Staphylococcus aureus Induces Expression of Receptor Activator of NF-κB Ligand and Prostaglandin E₂ in Infected Murine Osteoblasts

Abstract

Osteomyelitis is an inflammatory disease of the bone that is characterized by the presence of necrotic bone tissue and increased osteoclast activity. Staphylococcus aureus is responsible for approximately 80% of all cases of human osteomyelitis. While the disease is especially difficult to treat, the pathogenesis of S. aureus-induced osteomyelitis is poorly understood. Elucidating the molecular mechanisms by which S. aureus induces osteomyelitis could lead to a better understanding of the disease and its progression and development of new treatments. Osteoblasts can produce several soluble factors that serve to modulate the activity or formation of osteoclasts. Receptor activator of NF-κB ligand (RANK-L) and prostaglandin E₂ (PGE₂) are two such molecules which can promote osteoclastogenesis and stimulate bone resorption. In addition, previous studies in our laboratory have shown that osteoblasts produce inflammatory cytokines, such as interleukin 6, following infection with S. aureus, which could induce COX-2 and in turn PGE₂, further modulating osteoclast recruitment and differentiation. Therefore, we hypothesized that following infection with S. aureus, osteoblasts will express increased levels of RANK-L and PGE₂. The results presented in this study provide evidence for the first time that RANK-L mRNA and protein and PGE₂ expression are upregulated in S. aureus-infected primary osteoblasts. In addition, through the use of the specific COX-2 inhibitor NS 398, we show that when PGE₂ production is inhibited, RANK-L production is decreased. These data suggest a mechanism whereby osteoblasts regulate the production of RANK-L during infection.

Osteomyelitis is a severe infection of bone tissue that results in progressive inflammatory destruction of bone (41). The gram-positive organism Staphylococcus aureus is the most common causative agent of osteomyelitis, accounting for approximately 80% of all human cases (27). It is often necessary that infected bone be debrided, tissues reconstructed, and long-term antibiotic therapy utilized (18). The current treatment for osteomyelitis is often traumatic and expensive and often leads to further selection of antibiotic-resistant strains of S. aureus. The growing incidence of antibiotic-resistant S. aureus strains can explain the recurrent attacks of osteomyelitis in patients undergoing therapy. In addition, the pathogenesis of S. aureus-induced osteomyelitis is poorly understood. Elucidating the mechanisms by which S. aureus induces osteomyelitis could therefore lead to a better understanding of the disease, its progression, and development of new treatments.

Bone remodeling involves a continuous, coordinated equilibrium between bone synthesis and bone resorption, and two cell populations are responsible for the continual process (23). Osteoclasts drive the resorption of bone by acidification and release of lysosomal enzymes, while osteoblasts produce components of the bone matrix, principally type I collagen. Osteoblasts also produce factors which serve to modulate the activity or formation of osteoclasts (1). One such key molecule produced by osteoblasts is receptor activator of NF-κB ligand (RANK-L), which on binding to RANK, found on osteoclast precursors, triggers intricate and distinct signaling cascades that control lineage commitment and activation of osteoclasts (53). RANK-L is therefore thought to be the essential and final common signal required both in vitro and in vivo for full osteoclastic differentiation from multipotential hematopoietic precursor cells into mature multinucleated bone-resorptive osteoclasts (22, 24, 25). Thus, osteoblasts can regulate both net bone formation and resorption.

Prostaglandin E₂ (PGE₂) is produced in bone mainly by osteoblasts and acts as a potent stimulator of bone resorption by inducing osteoclast formation in both mouse bone marrow cultures and calvarial cultures (1, 7, 37, 40, 46, 48). PGE₂ synthesis is regulated by three metabolic steps: the release of arachidonic acid from the membranous phospholipids by phospholipase A₂, the conversion of arachidonic acid to prostaglandin H₂ (PGH₂) by cyclooxygenase (COX), and the synthesis of PGE₂ catalyzed by PGE synthase (9, 20, 26, 34, 42, 44, 50). Although both constitutive COX (COX-1) and inducible COX (COX-2) are expressed in mouse osteoblasts, the expression of COX-2 is markedly induced by several proinflammatory cytokines, such as interleukin 6 (IL-6). Bost et al. (5, 11) have reported that S. aureus-infected osteoblasts produce high levels of IL-6, which have been shown to modulate the activity of bone-resorptive osteoclasts either directly or indirectly (21, 45). It has also been determined that PGE₂ can upregulate RANK-L expression in osteoblasts (16, 17, 54, 57).

