Gram-positive bacteria cell wall-derived lipoteichoic acid induces inflammatory alveolar bone loss through prostaglandin E production in osteoblasts

Periodontitis is an inflammatory disease associated with severe alveolar bone loss and is dominantly induced by lipopolysaccharide from Gram-negative bacteria; however, the role of Gram-positive bacteria in periodontal bone resorption remains unclear. In this study, we examined the effects of lipoteichoic acid (LTA), a major cell-wall factor of Gram-positive bacteria, on the progression of inflammatory alveolar bone loss in a model of periodontitis. In coculture of mouse primary osteoblasts and bone marrow cells, LTA induced osteoclast differentiation in a dose-dependent manner. LTA enhanced the production of PGE2 accompanying the upregulation of the mRNA expression of mPGES-1, COX-2 and RANKL in osteoblasts. The addition of indomethacin effectively blocked the LTA-induced osteoclast differentiation by suppressing the production of PGE2. Using ex vivo organ cultures of mouse alveolar bone, we found that LTA induced alveolar bone resorption and that this was suppressed by indomethacin. In an experimental model of periodontitis, LTA was locally injected into the mouse lower gingiva, and we clearly detected alveolar bone destruction using 3D-μCT. We herein demonstrate a new concept indicating that Gram-positive bacteria in addition to Gram-negative bacteria are associated with the progression of periodontal bone loss.


LTA stimulated the production of PGE 2 via the upregulation of PGE 2 synthesis-related genes in osteoblasts.
To clarify the mechanisms of osteoclast differentiation induced by LTA, we examined the effect of LTA on the mRNA expression of Rankl, Cox2 and mPges1 in POBs by RT-qPCR. LTA markedly upregulated the mRNA level of these genes in POBs ( Fig. 2A). We further found that LTA induced the production of PGE 2 by POBs (Fig. 2B).
We previously reported that LPS stimulated the transcriptional activity of NF-κB, leading to the increased mRNA expression of Cox2 and mPges1 1 . To test whether LTA activates the NF-κB pathway, the protein level of IκBα, an endogenous inhibitor of NF-κB, was analyzed by Western blotting, since the activation of the NF-κB pathway by inflammatory molecules rapidly degrades IκBα via the ubiquitin-proteasome system. Adding LTA induced a notable reduction of IκBα (Fig. 2C) and promoted the transcriptional activity of NF-κB in a dual luciferase reporter gene assay (Fig. 2D).

Indomethacin blocked LTA-induced bone resorption.
To determine the roles of PGE 2 in osteoclast differentiation, BMC and POB were cocultured with or without LTA and indomethacin, a typical nonsteroidal anti-inflammatory drug. Indomethacin markedly suppressed the LTA-induced osteoclast differentiation in the cocultures. (Fig. 3A, B). In calvarial organ cultures, indomethacin completely blocked the bone resorbing activity induced by LTA (Fig. 3C). In POBs, LTA-induced PGE 2 production and the mRNA expression of RANKL were significantly blocked by the addition of indomethacin (Fig. 3D, E).
LTA prolonged the life span of mature osteoclasts. We examined the effect of LTA on the lifespan of mOCs. Raw264.7 cells, a mouse macrophage cell line, were differentiated into mOCs by sRANKL, and the mOCs were cultured with sRANKL or LTA. The removal of sRANKL on day 4 induced the cell death of mOCs, while in the absence of sRANKL, LTA prolonged the lifespan of mOCs in a dose-dependent manner (Fig. 4A, B). It is well-known that NFATc1 (Nuclear factor of activated T-cells, cytoplasmic 1) encoded by the Nfatc1 gene is the master transcription factor to osteoclast differentiation and survival, and cathepsin K encoded by the Ctsk gene is a major target gene of NFATc1 and relates to osteoclast activity 16 . To reveal the mechanism of LTAinduced osteoclast survival, we analyzed osteoclast marker genes, Nfatc1 and Ctsk, by RT-qPCR. In the absence of sRANKL, the expression of Nfatc1 and Ctsk in mOCs was increased by the addition of LTA (Fig. 4C).

LTA induced alveolar bone loss in a mouse ex vivo model of periodontal bone resorption and in vivo model of periodontitis.
We first examined an ex vivo model of periodontal bone resorption 1 .
Mouse mandibular alveolar bone was collected and cultured with LTA for 5 days. LTA induced alveolar bone resorption was shown in a dose-dependent manner (Fig. 5A), while indomethacin significantly inhibited LTAinduced bone-resorbing activity (Fig. 5B).
