Porphyromonas gingivalis lipopolysaccharide regulates ephrin/Eph signalling in human periodontal ligament fibroblasts

Objective EphrinA2‐EphA2 and ephrinB2‐EphB4 critically engage in bidirectional signalling to modulate alveolar bone remodelling. The present study aimed to investigate the effects of lipopolysaccharides (LPS) derived from Porphyromonas gingivalis on ephrin/Eph signalling in periodontal ligament fibroblasts (PDLFs). Material and Methods The primary cultured PDLFs were incubated in the absence (as a control) or presence of P. gingivalis LPS at 0.001‐10 μg/mL for 24 hours. The PDLFs were then stimulated with P. gingivalis LPS at the optimal concentration (0.1 μg/mL) for different periods (6‐48 hours). The expression of ephrinA2, ephrinB2, EphA2 and EphB4 was assessed by quantitative reverse‐transcription real‐time polymerase chain reaction and western blotting. The osteoblastic markers alkaline phosphatase, osteocalcin and Runt‐related transcription factor 2 (Runx2), and the osteoclastogenesis‐related factors receptor activator of nuclear factor kappa‐B ligand (RANKL) and osteoprotegerin were also evaluated. Results The ephrinA2 and EphA2 expression was upregulated and EphB4 expression was downregulated by stimulation of P. gingivalis LPS. EphrinA2 mRNA expression in the PDLFs was significantly upregulated from 12 to 48 hours (P<.05), whereas EphA2 exhibited no change for the first 24 hours, after which there was a significant increase at 48 hours (P<.05). EphB4 exhibited lower mRNA expression at 12 and 24 hours than did the control (P<.05), but the change was insignificant at 48 hours. In contrast, the expression of ephrinB2 remained unchanged. The expressions of ephrinA2, EphA2, ephrinB2 and EphB4 at the protein level showed a similar pattern to that at the mRNA level. The expression of Runx2 and osteocalcin significantly decreased, whereas that of RANKL/osteoprotegerin increased. Conclusion The present study suggest that P. gingivalis LPS would contribute to a dysregulation of bone remodelling, whereby ephrinA2/EphA2 expression is stimulated and EphB4 expression is inhibited.

