Compactin (mevastatin), which inhibits 3-hydroxy-3-methylglutaryl–coenzyme A (HMG-CoA) reductase, and thus biosynthesis of cholesterol and the prenylation of proteins, inhibits osteoclastic bone resorption. Although it has been suggested that compactin inhibits bone resorption by inducing apoptosis of osteoclasts, the pathway by which compactin inhibits resorption has not been established. We investigated the effect of compactin on the differentiation of osteoclasts and the relationship between the morphological changes elicited by compactin and its inhibitory effect on bone resorption. Compactin inhibited the differentiation of osteoclasts, interfering with the fusion process by which prefusion osteoclasts (pOCs) develop into multinucleated osteoclast-like cells (OCLs), and also disrupted the actin ring of OCLs. The potency of compactin to inhibit fusion of pOCs and to disrupt the actin ring of OCLs corresponded to that of compactin to inhibit bone resorption. The effects of compactin were prevented by the addition of MVA lactone or its downstream products farnesylpyrophosphate (FPP) and geranylgeranyl-pyrophosphate (GGPP) but not by squalene. Apoptosis of OCLs was not induced by the concentration of compactin that inhibited fusion of pOCs and disrupted the actin ring. The normal process of pOC fusion and the integrity of the actin ring were restored by the withdrawal of compactin from the cultures after they had been treated with compactin for 24 h, but they were not restored by the addition of zVAD-fmk, a caspase inhibitor. Compactin also reversibly inhibited interleukin-1β (IL-1β)-, 1α,25-dihydroxyvitamin D3 (1 α,25(OH)2D3)–, and parathyroid hormone (PTH)–stimulated 45Ca release in bone organ cultures. Our results indicate that the inhibitory effects of compactin on bone resorption result from the inhibition of fusion of pOCs into OCLs and disruption of actin ring in OCLs and that apoptosis of OCLs is not necessary for these inhibitory effects of compactin. These effects of compactin are likely to be a consequence of the inhibition of prenylation of proteins that play an important role in the fusion of pOCs and in maintaining actin ring integrity in OCLs.
Osteoclasts are multinucleated cells that dissolve mineralized bone matrix and play a critical role in bone remodeling.(1–3) Osteoclast-mediated bone resorption occurs by a multistep process: (a) differentiation of mono-nuclear preosteoclasts from hemopoietic cells of the monocyte-macrophage lineage, (b) fusion of mononuclear prefusion osteoclasts into mature multinuclear osteoclasts, (c) attachment of mature osteoclasts to RGD-sequence–containing proteins of the mineralized bone surface through vitronectin receptors, (d) polarization of cytoplasmic structures such as ruffled borders and clear zones, and (e) secretion of acids and lysosomal enzymes into the resorption space beneath the ruffled border, resulting in the dissolution of apatite crystals and degradation of organic matrix.(1–11) Natural compounds that specifically inhibit these steps could be developed not only as antiresorptive drugs for the treatment of metabolic bone disorders such as osteoporosis, hyperparathyroidism, hypercalcemia of malignancy, and metastastic bone disease, which are characterized by excessive osteoclastic bone resorption, but they could also provide useful tools for studies of the mechanisms of differentiation and function of osteoclasts. Many natural low molecular weight compounds have been reported to inhibit the differentiation and function of osteoclasts. Microbial metabolites such as c-Src tyrosine kinase inhibitors (herbi-mycin A and genistein), a phosphatidyl inositol 3-kinase inhibitor (wortmannin), a calcineurin inhibitor (cyclosporin A), vitamin K2, vacuolar type proton adenosine triphosphatase (ATPase) inhibitors (concanamycins, bafilomycins, and prodigiosins), protease inhibitors (E-64, leupeptin, and others), and plant metabolites such as isoflavone and phytoestrogens have been reported to inhibit osteoclast differentiation and/or function.(12–36) Among the above metabolites, vitamin K2 and a derivative of natural isoflavone ipriflavone are already used clinically as drugs for the prevention and treatment of osteoporosis. To gain new insights into the mechanisms of differentiation and function of osteoclasts and to develop natural compounds as osteoporotic drugs for clinical application, we have screened microbial metabolites for low molecular weight compounds that inhibit the differentiation and function of osteoclasts. In the process of this screening we found that compactin (mevastatin) inhibits osteoclastic bone resorption.
