Comparing the Healing Potential of Late-Stage Periodontitis and Peri-Implantitis

Peri-implantitis and periodontitis are defined as inflammatory diseases affecting the soft and hard tissues around osseointegrated functioning implants and teeth.¹  Peri-implantitis is a growing concern because it affects ∼45% of all patients with implants and ∼14% present with moderate to severe disease.²  Moreover, current treatment modalities for peri-implantitis, which are derived from therapeutical options for treating periodontitis,^3–5  are unpredictable.^6–8 

The etiology of peri-implantitis is multifactorial, resembling periodontitis,⁹  and includes a complex interaction between the bacterial biofilm and the host's protective immune response. In addition, risk factors associated with peri-implantitis have also been associated with periodontitis, including poor oral hygiene, tobacco use, diabetes, and genetic factors.⁹  Moreover, history or current periodontitis status has been reported as a predictive risk factor for peri-implantitis.^9–14  Although peri-implantitis and periodontitis have pathophysiologic similarities, several studies have suggested pathophysiologic differences between the 2 conditions.^15–17 

Considering the current knowledge, we hypothesize that peri-implantitis is a more aggressive disease and does not respond as well to treatment when compared to periodontitis. To address this question, we used an established murine model of experimental peri-implantitis and periodontitis^18,19  that resembles late-stage conditions to compare how peri-implantitis and periodontitis respond after the elimination of the insult (ligature).

Thirty-five C57BL/6J 4-week-old male mice (at least [n ≥ 3] mice/group/time point) (The Jackson Laboratory, Bar Harbor, Maine) were used according to the Chancellor's Animal Research Committee of the University of California, Los Angeles guidelines for all the experiments. In addition, the Animal Research: Reporting in Vivo Experiments guidelines were followed.²⁰  Mice were fed a soft diet (Bio-Serv; Frenchtown, NJ) ad libitum for the duration of the experiment.

Left maxillary molars were extracted from 35 one-month-old C57BL/6J male mice. Eight weeks after extraction, 1 mm × 0.5 mm titanium implants (DP Machining Inc, La Verne, Calif) were placed in the edentulous socket. The implants were allowed to osseointegrate for 4 weeks.^18,21  Following implant osseointegration (28 implants osseointegrated), mice were randomly separated into control (no ligature) and experimental (ligature) groups. In the experimental group, mice received 6-0 silk ligatures (P.B.N. Medicals, Stenløse, Demark) placed immediately apical to the implant head and on the contralateral second molar, in a split mouth design. To resemble late-stage disease, ligatures were kept for 4 weeks without intervention. After 4 weeks, the experimental group was randomly divided into ligature-retained and ligature-removed groups. The ligature-retained group kept the ligatures for the duration of the whole experiment, and the ligature-removed group had their ligatures removed. All mice were sacrificed at 2 time points: 1 and 2 weeks from the date of ligature removal (Figure 1a and b). Maxillae were harvested and imaged using a digital optical microscope (VHX-1000; Keyence, Osaka, Japan), fixed in 4% formalin for 24 hours, and stored in 70% ethanol.

Maxillae were imaged using micro-computed tomography (micro-CT; Model 1172; SkyScan, Kontich, Belgium) at 10 μm resolution.^18,21  Linear bone loss measurements were performed using Dolphin software (Navantis, Toronto, Canada). In brief, the second molars were oriented such that the cementoenamel junction (CEJ) was parallel in the coronal plane. The axial plane represented the middle of the second molar crown. Measurements were taken from the CEJ to the alveolar bone crest (ABC).^18,21  Implants were oriented such that the head and the shaft of the implant were perpendicular to each other in the sagittal plane.¹⁸  A single blind examiner recorded measurements in the mesial, distal, buccal, and palatal locations for both teeth and implants. Measurements were averaged per mouse to create a mean bone loss value for each implant or tooth.

Samples were decalcified in 15% EDTA for 4 weeks.^18,21  Samples were stained with hematoxylin and eosin (H&E), picrosirius red (Polysciences, Inc, Warrington, Pa), and tartrate-resistant acid phosphatase (TRAP; Sigma-Aldrich, St Louis, Mo), according to the manufacturer's protocol. Osteoclasts, TRAP⁺ cells with ≥2 nuclei in contact with the bone,²²  were counted in the area immediately adjacent to the teeth and implants along the ABC (n = 3/group/time point).

Thirty-five mice were included in the study. Each mouse had an implant placed in the left maxilla and the teeth on the contralateral right side were used for the experiments performed on the groups with teeth. However, prior to the ligature placement 7 implants were lost; therefore, those implant samples could not be used in the study. Figure 1b depicts the numbers of samples that were used for each experimental group.