Bacterial pathogens can stimulate osteoclastogenesis, which can potentiate bone resorption, leading to bone destruction (31). Okahashi et al. have also demonstrated that normal mouse osteoblasts infected with Streptococcus pyogenes increase expression of RANK-L mRNA and protein (36). In addition, elevated levels of RANK-L have been detected in the bone lesions of mandibular osteomyelitis (33). We therefore propose that following an intracellular challenge with S. aureus, osteoblasts can increase RANK-L expression, which could stimulate osteoclast activity and thereby potentiate the massive bone loss that occurs in S. aureus-infected bone tissue. The current study investigates RANK-L mRNA and protein expression during normal mouse osteoblast infection with S. aureus. Such a proposed upregulation of RANK-L from infected osteoblasts could be mediated either directly or indirectly. Since it has been determined that PGE₂ can upregulate RANK-L, we propose in the current study that PGE₂ production by S. aureus-infected osteoblasts will be increased and that inhibition of PGE₂ by the specific COX-2 inhibitor NS-398 will result in decreased levels of PGE₂ and RANK-L by infected osteoblasts.

MATERIALS AND METHODS

Bacterial strains.

Staphylococcus aureus strain UAMS-1 (ATCC 49230) is a human osteomyelitis clinical isolate (15). Staphylococcus carnosus (ATCC 51365) is a nonpathogenic species with a reduced ability to invade osteoblasts (11). Strains N1, N2, N3, N4, and N5 are Staphylococcus aureus nasopharyngeal isolates kindly provided by Vance G. Fowler, Jr. (Duke University Medical Center).

Isolation and culture of mouse osteoblasts.

Osteoblasts were isolated from the calvariae of 2-day-old CD-1 mice by sequential collagenase and protease digestions as described previously (38).

Invasion assay.

The invasion assay with live or UV-killed S. aureus strain UAMS-1, live S. carnosus, or live nasopharyngeal isolates was carried out as previously described (11, 12). Briefly, bacteria were grown overnight (16 h) in 5 ml of tryptic soy broth in a shaking water bath at 37°C. The bacteria were harvested by centrifugation for 10 min at 4,300 × g at 4°C and washed twice. The pellets were then resuspended in 5 ml of growth medium lacking antibiotics/antimycotics. Confluent cell layers of osteoblasts were washed three times to remove the growth medium. The cultures were then infected at a multiplicity of infection (MOI) of 25:1 or 75:1 with S. aureus strain UAMS-1 or with S. carnosus, UV-killed S. aureus, or the nasopharyngeal isolates at a MOI of 25:1 in 4 ml of growth medium lacking antibiotics/antimycotics. Following a 45-min infection period, infected cell cultures were washed three times and incubated for 1 h in growth medium containing 25 μg of gentamicin/ml to kill the remaining extracellular bacterial cells. The osteoblast cultures were washed and subsequently lysed at different time points by the addition of 1.2 ml of 0.1% Triton X-100 (Fisher Biotech, Fair Lawn, NJ) with incubation for 5 min at 37°C.

Quantitative real-time PCR.

