In vitro studies
The first indication that the RANKL-OPG system may be expressed in the pulp region came with the in vivo demonstration that these molecules are immunolocalized on the ameloblast, odontoblast and pulp layers of neonatal murine teeth (Rani & MacDougall 2000). Further in vitro investigation demonstrated that both RANKL and OPG are constitutively expressed in murine odontoblasts and dental pulp cells (Rani & MacDougall 2000). Human dental pulp cells are also shown to express RANKL (Uchiyama et al. 2009, Belibasakis et al. 2011) and OPG (Sakata et al. 1999, Mizuno et al. 2005, Uchiyama et al. 2009, Belibasakis et al. 2011) in vitro. Nevertheless, in coculture with appropriate osteoclast precursors, the murine cells failed to induce or even inhibited osteoclast formation (Rani & MacDougall 2000), whereas the human pulp cells supported this action (Uchiyama et al. 2009). These findings denote that the control of osteoclastogenesis by dental pulp cells could be species specific.
Bacteria are putative regulators of RANKL and OPG in dental pulp cells, with potential effects on osteoclast or odontoclast formation. Nevertheless, in vitro studies on this topic are very sparse. An in vitro biofilm model consisting of six species representative of the supragingival microbiota was investigated for its effects on RANKL and OPG gene expression in human dental pulp and periodontal ligament cells. This multibacterial challenge up-regulated RANKL expression in both dental pulp and periodontal ligament cells, but had 4-fold greater effect on the former. Consequently, the relative RANKL/OPG ratio was strongly up-regulated in dental pulp cells, an event which could favour bone or dentine resorption and contribute to the development of apical periodontitis or internal root resorption (Belibasakis et al. 2011). Porphyromonas endodontalis, a black-pigmented anaerobic species with high association to endodontic infections (van Winkelhoff et al. 1992) was investigated for its effects on RANKL in human osteoblasts, rather than dental pulp cells. A time-dependent induction of RANKL protein production was observed, which was evident after 4 h of challenge with P. endodontalis (Chen et al. 2009). As the regulation of OPG was not investigated, it is not possible to conclude on the net effects on the RANKL/OPG ratio. Experimental systems using human osteoblasts may still be of relevance to apical periodontitis, as endodontic pathogens can well reach the periapical osteoblasts and stimulate the production of bone resorbing factors.
Several inflammatory mediators can also regulate the RANKL-OPG system in dental pulp cells. In an early study, interleukin (IL)-1β and TNF-α enhanced OPG expression by human dental pulp cells at a concentration of 0.5 ng mL−1, after 12 h of stimulation (Sakata et al. 1999). However, this study failed to detect RANKL expression by the cells. These early results would imply a potential role of cytokines on inhibition of dentine or bone resorption, despite their pro-inflammatory capacities. However, later studies have confirmed the expression of RANKL by dental pulp cells and its regulation by pro-inflammatory mediators. Both IL-1α and TNF-α, at concentrations of 10 ng mL−1, were shown to stimulate RANKL gene and protein expression, in a time-dependent manner. Their coadministration had a synergistic effect on RANKL expression, but down-regulated OPG expression (Kim et al. 2010). These results in dental pulp cells denote that the relative RANKL/OPG ratio is up-regulated by pro-inflammatory cytokines, well in line with their capacity to stimulate events related to hard-tissue resorption. Further on, this work demonstrated that cytosolic phospholipase A2 was responsible for the induction of PGE2 and the subsequent induction of RANKL expression (Kim et al. 2010). Substance P, a sensory neuropeptide, was shown to stimulate PGE2 and RANKL production by dental pulp cells (Kojima et al. 2006). PGE2 partially mediated the induction of RANKL gene expression in that experimental system, and its subsequent production by the cells. Conditioned medium from these cell cultures stimulated pit formation by osteoclasts on dentine slices. As a mediator of pulpal inflammation (Kvinnsland & Heyeraas 1992), substance P may exert its role in root or periaplical bone resorption by stimulating RANKL in a PGE2-dependent mechanism. Importantly, the findings of this study imply that dental pulp cell-derived RANKL supports the formation of functional osteoclasts on dentine. Therefore, it is reasonable to speculate that the RANKL-OPG system could control odontoclast formation in the dental pulp, thus mediating internal root resorption. However, this remains to be further elucidated.
