Toll-Like Receptor 2 Stimulation of Osteoblasts Mediates Staphylococcus Aureus Induced Bone Resorption and Osteoclastogenesis through Enhanced RANKL

Severe Staphylococcus aureus (S. aureus) infections pose an immense threat to population health and constitute a great burden for the health care worldwide. Inter alia, S. aureus septic arthritis is a disease with high mortality and morbidity caused by destruction of the infected joints and systemic bone loss, osteoporosis. Toll-Like receptors (TLRs) are innate immune cell receptors recognizing a variety of microbial molecules and structures. S. aureus recognition via TLR2 initiates a signaling cascade resulting in production of various cytokines, but the mechanisms by which S. aureus causes rapid and excessive bone loss are still unclear. We, therefore, investigated how S. aureus regulates periosteal/endosteal osteoclast formation and bone resorption. S. aureus stimulation of neonatal mouse parietal bone induced ex vivo bone resorption and osteoclastic gene expression. This effect was associated with increased mRNA and protein expression of receptor activator of NF-kB ligand (RANKL) without significant change in osteoprotegerin (OPG) expression. Bone resorption induced by S. aureus was abolished by OPG. S. aureus increased the expression of osteoclastogenic cytokines and prostaglandins in the parietal bones but the stimulatory effect of S. aureus on bone resorption and Tnfsf11 mRNA expression was independent of these cytokines and prostaglandins. Stimulation of isolated periosteal osteoblasts with S. aureus also resulted in increased expression of Tnfsf11 mRNA, an effect lost in osteoblasts from Tlr2 knockout mice. S. aureus stimulated osteoclastogenesis in isolated periosteal cells without affecting RANKL-stimulated resorption. In contrast, S. aureus inhibited RANKL-induced osteoclast formation in bone marrow macrophages. These data show that S. aureus enhances bone resorption and periosteal osteoclast formation by increasing osteoblast RANKL production through TLR2. Our study indicates the importance of using different in vitro approaches for studies of how S. aureus regulates osteoclastogenesis to obtain better understanding of the complex mechanisms of S. aureus induced bone destruction in vivo.

Introduction primary osteoblasts and osteoblastic cell line MG63 [28,29]. In the mouse osteoblastic cell line MC3T3-E1, S. aureus similarly decreases osteogenic differentiation and induces apoptosis [30]. S. aureus also upregulates RANKL mRNA and protein in primary mouse and human osteoblasts and in the mouse osteoblastic cell line MC3T3-E1 [17,28,31,32]. Interestingly, S. aureus protein A binds to tumor necrosis factor recptor-1 on osteoblasts causing decreased expression of bone formation genes and increased expression of inflammatory cytokines such as interleukin-6 (IL-6) [32,33]. It has also been reported that S. aureus can upregulate the expression of death inducing receptors (DR4 and DR5) leading to osteoblast apoptosis and increased OPG release [34]. These in vitro observations suggest that increased osteoclast formation caused by S. aureus may be due to S. aureus primarily targeting osteoblasts, which respond with increased RANKL expression.
In addition to the studies showing that S. aureus can interact with osteoblasts, it has been shown that this bacterium can be recognized by osteoclast progenitors. Using either live S. aureus, or S. aureus cell wall peptidoglycan or lipoteichoic acid, it has been found that all these preparations inhibit RANKL induced osteoclast formation in mouse bone marrow macrophage cultures [21,35,36], while stimulating differentiation along the macrophage lineage [36]. When using RANKL primed bone marrow macrophages, however, S. aureus cell wall peptidoglycan stimulated osteoclast formation [21]. In crude bone marrow cell cultures, containing both stromal cells/osteoblasts and hematopoietic cells, addition of surface-associated material from S. aureus enhanced osteoclastogenesis [37].
Due to the severity of S. aureus septic arthritis and to the increased use of prosthetic joint replacement with a risk of S. aureus infections, it is important to understand the bone destructive mechanisms exerted by S. aureus in order to develop new treatment strategies. Studies on effects by S. aureus on osteoclast formation have been performed using osteoclast progenitors from bone marrow showing either inhibition or stimulation of osteoclast formation. However, mature osteoclasts are formed exclusively on periosteal and endosteal surfaces and we, therefore, have studied how S. aureus can regulate osteoclastogenesis in the periosteum/endosteum. For this purpose we have used either ex vivo organ cultures of mouse parietal bones, or cell cultures containing periosteal/endosteal osteoblasts and osteoclast progenitors. We found that bone resorption and osteoclast formation caused by stimulation of organ cultured parietal bones or periosteal/endosteal cell cultures with RANKL was not affected by S. aureus. In contrast, S. aureus abolished RANKL induced osteoclastogenesis in bone marrow macrophage cultures. These observations demonstrate that regulation of osteoclastogenesis is different using osteoclast progenitor cells from different tissues. Most importantly, S. aureus stimulated bone resorption and osteoclast formation in both organ cultured bone and periosteal/endosteal cell cultures, similar to in vivo observations in humans and rodents with S. aureus infections, and this response was dependent on TLR2-mediated increase of RANKL.

