Regulation of Osteoblast Differentiation by Pasteurella Multocida Toxin (PMT): A Role for Rho GTPase in Bone Formation

Authors


  • This work was presented in part in abstract form at the 23rd Annual Meeting of the American Society for Bone and Mineral Research, Phoenix, Arizona, USA, October 12-16, 2001.

    The authors have no conflict of interest.

Abstract

The role of the Rho-Rho kinase signaling pathway on osteoblast differentiation was investigated using primary mouse calvarial cells. The bacterial toxin PMT inhibited, whereas Rho-ROK inhibitors stimulated, osteoblast differentiation and bone nodule formation. These effects correlated with altered BMP-2 and −4 expression. These data show the importance of Rho-ROK signaling in osteoblast differentiation and bone formation.

Introduction: The signal transduction pathways controlling osteoblast differentiation are not well understood. In this study, we used Pasteurella multocida toxin (PMT), a unique bacterial toxin that activates the small GTPase Rho, and specific Rho inhibitors to investigate the role of Rho in osteoblast differentiation and bone formation in vitro.

Materials and Methods: Primary mouse calvarial osteoblast cultures were used to investigate the effects of recombinant PMT and Rho-Rho kinase (ROK) inhibitors on osteoblast differentiation and bone nodule formation. Osteoblast gene expression was analyzed using Northern blot and RT-PCR, and actin rearrangements were visualized after phalloidin staining and confocal microscopy.

Results: PMT stimulated the proliferation of primary mouse calvarial cells and markedly inhibited the differentiation of osteoblast precursors to bone nodules with a concomitant inhibition of osteoblastic marker gene expression. There was no apparent causal relationship between the stimulation of proliferation and inhibition of differentiation. PMT caused cytoskeletal rearrangements because of activation of Rho, and the inhibition of bone nodules was completely reversed by the Rho inhibitor C3 transferase and partly reversed by inhibitors of the Rho effector, ROK. Interestingly, Rho and ROK inhibitors alone potently stimulated osteoblast differentiation, gene expression, and bone nodule formation. Finally, PMT inhibited, whereas ROK inhibitors stimulated, bone morphogenetic protein (BMP)-2 and −4 mRNA expression, providing a possible mechanism for their effects on bone nodule formation.

Conclusions: These results show that PMT inhibits osteoblast differentiation through a mechanism involving the Rho-ROK pathway and that this pathway is an important negative regulator of osteoblast differentiation. Conversely, ROK inhibitors stimulate osteoblast differentiation and may be potentially useful as anabolic agents for bone.

INTRODUCTION

THE FORMATION OF bone by osteoblasts is a complex process involving the commitment of mesenchymal stem cells to the osteoblastic cell lineage and the proliferation and differentiation of these committed precursors to mature, bone-forming osteoblasts.(1, 2) Many factors regulate osteogenesis, from systemic hormones to peptide growth factors, cytokines, and locally acting auto/paracrine factors.(3) In addition, the functional roles of transcription factors, such as Cbfa1/Runx2, Osterix, and activator protein (AP)-1, have demonstrated that the differentiation, activity, and gene expression of osteogenic cells are also controlled at the nuclear level.(4, 5) However, the various intracellular signal transduction pathways that link growth and differentiation stimuli to the activation of nuclear transcription factors in osteoblasts are only beginning to be understood.

Osteogenic cells play a key role in bone metabolism that extends beyond synthesizing bone, such as controlling the activity of bone-resorbing osteoclasts. Bone formation and resorption are tightly coupled, and any disruption of this coupling results in bone remodeling disorders such as osteoporosis and osteopetrosis.(3) Abnormal bone remodeling is also seen in atrophic rhinitis, a disease characterized by excessive bone resorption and the progressive loss of nasal turbinate bones in pigs.(6, 7) The principle pathogenic agent of atrophic rhinitis is a bacterial protein toxin called Pasteurella multocida toxin (PMT),(7) although previous studies investigating whether PMT specifically affects cells of the osteoblast or osteoclast lineages have proved inconclusive. Some studies using various in vivo and in vitro systems from different species have indicated that PMT can stimulate osteoclastic bone resorption. However, others have indicated a decrease in osteoclast numbers.(8-11) Furthermore, co-culture experiments using chick cells have suggested that the effects of PMT on osteoclasts may be mediated by osteoblastic cells.(12) PMT has also been shown to stimulate DNA synthesis in some, but not all osteoblastic, cell lines and to reduce the levels of alkaline phosphatase activity and type I collagen synthesis.(13, 14) Moreover, unpurified preparations of PMT have also been suggested to inhibit the differentiation of porcine bone marrow stromal cells, although cytotoxicity and the presence of additional bacteria-derived factors could not be ruled out.(15) Thus, the specific effects of PMT on different bone cell populations are not clear.