To quantify RANK-L mRNA, real-time PCR was performed using a Roche LightCycler 2.0 instrument with the Sybr green reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Amplicons of RANK-L and glyceraldehyde-3-phosphate dehydrogenase (G3PDH) were used to develop a standard curve. Samples were subjected to 40 cycles of amplification consisting of 95°C for 15 s followed by 53°C for 20 s. Each assay was normalized to G3PDH mRNA. PCR primers were derived from published sequences (IDT, Coralville, IA) (36). Positive- and negative-strand primers were as follows: RANK-L (89 bp), 5′-TACTTTCGAGCGCAGATGGAT-3′ and 5′-ACCTGCGTTTTCATGGAGTCT-3′; G3PDH (63 bp), 5′-AACTACATGGTCTACATGTTCCA-3′ and 5′-CCATTCTCGGCCTTGACTGT-3′.

Protein isolation.

Osteoblasts were isolated and seeded at 10⁶ cells per well in six-well plates and infected once the cells had reached confluence (6 to 8 days). Following a 45-min incubation with or without bacteria and a 1-h treatment with 25 μg/ml gentamicin sulfate, total protein was isolated from the osteoblasts at 0, 6, and 24 h using a tissue protein extraction reagent (Pierce, Rockford, IL) with additional protease inhibitors, aprotinin (2 μg/ml) and leupeptin (2.5 μg/ml) (Sigma Chemical Company, St. Louis, MO). Two hundred μl of the tissue protein extraction reagent was added to each well at the appropriate time point, and cells were removed from the culture plates with a cell scraper (Costar). Samples were collected in 1.5-ml microcentrifuge tubes and then centrifuged at 10,000 × g for 2 min to pellet cell debris. Supernatants were collected, and placed in fresh tubes, and stored at −80°C.

Inhibitor study.

Osteoblasts were isolated and cultured in six-well plates as described previously. Forty-eight hours prior to infection, cells were treated with either 0.025% dimethyl sulfoxide (DMSO) or 20 μM NS-398 (Cayman Chemicals, Ann Arbor, MI), a specific COX-2 inhibitor, dissolved in DMSO at an optimal dose that was previously determined to provide inhibition of COX-2 (6, 58). Following the 48-h treatment, cells were rinsed one time with osteoblast growth medium (OBGM) containing no gentamicin sulfate or amphotericin B. One ml of OBGM, containing no gentamicin sulfate or amphotericin B and containing either DMSO or NS-398, was then added to each appropriate well. The osteoblasts were either uninfected (MOI = 0) or infected with the bacterial strains noted above using the method described previously. Following the 45-min incubation with or without bacteria, each well was rinsed one time with OBGM. One ml of fresh medium, containing either DMSO or NS-398 and 25 μg/ml of gentamicin, was then added to each well. Cells were then incubated for 1 h at 37°C in a 5%-CO₂ atmosphere to allow the remaining extracellular bacteria to be killed. Following the 1-h incubation period, protein samples were isolated as described previously (time zero). Samples were also isolated at 6 h and 24 h following time zero. In addition, culture supernatants were collected and stored at −80°C.

Quantification of RANK-L protein and PGE₂.

Using protein samples collected as described previously, a mouse TRANCE/TNSF11/RANK ligand enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN) was used to detect the levels of RANK-L in osteoblast whole-cell extracts. A PGE₂ correlate enzyme immunoassay (EIA) (Assay Designs, Ann Arbor, MI) was performed with the cell culture supernatants following the manufacturer's instructions.

Statistical analysis.

Data were analyzed using a two-way analysis of variance followed by the Tukey-Kramer multiple-comparison analysis. Results are presented as the means ± standard errors. For all tests, P values of <0.05 were considered significant.

RESULTS

Induction of RANK-L mRNA expression in osteoblasts following exposure to bacteria.

Staphylococcal osteomyelitis is associated with increased osteoclast activity, and RANK-L is a key factor required for osteoclastogenesis (24, 25, 46, 57). Therefore, it is important to determine whether osteoblasts infected with S. aureus produce RANK-L and if the levels of RANK-L increase in a time- and dose-dependent manner. Quantitative analysis by real-time reverse transcription-PCR demonstrated that RANK-L expression in osteoblasts infected with S. aureus strain UAMS-1 (MOI = 75:1) was 10-fold greater than RANK-L expression in uninfected cells at 4 h following infection (Fig. 1). RANK-L mRNA expression by osteoblasts also increased in a dose-dependent manner. In addition, we showed that there was a time-dependent increase in RANK-L mRNA (Fig. 1), with the expression of RANK-L mRNA by infected cells being 10-fold greater than the expression in the uninfected controls at 4 h.

