Dr Dunford states that he has no conflicts of interest.
Inhibition of Protein Prenylation by Bisphosphonates Causes Sustained Activation of Rac, Cdc42, and Rho GTPases
Article first published online: 23 JAN 2006
Copyright © 2006 ASBMR
Journal of Bone and Mineral Research
Volume 21, Issue 5, pages 684–694, May 2006
How to Cite
Dunford, J. E., Rogers, M. J., Ebetino, F. H., Phipps, R. J. and Coxon, F. P. (2006), Inhibition of Protein Prenylation by Bisphosphonates Causes Sustained Activation of Rac, Cdc42, and Rho GTPases. J Bone Miner Res, 21: 684–694. doi: 10.1359/jbmr.060118
- Issue published online: 4 DEC 2009
- Article first published online: 23 JAN 2006
- Manuscript Accepted: 19 JAN 2006
- Manuscript Revised: 23 DEC 2005
- Manuscript Received: 5 SEP 2005
N-BPs, which inhibit bone resorption by preventing prenylation of small GTPases, unexpectedly cause the accumulation of GTP-bound, unprenylated Rho family GTPases in macrophages and osteoclasts. In macrophages, this also leads to sustained, Rac-mediated activation of p38. The antiresorptive activity of N-BPs may therefore be caused at least in part, by the accumulation of unprenylated small GTPases, causing inappropriate activation of downstream signaling pathways.
Introduction: Nitrogen-containing bisphosphonates (N-BPs) are potent inhibitors of bone resorption that act by inhibiting farnesyl diphosphate synthase, thereby indirectly preventing the prenylation of Rho family GTPases that are required for the function and survival of bone-resorbing osteoclasts. However, the effect that these drugs have on the activity of Rho family GTPases has not been determined.
Materials and Methods: The effect of N-BPs on the activity of Rho family GTPases in J774 macrophages and osteoclasts was measured using a pull-down assay to isolate the GTP-bound forms. The effect of N-BPs, or decreasing Rac expression using siRNA, on downstream p38 activity was evaluated by Western blotting and apoptosis assessed by measurement of caspase 3/7 activity.
Results: Rather than inhibiting GTPase function, loss of prenylation after treatment with N-BPs caused an increase in the GTP-bound form of Rac, Cdc42, and Rho in J774 cells and osteoclast-like cells, which paralleled the rate of accumulation of unprenylated small GTPases. Activation of Rac also occurred with other inhibitors of prenylation of Rho-family proteins, such as mevastatin and the geranylgeranyl transferase I inhibitor GGTI-298. The Rac-GTP that increased after N-BP treatment was newly translated, cytoplasmic unprenylated protein, because it was not labeled with [14C] mevalonate, and the increase in Rac-GTP was prevented by cycloheximide. Furthermore, this unprenylated Rac-GTP retained at least part of its functional activity in J774 cells, because it mediated N-BP–induced activation of p38. Paradoxically, although risedronate induces apoptosis of J774 macrophages by inhibiting protein prenylation, the p38 inhibitor SB203580 enhanced N-BP–induced apoptosis, suggesting that Rac-induced p38 activation partially suppresses the pro-apoptotic effect of N-BPs in these cells.
Conclusions: N-BP drugs may disrupt the function of osteoclasts in vivo and affect other cell types in vitro by inhibiting protein prenylation, thereby causing inappropriate and sustained activation, rather than inhibition, of some small GTPases and their downstream signaling pathways.
