INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Osteomyelitis (OM) is an infection of bone that results from hematogenous seeding, spread of infection from a contiguous area such as the skin adjacent to a wound, surgical inoculation of bacteria into bone, or trauma coincident with staphylococcal infection (57). The presence of an inert, prosthetic orthopedic device increases the likelihood of disease, and removal of the implant may be required (16). Chronic or recurring OM can be a persistent clinical problem that is difficult to treat effectively and results in abnormal bone remodeling, leading to a vascular compromise in the infected area. Long periods of antibiotic treatment are utilized in an attempt to control OM recurrences; however, methicillin-resistant staphylococci are now commonplace, and therapeutic levels of antibiotics in necrotic bone are difficult to achieve unless antibiotic-impregnated beads or implants are used (9).

Staphylococcus aureus is a capable bone pathogen because it possesses several cell surface adhesion molecules that facilitate its binding to the bone matrix. These include fibronectin-binding proteins (18, 30), fibrinogen-binding proteins (6, 10, 36), elastin-binding adhesin (42), collagen-binding adhesin (43), and a broad-specificity adhesin (MAP) which facilitates low-affinity binding of S. aureus to several proteins, including osteopontin, collagen, bone sialoprotein, fibronectin, fibrinogen, and vitronectin (37). In addition, S. aureus contains surface proteins that are able to stimulate bone resorption (39) via increasing osteoclast activity (4). The resultant bone destruction facilitates bacterial invasiveness. S. aureus not only colonizes bone matrix but is internalized in vitro (17, 28, 29) and in vivo (49) by osteoblasts (bone-forming cells). With the notable exception of Listeria monocytogenes, very little work has been done to examine mechanisms of invasion and intracellular survival by gram-positive bacteria. The ability of S. aureus to invade osteoblasts as well as several other cell types (3, 5, 38, 55) may be critical to the pathogenesis of the organism.

Upon binding and stimulation of many eukaryotic surface receptors, the mitogen-activated protein kinase (MAPK) pathway can be activated (45). The MAPKs extracellular signal-regulated protein kinases (ERK 1 [p44 MAPK] and ERK 2 [p42 MAPK]) have been found to be activated during the invasion of Henle-407 cells by L. monocytogenes (52) and Salmonella enterica serovar Typhimurium (41). MAPKs are important mediators in many cellular functions, including cytokine, mitogenic, and stress responses and cytoskeletal rearrangement (12, 58). As a result of the involvement of ERK 1 and 2 in other bacterial invasion systems, the role of ERK 1 and 2 in the invasion of both normal mouse and human osteoblasts by S. aureus strain UAMS-1 was examined. As a potential negative control, the ERK 1 and 2 response was examined following infection with Staphylococcus carnosus, since this bacterium was reported to be unable to invade the human osteoblastic cell line MG-63 (29). In addition to ERK 1 and 2, other isoforms of MAPK exist, such as p38 MAPK (hyperosmolarity [HOG] kinase) (25) and p54-p46 MAPK (c-Jun N-terminal kinase [JNK], also known as stress-activated protein kinase [SAPK]) (15, 32). SAPK and p38 MAPK are phosphorylated in response to extracellular signals, including proinflammatory cytokines and cellular stresses (11, 23, 26, 47). These isoforms have also been implicated in the invasion of Henle-407 cells by L. monocytogenes (53) and S. enterica serovar Typhimurium (27). It is also clear that the transcription factors ATF-2 (a target of p38 MAPK and SAPK) (24, 56) and c-Jun (a target of SAPK and ERK 1 and 2) (15, 32, 46) are stimulated by proinflammatory cytokines. These transcription factors have been shown to become activated during S. enterica serovar Typhimurium invasion (27). Since it is known that S. aureus induces an inflammatory response, it is possible that the severe inflammation observed with OM is mediated by members of the MAPK family and the above-mentioned transcription factors. It has recently been demonstrated that p38 MAPK activation leads to induction of expression of the proinflammatory cytokines interleukin 12 (IL-12) p40 (2) and IL-6 (13) and that IL-6 production is induced via ATF-2 activation in osteoblasts (19). It is also known that IL-6 induces bone remodeling via osteoclastogenesis (14). Additionally, tumor necrosis factor alpha (TNF-) mediates bone resorption (51), and SAPK is subsequently activated via phosphorylation when cells are exposed to TNF- (26). The activation of SAPK and c-Jun may be the mechanism responsible for the increased TNF- activity and mRNA transcript levels observed in an experimental model of S. aureus-induced OM (35).