Material and Methods Bacteria
The two S. aureus isolates, one Toxic shock syndrome toxin 1 (TSST-1) and Staphylococcal Enterotoxin A (SEA) producing, and one non-toxin producing strain, used in this study were originally isolated from healthy Swedish infants as previously described [38]. After 24 h growth on horse blood agar plates, harvested bacteria were washed in phosphate buffered saline (PBS), inactivated by exposure to UV-light (280-315 nm), and suspended in sterile PBS before use. Complete UV inactivation was confirmed by control cultures. Bacterial preparations were stored at-70°C until use.

Mice
CsA mice from our inbred colony, CB57BL/6J and B6.129 Tlr2 tm1Kir /J mice were from Jackson Laboratories. The mice were maintained ( 10 in each cage) under standard conditions of temperature and light, and were fed with standard laboratory chow and water ad libitum. Adult mice were killed by cervical dislocation and newborn mice by decapitation. 298 adult and newborn mice were used for this study. The Ethical committee of Umeå University, Umeå, Sweden has approved the animal care and experiments.

Organ culture of mouse parietal bones
Parietal bones from 5-7 days-old mice were dissected and cut either into halves for most of the experiments, or into quarters for mineral release analyses. Subsequently, the bones were incubated for 24 h in serum free α-MEM containing BSA (0.1%) and indomethacin (1 μM) to prevent the initial effect of released prostaglandins due to the dissection trauma [39,40]. The bones were then washed extensively with sterile PBS and cultured in indomethacin free media with or without S. aureus or other test substances.

Bone resorption assays
Bone resorption was analyzed by assessing either release of mineral ( 45 Ca) or of matrix degradation fragment (CTX) from the bones to the culture media. 2-3 days-old mice were injected with 1.5 μCi 45 Ca 4-5 days prior to dissection, and the amounts of radioactivity in bone and culture medium were analyzed by liquid scintillation at the end of the culture period. For the time-course experiments, the mice were injected with 12.5 μCi 45 Ca, and the radioactivity was analyzed at different time points by withdrawal of small amounts of the culture media. Isotope release was expressed as the percent release of the initial amount of isotope (calculated as the sum of radioactivity in medium and bone after culture).
The release of collagen fragments (CTX) from the bone matrix into the media was analyzed by RatLaps™ ELISA kit.

Isolation and culture of parietal cells
Periosteal and endosteal cells were isolated from 2-3 days-old mouse parietal bones by sequential collagenase digestion [41]. Pooled cells from populations 1-10, containing both osteoblast and osteoclast progenitors capable of forming bone resorbing osteoclasts [41], were used for osteoclastogenesis experiments. These cells were initially cultured in 25 cm 2 flasks with α-MEM containing 10% FBS for 48 h to expand the number of cells. Cells were then washed and detached and subsequently seeded in 12-multiwells (10 4 cells/cm 2 ) and cultured with or without S. aureus or other test substances for 9 days. Cells were fixed and stained for TRAP.
Cells from populations 6-10 are enriched for osteoblastic cells and widely used for osteoblastogenesis experiments. These cells were expanded as described above, and seeded in 24-multiwells (10 4 cells/cm 2 ) and incubated with or without test substances for 48 h at which time point RNA was isolated and used for gene expression analyses.