Bisphosphonates are a class of drugs widely used to inhibit bone destruction in common diseases of the skeleton, such as postmenopausal osteoporosis, Paget's disease, and cancer-associated osteolysis. Although these agents are effective at blocking the bone-resorbing activity of osteoclasts, their exact molecular mechanisms of action have only recently become clear.(1) Several years ago, we showed that nitrogen-containing bisphosphonates (N-BPs) such as risedronate (RIS) and zoledronate (ZOL) prevent the post-translational prenylation of small GTPases such as Rac, Rho, Cdc42, Rap1, and Rab proteins in J774 macrophages.(2,3) This effect has since been described in osteoclasts in vitro and in vivo(4–8) and is the result of potent inhibition of FPP synthase,(6,9,10) an enzyme in the cholesterol biosynthetic pathway that is essential for the synthesis of the isoprenoid lipids FPP and GPP (the substrates used for posttranslational protein prenylation). Inhibition of FPP synthase therefore leads to loss of farnesylation and geranylgeranylation of newly translated small GTPases and hence to the intracellular accumulation of the unprenylated form of the proteins.(2–5,7,8,11) Prenylation of small GTPases (most of which are geranylgeranylated) is vital for their function as tightly regulated molecular switches, because it enables their localization to the correct subcellular membranes and interaction with regulatory proteins, hence allowing interaction with downstream signaling effectors. It is therefore likely that, by preventing their prenylation, N-BPs interfere with regulation of the activity of small GTPases that are required for osteoclast function. Although the roles of small GTPases in osteoclasts are only just becoming clear,(12) inhibition of protein geranylgeranylation by N-BPs in particular results in disruption of the osteoclast cytoskeleton and the resorptive activity of these cells, and induces osteoclast apoptosis.(4,13,14) Because FPP synthase is a ubiquitous enzyme in eukaryotic cells, inhibition of this enzyme by N-BPs, with subsequent disruption of the mevalonate pathway and loss of protein prenylation, also affects a wide variety of other cell types in vitro.(15–19)
Although it is clear that the antiresorptive effects of N-BPs are caused by inhibition of protein geranylgeranylation, the effect of these drugs and other inhibitors of protein prenylation on the activity of small GTPases and downstream signaling pathways has not been thoroughly studied, and it might be assumed that inhibitors of protein prenylation prevent the activation of small GTPases. In this study, therefore, we examined the effect of loss of prenylation caused by treatment with N-BPs (and other pharmacologic agents that inhibit protein prenylation, i.e., mevastatin and the protein:prenyl transferase inhibitors FTI-277 and GGTI-298), on the activity of the geranylgeranylated Rho family proteins Rho, Rac, and Cdc42 in J774 macrophages and osteoclasts.
MATERIALS AND METHODS
Clodronate (CLO), RIS, ZOL, NE58051, NE11808, and NE11809 were from Procter & Gamble Pharmaceuticals (Cincinnati, OH, USA) and prepared as described previously.(3) Mevastatin (Sigma) was converted from the lactone to the active acid as described previously.(3) [14C]mevalonic acid lactone was from Perkin Elmer Life Sciences, and solvent was removed before use by evaporating under nitrogen. FTI-277 and GGTI-298 were a kind gift from Dr Saïd Sebti, University of South Florida. Other reagents were from Sigma Chemical, unless stated otherwise.
J774 macrophages were obtained from the European Collection of Animal Cell Cultures and were maintained in DMEM supplemented with 10% (vol/vol) FCS (Harlan Sera-Laboratories), 100 U/ml penicillin, 100 μg/ml streptomycin, and 1 mM l-glutamine in a humidified atmosphere of 5% CO2. Macrophage-colony stimulating factor (M-CSF)–dependent murine bone marrow macrophages were obtained by flushing the bone marrow from the tibias and femora of adult male C57BL/6 mice into 10-cm petri dishes (Falcon) and cultured in α-MEM (Life Technologies) containing 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM glutamine, 10% (vol/vol) FCS, and 100 ng/ml recombinant mouse M-CSF (R&D Systems). After 3 days, the adherent cells were removed by trypsinization and reseeded into 6-well plates at 1 × 106 cells/well in fresh α-MEM containing 10% (vol/vol) FCS plus 100 ng/ml M-CSF. All further cultures of these cells were performed in the presence of 100 ng/ml M-CSF. Rabbit marrow cells were isolated from 3-day-old New Zealand white rabbits as previously described.(4) The cells were seeded into 10-cm plates (3.5 × 107 cells/plate) or 6-well plates (4.5 × 106 cells/well; Costar) in α-MEM supplemented as above, plus 1 × 10−8 M 1,25(OH)2 vitamin D3. Cells were maintained for 7–8 days with a 50% change of medium every third day and washed extensively with PBS. The cultures contained >95% TRACP+ cells (prefusion mononuclear cells and multinucleated osteoclast-like cells capable of bone resorption). Human osteoclast-like cells were generated from peripheral blood mononuclear cells (PBMCs) of healthy volunteers. The PBMCs were isolated over Lymphoprep (Axis-Shield) and seeded into 6-well plates in α-MEM supplemented as above at 2 × 106 cells/well and treated with 20 ng/ml M-CSF. The medium was changed after 3 days, and when the cells were 80% confluent, they were treated with 20 ng/ml M-CSF plus 100 ng/ml RANKL (Peprotech). The cells were fed by replacement of the medium 3 days later. After 4–5 days in RANKL, >80% of cells in the culture were multinucleated, TRACP+ osteoclasts, at which stage they were treated with prenylation inhibitors as outlined below.
