Concise Review: Mesenchymal Stromal Cells Used for Periodontal Regeneration: A Systematic Review

This systematic review provided good evidence of the capacity of mesenchymal stromal cells to regenerate periodontal tissues in animals; however, experimentally generated defects used in animal studies do not sufficiently mimic the pathophysiology of periodontitis in humans. Moreover, the safety of such interventions in humans still needs to be studied.

Abstract

Periodontitis is a chronic infectious disease of the soft and hard tissues supporting the teeth. Recent advances in regenerative medicine and stem cell biology have paved the way for periodontal tissue engineering. Mesenchymal stromal cells (MSCs) delivered in situ to periodontal defects may exert their effects at multiple levels, including neovascularization, immunomodulation, and tissue regeneration. This systematic review had two goals: (a) to objectively quantify key elements for efficacy and safety of MSCs used for periodontal regeneration and (b) to identify patterns in the existing literature to explain differences between studies and suggest recommendations for future research. This systematic review provided good evidence of the capacity of MSCs to regenerate periodontal tissues in animals; however, experimentally generated defects used in animal studies do not sufficiently mimic the pathophysiology of periodontitis in humans. Moreover, the safety of such interventions in humans still needs to be studied. There were marked differences between experimental and control groups that may be influenced by characteristics that are crucial to address before translation to human clinical trials. We suggest that the appropriate combination of cell source, carrier type, and biomolecules, as well as the inclusion of critical path issues for a given clinical case, should be further explored and refined before transitioning to clinical trials. Future studies should investigate periodontal regenerative procedures in animal models, including rodents, in which the defects generated are designed to more accurately reflect the inflammatory status of the host and the shift in their pathogenic microflora.

Introduction

Periodontitis is a chronic infectious disease of the soft and hard tissues supporting the teeth [1], affecting 15%–50% of adults in developed countries [2]. This disease leads to the formation of deep infrabony defects and soft-tissue crevices called “periodontal pockets” between the tooth and its bony socket. Left untreated, periodontitis can result in tooth loss [1], with a significant impact on oral health and overall quality of life [3, 4]. Because of its infectious and inflammatory sides, periodontitis has been associated with adverse pregnancy outcomes, cardiovascular events, pulmonary disease, or diabetes [5]. Conventional periodontal therapy involves debridement of the root surface to induce healing by repair or regenerative pathways. For decades, several regenerative procedures have been used, including enamel matrix-derived (EMD) proteins [6], guided tissue regeneration [7], and bone graft placement [6].

The treatment of infrabony defects is associated with a high degree of variability in clinical outcomes [8]. Consequently, the regeneration of bone, cementum, and an effective periodontal ligament (PDL) remains a challenge. New treatments for periodontitis have to be developed to better reflect its etiopathophysiology. Recent advances in regenerative medicine have paved the way for such improvements by presenting innovative and imaginative opportunities for periodontal tissue engineering [9]. The main goals of regenerative therapy are to enhance the migration, proliferation, and commitment of endogenous and/or exogenous progenitor/stem cells to appropriate terminally differentiated phenotypes and to favor the biosynthesis of the extracellular matrix components that support tissue recovery [9]. However, genetic determinants, physiological and systemic conditions, and local disease state (infection, scars) are predicted to impair regenerative potential [10, 11]. Grafting exogenous cells to favor the production of new tissues and/or to make the local microenvironment more suitable for stimulation of endogenous progenitors [12, 13] has been tested with promising results. Indeed, successful treatment of various diseases, including cardiovascular diseases and large bone defects treated by cell therapy based on the use of adult multipotent mesenchymal stromal cells (MSCs), has been reported [14–16]. MSCs have the capacity for mesenchymal lineage plasticity and display sensitivity to local paracrine activity [17]. Because of these unique properties, MSCs delivered in situ to periodontal defects may exert their effects at multiple levels, including neovascularization [18], immunomodulation [19, 20], and tissue regeneration.

Translation of experimental outcomes achieved in animals to human clinical trials requires that the design and results from animal studies should be analyzed with appropriate tools based on rational criteria and strategies. The use of systematic reviews in biology is growing but has never been applied in the field of cell therapy for periodontal regeneration, although some literature reviews devoted to oral tissue regeneration by stem cells have been published previously [9, 21, 22]. A systematic approach is able to minimize the risk of selection bias in the review process. The goals of this systematic review are to objectively quantify key elements for efficacy and safety of MSCs used for periodontal regeneration and to identify patterns in the existing literature to explain differences between studies.