It is well established that many bacterial protein toxins are potent biological agents that interfere with key eukaryotic cell signaling processes by chemically modifying specific molecular targets and altering their activity.(16) PMT is the most potent mitogen identified for fibroblasts, stimulating DNA synthesis at low picomolar concentrations.(17) It is a large, 146-kDa protein that is known to act intracellularly, binding to cell-surface ganglioside-type receptors and becoming internalized.(17-19) Several intracellular signal transduction pathways are affected by PMT, although the primary molecular target(s) of PMT are not yet known. There is clear evidence that the toxin stimulates signaling pathways linked to the Gq/11 family of heterotrimeric G proteins,(20-22) leading to activation of phospholipase Cβ and increased inositol phosphate levels, stimulation of protein kinase C activity, mobilization of intracellular calcium, and activation of the Ras/Raf/MAP kinase pathway.(20, 23-25) Treatment of fibroblasts with PMT also stimulates signaling pathways linked to the cytoskeleton, causing the formation of actin stress fibers, assembly of focal contacts, and tyrosine phosphorylation of p125FAK and paxillin.(26) This is mediated by the small GTPase Rho and one of its downstream effectors, the Rho-associated protein kinase, p160ROK (ROK), because cytoskeletal rearrangements are abolished in the presence of Rho inhibitors, such as Clostridium botulinum C3 transferase, and specific ROK inhibitors.(26, 27) PMT also exhibits some homology to the N terminus of the E. coli toxin, cytotoxic necrotizing factor (CNF), which is known to modify the Rho family proteins, Rho, Rac, and cdc42, leading to their constitutive activation.(28)

Because all toxins have a very limited and specific repertoire of targets, they have emerged as unique and extremely valuable tools for dissecting different signal transduction pathways. Despite our knowledge of the intracellular events activated in fibroblasts, the specific bone cell populations and the signal transduction pathways affected by PMT that trigger its dramatic effects on bone homeostasis have not been fully investigated. Here we have focused on the osteoblast lineage by dissecting the effects of recombinant PMT on osteoprogenitor cell differentiation and gene expression and have identified that the Rho-ROK signal transduction pathway is one major regulator of the PMT effects on osteoblast differentiation and bone formation in vitro.

MATERIALS AND METHODS

Cell culture, proliferation, and differentiation assays

Primary cultures of osteogenic cells were obtained from neonatal CD1 mouse calvariae (Charles River, Margate, UK), using a modification of the sequential enzymatic digestion protocol as described previously.(29) Calvarial cells were maintained in α-MEM (+ribonucleosides/deoxyribonucleosides; Sigma-Aldrich, Poole, UK), whereas Swiss 3T3 fibroblasts were cultured in DMEM (Gibco BRL, Paisley, UK), all containing 10% heat-inactivated FBS (M.B. Meldrum Ltd., Hants, UK), 1% L-glutamine, and antibiotics (Gibco; 50 U/ml penicillin, 50 μg/ml streptomycin; standard medium), and incubated at 37°C in a humidified atmosphere of 5% CO2 in air. All studies involving animals were performed under the strict guidelines of the UK Home Office for the use of animals in research.

Recombinant PMT was expressed and prepared as described previously.(30) The inactive mutant PMT, C1165S, containing a C-terminal cysteine to serine exchange, displays identical biochemical properties to the wildtype recombinant protein but has no biological activity in vivo or in vitro.(30) For growth curves, cells were plated in 48-well plates (Nunc, Roskilde, Denmark) at 4 × 103 cells/cm2 in standard medium, treated 24 h later with the appropriate concentration of PMT or mutant C1165S as indicated, and counted at the indicated times using a hemocytometer. Bone nodule differentiation assays were performed as described previously(29) in medium supplemented with 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate (Sigma; differentiation medium), in the continuous presence or absence of PMT, mutant C1165S, and the Rho inhibitor C3 transferase (purchased from Calbiochem, Nottingham, UK, or expressed as a glutathione S-transferase fusion protein in E. coli and purified on glutathione-agarose beads(31)) or the ROK inhibitors HA-1077 or Y-27632 (Calbiochem), as indicated. Cultures were fixed after 15 days with 4% paraformaldehyde/PBS and stained for alkaline phosphatase activity and mineral using the von Kossa technique as previously described.(29) Nodules were quantified either by counting the number of individual nodules per well using a dissecting microscope or the point-counting method if nodules had coalesced together, expressed as the percentage surface area mineralized.

For time-lapse low-light microscopy of developing bone nodules, early nodule condensations (approximately days 8-9) were cultured in medium supplemented with 15 mM HEPES, pH 7.4, in a humidified chamber at 37°C in the presence or absence of 20 ng/ml PMT. Images were acquired every 10 minutes with a Nikon Diaphot 200 inverted microscope equipped with a Nikon 10× objective, using a Hamamatsu C4742-95 Orca1 cooled charge-coupled device camera (Hamamatsu Photonic Systems, Hamamatsu, Japan), controlled by Kinetic AQM 2000 software (Kinetic Imaging Ltd., Liverpool, UK). Imaging was started 30 minutes after the addition of PMT and was continued for a maximum of 96 h.