Compactin is known to inhibit the rate-limiting enzyme in cholesterol biosynthesis, 3-hydroxy-3-methylglutaryl–coenzyme A (HMG-CoA) reductase, which catalyzes the synthesis of mevalonic acid (MVA) lactone from HMG-CoA.(37) Its derivatives such as lovastatin and pravastatin are clinically used as antihyperlipidemic drugs for the treatment of arteriosclerosis. These inhibitors block the prenylation (farnesylation and geranygeranylation) of guanosine triphosphate (GTP)–binding proteins such as Ras by inhibiting the synthesis of MVA and its downstream intermediates farnesylpyrophosphate (FPP) and geranylgeranylpyro-phosphate (GGPP), which are substrates for prenyl protein transferases that prenylate the GTP-binding proteins.(38) It has been reported that inhibition of protein prenylation by HMG-CoA reductase inhibitors induces morphological changes in cells, affects intracellular signaling pathways, and results in apoptotic cell death.(39–41) The effects of these inhibitors could result from the inhibition of the prenylation of target proteins by prenyl protein transferase, through preventing the production of FPP and/or GGPP.(41,42) Recently, compactin and nitrogen-containing bisphosphonates such as alendronate (ALN), which are used clinically as antiresorptive drugs for the treatment of osteoporosis, have been reported to cause osteoclast apoptosis and thus inhibit osteoclastic bone resorption.(43) It was shown that ALN as well as compactin elicit their apoptotic effects by inhibiting the prenylation of target proteins through inhibition of the production of the substrates (FPP and GGPP) of prenyl protein transferases.(43) Although compactin can inhibit bone resorption by inducing apoptosis of osteoclasts, effects of compactin on the specific steps in the process of osteoclast differentiation and function have not been determined, and the exact mechanism of the effect of these inhibitors on bone resorption has not been established. To determine this, we assessed whether compactin inhibits osteoclast differentiation in cocultures of mouse calvarial osteoblastic cells and bone marrow cells. We also investigated the relationship between the morphological changes in osteoclasts and inhibition of bone resorption by compactin. Our results reveal that this HMG-CoA reductase inhibitor reversibly inhibits the fusion steps of prefusion osteoclasts (pOCs) and disrupts the actin ring in osteoclast-like cells (OCLs). The inhibitory effect of compactin on parathyroid hormone (PTH)–stimulated resorption in organ cultures was also reversible. In addition, we determined that the inhibitory effects of compactin on fusion of pOCs and on actin ring integrity could be prevented by the addition of several components of the mevalonate pathway, that is, MVA, FPP, and GGPP, but not by squalene or the caspase inhibitor (zVAD-fmk). Our findings indicate that the inhibitory effects of compactin on bone resorption are a result of the inhibition of fusion of pOCs and the disruption of actin ring in OCLs. The results also suggest that the prenylation of target proteins such as Rho, Rac, and Rab is involved in the fusion of pOCs and in maintaining actin ring integrity in OCLs.
MATERIALS AND METHODS
Animal and chemicals
Newborn Std.ddY mice and 6- to 9-week-old male Std.ddY mice were purchased from Japan SLC Co. (Hamamatsu, Japan). FPP, GGPP, squalene, fast violet LB salt, and naphthol AS-MX phosphate were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). MVA lactone and collagenase were purchased from Wako Pure Chemicals Co. (Osaka, Japan). Collagen gel solutions (cell matrix, type IA) were purchased from Nitta Gelatin Co. (Osaka, Japan). Rhodamine-conjugated phalloidin was purchased from Eugene Co. (Eugene, OR, U.S.A.). The 1α,25-dihydroxyvitamin D3 [1α,25(OH)2D3] was purchased from Wako Pure Chemicals (Osaka, Japan). Dispase was purchased from Godo Shusei (Tokyo, Japan). Compactin (mevastatin) was a generous gift from Dr. A. Endo (Biopharm Research Laboratories, Tokyo, Japan). Lovastatin and simvastatin were a generous gift from Sankyo Co.(Tokyo, Japan). PTH, recombinant murine interleukin-1β (IL-1β), recombinant murine macrophage colony-stimulating factor (M-CSF), and soluble recombinant human receptor activator of nuclear factor-κB ligand (sRANKL) were from Bachem, Calbiochem (La Jolla, CA, U.S.A.), R&D Systems (Minneapolis, MN, U.S.A.), and Pepro Tech EC, Ltd (London, U.K.), respectively.
