Mechanisms Involved in Enhancement of Osteoclast Formation and Function by Low Molecular Weight Hyaluronic Acid*

Hyaluronic acid (HA) is a component of the extracellular matrix that has been shown to play an important role in bone formation, resorption, and mineralization both in vivo and in vitro. We examined the effects of HA at several molecular weights on osteoclast formation and function induced by RANKL (receptor activator of NF-κB ligand) in a mouse monocyte cell line (RAW 264.7). HA at Mr < 8,000 (low molecular weight HA (LMW-HA)) enhanced tartrate-resistant acid phosphatase-positive multinucleated cell formation and tartrate-resistant acid phosphatase activity induced by RANKL in a dose-dependent manner, whereas HA at Mr > 900,000 (high molecular weight HA (HMW-HA)) showed no effect on osteoclast differentiation. LMW-HA enhanced pit formation induced by RAW 264.7 cells, whereas HMW-HA did not, and LMW-HA stimulated the expression of RANK (receptor activator of NF-κB) protein in RAW 264.7 cells. In addition, we found that LMW-HA enhanced the levels of c-Src protein and phosphorylation of ERKs and p38 MAPK in RAW 264.7 cells stimulated with RANKL, whereas the p38 MAPK inhibitor SB203580 inhibited RANKL-induced osteoclast differentiation. This enhancement of c-Src and RANK proteins induced by LMW-HA was inhibited by CD44 function-blocking monoclonal antibody. These results indicate that LMW-HA plays an important role in osteoclast differentiation and function through the interaction of RANKL and RANK.

Hyaluronan, or hyaluronic acid (HA), 1 is a long polysaccharide chain that is made of repeating disaccharide units of N-acetylglucosamine and glucuronic acid. HA is the most abundant glycosaminoglycan in mammalian tissue and is present in high concentrations in connective tissues as well as in skin, the vitreous humor of the eye, cartilage, and umbilical cord tissue. The largest single reservoir is the synovial fluid of diarthroses joints, where concentrations of 0.5-4 mg/ml have bee found (1,2). In rheumatoid arthritis and osteoarthritis, the molecular weight and concentration of HA in synovial fluid are reduced, probably due to abnormal biosynthesis by synovial type B cells and free radical depolymerization of the HA chain (3). These changes in the rheological properties of synovial fluid in arthritic joints may contribute to disease progression because articular cartilage, subchondral bone, and synovial connective tissues are subjected to increased mechanical stress when the viscoelastic and lubricating properties of HA are diminished.
Many studies have reported the biological effects of HA on chondrocytes (4 -11). Furthermore, HA affects the catabolic activity of the cartilage matrix by inducing the production of TIMP-1 (tissue inhibitor of metalloproteinase-1) by bovine chondrocytes (6). Sakamoto et al. (12) reported that high molecular weight HA (HMW-HA) injected into the joint cavity becomes attached to the surface of articular cartilage, where it exerts chondroprotective effects. Extrinsic HA has also been shown to penetrate degenerated cartilage and to reach the surface of chondrocytes (13). In addition, in vitro studies using synovial fibroblasts have indicated that synovial type B cells, the major source of synovial HMW-HA, lead to de novo synthesis of HMW-HA (14).
In a previous study utilizing high performance liquid chromatography (15), we demonstrated that the molecular weights of HA in synovial fluid from temporomandibular joint samples from patients with internal derangement and osteoarthritis are decreased probably due to free radical depolymerization of the HA chain and/or abnormal biosynthesis by the synovium. The results of that study shed light on the mechanisms of the changes in joint lubrication and the pathology of the temporomandibular joint. Several researchers have reported the involvement of the interaction between CD44 and HA in bone resorption (2,16,17). However, less attention has been paid to the effect of low molecular weight HA (LMW-HA) on osteoclast formation induced by RANKL (receptor activator of NF-B ligand). In this study, we examined the effects of HA at several molecular weights on osteoclastogenesis induced by RANKL and found that LMW-HA enhanced both osteoclast formation and function in vitro.