Materials and Methods

This systematic review follows configurative logic that provides insight through an iterative literature search strategy that allows for new ways of understanding existing literature and suggests ways to improve future studies [23]. The methodology used in this review is based on the guidelines for scoping reviews, as suggested by Arksey and O’Malley, and was applied as follows: identify the research questions; search for relevant studies; select appropriate studies; chart the data; collate, summarize, and report the results [24].

Identifying the Research Questions

This systematic review was designed to answer the following questions: What are the different methods used to quantify the effect of MSCs on periodontal regeneration? What is the current evidence on the efficacy and the safety of MSCs used for periodontal regeneration? How can possible differences between studies be explained?

Searching for Relevant Studies

We used a combination of a protocol-driven method and a “snowballing technique” to search for relevant studies. Details of the electronic search strategy in eight databases are provided in the supplemental online data for this study. The search strategy was designed to include data from both animal and human studies. The languages of publication were restricted to English, French, Spanish, and Portuguese. Reference lists of query studies were inspected to identify any additional relevant published or unpublished data. Details of the literature search in the “gray literature” are also provided in the supplemental online data.

Study Selection

Configurative logic was chosen for the selection of relevant studies. By using an iterative process, we determined clear macro rules for inclusion of studies and assumed that specific detailed criteria for meeting those rules would become apparent through the process of doing the review.

All studies dealing with regeneration of deep periodontal tissues (cementum, periodontal ligament, and alveolar bone), with a control group for animal studies, were considered to be eligible. All animal species were considered. With regard to periodontal conditions, there was no restriction regarding the manner in which periodontal lesions were induced. All types of defined MSCs were considered. For outcomes, only studies with quantitative outcomes for both experimental and control groups were selected, and outcomes were listed to answer the research question on the different methods used to quantify the effect of MSCs on periodontal regeneration (supplemental online Fig. 1).

Data Extraction and Subgroup Analyses

We extracted data concerning animal model (gender and species), method used to generate periodontal defects, type of periodontal defect, cell source, type of cell carrier, randomization, study duration, and quality of reporting. We planned to determine whether differences in the efficacy or safety of periodontal regeneration between experimental and control groups were influenced by these characteristics. The chi-square test was used to perform subgroup analyses if at least three groups of two or more studies were available.

Statistical Analysis and Data Charting

Animal and human studies were analyzed separately. We performed meta-analyses on standardized mean differences using an inverse variance random-effects method. Review Manager 5.2 (The Cochrane Collaboration, Copenhagen, Denmark, http://www.cochrane.org) was used to plot the data. Only studies with an appropriate control group were considered for quantitative analysis. When a study had several experimental groups, each group was included separately in the quantitative analysis. The funnel plot was visually examined, and an Egger’s test was performed using the “metabias” command in Stata 11.1 (StataCorp, College Station, TX, http://www.stata.com) to determine publication bias. The studies determined to be significantly responsible for publication bias (e.g., increasing heterogeneity) were removed from the meta-analysis and discussed separately.

Results

Seventy publications from 2,986 records met the inclusion criteria and were included in the descriptive synthesis (supplemental online Fig. 2 and supplemental online Table 1 show details of the selection process). Finally, 56 studies (45 animal studies, 11 human studies) were included.

Periodontal Regeneration in Animal Studies

Characteristics of each study are shown in supplemental online Table 3 and summarized in supplemental online Table 2.

