LPS is a potent stimulator of bone resorption in inflammatory diseases. The mechanism by which LPS induces osteoclastogenesis was studied in cocultures of mouse osteoblasts and bone marrow cells. LPS stimulated osteoclast formation and PGE2 production in cocultures of mouse osteoblasts and bone marrow cells, and the stimulation was completely inhibited by NS398, a cyclooxygenase-2 inhibitor. Osteoblasts, but not bone marrow cells, produced PGE2 in response to LPS. LPS-induced osteoclast formation was also inhibited by osteoprotegerin (OPG), a decoy receptor of receptor activator of NF-κB ligand (RANKL), but not by anti-mouse TNFR1 Ab or IL-1 receptor antagonist. LPS induced both stimulation of RANKL mRNA expression and inhibition of OPG mRNA expression in osteoblasts. NS398 blocked LPS-induced down-regulation of OPG mRNA expression, but not LPS-induced up-regulation of RANKL mRNA expression, suggesting that down-regulation of OPG expression by PGE2 is involved in LPS-induced osteoclast formation in the cocultures. NS398 failed to inhibit LPS-induced osteoclastogenesis in cocultures containing OPG knockout mouse-derived osteoblasts. IL-1 also stimulated PGE2 production in osteoblasts and osteoclast formation in the cocultures, and the stimulation was inhibited by NS398. As seen with LPS, NS398 failed to inhibit IL-1-induced osteoclast formation in cocultures with OPG-deficient osteoblasts. These results suggest that IL-1 as well as LPS stimulates osteoclastogenesis through two parallel events: direct enhancement of RANKL expression and suppression of OPG expression, which is mediated by PGE2 production.
Osteoclasts are bone-resorbing multinucleated cells that originate from hemopoietic progenitors of the monocyte/macrophage lineage (1, 2, 3, 4). Osteoblasts or bone marrow stromal cells are involved in osteoclastogenesis through a mechanism involving cell-to-cell contact with osteoclast progenitors (4, 5). Studies of M-CSF-deficient op/op mice have shown that M-CSF produced by osteoblasts is an essential factor for osteoclastogenesis (6, 7). Receptor activator of NF-κB ligand (RANKL)3 (3) was also identified as another factor essential for osteoclastogenesis (8, 9, 10, 11). RANKL is a member of the TNF-ligand family that is expressed by osteoblasts/stromal cells as a membrane-associated factor. Osteoclast precursors express RANK, a receptor of RANKL; recognize RANKL through cell-cell interaction; and differentiate into osteoclasts in the presence of M-CSF (12, 13). Osteoblasts/Stromal cells also produce a soluble decoy receptor for RANKL, osteoprotegerin (OPG), which inhibits osteoclast formation in vivo and in vitro by interrupting the interaction between RANKL and RANK (14, 15).
In mouse cell cocultures, osteoclasts are formed in response to bone-resorbing factors such as 1,25-dihydroxyvitamin D3 (1,25(OH)2D3), parathyroid hormone (PTH), PGE2, and IL-11 (4). Almost all of the bone-resorbing factors stimulate expression of RANKL in osteoblasts/stromal cells (4). Three independent signals have been proposed to induce RANKL expression in osteoblasts/stromal cells: vitamin D receptor-mediated signals induced by 1,25(OH)2D3, cAMP/protein kinase A (PKA)-mediated signals induced by PTH or PGE2, and gp130-mediated signals induced by IL-11 (4). Among these signals, vitamin D receptor- and cAMP/PKA-mediated signals suppress OPG expression in osteoblasts/stromal cells. Recently, we reported that compounds that elevate intracellular calcium, such as ionomycin, A23187, cyclopiazonic acid, and thapsigargin, stimulated osteoclast formation in mouse cocultures (16). Treatment of primary osteoblasts with those compounds stimulated the expression of RANKL. Thus, the signal mediated by calcium and protein kinase C (PKC) is proposed to be another (fourth) signal that induces RANKL expression in osteoblasts/stromal cells.