FIG. 1.

Expression of RANK-L mRNA by osteoblasts. Primary mouse osteoblasts either were not infected (MOI = 0) or were infected with S. aureus strain UAMS-1 (MOI = 25:1 or 75:1). Following a 45-min incubation with or without bacteria (time zero hours), total RNA was extracted at various time points and subjected to quantitative real-time reverse transcription-PCR analysis for RANK-L mRNA. All values were normalized to G3PDH and are expressed as n-fold change over baseline levels (0 h; MOI = 0). Data are expressed as the means ± standard errors for three separate osteoblast cultures. *, P < 0.05 versus results for MOI of 75:1 at 0 min; #, P < 0.05 versus results for MOI of 0 at time zero min; $, P < 0.05 versus results for MOI of 25:1 at 4 h.

To confirm that the changes in RANK-L mRNA result in corresponding protein expression, cell-associated RANK-L was measured by specific-capture ELISA. As shown in Fig. 2, there was a time- and dose-dependent response to S. aureus strain UAMS-1. At 24 h and at an MOI of 75:1, there was a 2.5-fold increase in the RANK-L protein compared to results with the control. Although the documented effects of RANK-L have been shown to be mediated by the membrane-bound form, we still measured the levels of soluble RANK-L as well and found its levels to be undetectable (data not shown). Osteoblasts infected with S. carnosus and UV-killed S. aureus strain UAMS-1 demonstrated the same levels of RANK-L protein expression as uninfected osteoblasts (Fig. 2). Additionally, osteoblasts infected with S. aureus nasopharyngeal isolates N3 and N4 induced RANK-L production at levels similar to those for strain UAMS-1 at 0 h. Strains N4 and N3 sustained this increased RANK-L expression following 6 and 24 h of incubation, respectively.

FIG. 2.

Expression of RANK-L protein by osteoblasts. Primary mouse osteoblasts were either uninfected (MOI 0) or infected with either S. aureus strain UAMS-1 (MOI = 25:1 or 75:1), UV-killed strain UAMS-1 (MOI = 25:1), S. carnosus (MOI = 25:1), or S. aureus nasopharyngeal isolate N1, N2, N3, N4, or N5 (MOI = 25:1). Following a 45-min incubation with or without bacteria and a 1-h treatment with 25 μg/ml gentamicin sulfate, whole-cell extracts were taken at time zero hours, 6 h, and 24 h. RANK-L concentrations were determined using an ELISA, and the results are displayed in pg/ml. Data are expressed as the means ± standard errors for three separate osteoblast cultures. *, P < 0.05 versus results with MOI of 0 at all time points; #, P < 0.05 versus results with MOI of 75:1 0 h and 6 h; $, P < 0.05 versus results with MOI of 25:1 at 6 h or 24 h.

Induction of PGE₂ expression by osteoblasts following exposure to S. aureus.

Choi et al. demonstrated that PGE₂ is a main mediator in RANK-L-dependent osteoclastogenesis induced by several periodontal pathogens (8). Therefore, it is possible that RANK-L production in osteoblasts infected with S. aureus involves a PGE_2-dependent mechanism. The current study examined levels of PGE₂ following infection of osteoblasts by S. aureus. Figure 3 demonstrates that PGE₂ is upregulated by S. aureus strain UAMS-1-infected osteoblasts at all time points examined. Although there is no significant difference between infection at an MOI of 25:1 and that at 75:1, there is a significant difference in PGE₂ expression between S. aureus strain UAMS-1-infected and uninfected osteoblasts. Figure 3 also shows that the increase in levels of PGE₂ occurs early following S. aureus infection and the levels remain elevated throughout the 24-h duration of the experiment. PGE₂ expression in response to UV-killed S. aureus strain UAMS-1- or S. carnosus-infected osteoblasts is not significantly different from that observed for uninfected osteoblasts (Fig. 3). Although S. aureus nasopharyngeal isolates N1 to N5 significantly increase PGE₂ expression at time zero, PGE₂ levels following 6 and 24 h of infection are comparable to those for uninfected osteoblasts.