In the present study, the role of intracellular signaling mechanisms in the invasion of normal mouse and human osteoblasts by S. aureus strain UAMS-1, an OM clinical isolate, was examined. S. aureus induced a time-dependent and dose-dependent activation of several members of the MAPK family, including ERK 1 and 2 and SAPK, but not p38 MAPK, upon association with normal mouse and human osteoblasts. In addition, S. carnosus infection of normal mouse osteoblasts induced this same ERK 1 and 2 response, albeit at a lower level than with S. aureus. Active bacterial gene expression is required for full stimulation of ERK 1 and 2. Attachment of the bacteria to the osteoblast surface results in ERK 1 and 2 phosphorylation, and the kinetics of ERK 1 and 2 activation correlates with the time course of invasion. Finally, invasion of normal mouse osteoblasts by S. aureus resulted in activation of c-Jun but not ATF-2 or ELK-1. These studies are the first to examine intracellular signaling in normal osteoblasts in response to S. aureus infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. S. aureus strain UAMS-1 (ATCC 49230) is a human OM clinical isolate (21). S. carnosus (ATCC 51365) is a nonpathogenic species reported to be incapable of invasion of osteoblasts (29).

Isolation and culture of normal mouse osteoblasts. Osteoblasts were isolated from the calvariae of 1-day old BALB/c mice according to a method described elsewhere for chicken embryos (48). Bone-forming cells were isolated from mouse neonate calvariae by sequential collagenase-protease digestion. The periostea were removed, and the frontal bones were harvested free of the suture regions and incubated for 10 min at 37°C in 10 ml of digestion medium containing collagenase (375 U/ml; type VII; Sigma, St. Louis, Mo.) and protease (7.5 U/ml; Sigma). The digestion medium and released cells were removed and discarded. Ten milliliters of fresh digestion medium was then added, and the incubation was continued for 20 min. The cells were harvested by centrifugation and rinsed three times in 25 mM HEPES-buffered Hanks' balanced salt solution (HBSS) (pH 7.4). The digestion step was repeated twice, and the three cell isolates were pooled in mouse osteoblast growth medium (MOBGM) consisting of Dulbecco's modified Eagle medium containing 25 mM HEPES, 2 g of sodium bicarbonate/liter, 75 µg of glycine/ml, 10% fetal bovine serum (Sigma), 100 µg of ascorbic acid/ml, 40 ng of vitamin B₁₂/ml, 2 µg of p-aminobenzoic acid/ml, 200 ng of biotin/ml, and penicillin-streptomycin-Fungizone (Abx-Amx) (100 U/ml-100 µg/ml-0.25 µg/ml; Sigma) (pH 7.4). The cells were seeded at 10⁶/well in six-well plates and incubated in a humidified incubator at 37°C in a 5% CO₂ atmosphere until they reached confluence (6 to 7 days). The medium was changed every 48 h after being seeded.

Normal human osteoblast cultures. Normal human osteoblasts (Clonetics, San Diego, Calif.) were purchased and propagated according to the guidelines provided by the vendor. The cells were seeded in 25-cm² flasks and incubated in a humidified incubator at 37°C in a 5% CO₂ atmosphere in growth medium supplied by the manufacturer that contained 10% fetal calf serum, ascorbic acid, and gentamicin. After the cells reached 80% confluence (5 to 9 days), they were removed from the flasks by use of 0.025% trypsin-0.01% EDTA, washed in growth medium, and seeded in six-well plates. The osteoblasts were then maintained in growth medium and grown as described above until they reached confluence (6 to 7 days). The growth medium was changed every 48 h after being seeded. These commercially available cells have been extensively characterized as osteoblasts (20, 22).

