Abstract
Nitrogen-containing bisphosphonates are pyrophosphate analogs that have long been the preferred prescription for treating osteoporosis. Although these drugs are considered inhibitors of prenylation and are believed to exert their effects on bone resorption by disrupting the signaling pathways downstream of prenylated small GTPases, this explanation seems to be insufficient. Because other classes of prenylation inhibitors have recently emerged as potential antiviral therapeutic agents, we first investigated here the effects of bisphosphonates on simian virus 40 and adenovirus infections and, to our surprise, found that viral infections are suppressed by bisphosphonates through a prenylation-independent pathway. By in-house affinity-capture techniques, dynamin-2 was identified as a new molecular target of bisphosphonates. We present evidence that certain bisphosphonates block endocytosis of adenovirus and a model substrate by inhibiting GTPase activity of dynamin-2. Hence, this study has uncovered a previously unknown mechanism of action of bisphosphonates and offers potential novel use for these drugs.
Bisphosphonate (BP) drugs are pyrophosphate analogs that have long been the preferred prescription for treating resorption-related bone diseases, such as osteoporosis and different types of tumor-induced osteolysis (Heymann et al., 2004; Russell et al., 2007). The mechanism of action of the two types of BP drugs, non–nitrogen-containing BPs (NN-BPs) and nitrogen-containing BPs (N-BPs), are believed to be different. On the one hand, NN-BPs, such as etidronate (ETI), are metabolized into nonhydrolyzable ATP analogs that lead to apoptosis (Frith and Rogers, 2003; Roelofs et al., 2006). On the other hand, N-BPs, such as alendronate (ALN) and zoledronate (ZOL), are considered inhibitors of prenylation, which involves the transfer of an isoprenoid lipid moiety from farnesyl pyrophosphate or geranylgeranyl pyrophosphate onto a C-terminal cysteine residue of proteins with a CXXX box, a characteristic prenylation motif (Frith and Rogers, 2003; Roelofs et al., 2006). At the molecular level, N-BPs target farnesyl diphosphate synthase (FPPS), thereby inhibiting the production of the substrates required for prenylation of small GTPases (Reszka et al., 1999; Coxon et al., 2000).
Prenylation inhibitors other than N-BPs have emerged recently as potential antiviral therapeutic agents (Einav and Glenn, 2003). For example, statins such as lovastatin (LOV) inhibit HIV-1 infection through HMG-CoA reductase inhibition (del Real et al., 2004) or respiratory syncytial virus replication by inhibiting RhoA geranylgeranylation (Gower and Graham, 2001). Other examples include the farnesyl transferase inhibitors, such as BZA-5B and FTI-277, which have been shown to inhibit the prenylation of CXXX box found in the hepatitis δ virus large δ antigen, leading to the suppression of production of hepatitis δ virus particles in different systems (Glenn et al., 1998; Bordier et al., 2002). In theory, prenylation inhibitors could also exert antiviral effects on other viruses with key molecules containing the CXXX box, such as the RNA polymerase protein of hepatitis A virus and animal enterovirus (Einav and Glenn, 2003) and the UL32 protein of herpes simplex virus, which has a lower titer accompanied by decreased viral protein synthesis in prenylation inhibitor-treated cells (Farassati et al., 2001).
With respect to N-BPs, there are still few reports on the use of N-BPs in the context of antiviral therapy. One article showed that N-BPs inhibit the growth of human T-cell leukemia virus type I-infected T-cell lines by inhibiting prenylation through FPPS targeting (Ishikawa et al., 2007), whereas others report the clinical use of N-BPs to reverse HIV-associated osteoporosis without in-depth discussion of the role of prenylation therein (Qaqish and Sims, 2004; Negredo et al., 2005). Hence, there seems to be a potential for dual use of N-BPs as osteopathic and antiviral drugs, although the underlying mechanisms remain unclear.
