Experimental Formation of Periodontal Structure Around Titanium Implants Utilizing Bone Marrow Mesenchymal Stem Cells: A Pilot Study

An osseointegrated implant may closely resemble a natural tooth; however, the absence of periodontal ligament and connective tissue via cementum results in fundamental differences in the adaptation of the implant to occlusal forces. The character of the tissues around the titanium implant and the nature of the tissue attachment to the implant surface influence the biomechanical responses of this integrated system. Because of this structural difference, functional differences also exist between teeth and osseointegrated implants.1 

In function, the mobility of an osseointegrated implant is different from that of teeth, which may result in biomechanical problems when teeth and implants are combined for the support of a rigid prosthesis. Although many investigators have recommended using shock-absorbing elements to simulate the natural resilience of the periodontal ligament in the implant or its suprastructure, others have preferred to have anchorage of dental implants with the same functional mobility as natural teeth.2,^–4 

In addition, physiologic migration of natural teeth is a functional requirement of adaptation to maintain interdental and interocclusal relationships—a characteristic that pertains to healthy periodontal ligament and that it is impossible to achieve with dental implants.⁵

Additional differences include the fact that the potent cellular defense mechanism of the periodontal ligament protects the tooth in case of inflammation, a phenomenon that does not exist around dental implants because of differences in the wound healing mechanism around a natural tooth with a periodontal ligament vs an osseointegrated implant with direct bone-to-implant surface contact.6 

In addition, the periodontal ligament has a sensitive proprioceptive mechanism, which can detect minute changes in forces applied to the teeth. Forces applied to the teeth are dissipated through compression and redistribution of the fluid elements, as well as through the fiber system. Forces transmitted to the periodontal ligament can result in remodeling of tooth movement as seen in orthodontics, or in widening of the ligament space and an increase in tooth mobility in response to excessive forces (eg, occlusal trauma).7,^–9 

In osseointegration, a greater level of bone contact occurs in cortical bone than in cancellous bone, where marrow spaces are adjacent to the implant surface, which allows an initial period of healing after the surgical procedures have been completed. This results in bone resorption and is followed by bone deposition over time. Although this is a dynamic process in which bone turnover occurs, it is not an adaptive process, in that the same happens with the natural tooth surrounded by a periodontal ligament.10 

Thus, the qualities of the periodontal ligament (PDL) from the anatomic and functional stand points provide a number of potential advantages derived from its presence on a dental implant (eg, for the use of implants in growing patients). This can offer the potential application of dental implants for orthodontic tooth movement; hence, an implant with a PDL could be moved orthodontically to optimized positions so as to achieve more favorable esthetics and function.11,^–13 

Recent studies have shown the possibility of formation of periodontal ligament around titanium implants when a special model of application is used; this occurs when tooth-to-implant contact results from orthodontic movement or movement within a novel dentin chamber model. However, the methods used were far from clinical applications.5^,14 

In the present study, we investigated the possibility of engineering a periodontal structure around a titanium dental implant placed immediately in a fresh extraction socket in a goat experimental animal model. The tissue engineering principles utilized in the present work included seeding of bone marrow mesenchymal stem cells onto biodegradable porous scaffolds to be placed around the titanium implant fixture.

Three-dimensional (3D) hollow porous root-form scaffolds were prepared from 50:50 poly DL-lactide-co-glycolide (PLG) (Absorbable Polymers International, Pelham, Ala) with the use of the solvent casting/compression molding/particulate leaching technique, as described in our previous work.15^,16 PLG scaffolds were prepared from 50/50 PLG with 1 g of salt of particle size 180–300 μm. The hollow porous PLG scaffold was fabricated to fit around the fixture of the implant, that is, a hexed-head, threaded, 3.75 mm, hydroxyapatite-coated fixture with mount universal head diameter, and measuring 10 mm in length (IMTEC Corporation, Ardmore, Okla).