We next examined the in vivo model of periodontitis with alveolar bone resorption 1 . In the mouse model of periodontitis, we injected LTA or PBS into the left outer gingiva of the first molar in the lower jaw at a depth of 1-2 mm on days 0, 2 and 4, and the bone mass of the alveolar bone was analyzed by dual-energy X-ray absorptiometry (DXA) and μCT on day 7. In the DXA analysis, alveolar BMD was significantly reduced by the injection of LTA (Fig. 6A) www.nature.com/scientificreports/ cm 2 , a reduction that was similar to that noted following LPS injection in a mouse model of periodontal bone resorption 1 . Since the tooth root is exposed from the alveolar bone due to increased bone resorption, a 3D-μCT analysis is a definitive method for analyzing alveolar bone resorption in periodontitis 17 . We first examined the alveolar BMD around the areas of the tooth root and alveolar bone using 3D-μCT. The BMD of the alveolar bone was measured around the inter-tooth areas of the 1st and 2nd molars (Fig. 6B, left: [a]) and around the root trunk area of the 1st molar (Fig. 6B, right: [b]; indicated by the dotted square in the upper images of Fig. 6B). The alveolar BMD of the LTA injected group was significantly reduced by approximately 10% in both areas, indicating that the alveolar bone was resorbed due to periodontitis (Fig. 6B). We next clarified the alveolar bone loss using quantitative methods that are frequently used in clinical diagnoses. Analyzing the length of the gingival sulcus to the alveolar bone (inter-tooth space) for alveolar bone loss (Fig. 6C, lengths [a]), and the bottom of the molar root trunk to the alveolar bone (root trunk space) of the 1st molar (Fig. 6C, lengths [b]; indicated by the white line in the upper images of Fig. 6C). Due to the vertical loss of the alveolar bone, it appeared that the inter-tooth space (Fig. 6C, [a]) and the root trunk space of the 1st molar (Fig. 6C, [b]) were increased. We found that the LTA significantly increased both in the inter-tooth space The calcium concentration was measured to elucidate bone-resorbing activity. The data are expressed as the mean ± SEM of 5 wells. A significant difference between the two groups was indicated; ***P < 0.001 versus control; by one-way ANOVA and Tukey's post hoc test. www.nature.com/scientificreports/ (Fig. 6C, [a]) and the root trunk space of the 1st molar (Fig. 6C, [b]); this reflected the progression of alveolar bone loss due to LTA-induced periodontitis (Fig. 6C).

Discussion
In a healthy condition, the tooth root is embedded into a socket in the alveolar bone in periodontal tissue. Infection of mixed multiple Gram-negative bacteria resulted in alveolar bone resorption and tooth loss induced by severe inflammation in periodontal tissues. It is well-known that the major pathogens of periodontitis are dominantly Gram-negative bacteria, including Porphyromonas gingivalis, Aggregatibacter actinomycetemcomitans, Tannerella forsythia and Prevotella intermedia. LPS, as a ligand for TLR4, is an outer membrane component of Gram-negative bacteria and known to be a major definitive pathogenic factor of inflammatory bone resorption in periodontitis. On the other hand, Gram-positive bacteria have been known to contribute to the inflammation of the periodontal gums in the initial phase of periodontitis 18,19 ; however, there was little evidence to show that these pathogens contribute to the induction of inflammatory bone resorption in the late phase of periodontitis. Since LTA is a major cell-wall component of Gram-positive bacteria, we decided to examine the effects of LTA on osteoclast differentiation, inflammatory bone resorption, and alveolar bone resorption in a mouse model of periodontitis.
In this study, we demonstrated that LTA promoted PGE 2 production via the upregulation of the Cox2 and mPges1 mRNA expression in POBs, leading to PGE 2 -dependent osteoclast differentiation ( Figs. 1 and 2). These The transcription activity of NF-κB was measured with or without LTA (100 μg/mL). Plasmid pNFκB-TA-Luc (0.4 μg) and the pGL4.74[hLuc/TK] plasmid (40 ng) were transfected into POBs, and the luciferase activity was measured with the Dual-luciferase Reporter Assay system. The data are expressed as the mean ± SEM of 4 wells. A significant difference between the two groups was indicated; *P < 0.05 and **P < 0.01 versus control; by Welch's t-test. www.nature.com/scientificreports/ data suggest that LTA promotes COX-2-and mPGES-1-mediated PGE 2 production via NF-κB signaling, resulting in subsequent osteoclast differentiation. Indomethacin markedly suppressed LTA-induced osteoclast differentiation and calvarial bone resorption (Fig. 3). Furthermore, LTA maintained osteoclast cell survival, which was associated with the mRNA expression of Nfatc1 and Ctsk in the absence of sRANKL (Fig. 4). We clarified that LTA could induce alveolar bone loss in the ex vivo model of periodontal bone resorption for periodontitis (Fig. 5). Finally, we confirmed that LTA induced alveolar bone loss in the model of in vivo periodontitis. In the 3D-μCT analysis, the quantitative analysis frequently used in clinical diagnoses showed that LTA significantly increased bone resorption, which became apparent with an increase in both the inter-tooth space and the root trunk space of the 1st molar in the alveolar bone (Fig. 6).