Thus far, it has been assumed that N-BPs exert their effects on bone resorption by disrupting the signaling pathways downstream of prenylated small GTPases (Frith and Rogers, 2003). However, this explanation seems to be insufficient, because certain unprenylated small GTPases are found in their active form even in the presence of N-BPs (Dunford et al., 2006), whereas certain N-BPs exert effects that are not reversible by replenishing cells with substrates for protein prenylation, which suggests that N-BPs may have cellular targets other than FPPS (van Beek et al., 2003). Given the emerging roles of prenylation inhibitors in antiviral therapeutics and the wide but relatively unclear potential of N-BPs therein, here, we carried out an investigation on N-BPs in the context of antiviral therapy. First, we found that simian virus 40 (SV40) and adenovirus could be inhibited by N-BPs through a prenylation-independent pathway, which supports the existence of cellular N-BP targets other than FPPS. Hence, to search for this new N-BP molecular target, we used high-performance latex beads (glycidylmethacrylate-covered glycidylmethacrylate-styrene copolymer core beads) that we developed in-house, which is an affinity purification matrix that exhibits chemical and physical stability, high capacity for ligand fixation, low nonspecific protein binding, and high purification efficiency (Shimizu et al., 2000; Sakamoto et al., 2009). Using ALN-conjugated beads, we isolated dynamin-2 (DYN2) and sorting nexin 9 (SNX9) (Lundmark and Carlsson, 2003). Our data show that certain types of N-BPs and NN-BPs target GTPase activity of DYN2, leading to defective scission and release of endocytic vesicles into the cytoplasm. Hence, this study unveils a novel prenylation-independent pathway that is targeted by both types of BPs in inhibiting viral infections and may contribute to widening the scope of use of these efficacious drugs beyond bone diseases.
Materials and Methods
Cell Culture and Reagents.
Mouse macrophage cell line RAW264.7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS). Human 293FT cells were cultured in Dulbecco's modified Eagle's medium containing 10% FBS and G418 (Geneticin). Human HeLa cells were cultured in minimum essential medium containing 10% FBS.
ALN and ETI were obtained from Asahi Kasei (Shizuoka, Japan), and ZOL was obtained from Novartis Pharma (Tokyo, Japan). Small unilamellar vesicles (SUVs) were prepared essentially as described previously (Sweitzer and Hinshaw, 1998; Catimel et al., 2009), in the following way: bovine total brain extract (Sigma-Aldrich, St. Louis, MO) was dissolved in chloroform/methanol (1:2). The solvent was evaporated, and lipid was redissolved in the appropriate buffer to a final concentration of 25 or 1 mg/ml by sonication. Likewise, SUVs of defined compositions were prepared using phosphatidylcholine (PC) or PC/phosphatidylinositol (3,4,5)-triphosphate (PIP3) (9:1).
Affinity Purification Using ALN-Conjugated Beads.
The cytoplasmic extract of RAW264.7 cells was prepared as described previously (Dignam et al., 1983) and was dialyzed with buffer A [10 mM HEPES, pH 7.9, 10% glycerol, 100 mM KCl, 0.2 mM EDTA, 1 mM dithiothreitol (DTT), and 1 mM phenylmethylsulfonyl fluoride (PMSF)]. ALN was conjugated through the amino group to SGNEGDE beads by incubating 100 mM ALN with 5 mg of SGNEGDE at pH 11 at 70°C for 24 h, essentially as described previously (Shimizu et al., 2000). For affinity purification, ALN-fixed beads (0.5 mg) were incubated with 250 μl (1 mg protein/ml) of RAW264.7 cell cytoplasmic extract in buffer B (10 mM HEPES, pH 7.9, 10% glycerol, 50 mM KCl, 0.2 mM EDTA, 1 mM DTT, 1 mM PMSF, and 0.1% Nonidet P-40) at 4°C for 4 h. After the beads were washed five times with buffer B, bound proteins were eluted with buffer C (10 mM HEPES, pH 7.9, 10% glycerol, 300 mM KCl, 0.2 mM EDTA, 1 mM DTT, 1 mM PMSF, and 0.1% Nonidet P-40). Where indicated, various concentrations of free ALN, ZOL, or ETI were added to the cell extract before incubation with the beads. In Fig. 5, ALN-fixed beads (0.25 mg) were incubated with 10 μg of recombinant DYN2 proteins in 200 μl of buffer B in the absence or presence of the indicated SUVs for 4 h and processed as above.