Bone marrow macrophage isolation and cultures
Mouse bone marrow cells were isolated from tibia and femur as described [42]. The bone marrow macrophages were purified by incubating the cells on Corning dishes in the presence of M-CSF (30 ng/ml) for 48 h. The adherent bone marrow macrophages (BMM) were used as osteoclast progenitor cells. These cells do not contain T-or B-cells and all cells express the macrophage marker CD11b/Mac-1 [43]. After washing and detaching, cells were spot-seeded (5x10 3 cells in 10 μl) at the center of 96-multiwells and left to adhere for 10 min. Then, the wells were added 200 μl medium containing M-CSF (30 ng/ml; controls) or M-CSF (30 ng/ml) +RANKL (4 ng/ml) with or without S. aureus or other test substances and incubated for 96 h. In experiments with committed osteoclast progenitors, cells were primed with RANKL (4 ng/ml) in presence of M-CSF for 24 h. Cells were then washed and medium containing M-CSF with or without test substances was added. At the end of the cultures, cells were fixed and stained for TRAP.

TRAP staining
Cells were fixed, washed and stained for TRAP using the Naphtol AS-BI phosphate kit from Sigma Aldrich. TRAP + cells with at least three nuclei were counted as TRAP + multinucleated osteoclasts (TRAP + MuOCL).

Gene expression analyses
RNA was isolated from bone tissue or cells using RNAqueous-4 PCR 1 kit, according to manufacturer 0 s instructions. The RNA was quantified spectrophotometrically and single-stranded cDNA was synthesized from 0.1-0.5 μg of total RNA using High High-Capacity cDNA Reverse Transcription 1 Kit. To ensure absence of genomic DNA in the samples, negative controls with no MultiScribe™ reverse transcriptase were included. The following predesigned real-time PCR assays from Applied Biosystems were used for gene expression assays: . β-actin (4352341E) was used as a reference gene to normalize for variability in amplification due to possible differences in starting mRNA concentrations. ABI PRISM 7900 HT Sequence Detection System and Software were used for the amplifications.

RANKL and OPG protein analyses
Assessment of RANKL and OPG protein was made using ELISA kits after lysing the bones in 1 ml 0.2% Triton X-100. The sensitivities of the immunoassays are 5 pg/ml.

Prostaglandin E 2 analysis
The amount of released PGE 2 in the culture media was measured by a radioimmunoassay kit.

Neutralizing antibody experiments
Neutralizing antibodies against mouse interleukin-1β (IL-1β), IL-6, IL-11, leukemia inhibitory factor (LIF), oncostatin M (OSM) and tumor necrosis factor-α (TNF-α) were used to elucidate the role of these cytokines in S. aureus induced bone resorption. We first validated the efficacy of the antibodies by verifying that the antibodies neutralized the effect of IL-1β, IL-6+sIL-6R, IL-11, LIF, OSM or TNF-α (all used at maximally effective concentrations) on mRNA expression of Tnfsf11 in parietal bones. The parietal bones, after indomethacin pretreatment, were preincubated with the antibodies for 8 h prior to the stimulation with S. aureus. To eliminate the possibility that several cytokines were responsible for the effects, we added the antibodies all together to the cell and organ cultures in a mixture at final concentrations of 1μg/ml for anti-IL-6, anti-IL-11, anti-LIF and anti-TNF-α, 3 μg/ml for anti-OSM and 5 μg/ml for anti-IL-1β, and analyzed the response on the expression of Tnfsf11 mRNA or CTX release, respectively.

S. aureus stimulates bone resorption in parietal bones
Both the non-toxin producing S. aureus and toxin producing S. aureus (S. aureus Tox) enhanced the release of mineral ( 45 Ca) from cultured neonatal parietal bones in a time-dependent manner ( Fig 1A). The effect was statistically significant (P<0.05) already at 24 h. The stimulatory effect of S. aureus on 45 Ca release was concentration-dependent ( Fig 1B). S. aureus and S. aureus Tox also significantly enhanced the release of bone matrix degradation fragments (CTX) from these bones ( Fig 1C).
The S. aureus and S. aureus Tox induced CTX release from neonatal parietal bones was abolished by the bisphosphonate zoledronic acid (Fig 1D).
Mineral release and matrix degradation induced by S. aureus and S. aureus Tox were associated with time-dependent increased mRNA expression of Ctsk (encoding cathepsin K; Fig 1E) and Acp5 (encoding TRAP; Fig 1F). The increased mRNA expression of Ctsk and Acp5 was dependent on the concentration of S. aureus (Fig 1G and 1H).