Cell Culture-Mouse monocyte RAW 264.7 cells were maintained in ␣-minimal essential medium (Invitrogen) supplemented with 10% fetal * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Viability-RAW 264.7 cells were plated in 96-well plates at a concentration of 5 ϫ 10 2 cells/well 1 day before the experiment and then stimulated with RANKL and HA at several molecular weights. The stimulated cells were cultured for 1, 2, 5, or 6 days, after which a stock 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide solution (2.5 mg/ml, 20 l/well; Sigma) was added to the wells, and the plates were incubated for 4 h. Next, acid/isopropyl alcohol (100 l of 0.04 N HCl in isopropyl alcohol) was added and mixed thoroughly, and the plates were read using a Multiskan Bichromatic microplate reader (Labsystems, Helsinki, Finland), with a test wavelength of 540 nm and a reference wavelength of 620 nm (18).
Differentiation of Osteoclasts-Osteoclasts were detected by staining with tartrate-resistant acid phosphatase (TRAP) (Sigma) (19). In brief, RAW 264.7 cells were cultured in 96-well plates (5.0 ϫ 10 2 cells/well) in the presence of RANKL (40 ng/ml), HA at several molecular weights, or SB203580. After culturing for the indicated times, the adherent cells were fixed and stained for TRAP activity. TRAP-positive multinucleated cells containing three or more nuclei were considered to be osteoclast-like cells (OCLs) and were counted under a microscope.
TRAP Activity-For the TRAP activity assay, RAW 264.7 cells were cultured in 96-well plates (5.0 ϫ 10 2 cells/well) in the presence of RANKL (40 ng/ml), HA (100 ng/ml) at several molecular weights, or SB203580 for 6 days. The treated RAW 264.7 cells were suspended in 25 l of phosphate-buffered saline, pH 7.2, and then frozen and thawed three times. TRAP activities in the supernatants were analyzed using a phenyl phosphate substrate kit (Sanseiphospha KII-Test-Wako, Wako, Osaka, Japan) according to the manufacturer's instructions (20).
Bone Resorption Assay-To estimate bone resorption activity, RAW 264.7 cells were cultured for 14 days with RANKL (40 ng/ml) and HA (100 ng/ml) at several molecular weights on BD BioCoat TM Osteologic TM multitest slides, which consisted of submicron synthetic calcium phosphate thin films coated onto various culture vessels (BD Biosciences). The cells were removed using 6% NaOCl and 5.2% NaCl, and the number of the resorption pits formed in each well was counted under a microscope.
Western Blot Analysis-RAW 264.7 cells (1 ϫ 10 5 ) were cultured in ␣-minimal essential medium containing 10% fetal calf serum in the presence of RANKL (40 ng/ml) and HA (100 g/ml) at several molecular weights (M r ϭ 8,000 and 2,000,000), SB203580 (10 Ϫ6 M), or BRIC 235 (5 g/ml) on 6-well plates. Adherent cells were washed twice with phosphate-buffered saline and lysed in cell lysis buffer (75 mM Tris-HCl containing 2% SDS and 10% glycerol, pH 6.8). Protein contents were measured using a DC protein assay kit (Bio-Rad). The samples subjected to 10% SDS-PAGE and then transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). Nonspecific binding sites were blocked by immersing the membranes in 10% skim milk in phosphatebuffered saline for 1 h at room temperature, and the membrane was washed four times with phosphate-buffered saline, followed by incubation with diluted primary antibody for 2 h at room temperature. Anti-c-Src, anti-ERK, anti-phospho-ERK, anti-p38 MAPK, anti-phospho-p38 MAPK, and anti-RANK primary antibodies and horseradish peroxidase-conjugated anti-mouse, anti-rabbit, and anti-goat IgG secondary antibodies (Santa Cruz Biotechnology, Inc.) were used in this experiment. After washing the membranes, chemiluminescence was produced using ECL reagent (Amersham Biosciences) and detected with Hyperfilm-ECL (Amersham Biosciences). After exposure to film, the membranes were stained with Coomassie Brilliant Blue G-250 to confirm equal loading.
Statistical Analysis-Statistical differences were determined using an unpaired Student's t test with Bonferroni's correction for multiple comparisons. All data are expressed as the mean Ϯ S.E.