Plasmids and Recombinant Proteins.
The cDNA encoding DYN2 was obtained from Dr. Nakayama (Tsukuba University, Ibaraki, Japan). The DNA fragment encoding vesicular stomatitis virus-tagged DYN-2 was prepared by polymerase chain reaction using the primers 5′- GGAATTCTACACTGATATCGAAATGAACCGCCTGGGTAAGGCGATGGGCAACCGCGGGAT-3′ and 5′-CCCAAGCTTCTAGTCGAGCAGGGACG-3′, digested with EcoRI and HindIII, and ligated into the baculovirus expression vector pFastBac HT-A (Invitrogen, Carlsbad, CA). The DNA fragment encoding SNX9 was obtained from the NIH-3T3 cDNA library using the primers 5′-CGGGATCCATGGCCACCAAGGCTCG-3′ and 5′-CCGCTCGAGTTACATAACCGGGAAGCG-3′, digested with BamHI and XhoI, and ligated into the bacterial expression vector pGEX6T-1 (GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK). For mutational analysis, DNA fragments each encompassing the GTPase domain, the middle domain, the Pleckstrin homology (PH) domain, the GTPase effector domain (GED), and the proline-arginine-rich domain (PRD) of DYN2 were prepared by polymerase chain reaction using appropriate primer pairs and cloned into the bacterial expression vector pGEX6T-1. These domains were defined based on previous studies (Artalejo et al., 1997; Schmid et al., 1998).
Recombinant histidine- and vesicular stomatitis virus-tagged DYN2 protein was expressed in insect cells according to the manufacturer's instructions (Invitrogen) and purified with nickel-chelate resin (Qiagen, Hilden, Germany). GST-tagged SNX9, DYN2-GTPase, DYN2-middle, DYN2-PH, and DYN2-GED+PRD were expressed in Escherichia coli BL21 (DE3) and purified with glutathione Sepharose (GE Healthcare), and then the GST tag was removed by digestion with PreScission Protease (GE Healthcare).
GTPase Assay.
GTPase assay was performed as described previously (Damke et al., 2001) with some modifications. rDYN2 (0.1 μM) was incubated in the absence or presence of liposomes in GTPase buffer (0.1 M HEPES, pH 7.9, 1 mM EGTA, 1 mM MgSO4, 1 mM DTT, 1 mM sodium azide, and 0.2 mM PMSF) at room temperature for 10 min. Next, ALN was added where indicated and further incubated for 10 min. Then, 1 mM GTP containing 1 μCi of [α-32P]GTP was added and incubated for 30 min at 37°C. The reactions were stopped by the addition of 1 μl of 5% trichloroacetic acid, and 1.5-μl aliquots of the samples were spotted on polyethylenimine-cellulose plates and developed with 0.5 M LiCl/1 M formic acid. The plates were dried and analyzed with a BAS2000 image analyzer (Fuji Film, Tokyo, Japan).
Cellular Uptake of Transferrin or LDL.
HeLa cells serum-starved for 18 h were incubated with 30 μM chlorpromazine (CPZ) for 15 min or with 200 μM ALN, ZOL or ETI for 4 h, and then with 5 μg/ml Alexa-Fluor-546-transferrin or Alexa-Fluor-488-LDL (Invitrogen) for 15 min. To remove transferrin or LDL that was attached to the cell surface, the cells were washed with ice-cold buffer containing 0.2 M acetic acid and 0.5 M NaCl and then with ice-cold phosphate buffered-saline three times. Thereafter, the cells were collected with trypsin/EDTA and subjected to fluorescence-activated cell sorting analysis (BD Biosciences, San Jose, CA). For fluorescence confocal microscopy, the cells were fixed with 4% formaldehyde for 15 min and visualized under the fluorescence microscope IX-DSU (Olympus, Tokyo, Japan).