Assays for the GTP-bound form of small GTPases
Cells seeded into 6-well plates were treated with prenylation inhibitors for 12–72 h and lysed in 500 μl ice-cold lysis buffer (25 mM HEPES pH 7.5, 150 mM NaCl, 1% [vol/vol] Igepal CA-630, 10% [vol/vol] glycerol, 10 mM MgCl2, 1 mM EDTA, and 1 mM sodium orthovanadate) containing mammalian protease inhibitor cocktail. Before being lysed, osteoclasts generated from human PBMCs were starved of serum and M-CSF/RANKL for 2 h and stimulated with 100 ng/ml RANKL or vehicle for 15 minutes. In some cases, J774 macrophages were treated for 15 minutes with 100 ng/ml M-CSF (R&D Systems) before lysis. Lysates were cleared by centrifugation at 14,000g for 10 minutes, and 30 μl of these cell lysates was analyzed by Western blotting to detect the total levels (i.e., both active and inactive forms) of Rac, Cdc42, or RhoA. For the pull-down assays, 450 μl of lysate was incubated with 2.5 μg PAK-agarose to isolate Rac-GTP and Cdc42-GTP or rhotekin-agarose to isolate Rho-GTP (Upstate Biotech), at 4°C with shaking for 1 h. The agarose/protein complex was recovered by centrifugation and washed three times in ice-cold lysis buffer and resuspended in 10 μl 2× Laemmli sample buffer. Solubilized proteins were resolved on 12% polyacrylamide-SDS gels, transferred onto polyvinylidene fluoride (PVDF) membranes, and analyzed by Western blotting using 0.5 μg/ml anti-Rac (Upstate Biotech), anti-Cdc42 (BD Transduction Laboratories), or anti-Rho A (26C4; Santa Cruz Biotech) antibodies. After hybridization with peroxidase-coupled secondary antibodies (Calbiochem) and incubation with SuperSignal Extended Dura reagent (Pierce), bands were detected on a BioRad FluorSMax imager and quantified using BioRad Quantity One software. Quantified data were the amount of Rac-GTP expressed as percent of control, and statistical analysis was carried out by Mann-Whitney U-test.
[14C]mevalonate labeling of small GTPases
J774 cells were seeded at a density of 1 × 106/well in 6-well plates. The following day, cells were cultured in the presence of 275 KBq/ml [14C]mevalonolactone and 5 μM mevastatin for 24 h; 100 μM RIS or PBS (vehicle control) was added to the medium, and incubation was continued for a further 24 h. The cells were harvested and lysed as described above for the pull-down assays. To ensure precipitation of Rac/Cdc42-GTP in control lysates, 100 μM GTPγS was added to control lysates and incubated for 15 minutes at 30°C in the presence of 10 mM EDTA, and then MgCl2 was added to a final concentration of 60 mM. Rac/Cdc42-GTP was isolated using PAK-agarose beads, and blots were analyzed for Rac as described above. The blots were allowed to dry and [14C]-labeled Rac/Cdc42 on the blots was detected by exposure to a Kodak phosphorimaging screen and visualized using a BioRad FX Phosphor Imager.
Separation of cytosolic and membrane proteins
J774 cells were incubated in the presence or absence of 100 μM RIS in 10-cm dishes for 30 h, and then cytosolic and membrane proteins were separated by ultracentrifugation as previously described.(20) Samples were electrophoresed on 12% SDS-polyacrylamide gels, proteins were transferred to PVDF membranes, and Rac was detected by Western blotting.
Western blotting for phospho-p38
J774 cells were seeded at 0.5 × 106/well in 12-well plates. The following day, cultures were treated with prenylation inhibitors for 12–72 h, and the cells were homogenized on ice in 100 μl of buffer (10 mM potassium phosphate, pH 7.4, 137 mM NaCl, 0.1% [wt/vol] SDS, 0.5% [wt/vol] sodium deoxycholate, 1% [vol/vol] Triton X-100, 1 mM sodium orthovanadate, 1 mM EDTA, protease inhibitors, and phosphatase inhibitor cocktail 1). Lysates were cleared and assayed for protein using the BCA assay (Pierce, Rockford, IL, USA). Thirty micrograms of protein from each sample was resolved by SDS-PAGE on 12% polyacrylamide-SDS gels followed by transfer onto PVDF membranes. The blots were hybridized with anti-dual-phosphorylated p38 (Cell Signaling Technology 9211), and bands were detected as described above. Blots were stripped for 10 minutes at 50°C in 60 mM Tris-HCl, pH 6.7, 2% (wt/vol) SDS, and 1 mM DTT and reanalyzed for total p38 using 0.2 μg/ml anti-p38 antibody (Santa Cruz Biotechnology). Blots were quantified, and levels of phospho-p38 were expressed as percent of control. Data were analyzed by Mann-Whitney U-test.