A number of biological agents could potentially regulate RANKL and OPG expression in dental pulp cells. As such, inorganic polyphosphate, a biopolymer with several biological functions on eukaryotic cells, has been investigated for its potential effects. Long exposure (3 weeks) of human dental pulp cells to this agent has been shown to induce a 3.3-fold higher OPG expression, compared to the un-stimulated control (Kawazoe et al. 2008). Platelet-rich fibrin is a platelet concentrate that resembles a fibrin network and contains a large number of growth factors and cytokines, with the potential to enhance wound healing (Dohan et al. 2006). This agent up-regulated OPG production by dental pulp cells, already 1 day after stimulation, and persisted over a period of 5 days (Huang et al. 2010). Stem cells isolated from human dental pulp have also been considered with regards to OPG expression. These were seeded on silk scaffolds and underwent mechanical tension in a bioreactor environment. OPG expression was slightly increased in cells under tension (Han et al. 2010). The induction of OPG in these studies could indicate a protective effect against dentine or bone resorption, and a positive switch of the homoeostatic balance towards tissue formation. Nevertheless, RANKL was not studied in this context, and therefore, the overall regulation of the RANKL-OPG system cannot be deduced.
In conclusion, little in vitro evidence is available on the mechanisms of the regulation of the RANKL-OPG system in dental pulp cells, or in response to endodontic pathogens. Yet, the available information indicates that RANKL and OPG can be regulated by bacterial or pro-inflammatory stimuli in these cells, in a manner that favours bone resorption, and possibly via the mediation of PGE2. It should be reiterated that future studies in this direction must consider the concomitant investigation of RANKL and OPG. Moreover, relatively limited information is available on the capacity of dental pulp cells to stimulate odontoclast formation via the RANKL-OPG system. This is an open field of study, which could provide answers to the molecular mechanisms governing internal root resorption, and needs to be further pursued.
The first in vivo demonstration of the involvement of the RANKL-OPG system in periapical lesions came by immunolocalization studies, in experimental animal models. In particular, periapical localization of RANKL was initially investigated in a rat model whereby the pulp of the mandibular first molars was experimentally exposed (Zhang & Peng 2005a,b). Osteolcast-like cells, which are tartrate-resistant acid phosphatase (or short TRAP)-positive, and RANKL-positive cells were observed periapically, already 1 week after pulpal exposure. By end of the second week, dense inflammatory infiltrates and periapical bone resorption was observed, accompanied by a climax in RANKL-positive and osteoclast-like cells. After 4 weeks of exposure, although the chronic inflammatory infiltrates were established, the presence of these osteoclast-like and RANKL-positive cells was reduced to baseline levels, and periapical bone resorption was slowed down (Zhang & Peng 2005a,b). This study revealed the involvement of RANKL in periapical bone resorption, and identified that the peak of its expression is commensurate with a peak of osteoclast activity and bone resorption. This was further confirmed on the mRNA expression level, in a later study using a similar experimental system, which also took into account the expressions of OPG and RANK (Kawashima et al. 2007). The periapical expression of RANKL peaked after 2–3 weeks of pulpal exposure and maintained higher than baseline levels, over a period of 8 weeks. RANK and OPG expressions followed a similar kinetic pattern of increase, however, not as pronounced as RANKL. The enhancement of OPG expression was explained as a compensatory mechanism for the rapid induction of RANKL. Nevertheless, the relative RANKL/OPG expression ratio peaked over 3 weeks, and persisted at high levels over the 8-week observational period, suggesting an enhanced capacity for bone resorption. Accordingly, expression of pro-inflammatory cytokines IL-1α, IL-1β and TNF-α, with the capacity to induce RANKL, also peaked during the interval between 2 and 3 weeks. Immunohistochemically, RANKL-positive cells with diverse morphologies were detected in the periapical lesion, in close proximity to the alveolar bone and to osteoclast-like cells. These were identified as mainly fibroblastic, but a small proportion of T cells were also evident. Therefore, this study demonstrated that a local increase in the RANKL/OPG expression ratio, occurring 2–3 weeks after pulpal exposure, is concomitant with the expansion of the periapical lesion.