Liposome Scission Assay.
Unilamellar liposome was prepared as described previously (Sweitzer and Hinshaw, 1998; Catimel et al., 2009). Solution containing 100 μg of liposome was centrifuged at 12,000g for 10 min, and the pellet was resuspended in buffer containing 25 mM HEPES, pH 7.4, 25 mM KCl, 2.5 mM magnesium acetate, 150 mM potassium glutamate, and 10 mM calcium acetate with or without 400 nM DYN2 and 2 mM ALN, and then it was incubated at 4°C for 4 h. Thereafter, the samples were further incubated with 400 μM GTP at 37°C for 10 min and then centrifuged at 12,000g for 10 min. Grids were glow-discharged using an ion-coater, and 2 μl of the samples was mounted on the grids, stained with 2% uranium acetate, and observed with a H-7500 electron microscope (Hitachi, Tokyo, Japan) at 80 kV.
Analysis of Viral Infection.
In Fig. 1, B and C, 293FT cells were incubated with 100 μM ALN or 10 μM LOV in the presence or absence of various concentrations of geranylgeraniol (GGOH) for 4 h and then infected with recombinant SV40 or adenovirus each carrying the luciferase gene. Two hours after infection, the medium was replaced with fresh medium containing the same concentrations of chemicals, and 18 h after infection, luciferase assay was carried out according to the manufacturer's instructions (Promega, Madison, WI). SV40-luc was constructed by replacing the T-antigen gene with the firefly luciferase gene (Nakanishi et al., 2008). AxCA Luc+, recombinant adenovirus expressing the luciferase gene under the control of the CAG promoter, was obtained from the RIKEN BioResource center (RDB 2453; Ibaraki, Japan). In Figs. 2, A and B, and 7B, HeLa cells were incubated with 200 μM BPs or 10 μM LOV in the presence or absence of 10 μM GGOH for 4 h and were then infected with adenovirus for 18 h. Adenovirus infection was analyzed by immunoblotting or immunofluorescence microscopy using anti-E1A monoclonal antibody (NeoMarkers, Fremont, CA). To examine protein prenylation levels, we used anti-Rap1A antibody (Santa Cruz Biotechnology, Santa Cruz, CA), which reacts with the nonprenylated form of Rap1A. The infected cells were also observed with an H-7500 electron microscope (Hitachi).
Results
Inhibition of Viral Infections by ALN through a Prenylation-Independent Pathway.
The first step in this study was to test whether a popular N-BP, ALN (Fig. 1A), exerts an inhibitory effect on different types of viruses. LOV, known to inhibit the mevalonate pathway by a different mechanism from ALN (i.e., by blocking HMG-CoA reductase activity), was used as control. Immunoblot analysis of the nonprenylated form of the small GTPase Rap1A showed that under our conditions, prenylation of cellular proteins is inhibited by ALN and LOV (Fig. 1, B and C, lanes 2 and 7) and that this inhibition is restored by GGOH, an intermediate of the mevalonate pathway, in a concentration-dependent manner (lanes 3–6 and 8–11). Here, recombinant SV40 and adenovirus each carrying the luciferase gene were used to measure their infectivity. As shown in Fig, 1B, SV40 infection was inhibited by both ALN and LOV (lanes 2 and 7). We were surprised, however, by the finding that GGOH had only a small effect on SV40 infection of ALN-treated cells, whereas it completely rescued SV40 infection of LOV-treated cells (lanes 3–6 and 8–11). Even more puzzling is the observation that adenovirus infection was inhibited by ALN but not by LOV appreciably (Fig. 1C, lanes 2 and 7), and that GGOH had little effect on adenovirus infection of ALN-treated cells (lanes 3–6). These results suggest that although inhibition of the mevalonate pathway is partly responsible for the suppression of SV40 infection by ALN and LOV, there is a different pathway that is important to infection by these viruses and is targeted by ALN. From the data shown in Fig. 1D, IC50 for the antiviral effect of ALN was calculated to be 8.3 μM.