Bacterial LTA is known to be a natural ligand for TLR2/6 13,15 . In the present study, both osteoblasts and osteoclasts expressed the TLR2 and TLR6 mRNAs, and LTA induced osteoclast differentiation via PGE 2 production and osteoclast cell survival. LTA-induced PGE 2 production by osteoblasts upregulates the mRNA expression of Rankl via EP4-cAMP response element binding protein (CREB) signaling in an autocrine manner, since the promoter region of Rankl gene possess CRE 20,21 . In addition, a previous study reported that PGE 2 downregulated the expression of osteoprotegerin (OPG), a decoy receptor for RANKL 22 . Thus, LTA-induced PGE 2 production resulted in osteoclast differentiation. In bone marrow macrophages, LTA has been reported to increase the release of inflammatory osteoclastogenic cytokines, such as TNF-α, IL-1 and IL-6 23  The data are expressed as the mean ± SEM of 3 wells. (C) Mouse calvariae from newborn mice were cultured with LTA (100 μg/mL) and indomethacin (10 μM) for 5 days. The calcium concentration was measured to elucidate bone-resorbing activity. The data are expressed as the mean ± SEM of 5 wells. (D) Mouse POBs were cultured with LTA (100 μg/mL) and indomethacin (10 μM) for 24 h, and the level of PGE 2 in the conditioned medium was measured. The data are expressed as the mean ± SEM of 3 wells. (E) Mouse POBs were cultured with LTA (100 μg/mL) for 24 h and total RNA was extracted and the mRNA expression of RANKL was analyzed by qPCR. The data are expressed as the means ± SEM of triplicate from a representative experiment. β-actin was used as a normalized gene. A significant difference between the two groups was indicated; **P < 0.01 and ***P < 0.001 versus control, # P < 0.05 and ### P < 0.001 versus LTA; by one-way ANOVA and Tukey's post hoc test.  29 . These cytokines also promote PGE 2 secretion in osteoblastic cells 10,30,31 . We demonstrated that LTA increased PGE 2 production and Rankl expression in osteoblasts (Fig. 2), but these effects were suppressed by indomethacin (Fig. 3). LTA also upregulated the mRNA expression of Nfatc1 and Ctsk in osteoclasts (Fig. 4). These data suggest that TLR2 signaling activated by LTA derived from Gram-positive bacteria contribute to PGE 2 -mediated inflammatory bone resorption in periodontitis. Gram-negative periodontal bacteria, such as P. gingivalis, mainly proliferated in the anaerobic environments of deep periodontal pockets, while Gram-positive bacteria broadly formed bacterial microbiota in the aerobic environment of the oral cavity 32 . In addition, multi-species bacterial plaques are formed and produce LTA 33 . Since the percentage of proliferating Gram-positive bacteria in the oral environment is higher than that of Gram-negative bacteria, the periodontal bone resorption in periodontitis induced by Gram-positive bacterial  Osteoclasts were treated with LTA (100 μg/mL) or sRANKL (100 ng/mL). The mRNA expression of Nfatc1 and Ctsk was analyzed by qPCR. The data are expressed as the means ± SEM of triplicate from a representative experiment. β-actin (Actb) was used as a normalized gene. A significant difference between the two groups was indicated; *P < 0.05, **P < 0.01 and ***P < 0.001; by one-way ANOVA and Tukey's post hoc test.  34 .
Recently, other aspects of TLR2 ligand-induced inflammation were reported. One study reported that the activity of LTA varied among bacterial species, LTA from Staphylococcus aureus and Bacillus subtitis clearly activated the NF-κB pathway via TLR2, while LTA from Lactobacillus plantarum showed less potency in activating the same pathway 35,36 . In the oral cavity, the presence of several Gram-positive bacteria has been reported, including Streptococcus mutans, S. aureus, Streptococcus pneumoniae and Enterococcus faecalis 37 . LTA from E. faecalis was found to induce inflammatory responses via the production of TNF-α and nitric oxide (NO) 38 . In contrast, LTA from S. pneumoniae has been shown to possess lower activity than S. aureus and to fail to form heterodimers of the isolated ectodomains of TLR2 with TLR1 or TLR6 12,39 . These reports suggest that some Gram-positive bacteria inhabiting the oral cavity may contribute to inflammation. Another biological aspect of the TLR2 ligand that has been reported is that the recognition of LTA requires co-receptors, such as CD36 and mannose binding lectin 15,40,41 . LTA from S. aureus also acted as the ligand for TLR2/6 heterodimer and promoted inflammatory cytokines, including TNF [41][42][43] . The present study showed that LTA from S. aureus also contributed to periodontal diseases, as S. aureus has been reported be part of the oral microbiota 37,44 . Further studies are needed to elucidate the roles of LTA from specified oral microbiota in periodontal bone resorption and to clarify the roles of LTA receptors, the signaling cascade, inflammatory cytokines, and PGE 2 in periodontitis.