OCL formation assay
Multinucleated OCLs were derived with a coculture system as previously described.(44) Briefly, bone marrow cells were obtained from the tibias and femora of 6- to 9-week-old male mice and osteoblastic cells were isolated from calvariae of newborn ddY mice. Bone marrow cells (2 × 107 cells) were cocultured with the osteoblastic cells (1 × 106 cells) on culture dishes precoated with a 0.2% type IA collagen gel matrix in α-minimal essential medium (α-MEM) containing 10% fetal bovine serum (FBS) and 10 nM 1α,25(OH)2D3 for 7 days. After culturing for 7 days, cultures were treated with 0.2% collagenase and 0.1% dispase. Cells were collected and replated in culture plates and stained for tartrate-resistant acid phosphatase (TRAP) activity. TRAP staining was carried out as described previously. In brief, the cells fixed with 10% formalin and ethanol were reacted for 30 minutes in 0.1% sodium acetate buffer (pH 5) containing 50 mM sodium tartrate, 0.1 mg/ml naphthol AS-MX phosphate, and 1 mg/ml fast violet LB salt. After washing with water and drying, TRAP-positive cells were counted as OCLs.
Actin ring and pit formation assay
Actin rings of OCLs were detected by staining actin filaments with rhodamine-conjugated phalloidin. Collected OCLs (about 300 cells/well) were replated in culture plates for about 1 day and treated with compactin and its derivatives for 24 h. At the end of incubation, OCLs were stained for TRAP activity for 30 minutes at room temperature. TRAP-positive OCLs were stained with rhodamine-conjugated phalloidin in the dark and washed with cold phosphate-buffered saline (PBS). The distribution of actin rings of OCLs was visualized and detected under a fluorescence microscope. For determining the resorptive activity of OCLs, the OCLs were cultured on dentine slices in the presence of compactin and its derivatives for 24 h. After 24-h incubation, dentine slices were ultrasonicated in 1 M aqueous NH4OH to remove adherent cells. Resorption pits on the slice were stained with Mayer's hematoxylin solution. Resorption was quantified based on the number of pits.
Cell fusion assay
The pOCs were prepared from cocultures for osteoclastogenesis as previously described.(45) Briefly, mouse osteoblastic cells (1 × 106 cells) and bone marrow cells (2 × 107 cells) were cocultured in α-MEM containing 10% fetal calf serum in a 10-cm dish for 4 days. Cells were treated with 10−8 M 1α,25(OH)2D3 for the last 2 days of culture. Floating cells were washed off, and mononuclear cells attached to the cell layer of osteoblastic cells were retrieved by gentle pipetting with fresh α-MEM. Mononuclear cells were collected by centrifugation (1000 rpm, 5 minutes) and used for preparation of pOCs. Results of staining for TRAP showed that the population of pOCs constituted more than 50–60% of the total cells, and no multinucleated cells that had more than two nuclei were detected. In addition, no osteoblastic cells were detected by staining for alkaline phosphatase, a recognized marker for osteoblasts. The pOCs (1 × 105 cells/well) were cultured in the presence or absence of osteoblastic cells (1 × 104 cells/well) and in the presence or absence of M-CSF, IL-1β, and RANKL (each at 100 ng/ml), in 96-well cell culture plates. After culture for 24 h, cells were fixed and stained for TRAP. TRAP-positive cells with more than four nuclei were counted as OCLs.
Preparation of OCLs, excluding osteoblastic cells
The pOCs were prepared by pipetting out from the co-culture of bone marrow cells and osteoblastic cells, which had been maintained in the presence of 10 nM 1α,25(OH)2D3 for 5 days. The pOCs were detected by TRAP staining. The fusion of pOCs into OCLs occurred within 24 h in the presence of cytokines that support the survival and fusion of osteoclast precursors.
45Ca release assay
Nineteen-day fetal rat limb bones were prelabeled with 45Ca as described previously.(46) Bones were cultured for 72 h with or without compactin in the presence of IL-β, 1α,25-(OH)2D3, and PTH in Dulbecco's modified Eagle medium (DMEM) containing 15% heat-inactivated horse serum (GIBCO, Paisley, U.K.) and 100 U/ml penicillin. At the end of the culture period, bones were extracted in 0.1N HCl. Radioactivity in media and bone extracts was determined by liquid scintillation spectrophotometry. Resorption was quantified as the percent of the incorporated 45Ca that was released into the medium.