Northern blot and RT-PCR analyses

Primary osteogenic cells cultured in the presence or absence of PMT were harvested after 48 h, and Northern blot analyses were carried out using either poly(A)+ RNA(32) or total RNA (RNeasy; Qiagen Ltd., Crawley, UK) as specified. For Northern blotting, the following probes were used: Cbfa1/Runx2 (Dr LJ Suva, University of Arkansas), alkaline phosphatase, osteocalcin, type I collagen, and bone morphogenetic protein (BMP)-4 (Dr B Lanske, Forsyth Institute), and GAPDH (Dr J Beresford, University of Bath) was used as a control for mRNA loading. Probes were labeled with [32P]dCTP (NEN, Boston, MA, USA) using Ready-To-Go oligo-labeling beads and ProbeQuant G-50 microcolumns (Amersham, St Albans, UK). Autoradiographs were scanned using a modified UMAX UTA PowerLook III instrument (Amersham) and analyzed using Amersham Biosciences ImageMaster TotalLab version 1.0 software. RT-PCR analysis was carried out using 1 μg total RNA and Moloney-Murine Leukemia Virus Reverse Transcriptase (M-MLV RT; Promega, Southampton, UK) for cDNA synthesis. Primer sequences were as follows: BMP-2 (GenBank accession no. NM007553) sense (nucleotides 68-91) 5′-GGCCCACCCCAAGACACAGTTCCC-3′; BMP-2 antisense (nucleotides 509-530), 5′-GGACCGACCGGACTCACGGACG-3′; GAPDH (GenBank accession no. M32599) sense (nucleotides 566-585), 5′-ACCACAGTCCATGCCATCAC-3′; GAPDH antisense (nucleotides 998-1017), 5′-TCCACCACCCTGTTGCTGTA-3′. The PCR reaction was performed for 28 cycles at 94°C for 1 minute, 60°C for 2 minutes, and 72°C for 3 minutes. The amplified products were separated on a 2% agarose gel stained with ethidium bromide.

Immunofluorescence and microinjection

For actin staining, the cells were cultured on glass coverslips, made quiescent, and treated with PMT or C1165S for the periods indicated. For microinjection experiments only, PMT was purchased from Sigma, which exhibits a 100- to 1000-fold lower activity than the recombinant PMT used in all other studies,(30) and was therefore used at higher concentrations as indicated without any toxicity. Actin filaments were localized after fixation, permeabilization, blocking, and staining with either FITC- or TRITC-labeled phalloidin (Molecular Probes, Leiden, The Netherlands) as described.(33, 34) For microinjection and BrdU experiments, the neutralizing anti-Ras antibody, Y13-259,(35) was injected into quiescent Swiss 3T3 cells along with a marker protein (rat immunoglobulin [IgG]) as previously described,(31, 33) incubated with PMT and BrdU (Sigma), and processed for immunofluorescence using a monoclonal antibody against BrdU (Roche, Lewes, UK) followed by incubation with FITC-conjugated rabbit anti-rat IgG and TRITC-conjugated rabbit anti-mouse secondary antibodies.(33, 36) Cells were examined either on an Olympus BH-2 or a Zeiss axiophot microscope equipped for epifluorescence, and images were recorded on Kodak T-MAX 400 ASA film. Confocal microscopy was performed using a Leica TCS NT confocal laser scanning microscope (CSLM; Heidelberg, Germany) as described previously.(34)

Statistics

Differences between the appropriate groups were analyzed using the Students t-test and compared at p < 0.05 and p < 0.01 confidence levels as indicated.

RESULTS

PMT stimulates proliferation of primary calvarial osteogenic cells

Growth curve analysis showed that PMT (20 ng/ml) stimulated calvarial cell proliferation as measured by a decrease in doubling time and an increase in saturation density (Fig. 1A). Lower concentrations of PMT had no effect on cell number (data not shown). This inhibition was caused by active toxin, because treatment with a mutant PMT protein, C1165S, which has no biological activity in vivo or in vitro,(30) did not affect cell proliferation (Fig. 1B). These proliferative effects were observed in standard media containing 10% serum, which show the extreme potency of PMT because all previous studies have typically characterized the ability of PMT to stimulate DNA synthesis in quiescent cells.(17) We next investigated the intracellular signaling pathways stimulated by PMT that affect cell proliferation, focusing on the Ras/Erk/MAPK pathway, which is known to mediate the proliferative responses of eukaryotic cells to mitogenic signals.(37) Microinjection of a neutralizing Ras antibody, Y13-259, followed by assessment of DNA synthesis by BrdU labeling, showed that inhibition of Ras clearly blocked PMT-induced BrdU incorporation (Figs. 1C and 1D). These observations suggest that the mitogenic effects of PMT are linked to activation of the Erk/MAPK pathway.

Figure FIG. 1..