substances involved in the regulation of alveolar bone remodelling. 1 In the mechanism of bone-related activity, Runt-related transcription factor 2 (Runx2) is associated with osteoblast differentiation. 2 Alkaline phosphatase (ALP) is closely associated with bone mineralization, and together with osteocalcin, acts as a hallmark of osteoblast differentiation in the regulation of osteogenesis. In contrast, the receptor activator of nuclear factor kappa-B ligand (RANKL)-osteoprotegerin (OPG) axis is known to be a master factor regulating osteoclast formation. 3 Under this physiological condition, the maintenance of alveolar bone is achieved through a dynamic equilibrium between bone deposition and resorption. However, the balance of this process is disturbed by the presence of chronic periodontal inflammation, resulting in activation of osteoclastogenic factors and subsequent bone loss. 4 Recent research has found that bone cell interactions are modulated by bidirectional signalling, ephrin ligands and Eph receptors. 5 The receptor tyrosine kinases of the Eph family interact with the ephrin ligands at the cell surface because both are transmembrane proteins. 6 Eph and ephrin molecules function as a receptor-ligand pair in cell-cell communication processes, triggering bidirectional signal transduction; namely, forward signalling via Ephs and reverse signalling via ephrin ligands. 7 Both ephrins and Ephs are divided into two subclasses, A and B. Recent studies have indicated that various bone cells express ephrin ligands and Eph receptors, of which ephrinB2-EphB4 and ephrinA2-EphA2 differentially regulate the bone biology. 8,9 In that way, the bidirectional signalling between osteoclastic ephrinB2 and osteoblastic EphB4 suppresses osteoclastogenesis and enhances osteogenesis in osteoclast-osteoblast communication. In contrast, reverse signalling via ephrinA2 and forward signalling via EphA2 reinforce osteoclastic bone resorption and inhibit osteoblastic bone formation. 5 Studies on ehprin-Eph largely focused on bone cells until 2011, when Diercke et al. 10,11 reported that periodontal ligament fibroblasts (PDLFs) also express ephrin/Eph molecules. Their research also revealed that ephrinB2-EphB4 and ephrinA2-EphA2 signalling are particularly important for the regulation of bone remodelling during orthodontic force loading between PDLFs and osteoblasts. Beyond the aforementioned studies by Diercke et al., little research has elaborated on the ephrin/Eph signalling in PDL and alveolar bone. Therefore, further investigation of this bidirectional signalling in PDL is worthwhile, particularly its role in periodontal inflammation when the bone remodelling balance has been broken and the osteoclastogenic factors and subsequent bone loss are activated. The results are expected to clarify the mechanism and shed light on the prevention and treatment of periodontal inflammation diseases.
It is well established that periodontitis, or inflammation of the periodontium, is triggered by pathogenic bacteria (mainly gram-negative anaerobes such as Porphyromonas gingivalis 12 in the biofilm. It is characterized by the pathologic destruction of tooth-supporting tissues, which can cause loss of the tooth. Lipopolysaccharides (LPS), the major element in the outer membranes of gram-negative bacteria, act as prototypical endotoxins. LPS derived from P. gingivalis have been shown to promote the secretion of periodontal disease-related proinflammatory cytokines by PDLFs. 13 LPS are also involved in RANKL and OPG-regulated osteoclast formation and bone resorption in PDLFs, 14 and their inhibitory effect on the osteoblastic differentiation of osteoprogenitor cells. 15 LPS are thus believed important pathogens in periodontitis that can lead to alveolar bone resorption in most advanced cases. Although LPS have been extensively studied and shown to contribute to the pathogenesis of periodontal disease, 16 their possible effects on periodontitis bone resorption via ephrin/Eph signalling are still poorly understood. PDLFs are the dominant cell population in the PDL, and have been shown to express high levels of ALP activity and to secrete RANKL and OPG, indicating that they participate in bone remodelling. 17 Therefore, it has been argued that the PDL plays a crucial role in alveolar bone metabolism. Collectively, we hypothesized that ephrin/Eph expression is also regulated by LPS stimuli in PDLFs. This study was therefore designed to extend the roles of the ephrinB2-EphB4 and ephrinA2-EphA2 systems to include PDLFs during inflammatory events.

| Primary cell culture
Human PDLFs were cultured using the method described in earlier research with slight modifications. 18 Healthy premolars extracted for orthodontic purposes were collected from four healthy donors (aged 13.5±1.7 years) who gave signed informed consent according to the Institutional Review Board of the University of Hong Kong (IRB: UW13-120). Briefly, to obtain primary PDLF tissue fragments of the PDL were isolated from the middle third surface of the tooth root with a scalpel. The cells were cultured in modified Eagle's mediumalpha, supplemented with 10% foetal bovine serum and 1% antibiotics (100 U/mL penicillin and 100 μg/mL streptomycin) and incubated at 37°C in a humidified atmosphere with 5% CO 2 . The culture medium was replaced every 3 days after passaging, and the fifth passage cells cultured in monolayers were ready to be used for all experiments.

| Lipopolysaccharide challenge of periodontal ligament fibroblasts
After reaching confluence, the PDLFs were seeded at a density of The concentrations of P. gingivalis LPS (0.001-10 μg/mL) we selected, which could be refined to different experiment protocols, are based on previous research regarding the in-vitro PDLF cultures. [19][20][21] To investigate the time-dependent effects, P. gingivalis LPS of the optimal concentration (0.1 μg/mL) was used to stimulate the PDLFs for different experimental periods (6, 12, 24 and  that differentially mediate the expression of immuno-inflammatory cytokines. 24,25 The P. gingivalis LPS used in the present study (LPS-PG Ultrapure, InvivoGen Catalogue number: tlrl-ppglps) is a highly purified preparation of LPS from the gram-negative bacteria P. gingivalis, which is a generally used one obtained from InvivoGen and used in other studies. [26][27][28] However, the isoform and molecular weight of the P. gingivalis LPS have not been specified by the manufacturer.