Differences between groups in the fusion of pOCs or actin ring disruption in OCLs were analyzed by the Student's t-test. Differences in 45Ca release after treatments were analyzed using the Newman-Keuls multiple comparison test.
Compactin inhibits osteoclast formation in the cocultures and the fusion of isolated pOCs into OCLs
TRAP-positive OCLs formed in vitro in cocultures of mouse osteoblastic cells and bone marrow cells in the presence of 1α,25-(OH)2D3 (10 nM). The addition of compactin for 24 h on the last 3 days (3–4 days, 4–5 days, or 5–6 days) of a 6-day incubation period inhibited OCL formation (cells with more than four nuclei) in response to 1α,25-(OH)2D3 (Figs. 1A and 1B). However, mononuclear pOCs were increased by compactin compared with control (4–5 days; 2 μM compactin, 3008 ± 116; control, 1737 ± 68). The IC50 for the inhibitory effect of compactin on OCL formation was approximately 2 μM (Fig. 1B). In contrast, the addition of compactin for 24 h on the first 3 days (0–1 day, 1–2 days, or 2–3 days) had no effect on the OCL formation. The viability of the osteoblastic cells was not affected by compactin at concentrations up to 20 μM (data not shown). OCLs are formed by the fusion of mononuclear pOCs, the stage of differentiation just before fusion to OCLs. In the model used here, the fusion occurs during the late stage of osteoclastogenesis (the last 72 h). The observed time course of the inhibitory effect of compactin on OCL formation suggested that compactin inhibited the fusion of pOCs into OCLs, because the inhibitor was only effective when added during this late stage of osteoclastogenesis. To test this directly, we determined whether compactin inhibited the actual fusion process. When pOCs, isolated from the osteoclastogenesis cocultures by pipetting as described in the Methods section, were cultured in the presence of osteoblastic cells, the pOCs formed into multinucleated OCLs. Compactin inhibited the fusion of pOCs into OCLs in a dose-dependent manner (Fig. 1C); its IC50 on the fusion process was about 2 μM. IL-1β, M-CSF, and RANKL also induce the fusion of pOCs into OCLs in the absence of osteoblastic cells.(47,48) The multinucleation of pOCs by IL-1β, M-CSF, and RANKL also was inhibited by 7 μM of compactin (Figs. 1D–1G). Formation of large OCLs (>10 nuclei) was much more sensitive to compactin than formation of OCLs having 4–9 nuclei. The TRAP staining of OCLs induced by RANKL in the absence or presence of 7 μM of compactin is shown in Figs. 1F and 1G. OCLs having more than 10 nuclei were found rarely in cultures treated with compactin (Fig. 1G). The inhibitory concentration of compactin for the fusion of pOCs induced by either osteoblastic cells or cytokines was similar and also was in the same range as that affecting osteoclastogenesis in the osteoblastic cell/marrow cell cocultures.
Compactin induces morphological changes in OCLs and thus inhibits osteoclastic bone resorption
We determined the effects of compactin on osteoclastic bone resorption on bone slices. Compactin inhibited pit formation in the same concentration range (0.2–20 μM) at which it inhibited osteoclastogenesis (Fig. 2A). Pit size as well as pit number were decreased by the addition of 7 μM compactin as shown in Figs. 2B and 2C. To investigate the relationship between morphological changes in osteoclasts and inhibition of bone resorption by compactin, we determined the effect of compactin on morphology of OCLs, specifically, on the actin ring, which was visualized by staining actin with rhodamine-conjugated phalloidin. When OCLs isolated from collagen gel cultures for osteoclastogenesis were replated on culture dishes, more than 80% of the OCLs spread out and formed ringed structures of F-actin (actin rings) within 4 h. OCL spreading and actin ring formation were not affected by compactin in this 4-h period (data not shown). In contrast, when compactin was added to the cultures after actin rings had formed in OCLs, the size of OCLs was constricted and the actin rings were disrupted in a dose-dependent manner within 24 h (Figs. 3A–3F). Distinct structural changes in the F-actin stress fibers were not observed in osteoblastic cells (data not shown). To more fully characterize and establish the importance of these morphological changes in OCLs for the effects of compactin on resorption, we carried out a time course study. The treatment times required for 7 μM compactin to produce 50% and 90% disruption compared with control were approximately 12 h and 24 h, respectively, as shown in Fig. 3B. Dose-dependence studies showed that compactin-induced effects on actin ring were similar to effects on osteoclastic pit formation. Compactin derivatives such as lovastatin and simvastatin also induced the disruption of actin rings in OCLs (data not shown). The inhibitory activity of lovastatin and simvastatin was similar to that of compactin. Next, to confirm whether compactin directly affected OCLs, the effect of compactin on actin ring disruption was determined in the absence of osteoblastic cells. Compactin induced actin ring disruption in a dose-dependent manner within 24 h in OCLs, which had been fused from pOCs by RANKL, without osteoblastic cells (Fig. 3A). The compactin concentration producting 50% disruption was approximately 1 μM. When 7 μM compactin was used, only about 10% of OCLs had actin rings.