Invasion assay to enumerate intracellular bacteria. S. aureus and S. carnosus 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 in 5 ml of HBSS. The pellets were then resuspended in 5 ml of MOBGM lacking Abx-Amx. Confluent cell layers of osteoblasts were washed three times with 4 ml of HBSS to remove the growth medium. The cultures were then infected at a multiplicity of infection (MOI) of 75 or 250:1 with either S. aureus or S. carnosus in 4 ml of MOBGM lacking Abx-Amx. Following an appropriate infection period, the cell cultures were washed three times with 4 ml pf HBSS and incubated for 3 h in 4 ml of MOBGM containing 25 µg of gentamicin/ml to kill the remaining extracellular S. aureus or S. carnosus cells. The osteoblast cultures were washed as described above and subsequently lysed by the addition of 1.2 ml of 0.1% Triton X-100 (Fisher Biotech, Fair Lawn, N.J.) with incubation for 5 min at 37°C. To quantify the number of bacteria internalized, suspension dilutions of the lysates were plated in triplicate on tryptic soy agar plates followed by incubation at 37°C overnight.

Preparation of osteoblast lysates. Mouse or human osteoblasts were infected with live or UV-killed S. aureus strain UAMS-1 or S. carnosus at an MOI of 25, 75, or 250:1 for the live bacteria and 75 or 250:1 for the UV-killed organisms. Following an infection period, the osteoblasts were washed three times with 4 ml of ice-cold phosphate-buffered saline, pH 7.4, containing 0.4 mM Na₃VO₄ (Sigma), 1 mM NaF, and 0.1 mg of phenylmethylsulfonyl fluoride (Sigma)/ml and removed from the culture plates with a cell scraper (Costar). The cells were then concentrated by centrifugation at 10,000 × g for 1 min. The cell pellet was resuspended in 100 µl of lysis buffer (50 mM Tris-HCl [pH 7.6] containing 0.4 mM Na₃VO₄, 1 mM NaF, 0.1 mg of PMSF/ml, 0.01 mg of leupeptin [Sigma]/ml, and 1.0% Triton X-100). The resulting lysate was then subjected to centrifugation at 13,600 × g for 30 min. Equivalent amounts of protein from the Triton X-100-soluble fraction (10 to 20 µg) were mixed with concentrated 5× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer (0.313 M Tris-HCl [pH 6.8], 10% [vol/vol] SDS, 25% [vol/vol] 2-mercaptoethanol, 50% [vol/vol] glycerol, 0.01% [wt/vol] bromophenol blue), boiled for 5 min, cleared by centrifugation at 13,600 × g for 5 min, and subjected to SDS-PAGE analysis.

SDS-PAGE and immunoblot analyses. Osteoblast proteins were separated by SDS-PAGE as described by Laemmli (33). The proteins were then electrophoretically transferred to polyvinylidene difluoride membranes (Fisher Biotech) using the Mini Trans-Blot apparatus (Bio-Rad, Hercules, Calif.) at 300 mA for 1 h at 4°C according to the manufacturer's recommendations. The membranes were blocked with buffer A (Tris-buffered saline [Bio-Rad] containing 5.0% skim milk and 0.5% Tween 20 [Sigma]) for 1 h with gentle shaking and then washed with buffer B (Tris-buffered saline containing 0.1% Tween 20) three times for 5 min each time. A two-step detection method was used to identify phosphorylated proteins of interest. Blots were first incubated overnight with gentle shaking at 4°C with mouse anti-phospho-ERK 1 and 2 (p44-p42 MAPK) (New England Biolabs [NEB], Beverly, Mass.) diluted 1:5,000 in buffer A or with either rabbit anti-phospho-p38 MAPK (NEB), rabbit anti-phospho-MKK3-MKK6 (NEB), rabbit anti-phospho-SAPK (NEB), rabbit anti-phospho-c-Jun (NEB), rabbit anti-phospho-Elk-1, or rabbit anti-phospho-ATF-2 (NEB) diluted 1:1,000 in buffer A. The membranes were then washed three times the following day for 5 min each time in buffer B with gentle shaking, followed by incubation for 1 h with gentle shaking at room temperature in buffer A with either anti-mouse or anti-rabbit immunoglobulin G (1:5,000) conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, Pa.). Reactive proteins were then visualized by enhanced chemiluminescence (Amersham, Arlington Heights, Ill.) following exposure to X-ray film and subsequent film development.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