Pulp exposure models in animals allow for the ‘natural’ development of a periapical lesion. Molecules such as RANKL and OPG can thus be studied regarding their role in the development of hard-tissue lesions in real time. However, some of the animal studies reviewed here may not have taken stringent consideration of the bacterial engagement in this process. In a recent variation of this experimental model, the infectious agent was controlled by coadministration of Escherichia coli LPS (soaked on a paper-point) into the exposed distal root canal of the lower first molars, up to the apex (Chuang et al. 2012). Periapical bone resorption, expression and localization of RANKL and OPG, and osteoclast formation were followed over a period of 8 weeks. A boost in RANKL expression was observed after 3 weeks of exposure, culminating in 208% of the unexposed pulp control, after 8 weeks. OPG expression was fluctuating between weeks 1 and 8, but was detected at lower levels than the control group. Accordingly, a very strong induction of the relative RANKL/OPG expression ratio was evident at weeks 7 and 8, which was, respectively, 30 and 41 times higher than the control group. Increased localization of both RANKL and OPG in the periapical lesions was evident from week 3 onwards, accompanied by an increase in the presence of osteoclast-like cells. This model of LPS-induced periapical lesions in the rat provides insights to the involvement of RANKL and OPG in this infectious process (Chuang et al. 2012). Nevertheless, it is difficult to interpret the relative contribution of LPS, owing to the lack of a control group where the pulp is exposed, yet not treated with LPS.
Experimental animal models are amenable to technical manipulations, allowing for the study of specific factors in the development of the disease. For instance, ovariectomized rats have been used to study the relative effect of oestrogen deficiency on the synthesis of RANKL and OPG, in induced periapical lesions (Zhang et al. 2007). It was found that in the ovariectomized rat group, more osteoclasts and more RANKL-positive cells were present, than the sham-ovariectomized group, as early as 1 week after the induction of periapical lesion. Nevertheless, OPG production was also increased at early time points in the oestrogen-deficient mice, which by time was reduced to control levels. This is perceived as an early protective response of the tissue, to compensate against the overproduction of RANKL and, consequently, bone resorption. Overall, oestrogen deficiency appears to accelerate periapical bone loss, characterized by high local levels of RANKL production (Zhang et al. 2007).
Experimental periapical lesion models have also been utilized in genetically modified mice, to evaluate the involvement of specific genes in the RANKL or OPG-associated responses. In a mouse model of endodontic infections, the relative role of osteopontin (OPN) was considered (Rittling et al. 2010). OPN is a secreted integrin-binding protein that constitutes part of the extracellular matrix of bone. It is also considered to play a functional role in immunoregulatory responses. In this model, periapical tissue RANKL expression was increased already 3 days after infection, and was more than 2-fold higher in the OPN-deficient mice, compared to the wild-type counterparts. After 3 weeks, periapical bone loss was significantly more severe in OPN-deficient mice and was characterized by dense inflammatory infiltrates. This study revealed a potential negative association between OPN and RANKL in endodontic infection, which awaits further investigation (Rittling et al. 2010).
In mice lacking the CCR2 chemokine receptor, inoculation of four bacterial taxa (Porphyromonas gingivalis, Prevotella nigrescens, Actinomyces viscosus, Fusobacterium nucleatum) into the pulp resulted in larger periapical lesion formation, compared to the wild-type mice. After 1 week of infection, the periapical expression of RANKL was higher, whereas that of OPG was lower in the CCR2-deficient mice. Nevertheless, these differences were not significant after 2 weeks. This study concluded on a protective role of CCR2 against bacterially induced periapical osteolysis (Garlet et al. 2010).
The relative role of nitric oxide synthase (iNOS) or phagocyte oxidase (PHOX) in a pulp exposure model in mice was also studied (Silva et al. 2011). After 2 weeks, experimental (pulp exposed) iNOS-deficient mice exhibited greater periapical resorption area and number of osteoclasts, and higher RANK, RANKL, but not OPG, expression levels, compared to wild-type mice. On the contrary, PHOX-deficient mice did not exhibit any periapical osteoclasts and the periapical expressions of RANK, RANKL and OPG were not significantly different, compared to the wild-type mice. Collectively, the results of this study indicate that nitric oxide, but not reactive oxygen species, are involved in the progression of periapical bone resorption (Silva et al. 2011). The role of iNOS-inducible nitric oxide in periapical bone resorption has also been demonstrated in an earlier study involving bacterial infection of the root canal with the four previously mentioned bacterial species (Fukada et al. 2008). In this study, iNOS-deficient mice with exposed pulp exhibited higher number of osteoclasts, and higher RANK, but lower OPG expression, compared to the wild-type experimental mice. Periapical RANKL expression was not different between iNOS-deficient and wild-type mice. Hence, iNOS deficiency was considered as a factor causing an imbalance of the bone resorption modulating-factors, and leading to increased infection-induced periapical bone resorption.