Knockdown of Rac1 expression using siRNA
J774 cells were transfected with 50 pmol/106 cells Smartpool siRNA against Rac1 (Upstate) by electroporation using an Amaxa Biosystems Nuclefector. Control cells were transfected with a nonspecific control siRNA. Cells were seeded into 6-well plates at 1 × 106 cells/well and incubated for 24 h before treatment with 100 μM RIS for a further 30 h. Cells were lysed in the appropriate homogenization buffer for measuring Rac-GTP or blotting for phospho-p38, as described above.
Quantification of caspase activity
Caspase 3/7 activity was measured by detecting cleavage of a fluorogenic caspase substrate using the Apo-ONE homogenous caspase-3/7 assay kit (Promega). J774 cells were seeded at 1 × 104 cells/well into black, clear bottom 96-well plates (Corning International). The following day, cells were cultured for 24 h in 50 μl medium with 0–100 μM RIS in the presence or absence of 20 μM SB203580 (Calbiochem), a p38 inhibitor. To initiate the assay, 50 μl of caspase 3/7 substrate was added to each well. Plates were incubated at room temperature for 90 minutes, and fluorescence was measured at 530 nM (excitation wavelength of 460 nM) using a LabTek fluorescence plate reader.
Bisphosphonate inhibitors of protein prenylation cause activation of Rac, Cdc42, and Rho
GTP-bound Rac, Cdc42, and Rho in lysates of J774 cells were measured using a pull-down assay, using the binding domain of the effectors PAK or rhotekin conjugated to agarose beads. The level of Rac-GTP was significantly increased in J774 cells after treatment for 30 h with 100 μM RIS or ZOL. In contrast, treatment with up to 1 mM CLO did not increase the level of Rac-GTP (Figs. 1A and 1B). Moreover, pretreatment of J774 cells with RIS did not inhibit, but potentiated, the increase in Rac-GTP levels resulting from treatment with 100 ng/ml M-CSF for 15 minutes (Fig. 1C). RIS and ZOL, but not CLO, also increased the levels of Cdc42-GTP (Fig. 1D) and Rho-GTP (Fig. 1E) in J774 cells after treatment for 30 h. The increase in GTP-bound Rac, Cdc42, and Rho after treatment with N-BPs was not restricted to J774 cells, because treatment of M-CSF–dependent murine macrophages and rabbit or human osteoclast-like cells for 30 h with 100 μM potent N-BPs also increased levels of Rac-GTP, Cdc42-GTP, and Rho-GTP (Figs. 2C, 2D, and 5C and data not shown). Furthermore, the increase in GTP-bound Rac, Cdc42, and Rho after RIS or ZOL treatment was not the result of increased levels of total (i.e., both active and inactive) Rac, Cdc42, or Rho protein, because total levels of these GTPases in cell lysates usually decreased slightly after treatment with BPs, because of inhibition of cell proliferation resulting in a decrease in total protein content.
A comparison was made between two pairs of structurally related bisphosphonates with markedly different potencies for inhibiting FPP synthase (RIS/NE58051 and NE11808/NE11809(10)); 100 μM RIS or NE11808, both potent inhibitors of FPP synthase, caused a marked increase in Rac-GTP after 30-h treatment of J774 cells (Figs. 2A and 2B) or M-CSF–dependent macrophages (Fig. 2C). However, 100 μM NE58051 and NE11809, which are structurally similar to RIS and NE11808, respectively, but are weak inhibitors of FPP synthase,(10) did not affect Rac-GTP levels. Similar effects were observed for Cdc42 and Rho in J774 cells, M-CSF–dependent macrophages, and rabbit osteoclast-like cells (Fig. 2D and data not shown).
Other inhibitors of protein prenylation cause activation of Rac, Cdc42, and Rho
Like RIS and ZOL, treatment of J774 cells for 30 h with two structurally unrelated inhibitors of prenylation of Rho proteins, GGTI-298 (which specifically prevents protein geranylgeranylation) and mevastatin (which prevents geranylgeranylation and farnesylation), also caused an increase in Rac-GTP (Fig. 3A). However, a specific inhibitor of protein farnesylation, FTI-277, had no effect on the level of Rac-GTP.