S. aureus-induced osteoclast formation and bone resorption in parietal bones is mediated by enhanced RANKL
Osteoclastogenesis requires activation of c-Fms by its ligand M-CSF and activation of RANK by its ligand RANKL, with OPG being a decoy receptor for RANKL [22]. In addition, activation of the receptor OSCAR (Osteoclast-associated receptor) is important for osteoclast differentiation [44,45]. Downstream signaling includes activation of NFATC1 (Nuclear factor of activated T-cells c1) which is regarded as the master transcription factor of osteoclastogenesis [46]. We assessed the effect by S. aureus on these cytokines, receptors and transcription factor in the parietal bones.
S. aureus and S. aureus Tox time-dependently stimulated the mRNA expression of Oscar, Nfatc1 and Tnfsf11 (encoding RANKL) (Fig 2A-2C). The stimulatory effect on these transcripts was dependent on the concentration of S. aureus (Fig 2D-2F). The bacterium also In agreement with the mRNA analyses, S. aureus and S. aureus Tox significantly enhanced RANKL protein levels in the parietal bones (Fig 2G), without significantly affecting OPG protein ( Fig 2H).
The importance of increased RANKL/OPG ratio for the stimulatory effect on bone resorption in neonatal parietal bones was demonstrated by the observation that bone matrix degradation (CTX release) in parietal bones challenged by S. aureus and S. aureus Tox was abolished when recombinant OPG (300 ng/ml) was added (Fig 2I). OPG also inhibited S. aureus induced mRNA expression of Ctsk ( Fig 2J) and Nfatc1 (Fig 2K), indicating that OPG inhibited bone resorption through inhibition of osteoclast differentiation. The fact that OPG did not affect Tnfsf11 mRNA (Fig 2L) shows that OPG acted downstream RANKL formation to inhibit osteoclastogenesis.
The role of osteoclastogenic cytokines and prostaglandins in S. aureus induced RANKL and bone resorption S. aureus enhanced the mRNA expression of Il1b, Il6, Il11, Lif, Osm, Tnfsf2 (encoding TNF-α) and Ptgs2 (encoding cyclooxygenase-2) in the parietal bones in a dose-and time-dependent manner (Fig 3A-3H), as expected since TLR2 activation often results in enhanced expression of these proinflammatory molecules [47]. S. aureus as well as S. aureus Tox enhanced the release of PGE 2 from the parietal bones (Fig 3I). Since these cytokines and prostaglandins are osteoclastogenic and can promote bone resorption, we examined their possible role in the S. aureus induced bone resorption and enhanced Tnfsf11 mRNA expression by using specific antibodies neutralizing the cytokines and indomethacin to inhibit prostaglandin biosynthesis. We first confirmed the efficacy of the antibodies in the organ culture assay of parietal bones (S2A-S2C Fig), and then added a mixture of antibodies neutralizing IL-1β, IL-6, IL-11, LIF, OSM and TNF-α to S. aureus stimulated bones. Addition of antibodies did not affect bone resorption induced by the bacteria at optimal or suboptimal concentrations, as assessed by bone matrix degradation (CTX release; Fig 3J). Nor did the antibodies affect the increased Tnfsf11 mRNA expression induced by S. aureus ( Fig  3K). Despite the partial decrease of S. aureus induced Tnfsf11 mRNA expression (Fig 3L) in parietal bones by indomethacin, bone resorption was not affected (Fig 3M).
S. aureus stimulates Tnfsf11 in mouse parietal osteoblasts independent of cytokine induction but dependent on TLR2 Osteoblasts are resident cells that communicate and activate osteoclastogenesis in bone tissue by producing RANKL in response to a variety of bone resorbing hormones and cytokines [48,49]. We, therefore, investigated if S. aureus could induce production of RANKL in mouse parietal osteoblasts. Stimulation of these cells by S. aureus caused a time-and concentrationdependent increase of Tnfsf11 mRNA expression but had no effect on Tnfrsf11b mRNA expression (Fig 4A and 4B).
S. aureus and S. aureus Tox increased the mRNA expression of Il1b, Il6, Il11, Lif, Osm and Tnfsf2 in the parietal osteoblasts (Fig 4C). Neutralization of these cytokines by a mixture of antibodies neutralizing IL-1β, IL-6, IL-11, LIF, OSM and TNF-α, did not affect the S. aureus induced Tnfsf11 mRNA expression (Fig 4D). S. aureus and S. aureus Tox also enhanced the expression of Ptgs2 mRNA (Fig 4E). Inhibition of prostaglandin biosynthesis by indomethacin abolished S. aureus and S. aureus Tox induced mRNA expression of Tnfsf11 (Fig 4F).
Others and we [50][51][52] have previously shown that osteoblasts express TLR2 and we, therefore, assessed if S. aureus induced Tnfsf11 expression was due to stimulation of TLR2. Using osteoblasts isolated from Tlr2 deficient mice, we found that Tnfsf11 mRNA induced by S. aureus was entirely dependent on Tlr2 expression (Fig 4G). Similarly, S. aureus did not upregulate Ptgs2 mRNA in osteoblasts in cells isolated from Tlr2 deficient mice (Fig 4H).