The effects of PMT on calvarial cell proliferation. (A) Primary mouse calvarial cells were plated in standard media that was replaced after 24 h with either vehicle (▪) or PMT (20 ng/ml; ▵), and the cells were counted on the days indicated. The data represent the averages of duplicate samples. (B) Calvarial cells were cultured as in A in the absence or presence of PMT (20 ng/ml) or mutant PMT C1165S (20 ng/ml), and the cells were counted 10 days later. (C and D) The role of Ras in PMT-induced BrdU incorporation. Quiescent Swiss 3T3 fibroblasts were co-injected with the neutralizing Ras antibody, Y13-259 (10 mg/ml), and rat IgG as a marker. After microinjection, cells were treated with PMT (100 ng/ml) and BrdU (10 μg/ml), and cells were fixed 48 h later. (D) BrdU staining was visualized by indirect immunofluorescence. The injection marker is shown in C. Arrows in D indicate Y13-259-injected cells lacking nuclear BrdU expression after PMT treatment. Bar: 20 μm.

PMT inhibits osteoblast differentiation

Continuous treatment of calvarial cells with PMT in long-term differentiation cultures markedly inhibited the formation of bone nodules, whereas equimolar concentrations of the mutant C1165S had no effects on nodule formation (Fig. 2A). The PMT effects were dose-dependent, with doses as low as 0.2 ng/ml causing a >95% reduction in nodule number, and concentrations from 2 to 200 ng/ml completely inhibiting nodule formation (Fig. 2B). These effects were not species-specific; PMT caused similar effects on the proliferation and differentiation of rat calvarial osteoblasts (data not shown).

Figure FIG. 2..

The effects of PMT on bone nodule formation. Primary mouse calvarial osteoblasts were cultured in (A) differentiation medium in the absence (CON) or presence of PMT (20 ng/ml) or mutant PMT C1165S (20 ng/ml) or (B) cultured in the indicated concentrations of PMT. (C-E) The short-term effects of PMT on nodule formation. Mouse calvarial cells were plated as above and treated with PMT (20 ng/ml) for various intervals. (C) PMT was added at day 0 and removed at days 3, 6, 9, 12, or 15. (D) PMT was pulsed for 3-day intervals starting at days 0, 3, 6, 9, or 12. (E) Cells were treated with PMT for 0, 6, 24, or 72 h or cultured in its continuous presence (cont). All cultures were fixed at day 15 and stained with the von Kossa technique, and the percent mineralization or number of nodules was quantified as described in the Materials and Methods section. The data represent the mean ± SEM of triplicate wells (*p < 0.05, **p < 0.01 vs. control).

Primary calvarial cells were next exposed to PMT at different times in culture throughout the 15-day differentiation period, coinciding with the stages of proliferation (days 0-6), cell multilayering and early nodule formation (days 6-9), nodule growth and matrix deposition (days 9-12), and matrix mineralization (days 12-15). Addition of PMT from day 0 for 3, 6, 9, 12, or 15 days inhibited the number of nodules at all time points (Fig. 2C). During the 15-day culture period, the removal of PMT did not restore nodule formation (Fig. 2C), suggesting that a 3-day pulse during the early proliferative phase was enough to elicit the full PMT-mediated inhibition. This early effect was also observed by addition of PMT from days 0-3, 3-6, and 6-9, which completely inhibited nodule formation, whereas exposure to PMT after nodules had already started forming (day 9) had a slight inhibitory or no effect (Fig. 2D). These variations are likely caused by the differential effects of PMT on mature nodules versus nodules that are being initiated within this time window. It was difficult to assess the long-term effects of PMT (i.e., >15 days) when added after nodules had already formed because the cultures could not be maintained for longer periods of time because of detachment of the multilayers from the substrate (data not shown). We also investigated whether shorter incubation times were also effective. As little as a 6-h exposure to PMT was sufficient to reduce the number of nodules by ∼65% compared with untreated cells (Fig. 2E). Similarly, PMT treatment for 24 h caused a ∼90% inhibition of nodule formation, although in some experiments, 24-h PMT treatment caused a complete inhibition of nodule formation, as did 72-h and continuous treatment (Fig. 2E). These results clearly indicate that the early stages of the osteoblast differentiation program, during proliferation and early nodule formation, are sensitive to the inhibitory effects of PMT.

Molecular analysis of the PMT effect by Northern blotting demonstrated that PMT inhibited Cbfa1/Runx2 expression (∼65%) and caused a marked inhibition in the expression of osteocalcin and alkaline phosphatase (∼80%), whereas type I collagen expression was reduced by ∼50% (Figs. 3A and 3B). The inactive PMT mutant C1165S did not affect the expression of any of the marker genes tested (data not shown). Thus, the inhibition of osteoblastic marker gene expression by PMT correlates well with its potent inhibition of nodule formation.

Figure FIG. 3..