S. aureus infection results in phosphorylation of the mouse osteoblast proteins ERK 1 and 2. Osteoblast cultures infected with S. aureus at an MOI of 250:1 demonstrated a marked increase in ERK 1 and 2 phosphorylation over time (Fig. 1A). Immunoblots incubated with anti-phospho-ERK 1 and 2 antibodies demonstrated that ERK 1 and 2 are phosphorylated as early as 5 min following infection and are maximal at 30 min (Fig. 1A). Studies examining the kinetics of S. aureus invasion of normal mouse osteoblasts indicate that bacterial invasion correlates with observed ERK 1 and 2 phosphorylation (Fig. 1B). The numbers of intracellular S. aureus cells increased logarithmically up to approximately 30 min following infection.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   ERK 1 and 2 phosphorylation correlates with the kinetics of S. aureus invasion. (A) ERK 1 and 2 phosphorylation is induced by S. aureus infection. Normal mouse osteoblasts were either uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 250:1 for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the right show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by S. aureus infection. (B) Rate of osteoblast invasion by S. aureus. Normal mouse osteoblasts were infected with S. aureus strain UAMS-1 at an MOI of 250:1 for various times. The wells were then washed three times, followed by the addition of medium supplemented with gentamicin for 3 h to kill extracellular bacteria. Following a 3-h incubation, the wells were washed three times and the osteoblasts were lysed to enumerate intracellular S. aureus cells. The standard deviations are indicated. Both the exposed film and enumeration data are representative of results from three independent experiments.

Cytochalasin D does not affect ERK 1 and 2 phosphorylation; PD98059 reduces ERK 1 and 2 phosphorylation. Experiments have demonstrated that inhibition of osteoblast actin polymerization with 5 µg of cytochalasin D/ml decreases intracellular S. aureus numbers by as much as 99.8% compared to control cultures (17). To examine whether invasion or attachment of S. aureus is the stimulus for ERK 1 and 2 phosphorylation, normal mouse osteoblast cultures were treated with cytochalasin D (5 µg/ml), and the ERK 1 and 2 phosphorylation levels were determined. The increase in ERK 1 and 2 phosphorylation observed in response to osteoblast infection by S. aureus was unaffected by treatment of cultures with cytochalasin D (Fig. 2). The data therefore suggest that ERK 1 and 2 phosphorylation is induced by S. aureus attachment to the osteoblast surface.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   S. aureus attachment is sufficient to induce phosphorylation of ERK 1 and 2. Normal mouse osteoblasts were treated with cytochalasin D (5 µg/ml) or with the solvent dimethyl sulfoxide (Control) for 30 min. The osteoblasts were then either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 250:1 for 10 or 30 min. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the right show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

View larger version (41K): [in this window] [in a new window]   FIG. 3.   ERK 1 and 2 is not phosphorylated when osteoblasts are treated with the MEK 1 and 2 inhibitor PD98059 followed by infection with S. aureus. Normal mouse osteoblasts were treated with PD98059 (100 µM) or with the solvent dimethyl sulfoxide (Control) for 60 min. The osteoblasts were then either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 250:1 for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the right show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

ERK 1 and 2 phosphorylation is dependent on S. aureus MOI. Maximal ERK 1 and 2 phosphorylation was observed at 30 min following infection in response to an MOI of 75:1 (Fig. 4A). There was 25-fold less ERK 1 and 2 phosphorylation at the same time point when osteoblasts were infected at an MOI of 25:1 (Fig. 4B). Therefore, all remaining studies examined phosphorylation responses at an MOI of 75:1 at 30 min following infection.

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Dose-dependent S. aureus induction of ERK 1 and 2 phosphorylation. (A) Normal mouse osteoblasts were either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 75:1 for various times. (B) Normal mouse osteoblasts were infected with live S. aureus strain UAMS-1 at an MOI of 25:1 or with UV-killed S. aureus strain UAMS-1 at an MOI of 250 or 75:1 for 30 min. Equivalent volumes of Triton X-100-soluble protein samples were resolved on a single SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the left show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

Interaction of mouse osteoblasts with UV-killed S. aureus results in reduced ERK 1 and 2 phosphorylation compared to infection with viable bacteria. UV treatment of S. aureus was used to kill the bacteria while maintaining the integrity of surface structures, presumably including those that bind and interact with the osteoblast surface. When UV-killed S. aureus cells were incubated with normal mouse osteoblast cultures for 30 min (MOI, 75 and 250:1), ERK 1 and 2 phosphorylation was observed, albeit at lower levels than for live S. aureus cells (Fig. 4B). Viable S. aureus cells at an MOI of 75:1 induced eightfold-greater ERK 1 and 2 phosphorylation than UV-killed S. aureus at the same MOI (Fig. 4).

