue and improvements in tissue healing in the experimental group. In a randomized controlled clinical trial, Jhaveri et al. [198] compared human autologous fibroblasts seeded on acellular dermal matrix graft with a combination of a connective tissue graft (CTG) and coronally advanced flap (CAF) and did not observe statistical differences in clinical measurements among groups. Milinkovich et al. [196] also found conventional CTG technique more effective for gingival recession management in terms of keratinized gingiva width than the experimental GF-rich collagenous scaffold during a 12-month observation period.
From the clinical studies involving cells and scaffolds for periodontal tissue regeneration (Table 4) it can be summarized that the efficiency of MSCs-based therapies for periodontal regeneration is questionable and should be further evaluated in histological and long term randomized clinical trials. Apart from that, several important issues such as safety, immunogenicity of MSCs, potential risks and cost-efficiency should be considered before the implementation of cell-based periodontal therapies as a treatment option in clinical practice. Several clinical studies utilize autologous GFs seeded on bioresorbable scaffolds such as collagen [194][195][196], plasma mesh rich in growth factors [197] and acellular dermal matrix allograft (ADMA) [198] for management of gingival recession. This cell tissue engineering strategy does not require extensive grafting, as a very small piece of gingiva is collected, so it decreases patients' morbidity. Favorable results of GF-based therapy on collagenous matrix in terms of recession coverage and increase of keratinized tissue width were reported by Mohammadi et al. [195] and Dominiak et al. [194] at 3 and 6 months, respectively. Histological evaluation revealed complete resorption of collagenous membrane in 3 months, formation of new keratinized tissue and improvements in tissue healing in the experimental group. In a randomized controlled clinical trial, Jhaveri et al. [198] compared human autologous fibroblasts seeded on acellular dermal matrix graft with a combination of a connective tissue graft (CTG) and coronally advanced flap (CAF) and did not observe statistical differences in clinical measurements among groups. Milinkovich et al. [196] also found conventional CTG technique more effective for gingival recession management in terms of keratinized gingiva width than the experimental GF-rich collagenous scaffold during a 12-month observation period.
From the clinical studies involving cells and scaffolds for periodontal tissue regeneration (Table 4) it can be summarized that the efficiency of MSCs-based therapies for periodontal regeneration is questionable and should be further evaluated in histological and long term randomized clinical trials. Apart from that, several important issues such as safety, immunogenicity of MSCs, potential risks and cost-efficiency should be considered before the implementation of cell-based periodontal therapies as a treatment option in clinical practice.

Discussion and Concluding Remarks
Different materials have been used for the development of scaffolds for PDL regeneration and evaluated in vivo. Among them aliphatic polyesters such as PLA, PGA, their copolymer PLGA, and PCL have been extensively investigated [155]. PCL in particular is the most commonly used polymer, either applied in the form of membranes or in composite scaffolds, with other polymers such as PLGA or inorganic minerals such as HA or TCP. This preferable use of PCL derives from its availability, relatively low cost, and high modification potential [156]. Although it is a highly crystallized material (50-60%) with slow hydrolysis in vivo, its physicochemical and mechanical properties are easily tailored to meet different needs. Electrospun PCL membranes have been developed for combined drug delivery [9] and as barriers to cover periodontal defects [109], electrospun or printed PCL scaffolds, alone [12,132], combined with β-TCP [110,154] or HA [153] to provoke hard tissue formation and electrospun PCL/PLGA [123,130] or multiphasic PCL/COL/PEG scaffolds [160] for better mimicking periodontal tissue architecture and cementum formation. Calcium phosphate minerals or biomimetically developed calcium phosphate layers have been used in various forms to promote periodontal regeneration. Among them HA [13,72,92,123,126] and β-TCP [62,72,92,124,129,201] are the most popular, followed by bioactive glasses [10,121] and other ceramics [116]. Calcium phosphate coatings were applied in two in vivo studies [12,152] that yielded higher bone or cementum apposition. In clinical applications, β-TCP is the material used almost exclusively, despite that when not-combined with other scaffolding materials or growth factors does not seem to effectively regenerate periodontal tissues [163,[165][166][167]. Another calcium phosphate scaffold was developed by 3D-printing from biphasic calcium phosphate (BCP) and applied in one case study [181], while BCP was also used in a recent randomized clinical study [183]. This composite material consisting of HA and β-TCP at a proportion of HA to β-TCP of 60:40, provides a balance over the high degradation rate of β-TCP and the low dissolution of HA, in an effort to mimick bone resoprption rate [202]. Scaffolds in the form of hydrogels and sponges have also been used for periodontal regeneration [115,119,131,133,157,189,199]. In this respect, collagen is the predominant material [115,119,195] used combined with HA [123], FGF2 [125], cell binding peptide (P-15) [164] and BCP [183]. Collagen hydrogels have been used effectively as drug carriers and possesses significant properties, such as feasible synthesis, affordable cost, low toxicity, and ease of use [203]. Other hydrogel scaffolds [203] such as chitosan-based [131], gelatin [133], PEG-DA and DTT [157], and self-assembling peptide hydrogels [128] have also been used in in vivo studies.