Description of Included Studies

We found comparable numbers of studies on periodontal regeneration of fenestration defects, infrabony defects, and furcation defects. To induce a periodontal defect, periodontal bone was removed with surgical burs (74% of total selected articles) or induced by creating an additional deep inflammatory process within the periodontium using impression paste, gutta-percha, or a ligature around the cervical part of the tooth (21%). Experiments were conducted on dogs in 49%, pigs in 14% and rodents in 33% of included studies. Cells were derived from oral tissues in 63% of studies, and embryonic stem cells and induced pluripotent stem cells were used in four studies. Engraftment was autologous in 63%, allogeneic in 14%, and xenogeneic in 28% of included studies. Cells were carried within cell sheets in 23% of studies. Extracellular matrix proteins (collagen, fibrin, gelatin, or hyaluronic acid) were used in 49% of included studies, enamel matrix derivatives were used in 5%, plasma-rich platelet (PRP) concentrate or blood coagulum was used in 9%, and polymers (alginate, poly-ε-caprolactone, polyglycolic acid, silk fibroin, or pluronic F127) were used in 23%. A guided tissue regeneration technique [7] was used in 30% of included studies. Quality of periodontal regeneration cell therapy reported for animal studies was investigated and summarized in supplemental online Figure 3.

Methods Used to Quantify Periodontal Regeneration

A variety of methods were used by the authors to quantify periodontal regeneration. Histological techniques (cementum, bone regeneration) were reported in 91%, radiological measures (bone volume) were reported in 14%, biological analyses (percentage of cells, fluorescence intensity) were reported in 5%, clinical assessment (clinical attachment level [CAL], periodontal pocket depth [PPD]) was reported in 37%, and safety appreciation (postoperative complications, neoplasm formation) was reported in 63% of studies.

Efficacy of Periodontal Regeneration

Bone and cementum regeneration were the most frequently reported outcomes. Bone regeneration was investigated by microscopic and/or radiography examination in 28 studies. Funnel plot and Egger’s test (supplemental online Fig. 4A) suggested publication bias (p < .05) that could be attributed to four studies [25–28]; therefore, these were removed from analysis. The forest plot performed on the remaining 24 studies showed that alveolar bone regeneration was significantly enhanced by MSC therapy (mean difference: 0.66 [95% confidence interval (CI): 0.37–0.95]) compared with control groups treated without MSCs (Fig. 1). Cementum regeneration was investigated by microscopic examination in 18 studies. One study [27] had to be removed from analysis (Egger’s test, p < .05) (supplemental online Fig. 4B). Once removed, the forest plot performed on the remaining 17 studies (Fig. 2) showed that cementum regeneration was significantly promoted by MSC therapy (mean difference: 0.93 [95% CI: 0.62–1.25]).

Figure 1.

Forest plot representing the 24 studies on alveolar bone regeneration by cell therapy. The result of each individual study with its confidence interval was plotted, and then a weighted average was calculated. The pooled analysis was given a diamond shape; the widest aspect of this shape is located at the global estimate and the corresponding horizontal width is the confidence interval. A random effect meta-analysis with the inverse variance method was used to obtain the global standardized mean difference with a 95% confidence interval. Bone regeneration was significantly higher in the cell therapy group than in the control group. Abbreviations: CI, confidence interval; df, degrees of freedom; IV, inverse variance; Std., standard.

Figure 2.

Forest plot representing the 17 studies on cementum regeneration by cell therapy. A random effect meta-analysis with the inverse variance method was used to obtain standardized mean difference with a 95% confidence interval. Cementum regeneration was significantly higher in the cell therapy group than in the control group. Abbreviations: CI, confidence interval; df, degrees of freedom; IV, inverse variance; Std., standard.

Safety Outcomes

There were 19 studies that reported postoperative complications. In one study [54], wound dehiscence was observed twice in four animals from the experimental group. Nine studies reported bone ankylosis, although no difference was observed between intervention and control groups. Eight studies reported root resorption; in seven of these, the number of samples with resorption was the same in the experimental and control groups. One study [55] showed resorption events for the control group only. No formation of a neoplasm occurred in the four studies that investigated this kind of adverse effect. Nevertheless, we identified two conference presentations on neoplasm formation after implantation of human periodontium-derived stem cells in periodontal defects in rats. One of the presentations [56] reported observing anaplastic squamous cell carcinoma in 6 of 11 rats. Another study [57] reported tumor initiation in 50% of immune-deficient rats.