The effects of PMT on the expression of osteoblastic marker genes. (A) Primary mouse calvarial cells were cultured for 48 h in the absence (−) or presence (+) of PMT (20 ng/ml), and Northern blot analysis was performed using poly(A)+ RNA as described in the Materials and Methods section. Membranes were hybridized with specific fragments for Cbfa1/Runx2, osteocalcin (OC), alkaline phosphatase (ALP), type I collagen (coll I), and GAPDH. GAPDH was used as a loading control, and transcript sizes are indicated. (B) The relative expression of each marker gene compared with the respective GAPDH levels was analyzed semiquantitatively by scanning densitometry as described in the Materials and Methods section.

Finally, time lapse microscopy of early nodule condensations showed that untreated nodules continued to proliferate and deposit bone matrix over the 15-day culture period, whereas early nodules treated with PMT failed to differentiate further, and these changes were observed as early as 4-5 h after exposure to PMT (Fig. 4). Moreover, PMT treatment caused a striking morphological change, from the cuboidal and polygonal morphology typical of wildtype nodule-forming cultures to more elongated and spindle-shaped cells, and this was evident after ∼10 h of PMT treatment (Fig. 4). These data suggest that changes in cell morphology may underlie the inhibitory effects of PMT.

Figure FIG. 4..

Time lapse microscopy of bone nodule formation. Primary mouse calvarial cells were cultured in differentiation medium in the absence of PMT. PMT (20 ng/ml) was added at day 8 after clear early foci of developing nodules were evident, and images were captured every 10 minutes thereafter for a maximum of 96 h. Representative images after 0, 5, 10, and 60 h of PMT treatment are shown. Bars: 100 μm.

PMT-induced cytoskeletal changes, Rho GTPases, and inhibition of bone nodule formation

We further examined whether there was any correlation between PMT-induced stress fibers and its inhibitory effects on osteoblast differentiation. First, PMT treatment of undifferentiated calvarial cultures caused dramatic cytoskeletal rearrangements with highly organized actin stress fibers throughout the cell, whereas the cytoskeleton of untreated cultures comprised predominantly cortical actin with very few stress fibers (Figs. 5A and 5B). The PMT-induced stressed fibers were inhibited by the ROK inhibitor Y-27632 (data not shown), consistent with previous reports(26, 27) and confirming that the cytoskeletal effects of PMT in calvarial cell cultures are mediated by the Rho GTPase. F-actin distribution was also examined in mature osteoblasts situated on the surface of 3D bone nodules before mineralization. Untreated cultures revealed cuboidal osteoblast-like cells containing a prominent cortical actin cytoskeleton with few visible stress fibers (Figs. 5C and 5D). In contrast, PMT-treated cultures showed marked cytoskeletal rearrangements and formation of actin stress fibers in mature osteoblasts on the surface of nodules (Figs. 5E-5G). These results clearly implicated the small GTPase Rho, because actin rearrangements are mediated by Rho. The inactive mutant C1165S did not cause stress fibers in either undifferentiated cultures or in differentiated osteoblasts within the nodules (data not shown). Interestingly, microinjection of the neutralizing Ras antibody, Y13-259, which blocked the mitogenic effects of PMT in osteoblasts, did not attenuate its ability to induce cytoskeletal rearrangements (Figs. 5H and 5I).

Figure FIG. 5..

The effects of PMT on the actin cytoskeleton and morphology of primary mouse calvarial cells and bone nodules. Undifferentiated, quiescent cells were cultured for 24 h in the (A) absence or (B) presence of PMT (20 ng/ml) and were fixed and stained with FITC-phalloidin as described in the Materials and Methods section. (C-G) Differentiating nodule-forming cultures were treated in the absence of PMT until clearly defined, 3D nodules were evident (12 days). Thereafter, cultures were treated either in the (C and D) absence or (E-G) presence of PMT (20 ng/ml) for an additional (C, E, and G) 24 or (D and F) 48 h. The actin cytoskeleton was visualized after fixation and staining with TRITC-phalloidin and was analyzed using confocal microscopy. (H and I) Inhibition of Ras failed to prevent PMT-induced cytoskeletal changes. Swiss 3T3 fibroblasts were co-injected with the neutralizing Ras antibody Y13-259 (10 mg/ml) and rat IgG as described in Fig. 1. After microinjection, cells were treated with PMT (100 ng/ml) and fixed 48 h later. Actin filaments were visualized with (I) TRITC-conjugated phalloidin, and the injection marker is shown in H. Bars: A, B, D, E, G, H, and I, 20 μm; C and F, 100 μm.