In conclusion, the above-described animal studies have mostly involved pulpal exposure models to generate periapical lesions, for the study of RANKL and OPG localization and expression. These studies confirm that this molecular system is also implicated in pathological periapical bone resorption, and reveal a locally enhanced RANKL/OPG ratio, in line with observations in other bone-destructive pathoses. However, it is noteworthy that no animal experimentations exist in which the RANKL-OPG system is studied within the dental pulp, in relation to pulpitis. This is an important gap in our knowledge that needs to be fulfilled, as it could reveal mechanisms involved in odontoclast formation and internal root resorption.
The systematic search produced 13 publications, which fulfilled the criteria (Table 2). In human teeth, RANKL was first identified in pulps of deciduous molars (Lossdörfer et al. 2002). Immunohistochemistry revealed cytoplasmatic granular staining for RANKL in odontoblasts and pulp fibroblasts in roughly half of the investigated specimens. However, RANKL appears to be expressed at much higher levels in pulps of deciduous teeth undergoing resorption (i.e. shedding) than in counterparts of healthy permanent teeth (Yildirim et al. 2008). In healthy deciduous tooth pulps, OPG is expressed at higher levels than RANKL (Yildirim et al. 2006). Only one study was identified that investigated the levels of OPG in human pulps of different inflammatory states (Kuntz et al. 2001). This investigation revealed intense immunostaining for OPG in the odontoblastic layer of healthy pulps. OPG was not detected in the central area of these pulps. In contrast, in inflamed pulps challenged by gross dental decay, the central aspects showed intense immunostaining also. RANKL levels were not investigated. The steady-state expression of OPG in odontoblasts could constitute a protective response of the pulp against imminent inflammatory dentine resorption.
The remaining 10 publications on RANKL and/or OPG that were identified, dealt with periapical inflammation and its possible modulation. A first immunohistochemical study showed the presence of RANKL in radicular cysts (Tay et al. 2004). Immunolocalization of RANKL was in concordance with cells staining for TRAP, i.e. osteoclasts. A subsequent publication confirmed the presence of RANKL at the gene expression level in inflammatory periapical lesions of undisclosed nature (Sabeti et al. 2005). RANKL mRNA was also semi-quantified in periapical granulomas, whilst it was below detection limit in healthy periodontal ligament (Vernal et al. 2006). Expression of RANKL on infiltrate leucocytes was further investigated using flow cytometry: monocytes (CD14+) and dendritic cells (CD83+) were the main synthesizers of RANKL in granuloma lesions (Vernal et al. 2006). An immunohistochemical study compared RANKL and OPG levels between apical granulomas and radicular cysts (Menezes et al. 2006). It was found that the ratio of OPG+/total cells and that of RANKL+/total cells was higher in granulomas than in cysts. However, the ratio between RANKL and OPG-positive cells did not differ between these two types of periapical inflammatory lesion. Again, various cell types stained positive for RANKL and OPG, with macrophage-like cells (CD68+) showing the highest intensity. Another investigation compared RANKL (but not OPG) mRNA levels between granulomas and cysts, demonstrating that its expression was significantly higher in granulomas than in cysts (Fukada et al. 2009). One study, however, did not identify any difference between total RANKL or OPG protein levels or their ratio between granulomas and radicular cysts (Fan et al. 2008).
Few human studies have attempted to investigate the role of the RANKL-OPG system in endodontic disease initiation or progression. In one investigation, periapical lesions were graded according to their inflammatory status (Fan et al. 2011). The authors found significantly more RANKL-positive cells in severely inflamed lesions compared to lightly inflamed counterparts. However, the RANKL/OPG ratio was statistically similar between inflammations graded as light, moderate, or intense. The RANKL/OPG ratio at the gene expression levels was also compared between granulomas and periodontal ligament of orthodontically moved teeth (Menezes et al. 2008a). Whilst the compression sites of orthodontically moved teeth almost consistently showed RANKL mRNA levels to be higher than OPG counterparts, and tension sites the reversed ratio, i.e. OPG > RANKL, this ratio was inconsistent with granulomas. The upstream transcription of pro-inflammatory cytokine genes regulating bone resorption is modulated by a group of molecules termed suppressors of cytokine signalling (SOCS) (Starr et al. 1997). In granulomas, it was shown that RANKL gene expression is negatively correlated to SOCS1 mRNA levels (Menezes et al. 2008b).
Only one human study attempted to correlate the presence of infective agents, namely cytomegalovirus and Epstein-Barr virus, to RANKL gene expression in granulomas (Yildirim et al. 2006). Whilst there was a higher occurrence of these viruses in the periapical lesions compared to healthy pulp tissues, no correlation was found between their presence and RANKL.