Subgroup Analyses

In order to investigate whether outcomes—more precisely, the magnitude of the difference between experimental and control groups—could have been influenced by study characteristics, we performed several subgroup analyses. We found that bone and cementum regeneration was statistically greater in the experimental group (delivered cells) compared with controls when cells were grafted with bone substitute (mean difference: 0.81 [95% CI: 0.22–1.39] and 0.92 [95% CI: 0.35–1.49]) or with collagen, fibrin, hydrogel, gelatin, or hyaluronic acid carriers (mean difference: 0.79 [95% CI: 0.43–1.14] and 1.12 [95% CI: 0.65, 1.59]) than with PRP, EMD, blood clots, cell sheets, or polymers. Autologous grafts (mean difference: 0.71 [95% CI: 0.45–0.97]) and allogeneic grafts (mean difference: 1.86 [95% CI: 0.73–2.98]) enhanced significantly more bone regeneration than xenogeneic grafting (mean difference: 0.07 [95% CI: −0.77 to 0.91]) (p = .04). Whatever the cell source, experimental groups displayed statistically greater bone and cementum regeneration compared with control. Cementum regeneration by bone marrow stromal cells (BMSCs) or adipose-derived stromal cells (ADSCs) (mean difference: 1.03 [95% CI: −0.16 to 2.23]) was nearly significant. More details are provided in supplemental online Table 4A and 4B.

Periodontal Regeneration in Human Studies

We identified seven human case reports (supplemental online Table 5) in which periodontal regeneration was achieved. Studies investigated periodontal ligament stromal cells (two studies), BMSCs (one study), and cells derived from gingiva (three studies) or periosteum (two studies). Among the three studies using a control group, two reported data from MSCs in combination with hydroxyapatite (HA) versus HA alone, suggesting higher probing-depth reduction [58] and attachment gain [58, 59] in defects treated with gingival or periodontal ligament cells compared with controls. Improvement of CAL and PPD in the group treated with BMSCs compared with controls was also reported [60]. In addition, we identified four clinical trials from our search of the gray literature (supplemental online Table 5).

Discussion

This systematic review provides evidence of the capacity of MSCs to regenerate periodontal tissues in animals; however, the safety of such interventions in humans still needs to be investigated. Subgroup analyses showed marked differences between experimental and control groups, suggesting that periodontal regeneration may be influenced by variables (e.g., carriers or cells sources) that are crucial to address before translation to human clinical trials.

The efficacy of periodontal regeneration techniques has been predominantly assessed using a quantitative approach (amount of tissues), but now investigation with greater statistical rigor is required to study outcomes related to the inherent structural aspects of the formation of a new functional attachment, the bone quality (e.g., cellularity), the type of cementum (i.e., regeneration of acellular extrinsic fiber cementum), and the hierarchy and organization of the periodontal ligament fibers [61].

The question of which biomaterial to use to support cell delivery is critical in treating mineralized tissue defects because it is well known that the outcomes of surgical therapy on periodontal pockets may depend on the scaffold used [62]. Carriers should mimic the cell microenvironment, in which extracellular substrates are able to contribute to their control over cell fate acquisition [63] while permitting cell-to-cell and cell-to-matrix interactions [64, 65]. Interestingly, this review did not find evidence of significant bone or cementum regeneration when cells were delivered as cell sheets or with EMD or PRP adjunctive therapy when compared with carrier alone. Because EMD and PRP are already used when treating humans with the intent of periodontal regeneration, and because in vitro studies previously showed that these materials or delivery systems promoted MSC differentiation [66], their lack of efficiency in cell therapy for periodontal pockets is surprising. Consequently, we hypothesized that cell sheets, EMD, and PRP may not be suitable as carriers in the type of defects created in these particular studies, which were mainly furcation defects in dogs. We showed that bone substitutes enhanced the efficiency of grafted MSCs in experimental periodontal defects probably because they fill the wound, stabilize the blood clot, and confine the cells within the surgical site without rapid resorption, which could alter the local environment for regeneration [67]. For now, given the biomaterials in our possession for safe and routine use, we suggest that the best way of delivering cells may depend on the type of defect and, specifically, the number of alveolar bone walls involved. When lesions are retentive, liquid or gel scaffolds might be used without risking cell dispersal. When defects are larger, outcomes may improve when cells are associated with bone substitute that confines them to the surgical site. Our analysis pointed out that it is not currently possible to state whether the use of MSCs for periodontal regeneration gives better results than conventional therapies. We suggest that future animal studies should compare cell therapy with current conventional periodontal regenerative therapies, with special focus on the type of cell-supporting scaffolds, and should use a split-mouth strategy to reduce interindividual variability.