FIG. 3.

Expression of PGE₂ by osteoblasts. Primary mouse osteoblasts were either uninfected (MOI = 0) or infected with either S. aureus strain UAMS-1 (MOI = 25:1 or 75:1), UV-killed strain UAMS-1 (MOI = 25:1), S. carnosus (MOI = 25:1,) or S. aureus nasopharyngeal isolate N1, N2, N3, N4, or N5 (MOI = 25:1). Following a 45-min incubation with or without bacteria and a 1-h treatment with 25 μg/ml gentamicin sulfate, supernatants were collected at time zero, 6 h, and 24 h. PGE₂ concentrations were determined by an EIA and are expressed in pg/ml. Data are expressed as the means ± standard errors for three separate osteoblast cultures. *, P < 0.05 versus results with an MOI of 0 at 0, 6 h, or 24 h.

PGE₂ production mediates S. aureus-induced increases in RANK-L production by osteoblasts.

One of the steps in the PGE₂ synthesis pathway involves the conversion of arachidonic acid to PGH₂, which is then metabolized to PGE₂, by COX. The addition of NS-398, a selective inhibitor of COX-2, inhibits IL-1-induced PGE₂ synthesis (7, 48). Therefore, in the current study, it is hypothesized that inhibition of COX-2 by NS-398 will result in decreased levels of PGE₂ production by S. aureus-infected osteoblasts. Figure 4 demonstrates that NS-398 significantly attenuates PGE₂ production in both uninfected and S. aureus-infected osteoblasts. Infected control cells treated with DMSO show the same dose response to S. aureus strain UAMS-1 seen in Fig. 3.

FIG. 4.

Expression of PGE₂ and the effects of NS-398. Primary mouse osteoblasts were cultured in the presence of 0.025% DMSO or 20 μM NS-398 in DMSO 48 h prior to infection. Osteoblasts were either uninfected (MOI = 0) or infected with S. aureus strain UAMS-1 (MOI = 25:1 or 75:1). Following a 45-min incubation with or without bacteria and a 1-h treatment with 25 μg/ml gentamicin sulfate, supernatants were collected at time zero, 6 h, and 24 h. PGE₂ concentrations were determined by an EIA and are expressed in ng/ml. Data are expressed as the means ± standard errors for three separate osteoblast cultures. *, P < 0.05 versus results with MOI of 0 and treatment with DMSO; #, P < 0.05 versus results for all groups treated with NS-398 at all time points.

Since it is known that PGE₂ upregulates RANK-L, we also examined the effects of NS-398 on RANK-L expression by S. aureus-infected osteoblasts. When infected osteoblasts are treated with NS-398, RANK-L levels are attenuated, indicating S. aureus-induced PGE₂ plays a role in expression of RANK-L by infected cells (Fig. 5).

FIG. 5.

Expression of RANK-L and the effects of NS-398. Primary mouse osteoblasts were cultured in the presence of 0.025% DMSO or 20 μM NS-398 in DMSO 48 h prior to infection. Osteoblasts were either uninfected (MOI = 0) or infected with S. aureus strain UAMS-1 (MOI = 25:1 or 75:1). Following a 45-min incubation with or without bacteria and a 1-h treatment with 25 μg/ml gentamicin sulfate, whole-cell extracts were collected at time zero, 6 h, and 24 h. RANK-L concentrations were determined by an ELISA and are expressed in pg/ml. Data are expressed as the means ± standard errors for three separate osteoblast cultures. *, P < 0.05 versus results with MOI of 0 and treatment with DMSO at 0 h, 6 h, or 24 h; #, P < 0.05 versus results for all groups treated with NS-398 at all time points; $, P < 0.05 versus MOI of 25:1 at 24 h.