To gain more insight into the additional pathway, we tested another N-BP, ZOL (Fig. 1A), and an NN-BP, ETI (Fig. 1A), for their effects on adenovirus infection by determining the expression level of the adenovirus E1A protein (Fig. 2A). Consistent with the knowledge that NN-BPs exert their pharmacological effects without inhibiting the mevalonate pathway (Coxon et al., 2000), prenylation of Rap1A was inhibited by ALN and ZOL but not by ETI appreciably (lanes 3–5). Nonetheless, these drugs were equally effective in inhibiting adenovirus infection (lanes 3–5), and GGOH did not rescue the inhibition by ALN (lane 7). These results were further validated by immunofluorescence staining using anti-E1A antibody, which showed negligible fluorescence signal from ALN- and ETI-treated cells exposed to adenovirus (Fig. 2B). It is therefore likely that adenovirus infection is suppressed through a prenylation-independent pathway not only by N-BPs but also by NN-BPs, which are considered to be structurally similar to, but functionally distinct from, N-NPs.
Identification of DYN2 and SNX9 as ALN-Binding Proteins.
The above finding suggests that these drugs have common cellular targets other than FPPS. To identify such targets, we carried out an affinity purification of ALN-binding proteins from extracts of rat macrophage cells RAW264.7 using high-performance latex beads (Shimizu et al., 2000) conjugated with ALN. Consequently, two proteins, one 100 kDa and the other 70 kDa, were found to bind specifically to ALN-conjugated beads (Fig. 3A). Sequence analysis by quadrupole time-of-flight mass spectrometry revealed that the 100- and 70-kDa proteins are DYN2, a factor important in various types of endocytic pathways (Takei et al., 1995; Mayor and Pagano, 2007), and SNX9, a factor that interacts with dynamin and regulates its assembly during endocytosis (Lundmark and Carlsson, 2003; Soulet et al., 2005), respectively. The identities of these proteins were confirmed by immunoblotting using their respective antibodies (Fig. 3B). Next, recombinant DYN2 (rDYN2) and SNX9 (rSNX9) proteins were overexpressed in insect cells and E. coli, respectively, and the binding of these proteins to ALN-conjugated beads was examined. As shown in Fig. 3C, rDYN2 bound to the beads independently of rSNX9, whereas rSNX9 only bound in the presence of rDYN2, indicating that ALN bound to DYN2 directly and that SNX9 was copurified by associating with DYN2.
Next, binding specificity was examined by adding free ALN, ZOL, or ETI (Fig. 3D). All three drugs led to the release of both rDYN2 and rSNX9 from the beads in a concentration-dependent manner, suggesting that both N-BPs and NN-BPs can target DYN2. These data are consistent with DYN2 being a molecular target of BPs in adenovirus infection.
ALN Reverses Lipid-Mediated Enhancement of DYN2 GTPase Activity through Competitive Binding to DYN2.
Dynamins are large GTPases that mediate many forms of endocytosis (Mayor and Pagano, 2007), and their GTPase activity is stimulated upon binding to lipid membranes (Praefcke and McMahon, 2004). Thus we examined whether ALN has any effect on DYN2 GTPase, using rDYN2 and lipid prepared from bovine brain. In the absence of lipid, ALN had a negligible effect on GTP hydrolysis to GDP by rDYN2 at the concentrations examined (Fig. 4A). Lipid stimulated rDYN2 GTPase activity 3-fold, and ALN reversed lipid-mediated enhancement of the GTPase activity. Because ALN binds to magnesium ions very well, its inhibitory effect could be due to the chelation of magnesium ions in the reactions. To exclude this possibility, we increased the magnesium concentration from 1 to 2 mM. As a result, there was no appreciable change in the inhibitory effect of 500 μM ALN.