A mold to serve as the initial iteration was designed and produced out of Teflon. The dimensions of the mold were estimated based upon the previous dimensions of the titanium dental implant screw. The mold consisted of an insert, a hollow tube, and a solid cylinder (Figure 1a and b).

In forming the scaffold, the PLG with the salt particles was placed in the tube and was pressed, and the flat end of the solid cylinder was used to press the PLG between the end of the insert and the flat end of the solid cylinder (Figure 1c). This procedure would confirm a standardized size, length, and width of the PLG scaffold each time (Figure 1d and e).

Because the scaffold has to fit accurately around the endosseous implant in vivo after being seeded, it was noticed that the PLG scaffold might expand during the leaching procedure. To avoid such discrepancies, 8 samples were reproduced using the previous mold. Multiple measurements were made on each sample, and if any discrepancies in diameter measurements were noted, the numbers were averaged and the standard deviation was recorded.

Imaging of the scaffold was done with a scanning electron microscope (SEM) (JEOL 5510, JEOL Ltd, Tokyo, Japan) under low and high vacuum. Samples for SEM were prepared by fixation in osmium tetroxide at 4°C overnight; then the samples were subjected to dehydration in a graded series of alcohol solutions. For the low vacuum imaging, samples were viewed without coating, and then they were gold sputter coated for high vacuum imaging.

A total of 6 hollow root-form polymer scaffolds that were prepared from biodegradable PLG composite were exposed to porosity measurement by mercury intrusion porosimetry (Poresizer 9320 V2.08) (ASTM: D2873, Micrometrics, Norcross, Ga).

In all, 36 hollow porous 3D PLG scaffolds were incubated at 37°C in phosphate-buffered saline (PBS) for 6 weeks. At the end of each week, 6 samples were frozen then freeze dried and weighed for mass loss. The percentages of mass loss were recorded at the end of the second week, and the degraded samples were exposed to mechanical testing under dynamic loading. After the second week, the samples were too fragile to be tested.

Five goats, ranging in age from 6 months to 3.5 years, were used in the study. For each one, the following procedures were carried out.

Each goat was anesthetized with Xylazine HCl 2% in a dose of 3 mg/kg (body weight) intramuscular, and ketamine/HCl 2 mg/kg body weight (b.w.) intravenous. The bone marrow was collected from the femur with a 20 mL syringe with a 14 gauge cannula (Figure 2).

The aspirated bone marrow was cultured in supplemented media (DMEM) with 10% FBS, 1% Pen-strep (100.000 IU penicillin/10.000 μg streptomycin), 1% L-glutamine, and 2% N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES buffer). Cells were plated in 75 cm² tissue culture flasks and were placed in a CO₂ incubator (5% CO₂, 37°C, 95% humidity). On day 18, cells were trypsinized and subcultured at 90% confluence.

Second passage goat bone marrow derived mesenchymal stem cells were used for the seeding procedures. Prepared polymer scaffolds were sterilized by soaking in 95% ethyl alcohol for 30 minutes. They were then transferred to 24 well tissue culture plates, one scaffold in each well. The scaffold was washed with PBS for 1 hour, and the PBS was changed every 15 to 20 minutes. All PBS was aspirated and scaffolds were prewet with 2 mL supplemented media. Each animal received 2 scaffolds: a control scaffold, which was unseeded, and an experimental scaffold, which was seeded with bone marrow–derived mesenchymal stem cells.

Media were added to the control group. For the test group, after 30 minutes, the media were aspirated and 500 μL of media containing cells (9.4 × 10⁶ cells/mL) was added to each well (1 scaffold per well). The following day, the media and unattached cells were aspirated, then 500 μL of fresh supplemented media was added to each well. The number of unattached cells was calculated. The seeded scaffolds were monitored daily with an inverted light microscope. On day 2, the scaffolds were used for the animal experiment. For each experimental animal used, before the surgery was begun, 2 scaffolds (1 control and 1 seeded) were utilized for SEM examination using low and high vacuums.