DISCUSSION

Osteomyelitis is a disease associated with abnormal bone remodeling and massive bone resorption. The mechanisms responsible for such bone loss are largely unclear. The generation of proinflammatory cytokines and chemokines helps recruit immune cells, such as leukocytes, which could alter the balance between bone resorption and formation. Recent studies in our laboratory have demonstrated S. aureus-infected osteoblasts as a hitherto unrecognized source of soluble immune modulators both in vivo and in vitro (3-5, 14, 29, 30). Osteoblasts can also produce other soluble factors that can modulate the activity or formation of osteoclasts. RANK-L is one such key molecule that stimulates differentiation in osteoclasts and is an inducing agent involved in bone destruction (24, 25, 46, 49, 57). Following the binding of RANK-L to its receptor, RANK, on osteoclast precursors, the signal transduction can diverge into diverse pathways that can regulate distinct aspects of osteoclast functions (10, 18, 35, 39, 56).

Bacterial pathogens can stimulate osteoclastogenesis, either directly or indirectly, which can potentiate bone resorption, leading to bone destruction (31, 36). Ishida et al. have recently shown elevated RANK-L expression in S. aureus-infected NRG cells, a fibroblast cell line that has osteoblast-like features (19). Based on these findings, we undertook this study to evaluate the production of RANK-L from S. aureus-infected primary osteoblasts. Results outlined in Fig. 1 and 2 demonstrate that the viable S. aureus strain UAMS-1 is a potent inducer of membrane-bound, and not secreted, RANK-L mRNA and protein production by osteoblasts. In addition, UV-killed S. aureus strain UAMS-1 did not induce RANK-L expression, thus indicating that active bacterial gene expression may be required for maximal RANK-L production. Similarly, S. carnosus, an invasion-deficient species, did not stimulate increased RANK-L expression. Although osteoblasts infected with S. aureus nasopharyngeal isolates N3 and N4 induce RANK-L expression comparable to expression from osteoblasts infected with the osteomyelitis clinical isolate UAMS-1 at time zero and the strains maintain induction at 6 h in one case and at 24 h in the other, it is possible that strains N3 and N4 are capable of inducing osteomyelitis if they gain access to the bloodstream and subsequently to bone. The majority of the nasopharyngeal isolates are not capable of inducing RANK-L production. Taken together, these findings substantiate that live S. aureus is required for maximal RANK-L production from infected osteoblasts. An increase in RANK-L production by infected osteoblasts can further osteoclast differentiation and activation, which can potentiate the exacerbated bone destruction observed during osteomyelitis.

While RANK-L potentiates osteoclastogenesis, Osteoprotegerin (OPG), produced and secreted by osteoblasts, is a negative regulator of osteoclast formation. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) (43) is another key molecule produced by osteoblasts which can bind to OPG. Alexander et al. have demonstrated that when osteoblasts are infected with S. aureus, TRAIL expression is increased (2). Since OPG is known to bind both TRAIL and RANK-L, competition may limit the amount of OPG available to block the interaction between RANK-L and RANK during infection of osteoblasts by S. aureus. Taken together, it is highly likely that the increased levels of RANK-L contribute to the massive bone loss and aberrant bone remodeling associated with the pathology of staphylococcal osteomyelitis.