To further examine whether Rac activation was caused by inhibition of protein geranylgeranylation, J774 cells were treated with RIS in the absence or presence of FOH (which restores farnesylation) or GGOH (which restores geranylgeranylation). FOH or GGOH alone did not affect the level of Rac-GTP (Figs. 3B and 3C). Whereas FOH had little effect on RIS-induced Rac activation, GGOH completely prevented the RIS-induced accumulation of Rac-GTP (Figs. 3B and 3C).
Activated Rac is newly synthesized and unprenylated in RIS-treated cells
Because protein prenylation is a post-translational modification, inhibition of protein synthesis should prevent the RIS-induced accumulation of unprenylated proteins such as Rac. J774 cells were therefore treated with 100 μM RIS in the absence or presence of cycloheximide (CHX), and cell lysates were analyzed for the presence of Rac-GTP. CHX (0.25 μM; a concentration that did not affect viability of J774 cells) had no effect on Rac activation alone but completely prevented the RIS-induced accumulation of Rac-GTP (Fig. 4A), confirming that Rac activation by RIS was dependent on the presence of newly translated Rac. To confirm that the Rac-GTP detected after treatment with inhibitors of protein prenylation (Figs. 1–3) was the unprenylated form, J774 macrophages were metabolically labeled with [14C]mevalonolactone for 24 h to allow incorporation of [14C]mevalonate into prenylated proteins. The cells were cultured without or with 100 μM RIS for a further 30 h, and cell lysates were analyzed for the level of Rac-GTP. As expected, Rac-GTP was detected in lysates from RIS-treated cells and in lysates from untreated cells to which GTPγS had been added to promote binding to the PAK-agarose (Fig. 4Bi). In lysates from untreated cells, the GTPase protein precipitated with PAK-agarose (i.e., Rac and Cdc42) was strongly 14C-labeled (i.e., prenylated; Fig. 4Bii). However, in RIS-treated cells, the GTPase protein was not radiolabeled (i.e., it was unprenylated). Because prenylation is essential for membrane localization of Rac, it is likely that the newly synthesized, unprenylated Rac-GTP that accumulates after treatment with RIS was cytosolic. Indeed, we found that treatment of J774 cells for 30 h with 100 μM RIS substantially increased the amount of Rac in the cytosol compared with untreated cells (Fig. 4C).
Bisphosphonate inhibitors of protein prenylation cause activation of p38
Because p38 is a downstream effector of Rac signaling, we examined the effect of inhibitors of protein prenylation on p38 activation. Treatment of J774 cells for 30 h with 100 μM RIS or ZOL, but not 500 μM CLO, caused a significant increase (240% and 260% of control for RIS and ZOL, respectively; both p < 0.01; n = 5) in the level of phospho-p38, without affecting total p38 levels (Fig. 5A). The increase in p38 phosphorylation during RIS treatment was evident after ∼12 h and steadily increased up to 62 h of treatment (Fig. 5B). This slow and sustained activation closely correlated with the time-course for activation of Rac, which was detected only after ∼12 h or more of RIS treatment (Fig. 5B) and differed from the very rapid and more robust (10- to 15-fold increase) activation of p38 detected in response to LPS treatment or osmotic shock (data not shown). In contrast, 100 μM RIS had no effect on phosphorylation of p38 in human osteoclast-like cells, despite causing a robust increase in Rac-GTP levels (Fig. 5C). Conversely, 100 ng/ml RANKL had no effect on basal or RIS-stimulated Rac activity in human osteoclast-like cells, but markedly increased levels of phospho-p38. This activation of p38 was not substantially affected by either RIS or CLO (Fig. 5C). In accordance with the results in J774 cells, CLO was unable to activate either Rac or p38 in human osteoclast-like cells.
Bisphosphonates activate p38 by preventing geranylgeranylation of Rac1
Because Rac was activated only by treatment of cells with BPs that potently inhibit FPP synthase and prevent protein prenylation (Fig. 2), we examined the ability of structurally related pairs of bisphosphonates (RIS/NE58051 and NE11808/NE11808) to activate p38. Treatment of J774 cells with 100 μM of the potent FPP synthase inhibitors RIS or NE11808 caused a significant increase in p38 phosphorylation compared with the analogs (NE58051 and NE11809), which are weak inhibitors of FPP synthase (Figs. 6A and 6B). Furthermore, because restoring protein geranylgeranylation in cells with GGOH prevented RIS-induced activation of Rac (Figs. 3B and 3C), we also examined whether GGOH or FOH could prevent RIS-induced activation of p38. Neither FOH nor GGOH alone affected p38 phosphorylation (Fig. 6C). However, GGOH significantly reduced the RIS-induced activation of p38, whereas FOH had no effect, indicating that RIS activates p38 by preventing the geranylgeranylation of protein(s) such as Rac. In support of this, mevastatin and GGTI-298, which both inhibit protein geranylgeranylation, caused activation of p38 in J774 cells, whereas FTI-277 (which specifically prevents protein farnesylation) had no effect (Fig. 6D). To identify whether Rac was the geranylgeranylated protein mediating RIS-induced activation of p38, we measured p38 activation in J774 cells that had been transfected with siRNA to Rac1 to block the expression of this GTPase. Transfection of J774 cells with siRNA to Rac effectively decreased the total levels of Rac and significantly reduced the accumulation of Rac-GTP after treatment with 100 μM RIS (Figs. 7A and 7B). Moreover, siRNA to Rac completely prevented RIS-induced stimulation of p38 phosphorylation, indicating that this activation of p38 is mediated by Rac-GTP (Figs. 7C and 7D).