S. aureus differentially regulates osteoclast formation in bone marrow and periosteal cell cultures
Activation of TLR2 in RANKL stimulated BMM results in inhibition of osteoclast differentiation and formation [53,54]. We, therefore, asked ourselves why not S. aureus inhibited osteoclast formation in the ex vivo bone organ cultures. For this purpose, we compared the effects by S. aureus in three different osteoclastogenic systems, all stimulated by RANKL with or without S. aureus. S. aureus, similar to other TLR2 agonists (LPS P. gingivalis and Pam2), abolished osteoclastogenesis in RANKL stimulated BMM (Fig 5A). In contrast, S. aureus did not inhibit osteoclastogenesis induced by RANKL in periosteal/endosteal cell cultures (Fig 5B), or RANKL stimulated bone matrix degradation in parietal bone organ cultures (Fig 5C). These observations indicate that osteoclast progenitors in the periosteum/endosteum are different from those in bone marrow.
Interestingly, S. aureus stimulated osteoclast formation in the periosteal cells in the absence of exogenous RANKL (Fig 5D and 5E). The effect was associated with enhanced mRNA expression of Tnfsf11, Ctsk and Acp5 as markers of osteoclast differentiation (Fig 5F-5H).
Stimulation of TLR2 by its synthetic ligand (Pam2) in committed osteoclast progenitors (RANKL-primed for 24h) has been reported to promote osteoclastogenesis [17,54]. This finding suggests the possibility that one reason for the different response to S. aureus in osteoclast progenitors from periosteum and bone marrow co-treated with RANKL might be that the differentiation stage of osteoclast progenitors determines the response to S. aureus. We, therefore, assessed the effect by S. aureus in RANKL-primed BMM. In agreement with the previous observations [17,54], we found that Pam2 and Pam3 enhanced osteoclast formation in RANKLprimed BMM to the same degree as treatment with RANKL (Fig 6A and 6B). Unlike the synthetic ligands Pam2 and Pam3, but similar to other bacterial TLR2 ligands such as HKLM and LPS P. gingivalis, stimulation of RANKL-primed BMM with S. aureus resulted in formation of mainly mononuclear TRAP + cells with only some few osteoclast-like cells (Fig 5A and 5B). This observation indicated that S. aureus can stimulate differentiation of RANKL-primed osteoclast progenitors but not to the same degree as Pam2 and Pam3 and not to the level where the progenitors fuse to typical multinucleated mature osteoclasts. Further evidence for the view that S. aureus can induce osteoclast progenitor cell differentiation were the observations that the mRNA expression of Ctsk, Acp5 and Calcr was significantly induced by S. aureus (Fig 6C-6E), which was also true for the mRNA expression of the two osteoclastogenic transcription factors c-Fos and Nfatc1 (Fig 6F and 6G). However, the degree of upregulation of these genes induced by S. aureus was clearly less than that caused by Pam2, Pam3 and RANKL. Similarly, LPS P. gingivalis induced all these osteoclast genes but the degree of stimulation was less than that induced by the compounds stimulating mature osteoclast formation (Fig 6C-6G).