Infection of mouse osteoblasts by S. carnosus results in lower levels of ERK 1 and 2 phosphorylation than S. aureus infection. It has been reported that S. carnosus is incapable of invasion of the human osteoblast cell line MG-63 (29); however, we report that S. carnosus is capable of invasion of normal osteoblasts, albeit at a significantly lower level than S. aureus (Fig. 5A). The differential ability of S. carnosus and S. aureus to invade osteoblasts is also manifested in the ability to induce ERK 1 and 2 phosphorylation. Osteoblasts infected with S. carnosus demonstrated an increase in ERK 1 and 2 phosphorylation over time at the same MOI of 75:1 used in the S. aureus studies; however, the amount of phosphorylated ERK 1 and 2 induced by S. carnosus was threefold less than that induced by S. aureus (Fig. 5B).

View larger version (16K): [in this window] [in a new window]   FIG. 5.   Differences in S. aureus and S. carnosus invasion of osteoblasts. (A) S. carnosus is impaired in invasion of osteoblasts compared to S. aureus. Normal mouse osteoblasts were infected at an MOI of 75:1 with either S. aureus or S. carnosus for 45 min. The wells were then washed three times followed by the addition of growth medium containing gentamicin for 3 h. The wells were then washed, and the osteoblasts were lysed to enumerate intracellular bacteria. The standard deviations are indicated. (B) ERK 1 and 2 phosphorylation in response to S. aureus or S. carnosus infection. Normal mouse osteoblasts were either left uninfected (0 min) or infected with S. aureus strain UAMS-1 or S. carnosus, each at an MOI of 75:1, for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the right show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by S. aureus and S. carnosus. The exposed film is representative of results from three independent experiments.

S. aureus infection results in phosphorylation of the mouse osteoblast protein SAPK but not p38 MAPK. Two major stress-activated pathways in eukaryotic cells are SAPK and p38 MAPK (15, 25). Thus, it was hypothesized that one or both of these pathways would be activated during S. aureus infection of osteoblasts. Normal mouse osteoblast cultures infected with S. aureus at an MOI of 75:1 exhibited a marked increase in phosphorylation of the SAPK isoforms p54 and p46 (Fig. 6). Like ERK 1 and 2 phosphorylation observed in response to S. aureus infection, SAPK phosphorylation was maximal at 30 min following infection.

View larger version (35K): [in this window] [in a new window]   FIG. 6.   SAPK phosphorylation in response to S. aureus infection. Normal mouse osteoblasts were either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 75:1 for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for SAPK phosphorylation by immunoblotting and reaction with anti-phospho-SAPK antibody. The arrows on the right show the positions of the 54- and 46-kDa phosphorylated SAPK isoforms induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

S. aureus infection results in phosphorylation of the mouse osteoblast protein c-Jun. The transcription factor c-Jun is a known target of the SAPK pathway (15, 32). As shown in Fig. 7, there is a rapid and marked increase in c-Jun phosphorylation in normal osteoblasts in response to S. aureus infection. As early as 10 min postinfection, c-Jun phosphorylation is maximal, and it returns to near control levels by 45 min following infection. In contrast to c-Jun, the transcription factors ATF-2 and Elk-1 are not phosphorylated in response to S. aureus (data not shown). The c-Jun data suggest a potential mechanism whereby infection of osteoblasts with S. aureus leads to specific activation of host cell transcription.