Inhibition of p38 enhances RIS-induced apoptosis of J774 cells
In accord with our previous studies, treatment of J774 cells for 24 h with 25–100 μM RIS caused a concentration-dependent increase in apoptosis, detected by an increase in caspase3/7 activity (Fig. 7E). The p38 inhibitor 20 μM SB203580 alone had little effect on J774 apoptosis. However, 20 μM SB203580 significantly increased RIS-induced apoptosis, suggesting that activation of p38 by RIS treatment (Figs. 5 and 6) partially suppressed the pro-apoptotic effect of RIS. In human osteoclast-like cells, treatment for 48 h with RIS had no effect on caspase 3/7 activation in the presence or absence of SB203580 (data not shown).
N-BPs disrupt the function of bone-resorbing osteoclasts by inhibiting FPP synthase and thereby preventing the prenylation of small GTPases.(10,13) In particular, inhibition of geranylgeranylation of Rho family proteins seems to account for the effects of N-BPs, because inhibition of bone resorption can be overcome by replenishing osteoclasts with a substrate for geranylgeranylation and can be mimicked by blocking geranylgeranylation (using an inhibitor of GGTase I) or blocking the function of the Rho family GTPases Rho, Rac, and Cdc42 using toxin B from Clostridium difficile.(4,13,14,21) It might therefore be assumed that N-BPs, by preventing protein prenylation, disrupt the interaction of Rho proteins with cell membranes, thereby preventing their activation by guanine nucleotide exchange factors (GEFs) and consequently also preventing their participation in intracellular signaling processes required for osteoclast function. However, contrary to these expectations, we show in this study that N-BP drugs enhance, rather than inhibit, the M-CSF–stimulated increase in Rac-GTP in J774 macrophages. Moreover, N-BPs alone cause a marked increase in the active, GTP-bound form of Rho, Rac, and Cdc42 in J774 macrophages, M-CSF–dependent mouse macrophages, and both rabbit and human osteoclast-like cells. This effect seems to occur because of disruption of prenylation of these proteins, because only N-BPs that are potent inhibitors of protein prenylation, such as RIS, ZOL, and NE11808, activate the Rho-family GTPases. In support of this, we found that other inhibitors of prenylation of Rho proteins, such as mevastatin and GGTI-298, also increased Rac-GTP in J774 cells and that restoring protein geranylgeranylation by supplementing cells with GGOH prevented the N-BP–induced increase in Rac-GTP. Furthermore, the Rac-GTP that accumulated in cells treated with N-BP seemed to be newly synthesized, unprenylated Rac, because the increase in Rac-GTP occurred over a prolonged time period, consistent with the rate of accumulation of unprenylated protein. This was confirmed using metabolic labeling experiments, in which the Rac-GTP that was elevated after N-BP treatment had not incorporated [14C] mevalonate (and was therefore not prenylated), and by inhibiting protein synthesis with cycloheximide, which prevented the accumulation of unprenylated Rac and Rac-GTP. Together, these observations provide compelling evidence that loss of prenylation causes activation of Rac and other Rho-family GTPases.