In conclusion, our data indicate that LTA from Gram-positive bacteria contribute to periodontal bone resorption. Figure 7 shows a schematic illustration summarizing that LTA promoted alveolar bone loss via the upregulation of PGE 2 production and RANKL induced osteoclast differentiation and function. LTA may enhance inflammation in the early to late phases of periodontitis and periodontal bone resorption. We showed a new concept of the disease progression of periodontitis that was induced by LTA from Gram-positive bacteria.   Osteoclast differentiation in cultures of Raw264.7 cells. Raw264.7 cells (4 × 10 3 cells/well) were cultured in α MEM containing 10% FBS with or without sRANKL (100 ng/mL) for 4 days in a 96-well plate with a flat bottom. The medium was changed on day 3, and cells were continuously treated with sRANKL. After differentiation into osteoclasts in the presence of sRANKL on day 4, the medium was changed, and osteoclasts were cultured for another 1 day with or without sRANKL as well as with LTA (1, 10, 100 μg/mL) in the absence of sRANKL. Cells were stained for TRAP, and TRAP-positive multinuclear cells containing 3 or more nuclei per cell were counted as osteoclasts.
The qPCR was performed with SsoAdvanced SYBR Green Supermix (Bio-Rad Laboratories Inc., CA, USA). The relative normalized expression of genes was quantified by the ΔΔCq method and β-actin was used as a normalized gene.
Measurement of PGE 2 production. The concentrations of PGE 2 in the cultured medium were measured using an enzyme immunoassay (EIA) (GE Healthcare Japan Corp., Tokyo, Japan). The antibody had the following cross-reactivity determined by the bound to free ratio: PGE 2 , 100%; PGE 1 , 7.0%; 6-keto-PGF 1α , 5.4%; PGF 2α , 4.3%; and PGD 2 , 1.0%. In the ex vivo model of periodontal bone resorption, teeth extracted mouse mandibular alveolar bones were cultured for 1 day in BGJb medium containing 1 mg/mL BSA. Then, alveolar bone was transferred to new medium with LTA and cultured for 5 days. The calcium concentration in the conditioned medium was measured by the OCPC method to determine the bone-resorbing activity.
In the in vivo mouse model of periodontitis, 250 μg/mouse of LTA dissolved in PBS was injected into the left outer gingiva of the first molar in the lower jaw on days 0, 2 and 4 in 6-week-old female mice. PBS was injected into the lower gingiva at each time point as a control. Mice were anesthetized via 10 mL/kg intraperitoneal administration of three types of mixed anesthetic agents: medetomidine hydrochloride (Nippon Zenyaku Kogyo Co., Ltd., Fukushima, Japan), midazolam (Astellas Pharma Inc., Tokyo, Japan) and butorphanol (Meiji Seika Pharma Co., Ltd., Tokyo, Japan). The alveolar bone of the lower jaw was collected from mice and subjected to μCT on day 7. After μCT, the teeth were removed from the alveolar bone. The bone mineral density (BMD) of the alveolar bone with the teeth removed was measured by DXA.
An μCT analysis and the diagnosis of the mouse model of periodontitis. For a further analysis of alveolar bone using three-dimensional (3D) reconstruction images were obtained by micro-computed tomography (μCT) (inspeXio SMX-90CT, Shimadzu, Kyoto, Japan). The bone mass was measured on a 3D image of the inter-tooth areas of the alveolar bone between the 1st and 2nd molars and the root trunk of the 1st molar. The inter-tooth spaces between the 1st and 2nd molars and the root trunk space of the 1st molar were measured on a 3D image of the alveolar bone, respectively 17 . www.nature.com/scientificreports/ Statistical analyses. All data were expressed as the mean ± standard error of mean (SEM). Student's t-test was used to compare two independent groups with equal variance. Welch's t-test was used to compare two independent groups with unequal variance. A one-way ANOVA, followed by Tukey's test for post hoc analysis was used for comparisons among three or more groups. All statistical analyses were performed using the IBM SPSS Statistics software program (Ver. 25; Armonk, NY, USA).