Severe bone loss due to excessive bone resorption is observed in inflammatory diseases such as periodontitis and osteomyelitis and some types of arthritides (17). LPS, a major constituent of Gram-negative bacteria, is proposed to be a potent stimulator of bone loss in these inflammatory diseases (17, 18, 19, 20, 21, 22). Recently, Toll-like receptor 4 (TLR4) was identified as the signal-transducing receptor for LPS (23, 24). The cytoplasmic signaling cascade of TLR4 is similar to that of IL-1Rs. Both TLR4 and IL-1Rs use common signaling molecules such as myeloid differentiation factor 88 and TNFR-associated factor 6 (25, 26, 27, 28, 29). Macrophages, lymphocytes, and osteoblasts/stromal cells express TLR4, and produce PGE2 and proinflammatory cytokines such as TNF-α and IL-1 in response to LPS (30, 31). These inflammatory factors also stimulate osteoclastogenesis directly or indirectly (17, 18, 19, 20, 21, 22, 32).
LPS stimulates PGE2 production in the target cells through the induction of mitogen-inducible cyclooxygenase 2 (COX2) expression (33). NS398, a nonsteroidal anti-inflammatory agent, specifically inhibits COX2 without affecting COX1 activity (34). Therefore, NS398 has been used as a tool to explore the role of PGE2 in pathological processes involving COX2 activity. PGE2 exerts its biological actions through binding to four specific membrane receptors (EP1, EP2, EP3, and EP4) (35). Sakuma et al. (20, 21) reported that induction of osteoclast formation by LPS, TNF-α, and IL-1 was barely observed in cell cultures prepared from EP4 knockout (EP4−/−) mice, and that urinary excretion of deoxypyridinoline, a sensitive marker for bone resorption, was not increased in EP4−/− mice injected with LPS. These results suggest that PGE2 is a key factor in the enhancement of osteoclastogenesis by LPS in vivo and in vitro. However, it is still not known how PGE2 is involved in the induction of osteoclastogenesis by LPS.
In the present study, we examined the mechanism of the induction of osteoclast formation by LPS in cocultures of mouse osteoblasts and bone marrow cells. We showed that LPS promoted osteoclastogenesis through two parallel events: one was direct enhancement of RANKL expression, and the other was suppression of OPG production mediated by PGE2 in osteoblasts. In addition, IL-1 stimulated osteoclast formation in the cocultures in a manner similar to LPS.
Materials and Methods
Reagents and mice
LPS (Escherichia coli O26:B6) and PGE2 were purchased from Sigma-Aldrich (St. Louis, MO). NS398 was from Calbiochem (San Diego, CA). Human rOPG and mouse rIL-1β were obtained from PeproTech (London, U.K.). Mouse rTNF-α and mouse rIL-1 receptor antagonist (rIL-1ra) were obtained from R&D Systems (Minneapolis, MN). Anti-mouse TNFR1 Ab was obtained from Genzyme Diagnostics (Cambridge, MA). Six- to 9-wk-old male and newborn ddY mice were obtained from Sankyo Laboratory Animal Center (Tokyo, Japan). C57BL/6 (B6) mice and OPG-deficient (OPG−/−) mice (C57BL/6 (B6)) were obtained from Clear Japan Clea (Tokyo, Japan). This study was reviewed and approved by the Showa University Animal Care and Use Committee.
Cell preparation and osteoclast formation assay
Primary osteoblasts were obtained from calvariae of newborn ddY mice, C57BL/6 (B6) mice, and OPG−/− mice by the conventional method using collagenase (36). Bone marrow cells were collected from femora and tibiae of 6- to 9-wk-old male mice. Primary osteoblasts (1 × 104 cells) and bone marrow cells (2 × 105 cells) were cocultured for 5 days in α-MEM containing 10% FCS (CSL, Victoria, Australia) in 96-well tissue culture plates (Corning, Corning, NY) (0.2 ml/well). Cocultures were incubated in the presence of LPS (0.001–10 μg/ml), PGE2 (1 μM), IL-1β (10 ng/ml), or TNF-α (10 ng/ml) for the final 3 days. Some cocultures were pretreated with NS398 (1 μM), OPG (100 ng/ml), IL-1ra (10 μg/ml), and TNFR1 Ab (10 μg/ml) for 1 h before adding LPS, PGE2, IL-1β, or TNF-α. Then cocultures were fixed and stained for tartrate-resistant acid phosphatase (TRAP; a marker enzyme of osteoclasts). TRAP-positive cells containing >3 nuclei were counted as osteoclasts. The results obtained from a typical experiment of three independent experiments are expressed as the mean ± SD of four cultures.