We further examined the possible link between the activation of Rho GTPases and the PMT inhibition of bone nodule formation. Concomitant treatment of osteoblast differentiation cultures with PMT and C3 transferase, which inhibits all functions of Rho, abrogated the PMT-induced inhibition of differentiation (Fig. 6A). Moreover, inhibition of Rho kinase, which mediates PMT-induced stress fiber formation, using the ROK inhibitors HA-1077 or Y-27632,(38, 39) prevented the PMT-induced inhibition of nodule formation at low doses of PMT and partly inhibited the PMT effects at high doses (Figs. 6B and 6C). These results strongly suggest that PMT inhibits the differentiation of osteogenic cells through activating Rho and in part through a ROK-dependent pathway. This was further assessed by investigating the effects of another related bacterial toxin, CNF, that targets all Rho family GTPases.(28) The results showed that continuous treatment of primary calvarial cultures with CNF caused a dose-dependent inhibition in bone nodule formation (Fig. 6D). More importantly, Y-27632 blocked the inhibitory effects of CNF, further supporting our conclusion that Rho-ROK is one signaling pathway responsible for the inhibition of osteoblast differentiation by PMT.

Figure FIG. 6..

The effects of Rho inhibitors on PMT-induced inhibition of bone nodule formation. (A) Primary mouse calvarial cells were cultured in differentiation medium in the presence or absence of PMT (20 ng/ml), and C3 transferase (10 μM) was added for 24 h at the beginning of the culture period. Cells were also treated continuously in the absence or presence of the ROK inhibitors, (B) HA-1077 (10 μM) or (C and D) Y-27632 (10 μM), with the indicated concentrations of (B and C) PMT or (D) CNF. All cultures were fixed at day 15 and stained with the von Kossa technique, and the percent mineralization or number of nodules were quantified as described in the Materials and Methods section. The data represent the mean ± SEM of triplicate wells (*p < 0.05, ** p < 0.01 vs. respective control).

Inhibition of Rho-ROK stimulates bone nodule formation

The ROK inhibitors HA-1077 and Y-27632 not only prevented the PMT and CNF effects, but also stimulated bone nodules in the absence of toxin (Figs. 6B-6D), further implicating the Rho-ROK pathway in negatively regulating osteoblast differentiation. Indeed, treatment of primary osteogenic cells with only C3 transferase, HA-1077, or Y-27632 significantly stimulated the number of mineralized nodules (Figs. 7A-7C). Further examination using Y-27632 showed that this ROK inhibitor increased the number of alkaline phosphatase-expressing cells compared with control cultures (Figs. 7D and 7E), leading to enhanced nodule formation and earlier mineralization (Figs. 7F and 7G). This suggested that Y-27632 affected the differentiation of osteoblast precursors rather than matrix deposition and mineralization. These cellular effects were confirmed at the molecular level by Northern blot analysis, which showed that PMT treatment alone inhibited the expression of osteocalcin and alkaline phosphatase RNAs, whereas treatment with Y-27632 markedly upregulated their expression and partly restored the inhibitory effects of PMT (Fig. 8A). These effects correlated well with the effects of PMT and Y-27632 on nodule formation. We further investigated whether PMT and Y-27632 modulated BMP expression levels, which may provide a possible mechanism to explain their effects on bone nodule formation. Northern blot analysis showed that PMT decreased BMP-4 expression, whereas Y-27632 upregulated BMP-4 RNA levels (Fig. 8B). Because BMP-2 was not detectable at high levels by Northern blotting under our experimental conditions, we used RT-PCR analysis. The results clearly indicated that, similar to BMP-4, BMP-2 expression was inhibited by PMT and stimulated by Y-27632 (Fig. 8C). These data suggest that one mechanism by which PMT and the Rho-ROK pathway regulates osteoblast differentiation is by altered expression of BMPs.

Figure FIG. 7..

The effects of the Rho inhibitor C3 transferase and the ROK inhibitors HA-1077 and Y-27632 on osteogenic cell differentiation. Primary mouse calvarial cells were cultured in differentiation medium in the presence or absence of the indicated concentrations of (A) C3 transferase, (B) HA-1077, or (C) Y-27632. All cultures were fixed at day 15 and stained with the von Kossa technique, and the percent mineralization was quantified. (D-G) Bright field micrographs of confluent cultures (D and E) before and (F and G) after nodule formation, stained for alkaline phosphatase activity. Y-27632 markedly increases the number of (E) alkaline phosphatase-positive cells vs. (D) control cultures and stimulates earlier formation of nodules that mineralize earlier (arrows in G) compared with (F) control cultures. Data represent the mean ± SEM of triplicate wells (*p < 0.01 vs. respective control). Bars: 500 μm.

Figure FIG. 8..

The effects of PMT and Y-27632 on expression of osteoblastic marker genes and BMPs. Primary mouse calvarial cells were cultured for 48 h in the absence (−) or presence (+) of PMT (20 ng/ml) or the ROK inhibitor Y-27632 (10 μM), either alone or in combination as indicated. Expression of (A) osteocalcin (OC) and alkaline phosphatase (ALP) and (B) BMP-4 was carried out by Northern blot analysis, whereas expression of (C) BMP-2 was carried out by RT-PCR analysis on total RNA. GAPDH was used a loading control. RNA transcript and RT-PCR product sizes are indicated.