| Quantitative reverse-transcription real-time polymerase chain reaction
Total cellular messenger RNA (mRNA) was extracted from the PDLFs using an RNeasy Mini Kit (Qiagen, Hilden, Germany), and its concentration was determined by the absorbance measurement at 260 nm.  Table 1. The amplification reactions were implemented in 96-well plates with a mixture volume of 20 μL that contained 1 μL cDNA and 10 μL Power SYBR Green PCR Master Mix (Applied Biosystems) in each well with the following programme: initial heating to 95°C for 10 minutes, followed by 40 cycles of denaturation at 95°C for 15 seconds and annealing and extension at 60°C for 1 minutes.

| Western blotting
The PDLFs were collected using an RIPA buffer (Thermo Scientific, Rockford, IL, USA) containing a protease inhibitor cocktail. The Pierce BCA Protein Assay Kit (Thermo Scientific) was used for total protein quantification. A lysate with 25 μg (for ephrinB2 and ephrinA2 detection), 30 μg (for EphA2 detection) or 35 μg (for EphB4 detection) of protein was then fractionated on 10% sodium dodecyl sulphatepolyacrylamide gel and transferred on to a nitrocellulose membrane.
The membranes were blocked for 1 hour at room temperature with 5% non-fat milk in TBS-T (10 mM Tris, 100 mM NaCl and 0.1% Tween20), after which they were probed overnight at 4°C with rab-

| Statistical analysis
The experiments were performed in triplicate and the results were expressed as means±standard deviation (SD). The data were compared with independent two-tailed t tests. A P value of .05 was considered to show statistical significance.

| mRNA expressions of ephrinB2-EphB4 and ephrinA2-EphA2 in a P. gingivalis lipopolysaccharide dose-dependent manner
After culturing with P. gingivalis LPS at different concentrations for 24 hours, the mRNA expression levels of ephrinB2-EphB4 and T A B L E 1 Primer sequences of the target genes for quantitative reverse transcription-polymerase chain reaction  (Figure 1). A P. gingivalis LPS concentration of 0.1 μg/mL was therefore considered optimal for the subsequent experiments.

| Time-dependent mRNA Expression of ephrinB2-EphB4 and ephrinA2-EphA2
EphB4 expressed lower levels of mRNA at 12 hours (P<.05) and 24 hours (P<.01) than did the control group, but the change was insignificant at 48 hours ( Figure 2D). The expression of ephrinB2 remained unchanged with different concentrations of P. gingivalis LPS over the treatment time ( Figures 1C and 2C).
EphrinA2 expression significantly increased in a time-dependent manner from 12 to 24 hours (P<.05) and slightly decreased at 48 hours, but remained markedly higher than in the control group (P<.01) (Figure 2A). The expression of EphA2 showed no change over the first 24 hours, after which a significant increase was seen at 48 hours (P<.05) ( Figure 2B).

| Expression of ephrinB2-EphB4 and ephrinA2-EphA2 at the protein level
The results of western blotting confirmed the findings of the mRNA expression level. The changing tendency at the mRNA and protein levels was the same but was observed at different treatment time points.
The protein expressions of ephrinA2 and EphA2 were increased at 24 and 72 hours, respectively (Figure 3), whereas the upregulation of ephrinA2 mRNA expression was found from 12 to 48 hours and EphA2 was upregulated at 72 hours ( Figure 2B).

| Expression of osteoblastic markers (alkaline phosphatase, osteocalcin and Runt-related transcription factor 2) and osteoclastogenesis-related factors (receptor activator of nuclear factor kappa-B ligand and osteoprotegerin)
For the osteoblastic markers, the PDLFs expressed decreasing mRNA of Runx2 (P<.01) and osteocalcin (P<.001) after stimulation with 0.1 μg/mL P. gingivalis LPS for 24 hours. No significant variation in ALP was found (Figure 4). In contrast, the RANKL/OPG ratio was upregulated significantly (P<.001) with decreased OPG (P<.001).