Given the finding that ALN inhibited the GTPase activity only partially even at the highest concentration examined, we speculated that ALN inhibits only the lipid-activated fraction of DYN2, possibly through competitive binding of ALN and lipid to DYN2. As shown in Fig. 4B, rDYN2 binding to ALN-fixed beads were indeed inhibited by increasing concentrations of lipid. Collectively, these results suggest that ALN reverses the lipid-mediated enhancement of DYN2 GTPase activity through competitive binding to DYN2.
ALN Binds to DYN2 through Its Phosphoinositide Lipid-Binding PH Domain.
DYN2 is composed of the GTPase domain, the middle domain, the lipid-binding PH domain, the GED, and the PRD (Artalejo et al., 1997; Schmid et al., 1998). We carried out ALN binding assays using a series of DYN2 mutants and found that its GST-tagged PH domain (GST-PH) binds to ALN, although somewhat weakly (Fig. 5A). Because the DYN2 PH domain is known to bind preferentially to acidic phospholipid (Catimel et al., 2009), synthetic lipid composed of PC or PC and PIP3 (9:1) was used instead of native lipid as competitors of ALN binding assays. Consequently, PIP3-containing lipid was found to inhibit rDYN2 or GST-PH binding to ALN efficiently (Fig. 5B). These results led us to conclude that ALN binds to DYN2 through its phosphoinositide lipid-binding PH domain.
ALN, ZOL, and ETI Inhibit Endocytosis.
Because DYN2 and its GTPase activity play important roles in endocytosis, our findings raise the possibility that inhibition of viral infection is due to defective endocytosis. This hypothesis was first tested using a model system. After incubation of HeLa cells with BPs at 37°C for 4 h, uptake of fluorescently labeled transferrin, which is a marker of the clathrin-mediated pathway, was measured by flow cytometry and fluorescence confocal microscopy. Our results show that, in the presence of ALN, ZOL, or ETI, transferrin uptake was significantly decreased and that the suppression level was similar to that in the presence of CPZ, a clathrin-dependent endocytosis inhibitor (Fig. 6).
ALN Inhibits Liposome Vesiculation in Vitro and Scission of Endocytic Vesicles in Adenovirus-Infected Cells.
Dynamins self-assemble into ring-like complexes to carry out vesicle scission by tubulation and severance of lipid membranes using GTP hydrolysis as source of energy (Mayor and Pagano, 2007). This activity can be reconstituted in vitro using purified rDYN2, which also self-associates and binds to lipid bilayers to form tubular structures (Fig. 7A, +rDYN2, −GTP) that can be constricted to form small lipid vesicles upon GTP addition (+rDYN2, +GTP) (Sweitzer and Hinshaw, 1998). However, this vesiculation activity was inhibited in the presence of ALN (+rDYN2+ALN, +GTP). Because this observation is similar to that of a GTPase-deficient mutant rDYN2 (K44A) (Damke et al., 1994), which forms tubular lipid structures but is incapable of vesiculation as a result of a point mutation in the GTP binding site [+rDYN2 (K44A)], our results strongly suggest that ALN targets the GTPase activity of DYN2 to suppress endocytosis.
HeLa cells infected with adenovirus in the absence or presence of ALN were analyzed by electron microscopy. As shown in Fig. 7B, normal endocytic vesicles containing virus particles are round and are found detached from the plasma membrane, whereas in the presence of ALN, these vesicles seem to be trapped near the cytoplasmic membrane. Thus, these observations suggest that the cause of ALN-mediated inhibition of adenovirus infection may be defective scission as a result of DYN2 GTPase activity inhibition at the neck of the invaginated pit (arrow).