Another key issue is the availability of cells. Given that periodontitis is not a terminal disease, the risk-benefit balance requires that any therapy should be strongly justified. Thus, sampling should be easy and painless and have a low risk of complications. Our systematic review showed that periodontal therapy with xenogeneic stem cells did not bring about positive outcomes. The only way to bypass the lack of availability of MSCs at the time of treatment appears to be to use frozen autologous or allogeneic cells. Five studies suggested that cryopreservation did not alter periodontal regenerative potential of MSCs. Nevertheless, except for one study [27], no adequate control group was present (e.g., a cell group without cryopreservation) to draw any robust conclusion. Many studies have examined the influence of the source of MSCs on their capacity to participate in periodontal regeneration. Obviously, oral cells appeared to be good candidates for periodontal therapy; however, because these cells are derived from PDL, dental pulp, gingiva, alveolar bone, or dental follicle cells, they are not always conveniently available in clinical practice. It is crucial to have other nonoral MSC sources. Data analyzed in this review showed that cells from extraoral bone marrow sites or from adipose tissue were comparable to adult mesenchymal stem cells derived from dental tissues in their ability to regenerate periodontal bone. This result confirmed that the microenvironment and surrounding tissue were important factors that influenced the fate of MSCs ultimately used [63]. Indeed, BMSCs were shown to shift toward periodontal ligament-like cell features when cocultured with the periodontal ligament in vivo [68]. In addition, because they have been used successfully in extraoral connective tissue regeneration [69], adipose-derived stromal cells are expected to be a valuable source of cells. These ADSCs can be easily isolated from intact resected adipose tissue or by using liposuction, and their properties are similar to cells isolated from bone marrow [14, 15]. However, quantitative data regarding the capacity of ADSCs to contribute to periodontal regeneration were sparse at the time of writing [70, 71].

Interpretation of data from this systematic review should be made with caution and balanced according to the limitations of reported data contained in studies included in this review. We suggest that methodological aspects of animal experiments should be a focus for improvement. It is important to calculate the number of animals that should be treated to obtain a power analysis predictive of reliable results, yet such calculations were reported in only one study [55]. Randomization of the defects to be attributed to treatment groups was performed in only 49% of studies. Six studies lacked control groups or did not compare regeneration with MSCs with a similar experimental design without MSCs. Methods used to minimize subjective bias, the reporting of animal randomization, and blinding design should also be a focus for improvement, and with these improvements, the selection and detection bias should also be minimized [72]. Finally, experimentally generated defects used in most of the animal studies do not mimic the natural pathophysiology of periodontitis, and the method of lesion induction might bias therapeutic outcomes. For future studies, we suggest that periodontal regenerative procedures should be investigated in animal models, including rodents, in which the defects that are generated more accurately reflect both the inflammatory status and the pathogenic microflora shift seen in periodontitis [73]. Periodontal cell therapy should also be explored in large animals. In particular, the minipig model seems to be relevant because the morphology of teeth and anatomy of the periodontal region in swine are close phenocopies for human traits; moreover, in the minipigs, periodontitis occurs spontaneously from 16 months of age [25, 55].

Conclusion

Given the heterogeneity of present studies, narrative reviews are insufficient. A systematic approach with meta-analysis is essential to provide guidance to support future studies and should provide data that can be generalized. The challenge remains to identify the best combination of cells, biomaterials, and biomolecules for various clinical situations, using animal models that best represent the etiopathophysiology of human periodontitis. Particular attention should be given to methodology used in randomized clinical trials. Moreover, even if some clinical trials are already recruiting, animal research should be maintained in parallel.

Author Contributions

P.M.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; J.-N.V.: conception and design, data analysis and interpretation, manuscript writing; C.N.: conception and design, final approval of manuscript; M.S.: administrative support, financial support, final approval of manuscript; M.L.S.: data analysis and interpretation, manuscript writing, final approval of manuscript, manuscript editing; V.P.-B.: conception and design, manuscript writing, final approval of manuscript; L.C.: conception and design, financial support, manuscript writing, final approval of manuscript; P.K.: administrative support, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.