Measurement of PGE2 production
Primary osteoblasts (3 × 104 cells) and bone marrow cells (6 × 105 cells) were cultured separately or in combination with or without LPS (1 μg/ml) or IL-1β (10 ng/ml) in α-MEM containing 10% FCS in 48-well culture plates (Corning). After the cultures were incubated for 6 h, the concentration of PGE2 in the culture medium was determined using an enzyme immunoassay (EIA; Cayman Chemicals, Ann Arbor, MI). The Ab showed the following cross-reactivity determined by comparing the bond/free ratios with several eicosanoids: PGE2, 100%; PGE2 ethanolamide, 100%; PGE3, 43%; PGE1, 18.7%; 6-keto PGF1α, 1%; and 8-iso PGF2α, 0.25%.
Northern blot analysis
Primary osteoblasts (1 × 106 cells) were seeded in cell culture dishes (60 mm in diameter; Corning) and cultured in α-MEM containing 10% FCS for 3 days. After incubation in α-MEM containing 0.1% FCS for 3 h, the cells were incubated with LPS (1 μg/ml) or IL-1β (10 ng/ml). In some experiments, osteoblasts were cocultured with bone marrow cells (2 × 107 cells) in the presence of LPS (1 μg/ml) for 3 or 48 h. Some cultures were also treated with NS398 (1 μM) for 1 h before the addition of LPS. Total RNA was isolated from cultures using TRIzol (Life Technologies, Grand Island, NY). Northern blot analysis was performed using denaturing formaldehyde/agarose gels, as described (16). Double-stranded cDNA fragments encoding mouse RANKL OPG and COX2 were kindly provided by H. Yasuda (Snow Brand Milk Products, Tochigi, Japan). cDNA probes (RANKL, OPG, COX2, and β-tubulin) labeled with 32P were synthesized using a cDNA labeling kit (Takara, Tokyo, Japan). The RANKL, OPG, COX2, and β-tubulin probes were hybridized with membranes to which total RNA isolated from osteoblasts had been transferred. The membranes were exposed to Kodak BioMax MS film (Rochester, NY) for 3–48 h. Signals of RANKL, OPG, COX2, and β-tubulin mRNA were quantified using a radioactive image analyzer (BAS2000; Fuji Photo Film, Tokyo, Japan). Signals of RANKL, OPG, and COX2 mRNAs were normalized with the respective β-tubulin mRNA expression levels to calculate the relative intensity.
PGE2 is required for osteoclast formation induced by LPS
LPS stimulated TRAP-positive osteoclast formation in cocultures of primary osteoblasts and bone marrow cells in a dose-dependent manner (Fig. 1,A). The maximal number of osteoclasts was observed at 1 μg/ml of LPS. We then examined whether PGE2 is involved in the induction of osteoclast formation by LPS. LPS (1 μg/ml) as well as PGE2 (1 μM) induced TRAP-positive osteoclast formation in the cocultures (Fig. 1, B and C). NS398 (1 μM), a specific inhibitor of COX2, suppressed the induction of osteoclast formation by LPS, but not by PGE2 in the cocultures (Fig. 1, B and C). Both LPS- and PGE2-induced osteoclast formation in the cocultures was strongly inhibited by simultaneous addition of OPG (100 ng/ml) (Fig. 1, B and C). These results suggest that both PGE2 production and RANKL-RANK interaction are required for LPS-induced osteoclast formation in the cocultures. We next examined whether LPS induces osteoclastogenesis through IL-1 and TNF-α. Recombinant IL-1β (10 ng/ml) and TNF-α (10 ng/ml) induced osteoclastogenesis in the cocultures. IL-1ra (10 μg/ml) and anti-mouse TNFR1 Ab (10 μg/ml) strongly inhibited the osteoclast formation induced by IL-1β and TNF-α, respectively. However, neither IL-1ra nor TNFR1 Ab affected osteoclastogenesis induced by LPS (Fig. 1 D). These results suggest that PGE2 is a critical factor in LPS-induced osteoclastogenesis.