Discussion
In this study, we investigated the effects of BPs on SV40 and adenovirus infections and found that these viruses show different sensitivity toward the various drugs examined. The results obtained using the HMG-CoA reductase inhibitor LOV indicate that prenylation of a certain cellular or viral protein is important to SV40 infection, although the existence of such a protein is not known. On the other hand, the results obtained using ALN and GGOH indicate that ALN exerts its inhibitory effects on viral infection partly or entirely through a mevalonate-independent pathway, most likely through inhibition of DYN2-mediated endocytosis, as our biochemical data suggest here. ALN was found to target DYN2 GTPase activity, leading to its inability to pinch off endocytic vesicles. Moreover, DYN2 was found to be targeted by ETI, an NN-BP that is not known to target FPPS, which further supports our discovery of an additional nonmevalonate-related inhibitory mechanism for BPs. Adenovirus infection requires DYN2 in clathrin-dependent endocytosis, and expression of the GTPase-deficient mutant DYN2 (K44A) significantly reduces infectivity (Wang et al., 1998). SV40 enters cells via caveolar endocytosis (Anderson et al., 1996), a clathrin-independent but DYN2-dependent pathway. Thus, our identification of DYN2 as a target for antiviral activity of BPs is consistent with its key role in multiple endocytic pathways.
It may be surprising that using ALN as ligand, a highly charged molecule, DYN2 was purified almost exclusively. We have carried out a number of affinity purifications using various low molecular weight compounds as ligands. In some cases, numerous proteins were obtained in initial experiments, and subsequent optimization of various parameters, such as the amount of ligand immobilized, composition of binding and washing buffers, and elution conditions (e.g., competitive elution using free ligand), was necessary. But in the other cases, only one or a few proteins were isolated without the need for extensive optimization, leading to rapid identification of relevant molecular targets (Sakamoto et al., 2009). In principle, relative abundance of affinity-purified proteins is determined by the concentration of each protein in input material and its equilibrium dissociation constant for the ligand. Hence, a simple interpretation of our data are that there was sufficient difference in Kd between DYN2 binding and other “nonspecific” binding that might have occurred through ionic interactions. Besides, although it is not central to the discussion here, we believe that a key to success is the use of affinity beads called SG beads and FG beads, which we have developed over decades (Sakamoto et al., 2009). They are spherical organic-polymer particles of approximately 200 nm in diameter, which is 2 to 3 orders of magnitude smaller than the diameter of typical agarose beads. Their nonporous structure, large surface area per weight, and moderately hydrophilic surface seem to contribute to a high signal-to-background ratio in purification.
Below we discuss the possibility that clinically relevant doses of BPs contribute to treatments of diseases other than blocking bone loss. On the one hand, we determined the IC50 for the antiviral effect of ALN to be 8.3 μM (Fig. 1D). On the other hand, it is generally accepted that after bisphosphonate administration, a fraction of the drug becomes incorporated into the skeleton, whereas the remainder is cleared rapidly from the blood to the urine (Russell et al., 2007). After single oral administration of 70 mg of ALN to humans, its elimination half-life is 1.67 h and the maximum plasma concentration is 41 ng/ml or 0.16 μM (Yun and Kwon, 2006). It is therefore unlikely that by the regimen used in the study, plasma concentrations of ALN come close to the level that affects DYN2-mediated endocytosis. However, a few studies have shown that some of the drug also accumulates temporarily in soft tissues, such as liver and spleen, and is subsequently released over a period of days (Stepensky et al., 2003), although its precise pharmacokinetics remains to be determined. In another study, it was suggested that bisphosphonate concentrations at the aorta are much higher (up to 24-fold) than plasma levels (Ylitalo et al., 1996). In addition, the deposition in soft tissues may vary by the mode of drug administration (Stepensky et al., 2003). Furthermore, bisphosphonates are generally safe and well tolerated even at higher doses; for example, patients with Paget's disease receive up to 400 mg of bisphosphonates per day. In summary, local tissue concentrations of bisphosphonates may be higher than their plasma concentrations, and even higher local concentrations may be achieved by considering pharmaceutical formulations and mode of administration. For these reasons, we consider that our present findings may have potential clinical implications.