A likely explanation for the accumulation of unprenylated Rho-family proteins in the GTP-bound form may be the difference in the ability of the prenylated and unprenylated forms to interact with regulatory proteins. Although GEFs do not seem to interact with unprenylated Rac,(22–24) and in any case are thought to act at the membrane after dissociation of guanine nucleotide dissociation inhibitors (GDIs), unprenylated Rac is also unable to interact with Rho-GDI (because the isoprenyl group is required for binding to a hydrophobic pocket on Rho-GDI(25)) and therefore intrinsic nucleotide exchange (i.e., release of GDP and binding of GTP) may occur.(22,26,27) Because intracellular GTP levels are much greater than GDP levels,(28) binding to GTP rather than GDP will be favored. Therefore, over time, unprenylated Rho-family GTPases are likely to accumulate in the GTP-bound form as a result of uncatalyzed levels of nucleotide exchange, provided that the rate of nucleotide exchange exceeds the rate of hydrolysis of GTP to GDP. Molnar et al.(24) have shown that, although unprenylated Rho and Rac hydrolyze GTP, unlike their prenylated counterparts, they are unable to interact with GTPase activating proteins (GAPs) in vitro, so hydrolysis of GTP will occur only at uncatalyzed rates. Because the unprenylated form of Cdc42 (unlike Rac) can still interact with GAPs in vitro,(24) it seems surprising that inhibition of prenylation also caused an increase in Cdc42-GTP. However, there is evidence to suggest that GAP activity is greatest at the plasma membrane,(24,29) and it is therefore likely that cytosolic, unprenylated Cdc42 does not interact with GAPs because of differences in subcellular localization.
The effect of inhibitors of protein prenylation on the activity of small GTPases in other studies is controversial. Some studies also show activation of Rho GTPases; for example, atorvastatin and cerivastatin increase levels of GTP-bound Rac in HUVECs.(30) Moreover, Porter et al.(31) found that simvastatin blocked FCS-stimulated RhoA activation at the cell membrane, but stimulated RhoA activity in whole cell lysates in cardiac myofibroblasts. Our interpretation of this data is that unprenylated, cytosolic RhoA becomes activated, similar to our findings. In contrast, other reports have shown that mevastatin inhibits LPS-induced activation of Rac in monocytes but has no effect on Rac activation alone(32) and that ALN inhibits LPA-stimulated activity of Rho activity and downstream effectors in human ovarian cancer cells.(33) The reasons for these discrepancies are unclear but could relate to differences in the way that Rho-family GTPases are regulated in different cell types. However, we consistently detected the activation of Rac, Rho, and Cdc42 by N-BPs in a variety of cell types, including J774 cells, primary mouse macrophages, rabbit and human osteoclast-like cells, and in MCF7, MDA-MB-231, and PC3 tumor cell lines (data not shown).
Although in our study, active, GTP-bound Rho-family GTPases clearly accumulate after treatment with prenylation inhibitors, the question arises of whether these proteins are able to activate downstream signaling pathways, despite their lack of prenylation. To date, there is only limited evidence that unprenylated Rho-family proteins can interact productively with a physiological effector.(34,35) Prenylated Rac and Cdc42 are known to interact with p21-activated kinase (PAK) family members, MLK3 and MEKK1/4,(36) which have been implicated in the activation of the p38 MAP kinase and SAP/JNK kinase signaling pathways.(37) We found that p38 MAPK was activated in J774 cells by inhibitors of prenylation (potent N-BP inhibitors of FPP synthase, mevastatin, and GGTI-298), but not by weak N-BP inhibitors of FPP synthase. Moreover, the activation of p38 was delayed but sustained, mirroring activation of Rac by N-BPs, and addition of GGOH prevented N-BP-induced p38 activation, confirming the role of inhibition of protein geranylgeranylation in this response. We further showed that Rac1 mediates the increase in p38 activity, because N-BP–induced p38 activation could be abolished by blocking Rac1 expression using siRNA. However, the identity of the Rac effector that mediates p38 activation remains to be determined. In agreement with our data, Nègre-Aminou et al.(38) showed that simvastatin enhanced the levels of phosphorylated p38 in smooth muscle cells in response to oxidative stress in a mevalonate-dependent manner. In addition, N-BPs activated p38 in MDA-MB-231 breast cancer cells, although the mechanisms involved were not determined.(39) In contrast, other studies found that inhibitors of prenylation had little effect on p38 activity; for example, neither the N-BP YM529 nor fluvastatin activated p38 in HL-60 cells and vascular smooth muscle cells, respectively.(40,41)