RESULTS
Effects of RANKL and HA on Cell Growth-We examined the effects of RANKL and HA at several molecular weights on the proliferation of RAW 264.7 cells using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide viability assay. HA had no effect on RAW 264.7 cell growth after 1, 2, 5, and 6 days of culture with RANKL (data not shown).
LMW-HA Up-regulates Osteoclast Differentiation Induced by RANKL-Culturing LMW-HA activated the differentiation of RAW 264.7 cells into OCLs via RANKL in a dose-dependent manner (Fig. 1, A and B), whereas HMW-HA had almost no effect. The number of TRAP-positive multinucleated cells was maximized at higher concentrations (100 ng/ml) in this exper-iment; however, LMW-HA alone did not induce OCL formation (data not shown). Incubation with LMW-HA and RANKL for 0 -2 days was adequate to enhance OCL formation in RAW 264.7 cells (Fig. 1C).
LMW-HA Up-regulates TRAP Activity Induced by RANKL-We examined the effects of RANKL in the presence of HA at several molecular weights on the TRAP activity of RAW 264.7 cells using the phenyl phosphate substrate method. As shown in Fig. 2, RANKL alone enhanced TRAP activity in RAW 264.7 cells; however, when the cells were incubated with both RANKL and LMW-HA, the level of TRAP expression was 1.5fold higher than in cells treated with RANKL alone. In contrast, HMW-HA had no effect on RANKL-induced TRAP activity in RAW 264.7 cells.

LMW-HA Stimulates Bone Resorption Induced by RANKL-
To determine whether LMW-HA affects osteoclast function, differentiated RAW 264.7 cells were cultured on Osteologic TM multitest slides with RANKL (40 ng/ml) in the presence of HA (100 g/ml) at several molecular weights. LMW-HA enhanced the stimulatory effect of RANKL on bone resorption (Fig. 3), whereas it had no effect on basal bone resorption, with or without RANKL.
LMW-HA Enhances the Expression of RANK Protein in RAW 264.7 Cells-We also examined RANK expression in RAW 264.7 cells by immunoblot analysis. RAW 264.7 cells were cultured for 6 -72 h in the presence or absence of RANKL and HA at several molecular weights. Interestingly, treatment with LMW-HA alone induced RANK expression and enhanced the expression level in RAW 264.7 cells treated with RANKL, whereas HMW-HA had no effect on the level of RANK protein expression (Fig. 4).
LMW-HA Enhances RANKL-induced Expression of c-Src Protein in RAW 264.7 Cells-Next, we investigated the effect of RANKL and HA at several molecular weights on the expression of c-Src protein in RAW 264.7 cells by Western blotting. As shown in Fig. 5, the level of c-Src protein expression was increased in cells following stimulation with RANKL. Furthermore, c-Src protein expression was up-regulated upon the addition of LMW-HA in a time-dependent manner up to 72 h.
LMW-HA Enhances RANKL-induced Activation of MAPKs in RAW 264.7 Cells-The effects of RANKL and HA at several molecular weights on the activation of ERKs at M r ϭ 42,000 and 44,000 in RAW 264.7 cells were investigated (Fig. 6). Western blot analysis revealed that phosphorylated ERKs were detected within 15 min and reached a plateau at 30 min after the addition of RANKL. When the cells were incubated with both RANKL and LMW-HA, the level of ERK phosphorylation was higher than in cells treated with RANKL alone. In contrast, the total amounts of ERK protein were not affected by treatment with RANKL and HA.
We also examined the effects of RANKL and HA at several molecular weights on the phosphorylation of p38 MAPK in osteoclast precursors. Fig. 7 shows the time course of changes in the level of p38 MAPK phosphorylation in response to RANKL in RAW 264.7 cells. p38 MAPK was phosphorylated within 30 min in response to RANKL, and phosphorylation reached a maximum level within 60 min. LMW-HA stimulated the RANKL-induced phosphorylation of p38 MAPK; however, there was no change in the level of p38 MAPK phosphorylation upon treatment with LMW-HA alone. Furthermore, the total amounts of p38 MAPK protein in RAW 264.7 cells stimulated  with RANKL were unchanged in the presence and absence of HA at all time periods tested.