BPs have been prescribed for more than three decades in the treatment of bone diseases and are currently the most widely used bone resorption drugs. BPs mainly target osteoclasts but can also affect other cell types in various ways, including the suppression of proliferation, adhesion, migration, and invasion of tumor cells (Roelofs et al., 2006). It is interesting that the new BP target found here, DYN2, also plays diverse cellular roles. DYN2 is most famous for its role in membrane scission, the last step required to release invaginated endocytic vesicles into the cytoplasm (Marks et al., 2001; Mayor and Pagano, 2007). Hence, in addition to reducing osteoclast bone uptake indirectly through the inhibition of prenylation, N-BPs may also directly stop resorption in vivo by mechanically blocking the severance and release of loaded vesicles into the cytoplasm of osteoclasts, as our data suggest here. Furthermore, reports by others have shown that DYN2 also plays important roles in actin remodeling and the formation of podosomes, thus suggesting that BPs may also affect the motility of osteoclasts and how these cells secure contact between the membrane and the extracellular matrix (Bruzzaniti et al., 2005), which may also explain their effects on tumor cells.
Through the discovery of DYN2 as a target of BPs, this study has uncovered a previously unknown mechanism of action of BPs and offers potential novel use for these drugs. Our findings will probably be extendable to research on other viruses. One example is HIV-1, the infectivity of which is known to be enhanced by DYN2 (Pizzato et al., 2007). However, because of the complexity of the various pathways used by viruses to enter cells and the various cellular roles played by DYN2, much work is still needed to clarify how different viruses respond to BP treatments. Incidentally, investigations or trials of alternative uses of BPs, such as studies on the inhibition of growth of human T-cell leukemia virus type I-infected T-cell lines (Ishikawa et al., 2007) or the treatment of HIV-associated osteoporosis (Qaqish and Sims, 2004; Negredo et al., 2005), are emerging. The knowledge gained through these efforts may lead to better antiviral therapies through effective combinatorial treatments using new and old strategies.
Acknowledgments
We are grateful to Dr. Kazuhisa Nakayama (Kyoto University) for providing the DYN2 cDNA and to Dr. Hiroyuki Kouji (Institute for Life Science Research, Asahi Chemical Industry Co. Ltd.) for providing bisphosphonates.
Footnotes
-
This study was supported by Special Coordination Funds for Promoting Science and Technology from the Japan Science and Technology Agency [Grant KH217001]; and a Grant from the Global COE Program from the Ministry of Education, Culture, Sports, Science and Technology of Japan [Grant A-04].
-
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
doi:10.1124/mol.109.059006
-
ABBREVIATIONS:
- BP
- bisphosphonate
- ALN
- alendronate
- CPZ
- chlorpromazine
- DTT
- dithiothreitol
- DYN2
- dynamin-2
- ETI
- etidronate
- FPPS
- farnesyl diphosphate synthase
- GED
- GTPase effector domain
- GGOH
- geranylgeraniol
- LOV
- lovastatin
- N-BP
- nitrogen-containing bisphosphonate
- NN-BP
- non–nitrogen-containing bisphosphonate
- PC
- phosphatidylcholine
- PH
- Pleckstrin homology
- PIP3
- phosphatidylinositol (3,4,5)-triphosphate
- PMSF
- phenylmethylsulfonyl fluoride
- PRD
- proline-arginine-rich domain
- SNX9
- sorting nexin 9
- SUV
- small unilamellar vesicle
- SV40
- simian virus 40
- ZOL
- zoledronate
- PAGE
- polyacrylamide gel electrophoresis
- FBS
- fetal bovine serum
- GST
- glutathione transferase
- BZA-5B
- 2-amino-3-mercaptopropionate-(N-methyl)-3-amino-1-carboxymethyl-5-phenyl-benzodiazepine
- FTI-277
- N-[4-[2(R)-amino-3-mercaptopropyl]amino-2-phenylbenzoyl]methionine methyl ester trifluoroacetate salt.
- Received June 30, 2009.
- Accepted November 9, 2009.
- Copyright © 2010 The American Society for Pharmacology and Experimental Therapeutics