None of the materials used in scaffolds for periodontal regeneration has profound advantages over the others and in most situations the different materials combinations produce better results. The construction of multi phasic scaffolds with different compartments of biomimetic microtopography and patterning seems to be the optimum choice for scaffolds manufacture, as it successfully guides cells to develop fibrous PDL-like tissues with the appropriate orientation. Despite the enhancements of manufacturing technology for complex scaffold constructs, their morphology still is far away the normal architecture of periodontal tissues. The employment of new scanning systems that can transfer with high accuracy the actual dimensions of a periodontal defect to contemporary milling machines that can use sophisticated technology to produce desirable scaffold compartmentalization and surface micro-or nanopatterning and topolography can lay the foundation of personalized treatment of degenerated periodontal tissues. However, this approach has inherent drawbacks such as cost and limited resolution to accommodate the complicated nanotopography of the periodontal architecture.
Animal models have proven considerably important for evaluating different materials and approaches to regenerate periodontal tissues. However, they are unable to mimic clinical conditions as in most cases the defects are surgically created, not resembling the actual destruction sites after periodontitis establishment, and they can not include contributing factors such as the presence of aggressive bacteria species, systemic diseases, smoking, occlusal parafunction, or host response. Another point in evaluating scaffolds for PDL regeneration in models that do not include the in situ periodontal tissue environment, like subcutaneous placement or calvaria models, is the absence of clinically relevant mechanical loading that could affect the regeneration response by modulating the cellular and molecular pathways needed. Although the established protocols are more or less commonly applied in most of the studies, still there is a lot of heterogeneity among study methodologies and materials applied, that makes almost impossible the direct comparison of findings to draw safe conclusions.
Periodontal ligament regeneration requires the simulataneous regeneration of bone and cementum. Although new bone formation has been observed in most of the studies, cementum regeneration remains a challenging task in vivo, as the identification of expression markers and differentiation pathways has not been specific. Some of the markers that have been determined so far for the identification of cementoblasts are the cementum attachment protein (CAP) and the cementum protein-1 (CEMP-1) [204,205]. CEMP-1 has been identified in cementoblasts and its progenitor cells, overexpression of which in PDLCs promotes cementoblastic differentiation, while reduced expression indicates osteoblastic or periodontal differentiation [206]. During the early stages of cementogenesis, an increased expression of miR-628-5p, SPON1, and PTPLA has been observed, while CEMP-1 expression is restricted by the presence of miR-628-5p and miR-383 [207]. CEMP-1 and PDPLA expression is increased in the late stages of cementogenesis [207].
Cell sheet engineering has been proven an effective treatment strategy regarding the regeneration of the full periodontal complex, consisting of alveolar bone, cementum, and PDL tissue, mainly in preclinical models, but also in one clinical study. PDLSCs emerge as the most suitable cell source for periodontal cell sheet engineering. Cell sheet engineering application face challenges related to cell therapies, in general, as well as the specific technique. Cell isolation and in vitro expansion are costly procedures that must be controlled by specific regulations, thus limiting the availability of such therapies to the general population. Additionally, cell sheet engineering is a technically demanding procedure, where handling and stabilizing the construct can be challenging, while proper attachment in situ is crucial for the desired regenerative outcomes [58]. Autologous transplantation of PDLSC sheets has been proven effective and safe in preclinical studies, as well as in a clinical study. However, there are often limitations that hinder this therapeutic strategy, such as lack of cell source, presence of periodontitis, and patient's age [208]. As a procedure, cell isolation and expansion towards the fabrication of a consistent and stable product following quality standards with Good Manufacturing Practice (GMP) can be a time-consuming and expensive process, thus it does not always present as a favorable or attractive treatment choice for the patient [208]. Changing direction towards allogeneic cell source in order to create cell banks and readily available products, could be more time-and cost-efficient, while also providing a possible treatment for patients with the above-mentioned limitations. Application of allogeneic cell sheets have been proven a safe and effective alternative [104,106].
PDL regeneration on Ti surfaces with the application of cell sheet engineering has been proven feasible, and it is thought to be beneficial regarding the long-term survival of dental implants, through the prevention of bacterial invasion and protection against inflammatory peri-implantitis [68], and through the dispersion of increased occlusal load. Nonetheless, those assumptions have not been supported by clinical data yet, and the long-term stability of the interface between Ti surface and the newly regenerated PDL-like tissue needs further assessment.
The approach of tissue regeneration based on combination of scaffold material, stem cells MSCs, and/or bioactive molecules might be an alternative for conventional periodontal surgical treatments in clinical practice. Due to the divergence of treatment outcomes in reported clinical trials and clinical case series, the benefits of scafold-based therapies are ambiguous. Most of clinical trials are focused on alveolar bone regeneration and evaluate clinical parameters such as clinical attachment level, and radiographical bone defect fill. There are limited data about histological analysis to confirm periodontal regeneration in humans due to ethical isuues. Further assessment of risk benefit and adverse effects should be performed in large-cohort studies.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biom12030435/s1, Table S1: In vivo studies employed orthotopic models to assess the regenerative capacity of cell sheet transplantation in periodontal defect models, with or without biomaterials; Table S2: In vivo studies evaluating scaffolds for PDL regeneration employing the subcutaneous implantation model; Table S3: In vivo studies evaluating scaffolds for PDL regeneration employing other models than periodontal defect and subcutaneous placement model.

Conflicts of Interest:
The authors declare no conflict of interest.