View larger version (32K): [in this window] [in a new window]   FIG. 7.   c-Jun phosphorylation in response to S. aureus infection. Normal mouse osteoblasts were either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 75:1 for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for c-Jun phosphorylation by immunoblotting and reaction with anti-phospho-c-Jun antibody. The arrow on the right shows the position of the 42-kDa phosphorylated protein induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

S. aureus infection results in phosphorylation of the human osteoblast proteins ERK 1 and 2 and SAPK. To extend studies using normal mouse osteoblasts to human cells, immunoblot analysis was performed on cell lysates of normal human osteoblasts. S. aureus-induced phosphorylation responses in normal mouse osteoblasts were also evident in normal human osteoblasts. Normal human osteoblasts infected with S. aureus for 30 min demonstrated dose-dependent ERK 1 and 2 phosphorylation, with maximal activation at an MOI of 75:1 (Fig. 8). ERK 1 and 2 phosphorylation increased threefold at an MOI of 75:1 compared to that at 25:1. Interestingly, the amount of phosphorylated ERK 1 and 2 then actually decreased twofold at an MOI of 250:1 compared to levels observed at 75:1. As for mouse osteoblasts, UV-killed S. aureus induced significantly lower levels of ERK 1 and 2 phosphorylation (Fig. 8). The amount of phosphorylated ERK 1 and 2 was threefold higher with live bacteria than with UV-killed organisms at an MOI of 75:1. SAPK phosphorylation in human osteoblast cultures infected with S. aureus also mirrored the results observed with the mouse osteoblasts (Fig. 9). The SAPK response in human cells is also maximal at 30 min following bacterial infection.

View larger version (44K): [in this window] [in a new window]   FIG. 8.   Dose-dependent S. aureus induction of ERK 1 and 2 phosphorylation in normal human osteoblasts. Normal human osteoblasts were either left uninfected (MOI, 0) or infected with live S. aureus at an MOI of 250, 75, or 25:1 or with UV-killed S. aureus at an MOI of 250 or 75:1 for 30 min. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for ERK 1 and 2 phosphorylation by immunoblotting and reaction with anti-phospho-ERK 1 and 2 antibody. The arrows on the right show the positions of the 44- and 42-kDa phosphorylated ERK isoforms induced by live and UV-killed S. aureus. The exposed film is representative of results from three independent experiments.

View larger version (42K): [in this window] [in a new window]   FIG. 9.   SAPK phosphorylation in response to S. aureus infection of normal human osteoblasts. Normal human osteoblasts were either left uninfected (0 min) or infected with S. aureus strain UAMS-1 at an MOI of 75:1 for various times. Equivalent volumes of Triton X-100-soluble protein samples were resolved on an SDS-PAGE gel (12%), transferred to a polyvinylidene difluoride membrane, and analyzed for SAPK phosphorylation by immunoblotting and reaction with anti-phospho-SAPK antibody. The arrows on the right show the positions of the 54- and 46-kDa phosphorylated SAPK isoforms induced by S. aureus infection. The exposed film is representative of results from three independent experiments.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

When normal mouse osteoblasts were infected with S. aureus strain UAMS-1, phosphorylation of ERK 1 and 2 was increased, with the greatest activation at 30 min following infection. The maximal numbers of intracellular S. aureus cells correlated well with this maximal ERK 1 and 2 phosphorylation. These results clearly indicate that activation of ERK 1 and 2 occurs during invasion of osteoblasts by S. aureus. Activated ERK 1 and 2 probably then effects cytoskeletal rearrangement by inducing actin reorganization. Invasion of epithelial cells by L. monocytogenes correlates with greater phosphorylation of the p44 isoform (ERK-1) than the p42 isoform (ERK-2) (52); however, the situation is different with S. aureus invasion of osteoblasts. Infection of both normal mouse and human osteoblasts by S. aureus resulted in greater ERK-2 than ERK-1 phosphorylation. MAPKs are a group of threonine and serine kinases, and they have many diverse functions within a cell. ERK 1 and 2 are regulated by phosphorylation of a tyrosine and a threonine residue, and both sites must be phosphorylated for maximal activity (44). The ERK 1 and 2 signaling pathway is utilized by a variety of growth and differentiation factors (8, 31, 40), and activation of ERK 1 and 2 can result in phosphorylation of many different substrates. These include the transcription factors ATF-2, Elk-1 (1), and c-Jun (46). ERK 1 and 2 phosphorylation can also activate phospholipase A₂ (34). Phospholipase A₂ activation results in the production of leukotrienes after arachidonic acid release (50). Leukotrienes, which are hypothesized to open calcium channels on the host cell membrane, are important in the invasion of Henle-407 cells by S. enterica serovar Typhimurium (41). We have recently demonstrated that calcium channels are also important in the invasion of normal mouse osteoblasts by S. aureus (Ellington and Hudson, unpublished).