The inhibitory effects of compactin on fusion, actin ring integrity, and bone resorption were prevented by MVA and its intermediates FPP and GGPP
Inhibition of HMG-CoA reductase by compactin blocks the synthesis of mevalonate from HMG-CoA. This suggests that mevalonate or one of its downstream metabolites is required for the fusion of pOCs or actin ring integrity in osteoclasts. Therefore, we investigated whether the effects of compactin on the fusion of pOCs induced by RANKL and on actin ring integrity in OCLs were prevented by the addition of MVA or its downstream intermediates FPP and GGPP, cellular components whose concentrations are decreased by inhibition of HMG-CoA reductase. Although MVA alone had no effect on osteoclasts (data not shown), MVA (100 μM) completely prevented the effects of 7 μM compactin on the fusion of pOCs, actin ring integrity, and pit formation (Figs. 4A–4C). This result suggested that MVA or its downstream metabolites are required for the fusion of pOCs, actin ring formation, and pit formation. The effects of the downstream metabolites of MVA, FPP, GGPP, and squalene were therefore determined. These three metabolites had no effect on their own (data not shown). GGPP (20 μM), like mevalonate, completely prevented actin ring disruption and the inhibition of pit formation induced by compactin (Figs. 4B and 4C). The GGPP partially prevented the inhibition of the fusion of pOCs, restoring it to about 60% of the effect seen when MVA was added (Fig. 4A). FPP partially prevented the effects of compactin on the fusion of pOCs, actin ring integrity, and pit formation, restoring these effects to about 60% of the effect seen with MVA (Figs. 4A–4C). The effects of compactin were not prevented by the addition of squalene.
Apoptosis of OCLs is not necessary for the effect of compactin on fusion, actin ring formation, and bone resorption
To investigate whether the effects of compactin on fusion, actin ring integrity, and bone resorption resulted from apoptosis of pOCs in the presence of RANKL or OCLs in the presence of osteoblastic cells, we determined the reversibility of these actions of compactin. For these studies, OCLs were first treated with compactin (0.7–20 μM) for 24 h, after which compactin was washed out and the cultures maintained for an additional 24 h. Using this protocol, we found the effects of compactin to be reversible (Figs. 5A and 5B). Compactin markedly inhibited the formation of OCLs having more than 10 nuclei induced by RANKL but increased the number of OCLs having 4–9 nuclei 2-fold to 5-fold compared with control (Fig. 5A). When compactin was removed, the formation of OCL having more than 10 nuclei was completely restored in the cultures that had been treated with 2 μM of compactin for the first 24 h and partially restored in the cultures that had been treated with 7 μM of compactin for the first 24 h. The number of OCL with four to nine nuclei was restored to normal in cultures that had been treated with either 2 μM or 7 μM compactin for the first 24 h. Actin ring integrity in OCLs in the presence of osteoblastic cells was completely restored in cultures treated for the first 24 h with 7 μM compactin, a concentration that induced disruption of more than 90% of the actin rings (Fig. 5B). However, when the cultures were treated with 20 μM compactin for the first 24 h, fusion and actin ring integrity were only partially restored (Figs. 5A and 5B).