S. aureus inhibits the expression of osteoblast anabolic genes
Since S. aureus induced bone loss may not entirely depend on increased bone resorption we also assessed if the bacteria affected bone formation in the parietal bones. The mRNA expression of the bone matrix proteins osteocalcin (encoded by Bglap) and procollagen type I (encoded by procol1a1), as well as of the enzyme alkaline phosphatase (encoded by Akp1) was substantially decreased by S. aureus in organ cultured parietal bones (S3A-S3C Fig). This might be due to the decreased mRNA expression of the transcription factor Runx2 also observed (S3D Fig). Since Nfatc1 is expressed not only in osteoclasts but also in osteoblasts [55][56][57][58], and since S. aureus increased Nfatc1 mRNA in the parietal bones (Fig 2K), we assessed if increased Nfatc1 was involved in the decreased expression of osteoblast anabolic genes. We found, however, that inhibition of these genes induced by S. aureus was independent on Nfatc1 since the inhibition of Nfatc1 mRNA expression caused by OPG (Fig 2K) did not affect S. aureus induced down regulation of Bglap, Procol1a1, Akp1 or Runx2 mRNA expression (S3E-S3H Fig).

Discussion
It is well recognized that S. aureus infections can cause local and systemic bone destruction [15][16][17][18] but the mechanisms by which S. aureus induces bone resorption are still not fully understood. Although several reports have shown that S. aureus can target osteoblasts in vitro causing apoptosis, decreased bone formation and decreased expression of osteoblastic genes, as well as enhanced RANKL expression [17,28,31,32], the data regarding effects on osteoclasts are more diverse. S. aureus has been shown both to inhibit [21,35,36] and stimulate [21] osteoclastogenesis in mouse bone marrow macrophage cultures depending on if the bacterium is exposed to the cells simultaneously with RANKL or after RANKL pretreatment, respectively. Since mature osteoclasts are formed only at bone surfaces we have studied the effect of S. aureus on osteoclast formation and bone resorption using osteoclast progenitors present at periosteal/endosteal surfaces.
To mimic the microenvironment of bone tissue where osteoclast formation and bone resorption take place in vivo we used ex vivo organ cultures of mouse parietal bones, exhibiting a periosteum and a thin endosteum. We show that S. aureus enhances bone resorption in the parietal bones through a process inhibited by bisphosphonate, demonstrating the importance of osteoclasts. The finding that S. aureus increased osteoclastic genes such as those encoding TRAP and cathepsin K, and the osteoclastogenic transcription factor NFATc1, showed that S. aureus induced bone resorption is due to enhanced differentiation and activation of osteoclasts.
Since the RANKL/OPG ratio is crucial for osteoclastogenesis and bone homeostasis [59,60], we next investigated the effect of S. aureus on RANKL/OPG ratio. S. aureus enhanced this ratio by increasing the expression at both mRNA and protein levels of RANKL, without affecting those of OPG in the parietal bones. The inhibition of bone resorption and osteoclastic gene expression, caused by exogenous OPG added to S. aureus stimulated bone organ cultures, further supports the essential role of RANKL in bone resorption due to S. aureus infection. Using osteoblasts from wild type and Tlr2 deficient mice, we show that osteoblastic TLR2 is the receptor utilized by S. aureus in the bones to enhance RANKL. These data show that S. aureus stimulates periosteal/endosteal osteoclast formation and bone resorption in organ-cultured bones by enhancing RANKL/OPG in osteoblasts. Our observations further indicate that S. aureus does not inhibit osteoclast progenitors in these bones, in contrast to observations in bone marrow cell cultures. We cannot exclude, however, that increased osteoclast differentiation by S. aureus targeting osteoclast progenitors stimulated by endogenous RANKL produced in the periosteum/endosteum also may contribute to the enhanced bone resorption.