Bone loss measurements were represented as mean ± SEM. Osteoclast numbers were normalized to their respective controls (tooth experimental to tooth control groups, implant experimental to implant control groups). Significance between groups was compared using 2-way analysis of variance (ANOVA) and a Bonferroni correction post-test (Prism 5; GraphPad Software, La Jolla, Calif). Levels of significance were as follows: P < .05*, P ≤ .01**, P ≤ .001***.

Clinically, peri-implantitis and periodontitis ligature groups presented with generalized soft tissue edema in comparison with the respective control groups at the 1- and 2-week time points. Gingival edema decreased following ligature removal for peri-implantitis and periodontitis, at both time points (Figure 2a). However, the peri-implantitis ligature-removed groups appeared to still have some edema compared with the periodontitis ligature-removed groups at both time points.

There was statistically significant linear bone loss in the peri-implantitis and periodontitis ligature-retained and ligature-removed groups compared with their respective controls at both time points (Figure 2b and c). For the 2-week time point, the linear bone height for peri-implantitis and periodontitis were 0.431 ± 0.019 mm and 0.377 ± 0.027 mm, respectively, whereas the bone height for the controls were 0.218 ± 0.012 mm and 0.181 ± 0.005 mm. For the 1-week time points, the numbers were not too far different. Two weeks after ligature removal, there was statistically significant bone gain in the periodontitis and peri-implantitis ligature-removed groups compared with their respective controls. The linear bone heights measured 0.261 ± 0.013 mm for teeth and 0.375 ± 0.014 mm for implants (Figure 2c). However, the peri-implantitis ligature-removed group had less changes in bone height compared with the periodontitis ligature-removed group.

The tissues surrounding the implants were associated with higher levels of inflammation with or without the ligatures (Figure 3). Ligature-induced inflammation and bone resorption was observed in teeth and implants. However, peri-implantitis developed higher levels of inflammation and bone resorption compared with periodontitis. Closer examination also showed that new bone accumulated during active periodontitis (ligature-retained group), but not in peri-implantitis. In both diseases, removal of the ligature allowed new bone formation; nonetheless, this phenomenon was far more evident in the periodontitis ligature-removed group compared with the peri-implantitis ligature-removed group. New bone formation during active and recovering periodontitis was notably concentrated around the periodontal ligament (PDL) area, and the direction of the early bone formation appeared to be parallel with that of the PDL.

For both time points, the peri-implantitis ligature-retained groups presented with a statistically significantly greater osteoclast fold difference compared with their respective controls and ligature-removed groups (Figure 4a and b). Moreover, the peri-implantitis ligature-retained and ligature-removed groups presented with a greater osteoclast fold difference compared with the periodontitis ligature-retained and ligature-removed groups.

The collagen matrix around teeth and implants was assessed through picrosirius red staining under polarized light (Figure 5). For all groups, at all time points, the soft tissues around teeth presented with greater yellow birefringence than the soft tissues around implants. These signals are suggestive of type 1 collagen, which predominates around teeth. Control teeth, periodontitis ligature-retained, and periodontitis ligature-removed groups presented with collagen fibers organized parallel in between the molars and inserting into the cementum. In contrast, the peri-implantitis control groups presented with less type 1 collagen, and the collagen was loosely organized in a parallel manner. After the addition of ligatures, collagen integrity and organization further decreased in the peri-implantitis group. By the second week after ligature removal, collagen regained some organization in the peri-implantitis ligature-removed group by the second week, but was still not as organized as the periodontitis ligature-removed group.

As mentioned already, peri-implantitis is a growing concern because it affects approximately half of the patients with dental implants²  and current treatment modalities are unpredictable.^3–5  Therefore, to further our understanding of peri-implantitis, we evaluated treatment of late-stage peri-implantitis and compared it with periodontitis utilizing a murine model of ligature-induced peri-implantitis and periodontitis.

An animal model was chosen given the limitations of studying the pathophysiology of peri-implantitis in humans, such as lack of controlled environment, unknown disease onset,²³  and ethical dilemmas.^24,25  Among the animal models available to study peri-implantitis and periodontitis, the use of ligatures to induce disease has been widely accepted model.²⁶  The ligature model was first described by Rovin et al using rats,²⁷  and has since been used to successfully induce and study periodontitis in larger animals such as dogs and non-human primates.^28–32  Among the murine models of peri-implantitis, there are 2 models available, a ligature-induced peri-implantitis and P. gingivalis (P.g.)-induced model.³³  We elected to use the ligature-induced because it is more widely accepted and bypasses differences in colonization. To assess how late-stage peri-implantitis lesions respond to treatment, linear bone loss, inflammation level, osteoclasts, and collagen distribution/type were monitored at 1 and 2 weeks from ligature removal.