It has been reported that nitrogen-containing bisphosphonates or compactin induce apoptosis of macrophages by activating caspase, an essential enzyme in the pathway leading to apoptosis.(49) Therefore, we determined whether compactin induces caspase activation in OCLs and whether a caspase inhibitor (zVAD-fmk), which blocks the compactin-induced apoptosis of macrophages, prevents the effects of compactin on the actin ring integrity and pit formation. Compactin (7 μM) did not induce caspase activation in enriched OCLs (data not shown) and zVAD-fmk had no effect on compactin action on the actin ring integrity and pit formation (Figs. 6A and 6B). Figures 6D and 6E show that nuclear condensation did not occur in OCLs in which the actin ring had been disrupted by treatment with compactin. Most of the pOCs had died and disappeared from the plates in the absence of additives within 24 h.(45) Some of them died by apoptosis, as judged by Hoechst staining of DNA in OCLs (data not shown). The zVAD-fmk partially blocked the apoptosis of pOCs and allowed the survival and the fusion of pOCs (Fig. 6C). The zVAD-fmk slightly stimulated the fusion of pOCs induced by RANKL. However, the inhibitory effect of compactin on the fusion of pOCs by RANKL was not affected by the addition of zVAD-fmk (Fig. 6C). The treatment with compactin for 24 h did not induce apoptosis of pOCs.
Compactin inhibits IL-1β-, 1α,25(OH)2D3-, and PTH-stimulated 45Ca release in organ cultures
IL-1β, 1α,25(OH)2D3, and PTH stimulate bone resorption in organ cultures resulting in release of previously incorporated 45Ca. The effect of compactin on the 45Ca release elicited by these agents in fetal rat limb bones was examined. Compactin (0.1–2 μM) inhibited PTH-stimulated 45Ca release in a dose-dependent manner but did not significantly affect basal 45Ca release that probably results from the action of endogenous factors in the medium (Fig. 7A). Compactin (0.3 μM) also inhibited IL-1β- and 1α,25(OH)2D3-stimulated 45Ca release by about 50–60% (Fig. 7B). When bones were treated with or without compactin for 72 h and then the medium was replaced with fresh medium without compactin, the response to a subsequent 72-h treatment with PTH was not significantly different in bones that had been pretreated with compactin, again showing the reversibility of the inhibitory effect of compactin on bone resorption (Fig. 7C).
Simvastatin, pravastatin, and lovastatin, which are isolated from microbial metabolites, are used clinically as antihyperlipidemic drugs for the treatment of arteriosclerosis. The molecular target of these compounds is the rate-limiting enzyme, HMG-CoA reductase, which catalyzes the synthesis of MVA from acetyl-CoA.(37,38) These HMG-CoA reductase inhibitors were reported to induce apoptosis of many animal cells by inhibiting the synthesis of MVA and its downstream products FPP and GGPP, which are substrates of prenyl transferases, and subsequently prevent the prenylation of target proteins.(41,42) Nitrogen-containing bisphosphonates such as YM175 and ibandronate are known to inhibit sterol biosynthesis.(50,51) The compounds also were shown to induce apoptosis of J774 macrophage cell lines.(52) Luckman et al. recently have reported that nitrogen-containing bisphosphonates and compactin suppress osteoclastic bone resorption by inducing apoptosis of OCLs.(43) It was suggested that the apoptosis induced by these two classes of compounds is caused by the inhibition of the synthesis of MVA, FPP, and GGPP, preventing the prenylation of target proteins. In the current study, our findings showed that the HMG-CoA reductase inhibitor compactin reversibly suppresses bone resorption by inhibiting the fusion of pOCs into OCLs and by disrupting the actin rings in OCLs at concentrations that do not induce apoptosis of OCLs. This suggests that apoptosis of osteoclasts is not necessary for compactin to have an effect on the fusion of pOCs and on the mature osteoclasts function.