SB203580 Inhibits the Differentiation of RAW 264.7 Cells Induced by RANKL and LMW-HA-To further examine the role of p38 MAPK in RANKL-and HA-mediated osteoclast differentiation, RAW 264.7 cells were treated with RANKL, HA at several molecular weights, or SB203580 (a specific inhibitor of p38 MAPK). A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide viability assay showed that SB203580 had no significant effect on the proliferation of RAW 264.7 cells (data not shown), whereas it inhibited the induction of TRAPpositive multinucleated cells by RANKL and LMW-HA in a dose-dependent manner (Fig. 8). To confirm that p38 MAPK was inhibited by SB203580, RAW 264.7 cells were treated with RANKL and HA at several molecular weights in the presence and absence of SB203580. Western blot analysis revealed that SB203580 inhibited p38 MAPK activity mediated by RANKL and LMW-HA (Fig. 9).

CD44 Function-blocking Monoclonal Antibody Inhibits the Expression of c-Src and RANK Proteins in RAW 264.7 Cells
Induced by RANKL and LMW-HA-To further examine the role of CD44 as an HA receptor in this signaling pathway, RAW 264.7 cells were treated with RANKL and LMW-HA in the presence of CD44 function-blocking monoclonal antibody. Western blot analysis revealed that the monoclonal antibody inhibited the enhancement of c-Src and RANK protein expression mediated by LMW-HA (Fig. 10). DISCUSSION In this study, we used a homogeneous clonal population of murine monocyte RAW 264.7 cells to elucidate the direct effects of RANKL and HA on osteoclast differentiation and function. This cell line is known to express RANK and to differentiate into TRAP-positive cells when cultured with bone slices and RANKL (21). The main advantage of this system is that it does not contain any osteoblast/bone marrow stromal cells, which may also be targets of RANKL and LMW-HA actions. We focused on pre-osteoclast cells to examine the effects of RANKL and LMW-HA on differentiation and function. We also found that LMW-HA activated OCL formation induced by RANKL in mouse bone marrow culture (data not shown). These results suggest that the stimulatory effect of LMW-HA on OCL formation is involved in the RANKL-mediated signaling pathway in mouse bone marrow cells as well as RAW 264.7 cells.
RANKL, a member of the tumor necrosis factor family, triggers osteoclastogenesis by forming a complex with RANK, a member of the tumor necrosis factor receptor family. The binding of RANKL to RANK results in a cascade of intracellular events, including the activation of the intracellular adaptor protein family in pre-osteoclast cells (21)(22)(23). Our results show that LMW-HA markedly increases the level of RANK protein expression in RAW 264.7 cells. It is possible that LMW-HA changes RANK action, with an ultimate increase in osteoclast development in addition to increases in RANK production.
Bone resorption is a multistep process initiated by the proliferation of immature osteoclast precursors, which is followed by the commitment of those cells to the osteoclast phenotype and degradation of the organic and inorganic phases of bone by mature resorptive cells. Like their in vivo counterparts, in vitro generated osteoclasts are capable of bone resorption. When cultured with bone or dentin, osteoclasts excavate resorptive lacunae, or pits, which are similar to the structures formed when the cells degrade bone in vivo, and the number and size of the resorption lacunae formed in vitro are used as a quantitative measure of osteoclast activity (24). In this study, we used Osteologic TM slides coated with calcium phosphate substrate and observed the up-regulation of the pit-forming activity of OCLs stimulated with RANKL.
RANKL increases the level of c-Src protein, another marker molecule of osteoclast differentiation (25) that is a widely expressed non-receptor tyrosine kinase particularly abundant in platelets and neural tissues (26,27) and osteoclasts (28,29). c-Src plays an essential role in osteoclast function, as mice in which the src gene has been disrupted show normal osteoclast development, but fail to resorb bone, resulting in osteopetrosis (30). In the present TRAP activity results, the expression of c-Src induced by RANKL was enhanced by LMW-HA, suggesting that the enhanced osteoclast formation by LMW-HA is not caused by stimulating the proliferation of monocyte/macrophage lineage cells, but rather by increasing the number of cells committed to osteoclast lineage, which are responsive to RANKL-mediated terminal differentiation to mature osteoclasts.