Location: For each animal, lower right and left canine teeth were used.

The control was the right side, and the experimental side was the left side. On the control side, the canine tooth was extracted and the titanium fixture was inserted, then the hollow porous PLG scaffold was fitted around the fixture. On the experimental side, the canine was extracted and the fixture was inserted, then the PLG scaffold/seeded with bone marrow–derived mesenchymal stem cells (BMDSCs) was placed and fitted around the titanium fixture, as shown in Figure 3a through f.

Four animals were killed 4 weeks postoperatively, and 1 animal was killed at l0 days, the mandible of which was retrieved; occlusal and periapical radiographs were taken. The 10 day specimens for both control and experimental sides were radiographed and processed for non-decalcified sectioning with existing implant fixtures, then hard sections were obtained and stained with Van Geison picro-fuchsin and Stevenel's blue for light microscopy examination.17 For the 4 week animals, periapical radiographs were taken to determine the exact place of the implant fixtures. The mandibles were retrieved and sectioned on the right and left sides of the implant fixtures. The implant post was removed from the healed socket to allow study of the bone/implant interface and comparison of the healing mechanism in bony sockets that received cell/polymer construct around the implant vs control specimens, and to examine all surfaces of the fixtures by SEM.

Four-week specimens were processed and prepared for plastic resin embedding and non-decalcified sectioning. Hard sections were stained with Van Geison picro-fuchsin and Stevenel's blue to be examined under light microscopy (Olympus, Center Valley, Pa), and the implant fixtures were prepared for SEM examination, then for histologic evaluation.

Histologic sections also were analyzed with static cytophotometer microscopy (Leitz, Wetzlar, Germany) to measure the optical density of newly formed bone tissue in the peri-implant area. This was done to evaluate the quality of bone growth at the extraction socket lining wall following the implantation of different implant constructs.18 Control and experimental specimens were compared at different levels around the implant along its vertical axis, from the top coronal level down to its apical part, corresponding to the fundus of the extraction socket. At the horizontal level, measurements were taken to compare 2 different bone layers at the implant/socket wall interface: the innermost bony layer directly at the interface level (“the closest to the implant surface”) and the layer just next to it (“toward the outer surface of the jaw bone”). Collected data were analyzed by means of 3-factor analysis, with 3-factor interaction occurring as an error. The analysis of variance (ANOVA) procedure was performed with the Statistical Analysis System (SAS, SAS Institute, Inc, Cary, NC) program,19^,20 and comparisons among means were done to least significant difference .05 values.

Results of the expansion of the scaffolds are shown in Figure 4. This represents 1 iteration, which was used primarily to calculate the expansion coefficient; then a precise mold could be designed to produce the desired size polymer scaffold cylinder. With the optimized mold, the scaffolds were produced successfully to fit around the titanium screw satisfactorily.

The scaffold characteristics were 80% porosity, measured by mercury intrusion porosimetry that reflected open-celled structures, with good interconnection between the pores shown by SEM. The geometry of the 3D hollow root-form scaffold allowed distribution of masticatory load along the viscoelastic nature of the scaffold wall and interconnected pores.

Figure 5 shows the results of degradation indicating that weight loss increased over up to 6 weeks in vitro. Samples at the second interval demonstrated marked loss of their mechanical properties under dynamic compression loading, larger pores were obliterated, and the smaller-diameter pores were obliterated and disappeared as shown in Figures 6 and 7.

Cultured cells proliferated and reached confluence in 18 days (Figure 8). Figure 9 shows the cell/scaffold construct examination by inverted light microscope as well as SEM, which shows the adherence of cells to the scaffold with marked proliferation onto the porous surface.

Radiographic views for the retrieved goat mandible at 10 days are shown in Figure 10. Both sides; the control and experimental ones, demonstrated similar healing results radiographically (Figure 10A). The tissue surrounding the implant fixture resembled the tissue surrounding the natural tooth of the goat and distances previously occupied by the periodontal tissue had the same width of the natural periodontal tissue surrounding the tooth (Figure 10a through c).