Osteoblasts mainly produce PGE2 in response to LPS
We then measured the PGE2 concentration in the conditioned medium of cocultures incubated with or without LPS (1 μg/ml) for 6 h (Fig. 2,A). LPS significantly increased the PGE2 concentration in the culture medium (Fig. 2,A). The addition of NS398 (1 μM) to the cocultures completely blocked the induction of PGE2 production by LPS in the cocultures (Fig. 2,A). To determine the type of cells that respond to LPS in the cocultures, osteoblasts and bone marrow cells were cultured separately in the presence or absence of LPS for 6 h (Fig. 2,B). LPS stimulated PGE2 production in the cultures of osteoblasts, but not bone marrow cells. LPS-induced PGE2 production was strongly inhibited by the addition of NS398 (Fig. 2,B). Northern blot analysis showed that treatment of osteoblasts with LPS for 3 h stimulated the expression of COX2 mRNA (Fig. 2 C). These results suggest that osteoblasts in the cocultures produce PGE2 in response to LPS via up-regulation of COX2 mRNA expression.
LPS regulates RANKL and OPG gene expression in osteoblasts
We next analyzed the effects of LPS on RANKL and OPG mRNA expression levels in primary osteoblasts by Northern blot analysis (Fig. 3). Treatment of the osteoblasts with LPS increased RANKL mRNA expression with two peaks at 3 and 48 h. The expression of RANKL mRNA after treatment with LPS was increased within 1 h, and was still higher than that of the control cultures even after 72 h (Fig. 3,A). The expression of OPG mRNA in osteoblasts was also enhanced by the treatment with LPS for 3 h (Fig. 3,A). However, the expression of OPG mRNA in osteoblasts treated with LPS for 48 or 72 h was decreased to a level lower than that of the control culture (Fig. 3,A). NS398 (1 μM) had no effect on the level of RANKL mRNA induced by LPS at 3 h (Fig. 3,B). The LPS-induced up-regulation of RANKL mRNA expression at 48 h was slightly inhibited by the COX2 inhibitor, but the level of the mRNA was much higher than that in the control cultures. In contrast, the LPS-induced down-regulation of OPG mRNA expression in osteoblasts at 48 h after treatment with LPS was completely blocked by the addition of NS398, although the OPG mRNA expression at 3 h was not affected by the COX2 inhibitor (Fig. 3 B). The expression levels of RANKL and OPG mRNAs in bone marrow cells were lower than those in primary osteoblasts, and were unchanged even after treatment with NS398 for 3 or 48 h (data not shown). These results suggest that PGE2 produced by osteoblasts plays an important role in the down-regulation of OPG expression, but not the up-regulation of RANKL expression in osteoblasts treated with LPS.
Suppression of OPG expression is involved in induction of osteoclast formation by LPS
We next examined how PGE2 production is involved in LPS-induced osteoclast formation using osteoblasts from OPG-deficient (OPG−/−) mice. Primary osteoblasts prepared from OPG−/− mice were cocultured with bone marrow cells from wild-type mice in the presence or absence of NS398, OPG, and/or LPS (Fig. 4). In agreement with previously reported findings (37), TRAP-positive osteoclasts were formed in cocultures containing OPG−/− osteoblasts even in the absence of any stimulus (Fig. 4). The number of osteoclasts was further increased in the LPS-treated cocultures containing OPG−/− osteoblasts. NS398 strongly suppressed the spontaneous osteoclast formation in the control cocultures containing OPG−/− osteoblasts. This suggests that endogenous production of PGE2 plays an important role in the osteoclast formation in cocultures containing OPG−/− osteoblasts. NS398 slightly, but not completely, inhibited LPS-induced osteoclast formation in cocultures containing OPG−/− osteoblasts (Fig. 4). OPG completely suppressed osteoclast formation in the cocultures treated or not treated with LPS (Fig. 4). These results suggest that the down-regulation of OPG expression by PGE2 is crucially involved in the osteoclast formation induced by LPS in the cocultures.