MAPK family members are proline-directed serine/threonine kinases that are important for cell growth, differentiation, and apoptosis (31)(32)(33)(34) and become activated by phosphorylation of threonine and tyrosine in response to external stimuli. MAPK family members are classified into the ERK, JNK, and p38 MAPK groups, and it is widely accepted that peptide growth factors and phorbol esters preferentially activate ERKs, whereas cellular stress, such as that caused by hyperosmolarity or reactive oxygen species, potently activates JNKs and p38 MAPKs (35)(36)(37). A previous study has shown that activation of the p38 MAPK pathway plays an important role in the RANKL-induced osteoclast differentiation of precursor bone marrow cells (38). In this study, RANKL-induced activation of ERKs and p38 MAPK was clearly detected, and LMW-HA enhanced the activities of both kinases in RAW 264.7 cells.
The pyridinyl imidazole SB203580, a specific inhibitor of p38 MAPK (39), has been widely used to investigate the roles of p38 MAPK in the regulation of cell differentiation and function (37,40,41). p38 MAPK-mediated signals were shown to be involved in osteoclast bone resorption induced by interleukin-1 and tumor necrosis factor-␣ in fetal rat long bones using SB203580 (41). These results suggest that p38 MAPK-mediated signals regulate osteoclast differentiation or function. In this study, SB203580, but not PD98059, inhibited differentiation and TRAP activity induced by RANKL and LMW-HA, whereas it inhibited only p38 MAPK activity in RAW 264.7 cells (Fig. 9). Our findings suggest that p38 MAPK may play a critical role in RANKL/LMW-HA-induced osteoclast formation in RAW 264.7 cells. However, we cannot rule out the possible involvement of ERK in osteoclastogenesis.
In our experiments, LMW-HA enhanced both the differentiation and function of RAW 264.7 cells when cultured with RANKL. Interestingly, HMW-HA had almost no effect on osteoclast formation and function. Previous studies (4 -11) have confirmed the physiological effects on chondrocytes; however, relatively little attention has been directed toward the differentiation and function of osteogenic cells in bone metabolism. Furthermore, there are no known reports concerning the pathological effects of different molecular weights of HA on osteoclastogenesis or its signal transduction in osteoclasts. Our results suggest that HA at low molecular weights is effective in enhancing osteoclast formation and function under inflammatory conditions (15). We have no ready explanation for this phenomenon, but suggest that it is caused by the interaction of HA and HA-binding protein in joint tissues under pathological conditions.
HA is a major component of synovial fluid and plays impor- RAW 264.7 cells were exposed to RANKL (40 ng/ml) and LMW-HA (100 g/ml) in the presence or absence of CD44 function-blocking monoclonal antibody BRIC 235 (5 g/ml). Whole cell lysates were subjected to Western blot analysis. tant roles in the joint cavity. CD44, one of the major HAbinding proteins, is expressed in several human cells, including lymphocytes, alveolar macrophages, and fibroblasts, as well as in several kinds of tumor cells (42). Culty et al. (43) demonstrated that mature alveolar macrophages may be involved in regulating HA levels in the lung and suggested that the binding of HA and CD44 leads to the efficient degradation of HA. Furthermore, it has been reported that CD44-mediated degradation of HA may play an essential role during embryonic morphogenesis and the development of some organs, such as the liver, spleen, and thymus; bone marrow; and hair follicles. More recently, Cao et al. (44) reported that HA activates CD44 to stimulate RANKL expression in bone marrow stromal cells. In this study, we found that CD44 function-blocking monoclonal antibody remarkably inhibited the effect of HA on the signal introduction of c-Src and RANK in RAW 264.7 cells. On the basis of these findings, we speculate that the expression of CD44 in osteoclasts, which have the same lineage as alveolar macrophages, leads to the localized degradation of HA in the joint cavity, which in turn leads to induction of osteoclast formation and activation mediated by LMW-HA in the surrounding tissues. Further study is needed to examine the correlations between LMW-HA and CD44 regarding the enhancement of osteoclast formation and activation induced by RANKL in bone marrow stromal cells as well as signal transduction during osteoclastogenesis in joint tissue destruction.