ERK 1 and 2 phosphorylation in response to different MOIs was examined. When the MOI increased from 25:1 to 75:1, ERK 1 and 2 phosphorylation increased 25- and 3-fold in mouse and human osteoblasts, respectively; however, at an MOI of 250:1, ERK 1 and 2 phosphorylation actually decreased twofold in human cells compared to that at an MOI of 75:1. This reduced ERK 1 and 2 phosphorylation observed at the highest MOI used may be attributed to cell death via necrosis and/or apoptosis. We have recently demonstrated that S. aureus does induce osteoblast apoptosis (54). When osteoblasts were pretreated with the MEK 1 and 2 inhibitor PD98059, ERK 1 and 2 phosphorylation was undetectable in infected osteoblast cultures. This correlates with data from our laboratory demonstrating that following inhibition of MEK 1 and 2 with PD98059, the numbers of intracellular S. aureus cells decreased by as much as 80.5% (Ellington and Hudson, unpublished). These findings indicate that the activation of ERK 1 and 2 is a significant event in S. aureus invasion of osteoblasts.

To examine whether bacterial attachment was responsible for activation of ERK 1 and 2, osteoblast cultures were pretreated with cytochalasin D, which inhibits S. aureus invasion of osteoblasts (17). The data demonstrate that ERK 1 and 2 phosphorylation is still induced when osteoblasts are treated with cytochalasin D, indicating that ERK 1 and 2 activation occurs prior to bacterial invasion. In addition, UV-killed S. aureus cells were able to induce ERK 1 and 2 phosphorylation in both normal mouse and human osteoblasts, albeit at lower levels than with live S. aureus. This indicates that active bacterial gene expression may be required for maximal ERK 1 and 2 activation and further suggests that interaction of a bacterial ligand with an osteoblast receptor results in an increase in ERK 1 and 2 phosphorylation. Alternatively, data from experiments with killed bacteria could suggest that active secretion of soluble factors by bacteria plays a role in osteoblast responses. Experiments are currently in progress to examine whether secretion of soluble factors by S. aureus results in ERK 1 and 2 activation.

S. carnosus was utilized to examine how osteoblasts respond to a presumably noninvasive species of staphylococcus (29). Live and UV-killed S. carnosus did induce ERK 1 and 2 phosphorylation in normal mouse osteoblasts, albeit at significantly lower levels than S. aureus. Viable intracellular S. carnosus cells were also recovered from osteoblasts, but the numbers of intracellular bacteria were also significantly lower than the numbers of intracellular S. aureus cells. Differences in numbers of viable intracellular bacteria between species appear to be much greater than the differences in the levels of ERK 1 and 2 phosphorylation observed. These data may indicate that S. carnosus is debilitated in the ability to survive in the osteoblast intracellular environment compared to S. aureus.

In addition to the phosphorylation of ERK 1 and 2 observed during challenge with S. aureus, the phosphorylation of SAPK was also increased in both normal mouse and human osteoblasts; however, another stress pathway including p38 MAPK was not activated in response to infection by S. aureus. These data indicate that osteoblasts respond to S. aureus infection by activation of the SAPK pathway but not the p38 MAPK pathway.

To examine downstream activity in normal mouse osteoblasts challenged with S. aureus, possible transcription factors that are targets of ERK 1 and 2 and SAPK were examined. During osteoblast infection by S. aureus, a rapid increase in c-Jun phosphorylation was observed. The kinetics of c-Jun activation could indicate that the SAPK pathway (15, 32) and the ERK 1 and 2 pathway (46) both converge on c-Jun, resulting in its rapid phosphorylation in response to S. aureus infection; however, it is possible that another factor is responsible for c-Jun activation. In any case, phosphorylation of c-Jun probably then affects osteoblast gene expression, possibly inducing proinflammatory cytokine expression. Such cytokine expression has been observed in response to S. aureus invasion in both normal mouse and human osteoblasts (7).

The overall mechanism of S. aureus invasion of osteoblasts clearly involves a variety of converging signal transduction pathways. S. aureus invasion of osteoblasts is a complex process, in which the bacterium exploits the host cell machinery and escapes the extracellular environment and humoral immune response. Further investigation of eukaryotic cellular processes that become activated in response to S. aureus may allow modified treatment strategies for bone infection and possibly other S. aureus-induced diseases.