| DISCUSSION
In the present study, we investigated the ephrin/Eph expression EphA2 is likely to be preferred among osteoclast precursors during early differentiation. 9 In this case, the ephrinA2-EphA2 interaction facilitates the initiation phase of bone remodelling by enhancing osteoclast differentiation and suppressing osteoblast differentiation.
Both reverse ephrinA2 signalling and forward EphA2 signalling enhance osteoclast differentiation, but it is suppressed only by forward signalling via EphA2. 9 We found that reverse ephrinA2 signalling was There is also evidence that blocking ephrinB2-EphB4 interaction inhibits mineralization in osteoblasts. 35 Tonna et al. 36 even ascribed the influence of ephrinB2 in the promotion of osteoblast differentiation to its limiting effect on cell apopto- It has been reported that applying compressive force induces the expression of ephrinA2-EphA2 and reduces that of ephrinB2-EphB4 in PDLFs. 11 The bone resorption on the compression side during orthodontic tooth movement is mediated by the ephrin/Eph system.
In our research on LPS-stimulated PDLFs, ephrinA2-EphA2 and ephrinB2-EphB4 showed changing trends similar to that in the study on compressive force, 11 which suggests that ephrin/Eph can also be modulated by an inflammatory reaction and affects the regulation of bone remodelling.
To gain a more profound profile about the behaviour of PDLFs under inflammatory challenge, we also stimulated the cells with the optimal concentration (0.1 μg/mL) of P. gingivalis LPS for 24 hours, which is the same condition as the evaluation of ephrin/Eph signalling, and assessed the mRNA expression of osteogenic and osteoclastogenic genes. The interplay between RANKL and RANK is critical for osteoclast formation and can also enhance osteoclast activity.
The OPG receptor, other than RANKL functioning as an osteoclast differentiation factor, is an osteoclastogenesis inhibitory factor that can also combine with RANK to prevent a RANK-RANKL interaction and the consequent bone resorption. RANKL and OPG are related to periodontal disease, and a balance between them is essential in bone destruction processes. 40 In this study, the ephrin/Eph system was modulated by P. gingivalis LPS challenge with a concurrent significant decrease in the RANKL/OPG ratio, which is in accordance with the report 14  PDLFs. 11 Furthermore, EphB4 knockdown in osteoblasts displayed a lowered ALP activity. 8 In the present study, the mRNA expression of Runx2 and osteocalcin was significantly downregulated, which suggests that the regulation of ephrins and Ephs in our setting might be responsible for the suppression of osteogenesis. It has been shown that the expression of Runx2 and osteocalcin in LPS-treated osteoblasts is significantly suppressed, 41 which is consistent with the present study.
To date, the P. gingivalis LPS has been demonstrated to regulate the expression of proinflammatory cytokines through activating tolllike receptor (TLR) -2 or TLR-4, 42 while LPS from other gram-negative bacteria is TLR-1. The TLR-2 activity of P. gingivalis LPS is ascribed to the result of lipoprotein contamination. 43 The P. gingivalis LPS used in the present study was a commercial preparation purified by the supplier and has undergone enzymatic treatment to remove lipoproteins so that TLR-2 is not activated according to the manufacturer's instructions. Whether TLR-4 is dependent on the modulation of ephrin/Eph signalling needs further investigation in our future studies on the pathway regarding to ephrin/Eph signalling.
Positive regulation by means of enhancing the expression or negative regulation through attenuating the expression of ephrins or Ephs, and the determination of bone resorption activity, which is used to elucidate the functional consequences, were not performed in this study and will be investigated in a future study.In conclusion, the expression of ephrinA2-EphA2 significantly increased and that of EphB4 decreased in PDLFs upon stimulation with P. gingivalis LPS. These changes occurred with a corresponding expressive variation in the osteoblast differentiation genes (Runx2 and osteocalcin) and osteoclastogenesis-related genes (OPG and RANKL), indicating that ephrinA2-EphA2 and ephrinB2-EphB4 interaction might regulate inflammatory bone resorption. The results of this study will shed light on the development of therapeutic interventions on periodontitis.