It has been reported that activation of TLR2 inhibits RANKL-induced osteoclast formation in BMM cultures [21,35,36]. We, therefore, wondered how S. aureus could increase osteoclastogenesis and bone resorption in the ex vivo parietal bone organ cultures. To investigate if osteoclast progenitors in parietal bones were different from those in bone marrow we next used cell cultures of periosteal/endosteal cells from parietal bones and mouse bone marrow cultures and compared the effect S. aureus on non-stimulated and RANKL stimulated cells, respectively. When S. aureus was added together with RANKL to mouse bone marrow macrophage cultures we could confirm observations made by others [21,35,36] showing that the bacterium can abolish osteoclast differentiation. In contrast, when S. aureus was added together with RANKL to periosteal/endosteal cell cultures, no inhibition of osteoclastogenesis was observed. Similar to this finding, S. aureus did not affect bone resorption in the parietal bones stimulated by exogenous RANKL. These findings show that osteoclast progenitors in bone marrow and at bone surfaces are fundamentally different in their response to S. aureus and explain why S. aureus can stimulate osteoclast formation in intact bones despite its inhibitory effect on RANKL-stimulated bone marrow macrophages. Previously, we have similarly shown that also vitamin A and LPS P. gingivalis stimulate bone resorption in parietal bones and increase formation of bone resorbing osteoclasts in periosteal/endosteal cell cultures, while also inhibiting RANKLstimulated osteoclast formation in bone marrow macrophage cultures [52,61,62]. All together, these observations indicate that studies on osteoclast formation should not only be based upon osteoclastogenesis in bone marrow macrophages but also include experiments using osteoclast progenitors present at the surfaces of bone.
When S. aureus was added to periosteal/endosteal cell cultures not stimulated with RANKL, we observed increased formation of osteoclasts, similar to the observations in the calvarial bones. This response was associated with increased mRNA expression of the osteoclastic genes Acp5 and Ctsk as well as with Tnsf11 mRNA, indicating that increased number of mature osteoclasts was due to increased differentiation of osteoclast progenitor cells due to increased RANKL in osteoblasts which are abundant in these cultures.
The reason for the different responsiveness of osteoclast progenitors in bone marrow and periosteum/endosteum is not known but could be due to differences in differentiation stage and/or to the microenvironment. If bone marrow macrophages are primed with RANKL before subsequent stimulation by LPS E. coli or S. aureus cell wall peptidoglycan, in the absence of RANKL, formation of mature osteoclasts is induced [21]. Similarly, the synthetic TLR2 agonists Pam2 and Pam3 [17,21], and the periodontal pathogen P. gingivalis acting through TLR2 [54], stimulate osteoclast formation in RANKL-primed bone marrow macrophages. These findings indicate that the differences between bone marrow macrophages and periosteal/endosteal osteoclast progenitor responses to stimulatory ligands may depend on the differentiation level of osteoclast progenitors. When we added S. aureus to RANKL-primed bone marrow macrophages, the cells started to differentiate and became TRAP + but the mononuclear cells did not fuse to mature osteoclasts. In contrast, addition of the two synthetic TLR2 agonists Pam2 and Pam3 stimulated formation of mature osteoclasts to the same degree as RANKL. Similar to S. aureus, LPS P. gingivalis and HKLM, two other TLR2 agonists induced differentiation of TRAP + mononuclear cells but not formation of mature osteoclasts. Pam2 and Pam3 robustly upregulated osteoclastic and osteoclastogenic genes such as Ctsk, Acp5, Calcr, c-Fos and Nfatc1, a response also observed after treatment with S. aureus and LPS P. gingivalis but to a much lesser degree. We do not know if the reason why S. aureus was unable to induce mature osteoclast formation was due to the quantitative differences in gene expression, or if the TLR2 in the RANKL-primed osteoclast progenitors are not fully activated by the bacterial agonist, in contrast to the synthetic ligands. Interestingly, we have found that LPS P. gingivalis, HKLM, Pam2 and Pam3 stimulate bone resorption in ex vivo parietal bones and Tnfsf11 mRNA expression in osteoblasts through TLR2 to the same degree [52], indicating that TLR2 in osteoblasts and RANKL-primed bone marrow macrophages are not entirely similar. The fact that multinucleated osteoclast formation was observed in RANKL-primed bone marrow macrophages treated with P. gingivalis [54], in contrast to our findings showing differentiation of mononuclear osteoclasts, may be due to that whole bacteria was used instead of the LPS preparation used by us.