The 10 day histologic specimen of the control side, where the implant fixture was surrounded for 10 days by the polymer scaffold without BMDSCs, showed some areas with spaces resembling periodontal tissue width (Figure 11a). At larger magnification, connective tissue fibers seemed to be extending from the original socket wall through an opening in the newly formed bone toward the surface of the titanium fixture (Figure 11b). In addition, fibers from these previously observed bundles extended to the titanium surface of the serration, while other bundles of fibers extended toward each other, forming a network and filling the space between the fixture and the newly formed bone (Figure 11c and d).

In Figure 12, we observed at the control side newly formed woven bone extending onto the titanium surface in a way similar to osseointegration; this phenomenon was noticed in some areas throughout the whole titanium fixture (Figure 12a through d). In all control specimens, no polymer scaffold remained or was noticed histologically.

On the experimental side of the 10 day specimen (Figure 13a), bone regenerated along almost the whole length of the implant fixture, maintaining the space between the titanium screw surface and the bone. The width of the space appeared the same all around the implant and resembled the space occupied by PDL around the natural teeth. Upon examination of the specimens, a heavy bundle of connective tissue fibers was seen extending from the original socket wall through an opening in the newly formed bone while tightly adhering to the bone and extending to meet other circular fibers running toward and parallel to the titanium surface of the fixture (Figure 13a through c). Signs of remodeling were noticed with the presence of osteoclasts away from the implant fixture, while bone-forming areas were seen on the newly forming bone surface toward the implant fixture (Figure 13d).

Figure 14 shows obvious preservation of the width of the periodontal ligament space, and the newly formed connective tissue fiber bundles appeared to run parallel to the titanium fixture in the coronal part (Figure 14a), similar to the direction of the natural gingival/periodontal fibers. Toward the middle part of the fixture, we noticed the presence of circular collagen fibers that adhered well to the serration of the implant screw, then started to fill the space between the implant screws extending toward the newly formed bone (Figure 14b through d). These new collagen fibers were highly cellular and vascular. No remaining degraded polymer scaffold was seen in any area. Figure 15 demonstrates the dense collagen bundles around the coronal part of the fixture, extending from the implant surface and running to be inserted into the surface of the newly formed woven bone (Figure 15a through d).

The deposited cellular cementum-like layer on the titanium surface is demonstrated in Figure 16, which reveals an evenly thick layer of cementum. In Figure 16a and b, the layer appears to be adherent to the titanium surface, while the periodontal-like fiber bundles are inserted into this layer (Figure 16c). The connective tissue fibers appeared to be attached to this cementum-like layer. Higher magnification in Figure 16d shows the cementum-like layer.

In the 1 month specimens, the phenomenon of periodontal tissue regeneration around the titanium implant was observed clearly. The titanium screws were retrieved from the sockets so their surfaces could be examined via SEM and light microscope. In Figure 17, where the titanium screw was surrounded by PLG scaffold alone in the control specimen, very few bony spicules were regenerated in the space surrounding the retrieved screw, and no epithelium was seen growing in the space at any time. Healing at the socket area did not reveal any periodontal tissue–like fibers, yet new bone projections were seen to extend toward the space where the screw had existed (Figure 17a through c).

On the experimental side, where the implant fixture was surrounded for 4 weeks by PLG scaffold seeded with BMDSCs, peri-implant wound healing was similar to that seen in the 10 day specimens (Figure 18). Newly regenerated bone trabeculae were observed surrounding the titanium fixture, so the socket appeared to have double walls of bone: the outside layer, representing the original socket wall, and the inside layer, representing the newly formed wall (Figure 18b), as well as new bone trabeculae that have been extensively distributed between these two walls (Figure 18c). In addition, finer bony spicules appeared to start from the inside bony wall, extending in a horizontal direction toward the space of the relieved fixture (Figure 18c). These horizontally arranged spicules seemed very active and were lined by osteoblasts (Figure 18d), showing no remnants of the degraded scaffold (Figure 19a).