DISCUSSION

We have used the unique and potent bacterial toxin, PMT, which perturbs bone remodeling in vivo, to investigate the potential signal transduction pathways involved in osteoblast differentiation and have shown that cells of the osteoblast lineage are bona fide cellular targets for PMT. More importantly, the data show that PMT stimulates the Rho-ROK pathway, resulting in inhibition of osteoblast differentiation and bone formation in vitro and that specific inhibition of Rho and ROK stimulates bone formation.

PMT is a potent inhibitor of osteoprogenitor cell differentiation

Continuous treatment of primary mouse calvarial cells with PMT inhibited the formation of mineralized bone nodules, and this inhibitory effect was manifested early during differentiation, where as little as a 6-h exposure to PMT was sufficient to inhibit nodule formation. These data suggest that the effects of PMT on osteoprogenitors are rapid and that PMT may decrease the proportion of precursors that are capable of forming nodules. This was further supported by the time-lapse studies where early nodule condensations treated with PMT exhibited morphological changes from ∼10 h of treatment and failed to progress further to form 3D bone nodules.

Our data indicate that PMT is a mitogen that inhibits bone formation. Numerous hormones and growth factors that act through diverse receptor signaling pathways are mitogenic for osteoblasts and can inhibit bone formation in vitro.(3, 40) However, it is difficult to establish a simple correlation between proliferation and differentiation because primary calvarial cultures are heterogeneous, and nodule-forming osteoprogenitor cells exist at a very low frequency; thus, changes in the proliferative potential of these cells may not be observed.(41) Nevertheless, at the mass population level, several lines of evidence support the notion that these events may not be sequential in PMT-treated cultures. First, we observed clear inhibitory effects on bone nodule formation at PMT concentrations that failed to stimulate cell proliferation. Second, a short-term pulse of PMT inhibited the number of nodules but did not stimulate proliferation (data not shown). Third, inhibition of Ras, which is known to mediate growth factor-induced proliferation through the Erk/MAPK pathway, blocked PMT-induced BrdU incorporation but not PMT-induced stress fiber formation, which is mediated by Rho.(26, 27) Furthermore, preliminary data showed that the MEK inhibitor, PD98059, failed to block the inhibitory effects of PMT and had no effect on its own on osteoblast differentiation (data not shown), although inhibition of MEK has recently been shown to enhance BMP-2 effects in the MC3T3-E1 cell line.(42) Thus, it is tempting to speculate that the effects of PMT on osteoblastic cell proliferation and differentiation may be mediated by different signaling pathways, although this needs to be further clarified.

It is not yet known whether the inhibition of osteoblast marker gene expression by PMT is direct or indirect, although notably, the reduction in Cbfa1/Runx2 expression, which is essential for osteoblast differentiation,(2) correlated well with the inhibition of differentiation. Thus, one possibility is that the reduction in the expression of Cbfa1/Runx2 may represent one mechanism by which PMT inhibits osteoblast differentiation. Indeed, the downregulation of BMP-2 and −4 expression by PMT would suggest that the inhibition of Cbfa1/Runx2 expression may be indirect, because these BMPs are known to regulate Cbfa1/Runx2 transcription.(43) However, because we only observed ∼50% inhibition of Cbfa1/Runx2 expression, it remains possible that the PMT-dependent signaling events that inhibit osteoblast differentiation lie downstream or are independent of Cbfa1/Runx2.

We have eliminated the possibility that the effects observed are caused by possible contamination of our PMT preparations by bacterial products, such as a lipopolysaccharide (LPS), which can synergize with some bacterial toxins. LPS was the first bacterial component shown to induce osteoclastic bone resorption in vitro,(44) and it has been shown that LPS from Porphyromonas gingivalis inhibits bone nodule formation and gene expression in primary rat calvarial osteoblasts.(45) However, in contrast to LPS, PMT is heat labile,(17) and we have confirmed that heat inactivation of PMT abolished the inhibitory effects on osteoblast differentiation (data not shown). Moreover, removal of any potential LPS contamination using polymixin B did not inhibit the PMT effects, suggesting there is no synergy between LPS and PMT (data not shown). Together with the fact that the inactive mutant C1165S had no effects on nodule formation or proliferation, this shows that the effects of PMT on osteoblast differentiation are caused by the recombinant protein and not a result of other bacterial factors.