IL-1 stimulates osteoclast formation in a manner similar to LPS
Because the signaling pathway of IL-1Rs is quite similar to that of TLR4 (26, 27, 28, 29, 30), we finally examined whether IL-1 stimulates osteoclastogenesis in the cocultures in a manner similar to LPS. IL-1β (10 ng/ml) induced osteoclast formation in the wild-type cocultures, and the induction was inhibited by NS398 (1 μM) and OPG (100 ng/ml) (Fig. 5,A). IL-1β (10 ng/ml) also stimulated PGE2 production in osteoblasts, but not in bone marrow cells after treatment for 6 h (Fig. 5,B). Northern blot analysis showed that IL-1β up-regulated COX2 mRNA expression in osteoblasts at 3 h (Fig. 5,C). IL-1β also stimulated RANKL mRNA expression at 3 h (data not shown). IL-1β stimulated osteoclast formation in the cocultures of OPG −/− osteoblasts and wild-type bone marrow cells (the control: 53 ± 11, the mean ± SD of four cultures) (Fig. 5,D). NS398 (1 μM) did not completely suppress LPS-induced osteoclast formation in the cocultures with OPG −/− osteoblasts, but OPG did (Fig. 5 D). These results suggest that IL-1 and LPS stimulate osteoclast formation in the same manner in the cocultures.
In vivo and in vitro experiments have shown that PGE2 is crucially involved in the induction of osteoclastic bone resorption by IL-1, TNF-α, and LPS (20). EP4 subtype-mediated signaling has been shown to be particularly important for the induction of bone resorption by such inflammation-related factors as well as PGE2 (20). The present study showed that LPS stimulated COX2 expression and PGE2 production in osteoblasts, and NS398, a specific inhibitor of COX2, strongly blocked the LPS-induced osteoclast formation in cocultures containing wild-type osteoblasts (Figs. 1 and 2). These results suggest that PGE2 is somehow involved in LPS-induced osteoclast formation in the cocultures through PGE2 receptors of EP4 subtype.
It was reported that LPS stimulated the expression of RANKL mRNA in osteoblasts obtained from EP4−/− mice, and that COX inhibitors did not block this stimulation (21). In agreement with this finding, NS398 failed to inhibit the induction of RANKL expression by LPS in osteoblasts (Fig. 3). These results suggest that LPS induced RANKL expression in a manner that was independent of PGE2 production in osteoblasts. In contrast, the treatment of osteoblasts in the cocultures with LPS together with NS398 blocked the down-regulation of OPG mRNA expression at 48 h (Fig. 3). This suggests that suppression of OPG by PGE2 is an important event in osteoclast formation in the cocultures treated with LPS. This notion was further supported by the finding that LPS stimulated osteoclast formation even in the presence of NS398 in cocultures containing osteoblasts derived from OPG−/− mice (Fig. 4). Thus, PGE2 appears to play an important role as a suppressor of OPG expression rather than an activator of RANKL expression in LPS-induced osteoclast formation (Fig. 6). Recently, Fu et al. (38) reported that the activation of CREB by PTH is required for PTH-induced down-regulation of OPG expression. This suggests that the cAMP-PKA signals play a role in PGE2-induced suppression of OPG mRNA expression. Further study will elucidate the detail mechanism of the down-regulation of OPG expression by PGE2.
PGE2 has been shown to induce RANKL mRNA expression in osteoblasts (10). Suzawa et al. (39) reported that PGE2-induced RANKL expression is mediated through the cAMP signaling pathway. In our experiments, NS398 failed to inhibit RANKL expression in osteoblasts treated with LPS for as long as 48 h (Fig. 3). This suggests that LPS induces RANKL expression by the mechanism independent of PGE2 production. Kikuchi et al. (40) reported that LPS induces RANKL through extracellular signal-regulated kinase (ERK) and PKC. We also confirmed that calcium/PKC inhibitors, such as BAPTA-AM (an intracellular calcium chelator) and Ro-32-0432 (a PKC inhibitor), and ERK inhibitor PD98059 inhibited LPS-induced RANKL mRNA expression in osteoblasts (K.S., unpublished observation). PD98059 failed to inhibit the induction of RANKL mRNA expression by PGE2 and the induction of osteoclast formation in cocultures treated with PGE2 (data not shown). These results suggest that LPS directly stimulates RANKL expression through calcium/PKC signals, followed by ERK signals in osteoblasts. It is unlikely that PGE2-induced signals directly cross talk with LPS-induced signals in the induction of RANKL expression in osteoblasts.