At larger magnification, dense bundles of connective tissue fibers were seen between the active bone trabeculae also directed horizontally. These bundles appeared highly cellular and vascular and extended toward the implant fixture space from all around the inside bony wall (Figure 19c through f).

Upon examination, the retrieved fixture surface by SEM showed that some areas in the serrations had bone segments adhering to them, as shown in Figure 20a and b, while the entire fixture surface was covered with a cementum-like layer of even thickness. In other areas, particularly in the middle (third to sixth serrations), we observed periodontal-like tissue with obvious bundles; although many of them were cut during fixture retrieval, the remaining ones demonstrated the characteristic appearance of connective tissue fiber bundles anchored to the calcified tissue surface, resembling Sharpey's-like fibers (Figure 20c and d). Although the fibers of new cementum were running parallel to the fixture surface, the periodontal-like tissue fibers were seen to originate from the cementum layer, running away from the fixture surface toward the new bone spicules and confirming what was seen in the previous figure (Figure 19). The histology section of the fixture demonstrated the tightly adhered cementum-like tissue to the implant serration surface that was stained with Van Gieson and Stevenel's blue (Figure 21a and b).

According to the analysis of variance (ANOVA) test for bone density evaluation at the implant/socket wall interface, no significant differences in bone density were found at any level, along the vertical axis of the implant nor over the bone layers at the horizontal plane, between the experimental socket wall interface and the control one in 10 day specimens (Figure 22). Although the newly formed bone on the experimental side regenerated in the form of a continuous thin bony plate along the whole length of the implant, a tiny consistent distance at the interface area was found to be filled with connective tissue fibers that had been growing and extending between the growing bone and the implant surfaces.

On the other hand, the 4 week specimen analysis revealed differences in the bone density measurement at the bony socket wall between the side that received the implant/polymer construct with BMDSCs and that of the control socket; this difference was found to be highly significant (P < .01). These differences were position dependent, in that the experimental group showed less bone density at the apical half around the implant than did controls (P < .05), while no significant differences have been found between groups at the implant coronal half. Furthermore, in both groups, the innermost bone layer (L1) at the interface showed significantly less density than was noted in the next layer (outward L2) (P < .05) (Table 1).

Recent advances in the mechanisms of tissue regeneration have laid the foundation for novel approaches in clinical tissue engineering as a product of biotechnological advances in the field, and one of the most significant discoveries in postnatal life.

In dentistry, over the past few decades, the use of endosseous implants has increased as a means of providing a foundation for intraoral prosthetic devices, from single crowns to full arch dentures or other devices, because of the increasingly convincing data of long-term clinical success rates, which promoted the use of implants in more challenging clinical situations than were previously envisioned.21,^–23 

Although osseointegrated implants provide high survival rates on a long-term clinical basis, their supporting mode is not considered ideal. Several investigators have shown evidence of periodontal ligament formation around dental implants or other artificial surfaces.3,^–5 It has been shown that healing characteristics of the wound created from implant placement are determined by the cells derived from the bone compartment: bone cells, bone marrow cells, and blood cells, all of which determine the pattern of healing that results in direct bone apposition on the titanium surface (osseointegration).24^,25 

Previous studies have demonstrated that only cells residing in the periodontal ligament are capable of forming new cementum by inserting collagen fibers into exposed root surfaces. These studies focused on the use of periodontal ligament cells after they had been grown in culture in vitro. Results showed that these cells were able to retain their viability and phenotype, which led to new attachment formation after seeding in vivo.5^,14^,26^,27 