Activation of Rho-ROK mediates the PMT effects on osteoblasts and inhibits bone formation

The marked cytoskeletal rearrangements induced by PMT, particularly in cuboidal cells lining bone nodules, clearly indicated that PMT stimulates the Rho pathway in osteoblastic cells, because actin rearrangements and stress fiber formation are mediated by Rho GTPases. Thus, we have demonstrated a novel link between osteoblast differentiation, PMT, and Rho-induced actin rearrangements. However, it is not yet clear whether the altered cytoskeleton is causally involved in regulating osteoblast differentiation, because Rho GTPases can also activate other signaling pathways and transcription factors.(46) Indeed, a recent report has demonstrated that Rho-ROK can regulate the lineage commitment of mesenchymal cells and that this may be independent of the cytoskeleton.(47) Nevertheless, the activation of Rho played a functional role in osteoblast differentiation because the Rho inhibitor, C3 transferase, abrogated the PMT inhibition of bone nodule formation. It is not yet known how PMT induces Rho-dependent signaling in osteogenic cells, although it is well accepted that Rho is not a direct target of PMT. Rather, it is likely that heterotrimeric G proteins are involved, in particular the Gq/11 family, which are known to be activated by PMT and which have been shown recently to activate Rho, or alternatively, the G12/13 family, which is known to activate Rho through RhoGEFs.(20-22, 48, 49) We also cannot rule out that other signaling pathways stimulated by PMT downstream of Gq/11 or G12/13 may contribute to its inhibitory effects, although our data would suggest that it is unlikely that other Rho GTPases are involved (e.g., Rac, cdc42) because the inhibitory effects of CNF were reversed by the specific ROK inhibitor Y-27632. Last, previous reports have suggested that the increase in osteoblast activity after mechanical stimulation is linked to cytoskeletal rearrangements and induction of cyclo-oxygenase-2 (Cox-2) and may be blocked by inhibiting Rho activity.(50-52) We have recently demonstrated that CNF, which activates Rho, Rac, and cdc42, also stimulates Cox-2 expression in a manner that is dependent on Rho activation but independent of ROK, indicating that other Rho effectors mediate the transcriptional regulation of Cox-2.(53) Our preliminary evidence has indicated that inhibition of Cox-2 and prostaglandin synthesis by indomethacin does not block the inhibitory effects of PMT on osteoblasts (data not shown), suggesting that PMT inhibits bone nodule formation through a prostaglandin-independent pathway.

While Rho is important for mediating the PMT effects, we have shown additionally that the Rho effector, ROK, is important. Because nodule number was not fully restored by the ROK inhibitors, it remains possible that there exist different subpopulations within these cultures that respond differently to PMT and HA-1077/Y-27632 or that another Rho effector may play an important role.(46) However, the importance of ROK as a critical regulator of bone formation was highlighted by the fact that the ROK inhibitors, HA-1077 and Y-27632 alone, markedly stimulated osteoblast marker expression and caused precocious formation of bone nodules that mineralized earlier than control cultures. Moreover, BMPs may play a key role downstream of the Rho-ROK pathway because Y-27632 stimulated, whereas PMT inhibited, both BMP-2 and −4 expression, consistent with their stimulatory and inhibitory roles on nodule formation, respectively. These results are interesting in view of the observed stimulation of bone formation by statins,(54) which inhibit HMG-CoA reductase, eventually leading to inhibition of small GTPases by preventing their prenylation, although it is not established whether the observed anabolic effects by statins are directly caused by inhibition of Rho activity. Indeed, the possible mechanisms may differ as statins have been reported to stimulate BMP-2, but not BMP-4, expression.(54) Finally, a recent study has indicated that inhibition of ROK stimulates osteoblast gene expression, although bone formation was not assessed in that study.(55) Our data show for the first time that direct inhibition of Rho and ROK stimulates bone nodule formation and that this may be through enhanced BMP expression; we have also implicated specific heterotrimeric G proteins in the regulation of bone formation by PMT.

In conclusion, we have established that the bacterial toxin, PMT, targets osteogenic cells. There are a large number of bacterial products, some of which are protein toxins, that affect bone cell function, but the majority of them lead to bone destruction either by targeting osteoclastic bone resorption, inhibiting osteoblast proliferation, and/or stimulating apoptosis.(44) The main in vivo effect of PMT is bone loss, suggesting the involvement of osteoclasts, and indeed, our recent data has demonstrated that PMT has direct independent effects on the osteoclast lineage (unpublished data). However, the data presented here would suggest that a major contributing factor in the pathogenesis of atrophic rhinitis may be the inhibition of osteoblast differentiation and bone formation. The use of PMT as a tool to study intracellular signal transduction pathways has led to the demonstration that activation of the Rho-ROK pathway potently inhibits osteoblast differentiation. Moreover, inhibition of Rho-ROK signaling results in marked stimulatory effects on bone nodule formation, thus providing potential therapeutic targets for metabolic diseases characterized by bone loss.

Acknowledgements

The authors thank Dr Gillian Pullinger for providing the recombinant PMT and mutant C1165S and for helpful discussions, Dr Michael Horton for the confocal microscope facilities, and Ioannis Anagnostopoulos for RT-PCR analysis. We also thank Drs Caroline Damsky and Michael Horton for critical reading of the manuscript. GS is a recipient of a postdoctoral fellowship from the Arthritis Research Campaign. This work was supported in part by the Dental Funds Committee of the Guy's, King's and St. Thomas' School of Dentistry and in part by the Department of Orthodontics at the Guy's, King's and St. Thomas' School of Dentistry.

Ancillary