In agreement with the well-known consequence of TLR activation, S. aureus stimulated the expression of several cytokines such as IL-1β, IL-11, IL-6, LIF, OSM and TNF-α. Since these cytokines are potent stimulators of RANKL expression, osteoclast formation and bone resorption [22,[63][64][65], we assessed if RANKL and osteoclastogenesis induced by S. aureus was secondary to induction of these cytokines. Using neutralizing antibodies, we show, however, that the effect of S. aureus on bone resorption and RANKL formation in parietal bones and isolated osteoblasts is not mediated by these cytokines. S. aureus also enhanced the mRNA expression in parietal bones and isolated osteoblasts of Ptgs2, a key enzyme in prostaglandin biosynthesis. Although inhibition of prostaglandin biosynthesis in the parietal bones and osteoblasts decreased S. aureus induced Tnfsf11 mRNA expression, no effect on bone resorption was observed; most likely due to the robust stimulation of Tnfsf11 still observed in S. aureus stimulated bones co-treated with the prostaglandin inhibitor. The reason why inhibition of prostaglandin biosynthesis abolished S. aureus induced Tnsf11 mRNA expression in calvarial osteoblasts, while only partially decreased this response in calvarial bones, might be due to that S. aureus can induce RANKL in cells present in the calvarial bones, but not in the osteoblasts cultures, and that the calvarial cells are insensitive to prostaglandins. One such possibility is osteocytes which have been shown to be more important producers of RANKL than osteoblasts [26,27].
S. aureus may not cause decreased bone mass only by increasing bone resorption but also by decreasing bone formation. Several studies using human and mouse osteoblasts have shown that S. aureus can inhibit expression of genes associated with osteoblast differentiation and bone formation [28][29][30][31][32]. We observed a similar effect using bone organ cultures in which S. aureus decreased the mRNA expression genes encoding osteocalcin, procollagen type I, alkaline phosphate and Runx2. Nfatc1 is most well known as a master regulator of osteoclast differentiation [46], but is also expressed in osteoblasts. The role of Nfatc1 in bone formation is controversial with both stimulatory [55,56] and inhibitory [57,58] effects observed. We found, however, that stimulation of Nfatc1 by S. aureus in the calvarial bones was not involved in the inhibition of the osteoblast anabolic genes.
Since the array of symptoms displayed by patients with S. aureus infection is correlated to the arsenal of virulence factors exhibited by S. aureus, we used two different strains of S. aureus and evaluated the role of toxins in S. aureus induced bone resorption. Our findings demonstrate that the ability of toxin production has no significant effect on bone resorption stimulated by S. aureus in isolated in vitro assays. Most likely, the toxin production characteristics of certain S. aureus strains have favorable effects on invasiveness, escape and damage of the immune system and exacerbating the infection and inflammation in vivo.
In summary, S. aureus targets osteoblasts (or maybe osteocytes) through TLR2 causing increased RANKL and periosteal/endosteal osteoclast formation and bone resorption with no signs of S. aureus targeting the subpopulation of osteoclast progenitors present at the surfaces of bone. The finding that activation of TLR2 in a subpopulation of osteoclast progenitors present in bone marrow which have been primed by RANKL results in osteoclast differentiation indicate the possibility that S. aureus might increase bone resorption also through activation of osteoclast progenitors at a certain differentiation level. The relative importance of osteoblasts/ osteocytes and osteoclast progenitors for the bone resorptive effect by S. aureus has to be assessed in mice (and/or bone organ cultures) with cell specific deletion of TLR2.  H). In A, effects were statistically significant at 12 and 48 h (P<0.01) and at 24 h (P<0.001). In B, effects were statistically significant by 10 6 and 3x10 7 (P<0.01) and by 3x10 6 and 10 7 (P<0.001) CFU/ml. In C, effects at 1-48 h were statistically significant (P<0.001). In D, effects were statistically significant by 3x10 5 -10 7 (P<0.001) and by 3x10 7