The research model of the present work was designed to be in the extraction sockets immediately after tooth removal. The hypothesis was to benefit from the remaining periodontal ligament with its cells. Circumstantial evidence suggests that during wound healing in the periodontal ligament, different cell groups originating from the periodontal ligament possess the potential to form new/reparative cementum while simultaneously inserting new attachment ligament fibers into the newly formed tissue matrix. However, these observations have been confined to the areas closest to the source of progenitor cells at the wound periphery.28 

In the present work, we observed growth of periodontal tissue–like bundles of connective tissue fibers extending from the periphery of the wound toward the surface of the titanium fixture (Figures 11 through 15). These observations were seen at both the control and experimental sides in the 10 day specimens; this confirms previous study findings, indicating that progenitor cells from the remaining periodontal ligament have the capacity to differentiate into formative cells like cementoblasts and fibroblasts as long as the peri-implant wound area is close to the periphery of the extraction socket24^,25 (Figures 11 through 15).

Because of the anatomic position of the canines in the goat, which are tipped toward the orifice of the oral cavity, and because of the fact that the endosseous implant fixtures were placed in the same position as the extracted tooth, 1 surface of the implant fixture was closer than the other to the wound periphery, and these were the same surfaces in which we observed extension of new periodontal-like ligament structure.29^,30 Previously, cultured cells derived from periodontal ligament and alveolar bone were found to be basically capable of synthesizing periodontal tissue after their in vivo reimplantation, and these cells retained their capability to differentiate and participate in the regeneration of periodontal structure.31,^–34 

In the present work, the ability of periodontal tissue to regenerate on both control and experimental sides at the periphery of the wound confirmed the previous findings of these authors that progenitor cells in the remaining periodontal tissue at the wound periphery maintain their phenotype to be differentiated under environmental factors such as specific glycol proteins and under growth factors into osteoblasts, cementoblasts and fibroblasts.

Histologic findings for the 10 day specimen at the control side showed areas of osseointegration where new woven bone was in direct contact with the fixture serrations (Figure 12). All of these areas were away from the wound periphery, and the fact that the PLG scaffold degraded allowed the osteoblasts to reach the surface of the titanium fixture. This direct bone apposition or osteoconduction in areas away from the wound periphery has been recognized by others in peri-implant healing around endosseous implants. During this cascade of wound healing, the differentiated osteogenic cells would reach the bone/implant interface to start the de novo bone formation.35^,36 

In the experimental sides that received seeded PLG scaffolds with stem cells at the time of fixture placement, we observed new continuous woven bone all around the implant fixture; this was separated from the surface of the implant at the bone/implant interface with a continuously even space that appeared filled with periodontal-like tissue (Figures 13 through 15). These cells along with the progenitor cells from remaining periodontal tissue in the wound differentiated into cementoblasts, fibroblasts, and osteoblasts which were all responsible for periodontal tissue regeneration and the maintenance of periodontal ligament width. The presence of newly formed woven bone, the clear bundles of collagen fibers, and the presence of the cementum-like layer observed on the titanium surface were all obvious indications of periodontal-like structure regeneration around the titanium fixture in the 10 day results (Figures 13 through 16).

Previous investigators have used BMDSCs to regenerate periodontal tissue in periodontal defects around natural teeth. These investigators have shown new cementum, greater cellularity, and sparse extrinsic Sharpey's fibers upon implantation of BMDSCs into periodontal defects.37,^–41 

In this study of 10 day experimental specimens, we observed the extension of periodontal tissue toward the implant fixture in the area near the wound periphery to separate new woven bone from the titanium surface. Although progenitor cells from the remaining tissue residing in the extraction socket might be responsible for this previous finding at areas near the wound periphery, we observed the dense collagen bundles and the extrinsic Sharpey's-like fibers all around the fixture. The differentiation of BMDSCs into cementoblasts, osteoblasts, and periodontal ligament fibroblasts might be responsible for this previous finding. In addition, a bioactive hydroxyapatite layer on the titanium surface of the fixture has been identified to be a suitable layer for periodontal regeneration and cementogenesis, because it provides a low contact angle and superior wettability as well as adhesion characteristics when compared with the titanium surface alone.42^,43 