p38 has been implicated in apoptosis downstream of Rac signaling, in a variety of cell types, for example in UV-stimulated apoptosis in fibroblasts, oxidative stress-induced apoptosis in vascular smooth muscle cells, and radiation-induced apoptosis in Rat2 cells.(38,42,43) Sustained activation of p38 downstream of Rac could, therefore, also be a plausible pathway by which N-BPs induce apoptosis in J774 cells and osteoclasts.(44) However, we found that, in contrast to their effects of J774 cells, N-BPs had no effect on basal or RANKL-stimulated p38 activity in human osteoclast-like cells, indicating that activation of p38 is not involved in BP-induced osteoclast apoptosis. In support of this, it has been shown that p38 plays a crucial role in regulating osteoclast differentiation induced by a variety of factors, including RANKL, IL-1, and TGF-β,(45) but does not seem to play a role in regulating normal osteoclast function or apoptosis.(46) These results raise the possibility that unprenylated Rac cannot activate downstream effectors in osteoclasts, unlike in J774 cells. However, because N-BPs activated Rac but not p38, and RANKL induced p38 but had no effect on Rac activation in human osteoclast-like cells, it is possible that p38 does not lie downstream of Rac signaling in osteoclasts. With regard to J774 cells, we found that pharmacological inhibition of p38 with SB203580 actually led to a significant increase in N-BP–induced caspase activity in J774 cells, indicating that p38 activation seems to have an anti-apoptotic effect in these cells and that N-BP–induced apoptosis must occur through changes in signaling pathways other than p38. In agreement, it has also been found that in some other cell types that Rac/p38 signaling can exert anti-apoptotic effects, for example, in IL-3–dependent Baf3 cells.(47) Our results are also similar to those of Merrell et al.,(39) who found that blocking p38 activation augmented N-BP–induced growth inhibition in MDA-MB231 cells. The ability of p38 activation to act in a pro- or anti-apoptotic manner probably depends on the cell type and on the manner in which it is activated (i.e., a low level of chronic activation compared with a greater, short-lived activation). Our data therefore show that, at least in macrophages, the ability of N-BPs to induce apoptosis(44) is not caused by the increased p38 activation but must involve effects on other signaling pathways after loss of prenylation of small GTPases. At this stage, it remains unclear whether it is the unprenylated, GTP-bound form of Rho, Rac, or Cdc42 that mediates N-BP–induced apoptosis in J774 cells and osteoclasts or whether it results from loss of signaling from prenylated forms of these GTPases. In any case, these results indicate that N-BPs are able to induce conflicting signals (e.g., both pro- and anti-apoptotic signals), and the cellular outcome is likely to depend on the balance of these signals. It is therefore possible that N-BP treatment in combination with other agents (such as p38 inhibitors) could prove beneficial by shifting this balance to the most favorable outcome (e.g., potential antitumor effects of N-BPs).
In summary, we showed that N-Bs, which prevent the prenylation of small GTPases, unexpectedly cause the accumulation of GTP-bound, unprenylated Rho, Rac, and Cdc42 in macrophages and osteoclast-like cells. In J774 cells, this leads further to sustained, Rac1-mediated activation of p38. This extends our understanding of the molecular mode of action of this important class of drugs and raises the possibility that the effects of BPs may be caused, at least in part, by the accumulation of the unprenylated form of small GTPases (rather than the loss of pre-existing, prenylated small GTPases), thereby causing inappropriate activation of downstream signaling pathways. Another possible explanation is that the unprenylated Rac-GTP that accumulates exerts a dominant negative effect on prenylated Rac signaling by sequestering effectors in nonproductive cytoplasmic complexes. The possible role of unprenylated proteins in mediating the effects of N-BPs are supported by our previous finding that cycloheximide, which completely prevented the N-BP–induced accumulation of unprenylated Rac-GTP, effectively inhibited N-BP–induced apoptosis in J774 cells.(48) Further characterization of the effects of N-BPs on small GTPase-regulated signaling pathways in osteoclasts and other cells may help to identify novel pharmacological approaches for inhibiting bone resorption or for potentiating the antiresorptive or antitumor effects of these drugs.
This work was funded by Procter & Gamble Pharmaceuticals.
- 21998 Heterocycle-containing bisphosphonates cause apoptosis and inhibit bone resorption by preventing protein prenylation: Evidence from structure- activity relationships in J774 macrophages. J Bone Miner Res 13: 1668–1678., , , ,
- 72001 The molecular mechanism of action of the antiresorptive and antiinflammatory drug clodronate: Evidence for the formation in vivo of a metabolite that inhibits bone resorption and causes osteoclast and macrophage apoptosis. Arthritis Rheum 44: 2201–2210., , , ,
- 91999 Nitrogen-containing bisphosphonates inhibit isopentenyl pyrophosphate isomerase/farnesyl pyrophosphate synthase activity with relative potencies corresponding to their antiresorptive potencies in vitro and in vivo. Biochem Biophys Res Commun 255: 491–494., , , ,
- 141999 The role of geranylgeranylation in bone resorption and its suppression by bisphosphonates in fetal bone explants In vitro: A clue to the mechanism of action of nitrogen-containing bisphosphonates. J Bone Miner Res 14: 722–729., , ,