We next attempted to delineate the mechanism involved in upregulation of RANK-L expression in osteoblasts following exposure to S. aureus. PGE₂ is produced in bone mainly by osteoblasts and acts as a potent stimulator of bone resorption (37, 40, 46). It has been demonstrated that PGE₂ can mediate RANK-L-dependent osteoclastogenesis in periodontal diseases (8). Furthermore, the production of PGE₂ by osteoblasts is regulated by several cytokines and is also reported to induce osteoclast formation in mouse bone marrow cultures and stimulate bone resorption in mouse calvarial cultures (1, 7, 48). We therefore hypothesized that increased RANK-L expression in infected osteoblasts occurs as a result of PGE₂ activation upstream in the RANK-L pathway. The current study demonstrates that live S. aureus strain UAMS-1 induces significant PGE₂ production in infected osteoblasts (Fig. 3). However, UV-killed S. aureus strain UAMS-1 and S. carnosus do not induce PGE₂ production by osteoblasts. Osteoblasts infected with S. aureus nasopharyngeal isolates N1 to N5 demonstrate increased PGE₂ expression initially and decreased expression subsequently comparable to uninfected osteoblasts. Increased and sustained levels of PGE₂ production are induced only by live osteomyelitis clinical isolate UAMS-1, indicating the strain is capable of potentiating the bone damage observed in osteomyelitis. These results, along with the RANK-L data, are consistent with our previous findings that live S. aureus strain UAMS-1 is more effective at eliciting immune modulator production by cultured osteoblasts than UV-killed bacteria or nonpathogenic species (3-5, 11).

The functions of PGE₂ in target cells are mediated by four different G-protein-coupled receptor subtypes, EP1, EP2, EP3, and EP4 (28, 47). Among these PGE₂ receptor subtypes, EP4 is the main receptor expressed on osteoblasts and is responsible for mediating PGE₂-induced RANK-L expression by osteoblasts (13, 47, 52). In addition, PGE₂ synergistically promotes the differentiation of bone marrow macrophages into osteoclasts induced by RANK-L and macrophage colony-stimulating factor (51, 55). Thus, PGE₂ stimulates osteoclastic bone resorption through the following two different pathways: the induction of RANK-L expression in osteoblasts and the direct enhancement of RANK-L-induced differentiation of osteoclast precursor cells into osteoclasts. The mechanism of the synergistic effect of PGE₂ on the RANK-L-induced osteoclastic differentiation of precursor cells has not yet been explained (23).

PGE₂ synthesis is regulated by three metabolic steps: the release of arachidonic acid from the membranous phospholipids by phospholipase A₂, the conversion of arachidonic acid to PGH₂ by COX, and the synthesis of PGE₂ by PGE synthase (26, 34, 44). Both constitutive COX (COX-1) and inducible COX (COX-2) are expressed in mouse osteoblasts, with COX-2 being markedly induced by several bone-resorbing, inflammatory factors, including IL-1 and IL-6, in osteoblasts (32, 47). To establish the role of COX-2 with respect to production of PGE₂ during infection of osteoblasts with S. aureus, we used NS-398 to inhibit COX-2 and examined its effect on PGE₂ expression. When osteoblasts were treated with NS-398, PGE₂ levels were significantly attenuated in both uninfected and S. aureus-infected osteoblasts (Fig. 4). These results indicate that COX-2 is important in the regulation of PGE₂ production during S. aureus infection of osteoblasts.

We next examined the effects of NS-398 on the production of RANK-L. Although NS-398 significantly reduced the amount of RANK-L produced by S. aureus-infected osteoblasts, levels of RANK-L still remained detectable and were still responsive to S. aureus infection (Fig. 5). These data suggest that COX-2 and PGE₂ play a significant role in potentiating RANK-L production by S. aureus-infected osteoblasts.

The increasing incidence of bacterial infections caused by S. aureus and the emergence of antibiotic-resistant strains of this organism have made it imperative that we understand the mechanisms of initiation and maintenance of inflammation, as well as those of bone loss observed in sites of infection, in order to develop safe and effective therapies. Data presented here clearly outline a mechanism by which RANK-L, an essential factor for osteoclastogenesis, is upregulated by S. aureus strain UAMS-1-infected osteoblasts, further substantiating the role that osteoblasts may have in the aberrant bone remodeling observed in staphylococcal osteomyelitis. In addition, this upregulation of RANK-L appears to involve a COX-2-mediated PGE₂-dependent pathway. Potentially, agents that inhibit COX-2 or block RANK-L production could ameliorate bone loss. Further studies are warranted to investigate the in vivo significance of RANK-L in S. aureus-induced osteomyelitis.