At 4 weeks, we could not detect any periodontal-like tissue in the socket created by implant fixture removal, or at the surface of the fixture on the control side, which shows that the presence of progenitor cells from remaining periodontal tissue was not enough to maintain the newly regenerated collagen fibers that appeared at the control side in the 10 day specimen (Figure 17). Instead, bony spicules appeared to surround the implant fixture, yet they did not fill the entire area at the bone/implant interface. The incomplete regeneration of new connective tissue is believed to be due to the limited capacity of the progenitor cells from the remaining periodontal ligament to form and maintain the new attachment. This fact has been shown by other investigators to be caused by decelerated restricted migration and premature differentiation of progenitor cells.29^,30 

In addition, statistical analysis of the histophotometric measurements showed that the innermost bone layer (L1) density of the socket wall on the experimental side was significantly less than that of the control side 1 month after implant insertion. This confirms the histologic picture of the widely distributed newly forming and active bone spicules all around the lining wall of the extraction socket. These spicules are obviously of the immature woven bone type, which is characterized by high cellularity and less calcium content, compared with mature lamellated bone, which eventually will be replacing the new immature form. Such structure could explain the lower bone density at the experimental socket/implant interface than in the control group. This phenomenon was obvious around the implant apical half that showed the pattern of bone regeneration starting from the socket fundus up toward the coronal level. This reflects the role of seeded BMDSCs in new bone formation. Furthermore, these cells have shown in the present study their high potential and plasticity characteristics as mesenchymal unspecialized cells capable of being differentiated to variable cell groups and producing different tissue types of mesenchymal origin. Consequently, at the implantation site where the bony socket was originally formed of alveolar bone lining containing anchored periodontal ligament fibers at one extremity, while the other fiber extremities were inserted into the root cementum of the extracted tooth, it was found that implanted BMDSCs demonstrated signs of differentiation and regeneration of new tissues that had existed originally at the same site. These tissues represented the main elements of the tooth-supporting structures surrounding and attached to the implant fixture surface.44^,45 

In our previous work, we have shown that the presence of BMDSCs seeded onto porous PLG scaffolds influences the rate of degradation and enhances bone formation.15^,16 The 4 week specimens at the experimental side confirmed our previous observations.

All experimental specimens showed a space of even thickness at the bone/implant interface that was filled with new connective tissue fiber bundles all around the implant fixture. These fibers extended from the bone to the implant surfaces. The histologic sections showed the connective tissue bundle arrangement emerging between new bone spicules in an organized direction toward the implant fixture. Other investigators have shown regeneration of dense collagen bundles on the root surface 1 month after transplantation of BMDSCs into furcation defects.41 

The surface of the titanium fixture seemed to be covered with an even layer of cementum-like tissue that appeared to be adhered well to the fixture surface. Scanning electron micrographs showed the cementum-like layer to consist of densely packed collagen fibrils parallel to the fixture surface, while the collagen bundles of the periodontal tissue–like structure appeared to be distinct and loose as they left the cementum layer. The 1 month retrieved implant fixtures appeared to be covered with cementum matrix, which apparently was not fully mineralized, with collagen bundles simulating periodontal ligament fibers (Figures 20 and 21) being inserted into and extending toward the newly formed bone trabeculae.

The formation of periodontal-like structures as noticed in the present study may open new prospects in our understanding of wound healing in peri-implant tissues and may provide a novel approach for periodontal tissue regeneration. The results of this work encourage more research in this promising area.

We are grateful to Prof. Mamdouh El-Rouby for his valuable contribution to the statistical analysis. We also thank Rami Abdelaty, Ahmed Saad, and Mahmoud Hammad for their technical assistance in manuscript preparation. This work was funded by the Academy of Scientific Research and Technology-Egypt and the US-Egypt joint program.