Bone loss is one of the characteristic features of peri-implantitis and periodontitis. Based on the micro-CT analysis and histologic data, striking differences were observed between the responses of peri-implantitis and periodontitis to treatment (ligature removal). Insult removal led to significant new bone formation in periodontitis, and minimal bone formation in peri-implantitis. However, even the ligature-retained periodontitis group appeared to have bone formation, suggesting that periodontitis has bone formation even in the presence of ligature-induced inflammation. The bone formation in the periodontitis ligature-retained and periodontitis ligature-removed groups was adjacent to the PDL cells, suggesting that the reduced bone formation potential in the peri-implantitis groups could potentially be explained due to the lack of PDL around implants. The PDL not only possesses cells that can readily differentiate and assist in bone formation, but these cells are ready to produce bone when proper signals are present.³⁴ 

Bone formation and resorption involves an intimate interaction between osteoblasts and osteoclasts (OCs).³⁵  Given that OCs are responsible for bone resorption, we assessed OCs numbers through histologic staining. The peri-implantitis lesions contained a higher number of OCs compared to periodontitis, which is in line with previously published research.³⁶  The ligature-retained periodontitis lesions, however, did not reflect a fold increase of osteoclasts when compared with their respective control groups. This could partially be attributed to the length of the experiment, since these lesions resembled more of a chronic condition and less cells were present to induce osteoclasts. In terms of peri-implantitis treatment (peri-implantitis-ligature removal), there was a decrease in OC at both time points; however, the fold difference was still statistically greater than the control implant group after 1 week. This finding was not observed in the periodontitis ligature-removed group, as compared with the periodontitis with ligature-retained, suggesting that peri-implantitis experiences a delay in the signals to halt osteoclast differentiation when compared with periodontitis, which could partially explain the pathophysiologic differences between the 2 conditions.

The evaluation of soft tissue inflammation in our study demonstrated that the composition of the inflammatory infiltrate between peri-implantitis and periodontitis differs, which is in agreement with previously published research.³⁷  To further characterize disease resolution after treatment of late-stage disease, extracellular matrix differences, collagen composition, and organization were evaluated with picrosirius red. Picrosirius red staining allows for the visualization of collagen fibers under polarized light³⁸  and key differences between peri-implantitis and periodontitis were noted. The control peri-implantitis group showed less collagen fibers compared with the control periodontitis group, which correlates with previous studies,³⁷  and corroborates with the notion that, in health, the absence of the PDL and transseptal fibers leads to a weaker biologic barrier. Furthermore, the peri-implantitis ligature-retained group presented critical structural differences. The collagen bundles appeared thinner and fiber distribution was highly affected by inflammation, evident by a weaker and more disorganized birefringence. On the other hand, thick collagen fibers were produced supracrestally around teeth, separating the area of ligature insult from the bone, which could potentially play a protective feature that is absent around implants. After insult removal, collagen remained largely disorganized by the first week, but somewhat reorganized by the second week in both groups; however, collagen organization in the peri-implantitis ligature-removed group remained less prominent. Our data, in conjunction with previously published work, support the idea that implant connective tissue adhesion has poorer mechanical resistance compared with natural teeth.¹⁵ 

In conclusion, late-stage peri-implantitis treatment does not respond as well compared with periodontitis as seen by less bone formation, higher inflammation levels, delay in the reduction of OC numbers, and less organized type 1 collagen. Though peri-implantitis and periodontitis share some etiologic and clinical characteristics, our data showed that higher levels of inflammation/bone resorption associated with peri-implantitis and the intrinsic ability of the PDL to repair bone around the teeth are some of the main features of why periodontitis responds significantly better than peri-implantitis to treatment. These differences should be taken into consideration when designing a treatment protocol for peri-implantitis. This model should allow for future investigation to evaluate disease progression and healing, as well as aid in developing therapeutic approaches.

Abbreviations

ANOVA

analysis of variance

ABC

alveolar bone crest

C

control

CEJ

cementoenamel junction

H&E

hematoxylin and eosin

L

ligature-retained

micro-CT

micro-computerized tomography

OCs

osteoclasts

PDL

periodontal ligament

Rx

ligature-removed

TRAP

tartrate-resistant acid phosphatase

This work was supported by the University of California-Los Angeles School of Dentistry Seed Grant. RW was supported by the American Academy of Implant Dentistry and the UCLA Clinical and Translational Science Institute, NIH 5TL1TR000121-05. SH was supported by NIH/NIDCR T90 DE022734-01. We would like to thank the Translational Pathology Core Laboratory from the UCLA David Geffen School of Medicine for assistance with preparing the decalcified histologic sections.

The author(s) declare no potential conflicts of interest with respect to the authorship and or publication of this article.