The intracellular signaling pathway of TLR4 is quite similar to that of IL-1Rs (25, 26, 27, 28, 29). Like LPS, IL-1 stimulated COX2 mRNA expression at 3 h and PGE2 production at 6 h in osteoblast cultures (Fig. 5). IL-1β also induced RANKL mRNA expression in osteoblasts, as previously reported (data not shown) (32). NS398 inhibited IL-1-induced osteoclast formation strongly in cocultures containing wild-type osteoblasts, but only partially in cocultures containing OPG−/− osteoblasts (Fig. 5). These results suggest that PGE2 produced by osteoblasts in response to IL-1 plays a similar role to LPS in osteoclast formation through the suppression of OPG expression (Fig. 6).
In cocultures containing OPG−/− osteoblasts, osteoclasts were formed even in the absence of any stimulus (Fig. 4). The spontaneous osteoclast formation was strongly inhibited by the addition of either OPG or NS398 (Fig. 4). These results suggest that RANKL is involved in the spontaneous osteoclast formation, and that PGE2 constitutively produced in the cocultures stimulates RANKL expression in osteoblasts. LPS and IL-1β further enhanced osteoclast formation in cocultures containing OPG−/− osteoblasts (Fig. 4), suggesting that the up-regulation of RANKL expression by LPS and IL-1 enhances the osteoclast formation. The induction of osteoclast formation by LPS and IL-1 in cocultures containing OPG−/− osteoblasts was partially inhibited by the addition of NS398. Therefore, PGE2 induced by LPS and IL-1 appears to be involved in RANKL expression in osteoblasts. Our results indicate that the full inhibition of LPS- and IL-1-induced osteoclast formation by NS398 requires PGE2-dependent suppression of OPG production (Fig. 6).
The previous studies have shown that OPG production by osteoblasts is down-regulated by bone-resorbing factors such as 1,25(OH)2D3, PTH, and PGE2 (38, 41, 42, 43, 44). Our results confirmed the previous finding that osteoclasts spontaneously form in the control cocultures containing OPG−/− osteoblasts. The decrease in OPG production by osteoblasts was a key event for the induction of osteoclastogenesis by LPS and IL-1. OPG−/− mice exhibited severe osteoporosis caused by enhanced osteoclast formation and function (14, 15, 45, 46). These results suggest that OPG is a physiological regulator of bone resorption, and that the balance between RANKL and OPG expressions at bone is particularly important for the regulation of bone resorption in vivo and in vitro.
We previously reported that LPS and IL-1 directly stimulated the survival, fusion, and pit-forming activity of osteoclasts (47). Those results together with the results shown in this study suggest that LPS and IL-1 are involved in the stimulation of osteoclastic bone resorption in several ways: LPS and IL-1 directly stimulate osteoclast function, induce RANKL expression in osteoblasts, and suppress OPG expression through enhancement of PGE2 production. Further studies will be necessary to elucidate the precise mechanism of the regulation of osteoclastic bone resorption induced by these inflammatory factors.
This study was supported in part by grants-in-aid (12137209, 13557155, 13470394, and 14207075) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. K.S. is Research Fellow of the Japan Society for the Promotion of Science.
Abbreviations used in this paper: RANKL, receptor activator of NF-κB ligand; 1,25(OH)2D3, 1,25-dihydroxyvitamin D3; COX2, cyclooxygenase 2; EIA, enzyme immunoassay; ERK, extracellular signal-regulated kinase; IL-1ra, IL-1 receptor antagonist; OPG, osteoprotegerin; PKA, protein kinase A; PKC, protein kinase C; PTH, parathyroid hormone; RANK, receptor activator of NF-κB; TLR, Toll-like receptor; TRAP, tartrate-resistant acid phosphatase.