Although compactin was reported to inhibit osteoclastic bone resorption by inducing apoptosis of OCLs, the mechanism of the inhibitory effect was not established.(43) We therefore investigated the effect of compactin on OCL differentiation and function. Compactin inhibited OCL formation in the coculture of bone marrow cells and osteoblastic cells with 1α,25(OH)2D3. In the cocultures, OCLs are differentiated through two stages.(44,53) The first stage is the differentiation of precursors into mononuclear TRAP-positive pOCs, and the second later stage is the multinucleation of pOCs into OCLs. Compactin inhibited the late stage but not the early stage in the coculture for osteoclastogenesis. In the fusion assay system using pOCs isolated from cocultures for osteoclast formation, compactin also inhibited the fusion of pOCs induced by osteoblastic cells as well as by cytokines such as RANKL, M-CSF, and IL-1β in the absence of osteoblastic cells. Although compactin was more effective on OCL (>4 nuclei) formation in the presence of osteoblasts than in the presence of cytokines alone, the formation of large OCLs (>10 nuclei) was inhibited by compactin at the same concentrations that inhibited the fusion of pOCs induced by osteoblastic cells. The reason why there are more large OCLs under conditions in which cytokines are used may be the high concentrations of cytokines (100 ng/ml), which promote the fusion of pOCs. These results suggest that compactin acts directly on pOCs without a dependence on osteoblastic cell-mediated effects and then inhibits the fusion stage of pOCs into OCLs. Fisher et al. recently have reported that addition of lovastatin, during the last 5–7 days of the cocultures of mouse bone marrow and osteoblastic cells (MB1.8 cell line) for osteoclastogenesis eliminates large OCLs (>250 μm) from the cultures.(54) This effect of lovastatin is likely a result of inhibition of the fusion of pOCs into large OCLs. MVA completely prevented the inhibitory effect of compactin on the fusion of pOCs, indicating that the inhibitory effect of compactin results from the inhibition of the synthesis of MVA. Of its downstream metabolites, FPP and GGPP also prevented the effect of compactin; however, squalene did not. This suggests that the inhibition of the fusion by compactin is a consequence of the inhibition of the farnesylation and geranylgeranylation of target proteins by protein transferases and further suggests that signaling pathways related to the prenylation of target proteins are involved in the process of fusion of pOCs.
In the previous studies of apoptosis of osteoclasts induced by compactin or nitrogen-containing bisphosphonates, the exact mechanism by which the morphological changes occur was not established. One feature of morphological changes induced by inhibitors of protein prenylation is the breakdown of the actin cytoskeleton.(55,56) Therefore, we investigated the relationship between the morphological changes in osteoclasts and inhibition of bone resorption by compactin. When OCLs together with osteoblastic cells were replated on culture dishes, more than 80% of the OCLs spread out and formed ringed structures of F-actin (actin rings) within 4 h. We found that compactin caused actin ring disruption in OCLs and constriction of their size at the same concentration that inhibited bone resorption but had no effect on spreading and actin ring formation in OCLs just after OCLs were replated. An actin ring–disruptive effect of compactin also was observed in OCLs formed by RANKL from pOCs in the absence of osteoblastic cells. Distinct structural changes in the stress fibers of F-actin in osteoblasts were not observed. These facts indicate that compactin directly and specifically disrupts the actin rings in OCLs. Compactin derivatives such as lovastatin and simvastatin also induced disruption of the actin rings in OCLs. The inhibitory activity of these derivatives has a similar concentration dependence to that of these compounds on the enzymatic activity of HMG-CoA reductases, indicating that the decrease of synthesis of MVA and its downstream metabolites is involved in the disruption of actin ring by these inhibitors. The disruptive effect of compactin was prevented completely by MVA and GGPP, respectively, but only partially (to 60% of MVA and GGPP restoration) by FPP. The effects of these three metabolites on the inhibition of bone resorption by compactin are the same as those on disruption of actin rings by compactin. MVA and geranylgeraniol (GGOH which is metabolized to GGPP) were also reported to prevent the inhibition of bone resorption by lovastatin.(54) These facts indicate that the effect of compactin on the actin ring is mainly a result of the inhibition of prenylation by geranylgeranylated proteins. GGPP were known to involve in the prenylation of GTP-binding proteins such as Rac, Rho, and Rab, which play a key role in the regulation of cytoskeletal function and vesicular trafficking.(57) Rho and Rab proteins have been reported to be involved in bone resorption by regulating cytoskeletal organization in osteoclasts.(58,59) Cytoskeletal rearrangement in osteoclasts is required for their attachment onto bone surfaces and the polarization of cytoplasmic structures leading to the formation of actin rings and the ruffled border to resorb mineralized bone.(11) The potency of the effect of compactin on the actin ring in OCLs corresponded with that on osteoclastic pit formation. Our findings indicate that compactin inhibits bone resorption by disrupting actin rings in osteoclasts. Further, the findings suggest that the geranylgeranylation of GTP-binding proteins is probably required for actin ring stability, or that the accumulation of nongeranylgeranylated proteins somehow disrupts actin ring formation and in this way prevents osteoclastic bone resorption. In a study to examine the downstream events of signal transduction of geranylgeranylated proteins, Fisher et al. found that lovastatin activates a 34-kDa kinase in murine osteoclasts and the activation is blocked by GGOH or MVA.(54) Activation of 34 kDa kinase may play an important role in the effect of lovastatin on bone resorption. Further studies are required to determine which geranylgeranylated proteins and which signaling pathways downstream of the proteins are involved in actin ring stability in osteoclasts.
Many reports have shown that the inhibition of prenylation of GTP-binding proteins can lead to both actin cytoskeletal changes and apoptotic cell death.(41,42) Recently, Coxon et al. reported that ALN- or compactin-induced apoptosis of macrophage J774 cell lines is accompanied by the activation of caspase 3–like enzymatic activity (a common feature of apoptosis), which indicates an ongoing irreversible process leading toward cell death(49) Although ALN-induced apoptosis in osteoclasts is prevented by zVAD-fmk, the inhibitory effect of ALN on bone resorption is not.(60) Koshihara et al. also reported that ALN reversibly induces cytoskeletal changes (disruption of actin rings) in osteoclasts.(61) These facts suggest that disruption of actin ring integrity in osteoclasts and apoptosis of osteoclasts can each result in the inhibition of bone resorption but the processes are not necessarily linked. Our results also suggest that the inhibitory effect of compactin on the bone resorption is not a result of apoptosis. Treatment of OCLs with 7 μM compactin for 24 h did not induce caspase activation in enriched preparation of mature OCLs. The effect of compactin on actin ring and pit formation was not prevented by zVAD-fmk. We also did not observe nuclear condensation, a common characteristic of apoptotic cells, in OCLs in which the actin ring had been disrupted by treatment with 7 μM compactin. The pOCs are partially lost by apoptosis. Although the zVAD-fmk partially could prevent the apoptosis of pOCs, allowing the survival of and the fusion of pOCs, it could not prevent the effect of compactin on the fusion. Thus, these results indicate that the effect of compactin on bone resorption is mainly a result of the disruption of the actin ring and the inhibition of fusion in our experiment condition and that apoptosis is not necessary for compactin effects on bone resorption. This conclusion is supported by the reversibility of the effect of compactin on the fusion of pOCs and actin ring integrity. In the organ cultures the stimulation of 45Ca release by IL-1β, 1α,25(OH)2D3, and PTH is the end result of a resorptive process that results from the fusion of pOCs and subsequent OCL formation. Because compactin did not significantly affect basal 45Ca release, the inhibitory effect of compactin on 45Ca release in this model may be a consequence of the inhibition of the fusion steps for osteoclastogenesis. The reversibility of the effect on resorption in the organ cultures also supports the conclusion that apoptosis also is not necessary for compactin inhibition of the fusion. The effects of FPP and MVA on the compactin-induced apoptosis of J774 macrophages are similar to those of FPP and MVA on actin rings in OCLs, but GGPP shows a different pattern.(43) GGPP completely prevented the compactin-induced disruption of actin rings but only partially (less than 50% of MVA) prevented compactin-induced apoptosis of J774 macrophages. This fact also suggests that the disruption of actin rings and apoptosis of OCLs induced by compactin are independent events.
In summary, our findings indicate that inhibitory effects of compactin on bone resorption (a) result from both the inhibition of the fusion of pOCs and the disruption of actin ring in OCLs; (b) that these effects are a consequence of the inhibition of prenylation of target proteins by prenyl protein transferases through a decrease in their substrates, including FPP and GGPP; and (c) that the prenylation of specific proteins modulates signaling pathways related to the fusion of pOCs and maintenance of actin ring integrity, and that alterations in these processes and structures can inhibit bone resorption without resulting in apoptosis.
We thank Dr. A. Endo (BioPharm Research Laboratories, Tokyo, Japan) and Dr. N. Nakamura (Sankyo CO., Tokyo, Japan) for the generous gifts of compactin and its derivatives (lovastatin and simvastatin), respectively.