ARTHRITIS & RHEUMATISM Vol. 63, No. 12, December 2011, pp 3908–3917 DOI 10.1002/art.30646 © 2011, American College of Rheumatology Distinguishing the Proapoptotic and Antiresorptive Functions of Risedronate in Murine Osteoclasts Role of the Akt Pathway and the ERK/Bim Axis Takumi Matsumoto, Yuichi Nagase, Mitsuyasu Iwasawa, Tetsuro Yasui, Hironari Masuda, Yuho Kadono, Kozo Nakamura, and Sakae Tanaka Objective. Nitrogen-containing bisphosphonates are one of the most successful therapeutics for osteoporosis. The aim of this study was to elucidate the functional mechanism of one of the typical nitrogencontaining bisphosphonates, risedronate. Methods. Osteoclasts generated from murine bone marrow macrophages were treated with risedronate in vitro, and its effects on apoptosis and boneresorbing activity were examined. The mechanism of action of risedronate was examined by gene induction of constitutively active Akt-1 and constitutively active MEK-1, and by gene deletion of Bim. Bimⴚ/ⴚ mice, in which osteoclasts were resistant to apoptosis, were treated with risedronate and analyzed radiographically, biochemically, and histologically. Results. Risedronate induced osteoclast apoptosis through the mitochondria-dependent pathway with an increased expression of Bim, and the proapoptotic effect of risedronate was suppressed by Bim deletion and constitutively active MEK-1 introduction. In contrast, the risedronate-induced suppression of bone resorption was completely reversed by inducing constitutively active Akt-1, but not by Bim deletion or constitutively active MEK-1 introduction. These results suggested that apoptosis and bone-resorbing activity of osteoclasts were regulated through the ERK/Bim axis and the Akt pathway, respectively, both of which were suppressed by risedronate. Although osteoclast apoptosis in response to risedronate administration was suppressed in the Bimⴚ/ⴚ mice, risedronate treatment increased bone mineral density in Bimⴚ/ⴚ mice at a level equivalent to that in wild-type mice. Conclusion. Our findings indicate that the antiresorptive effect of risedronate in vivo is mainly mediated by the suppression of the bone-resorbing activity of osteoclasts and not by the induction of osteoclast apoptosis. Bisphosphonates, stable analogs of pyrophosphate, strongly inhibit bone resorption and have been used to treat various diseases driven by increased bone resorption, such as postmenopausal osteoporosis, Paget’s disease, and tumor bone metastases (1). Although bisphosphonates are poorly absorbed from the intestine, they are quickly deposited on the bone surface once absorbed (2). The acidic environment produced by osteoclasts reduces the ability of bisphosphonates to chelate Ca2⫹ and releases bisphosphonates from the bone surface, and bisphosphonates are then ingested into osteoclasts by endocytosis (3,4). It was speculated that the concentration of bisphosphonates reaches as high as 0.1–1 mM in the resorption lacuna (3). Bisphosphonates are divided into 2 groups according to the structure of the side chains, a nitrogen- Supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, by Health Science research grants (21591932 to Dr. Kadono and 20249024 and 21659350 to Dr. Tanaka) from the Ministry of Health, Labor, and Welfare of Japan, and in part by funding for the Global COE Program, Medical System Innovation on Multidisciplinary Integration, from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Takumi Matsumoto, MD, Yuichi Nagase, MD, PhD, Mitsuyasu Iwasawa, MD, PhD, Tetsuro Yasui, MD, PhD, Hironari Masuda, MD, Yuho Kadono, MD, PhD, Kozo Nakamura, MD, PhD, Sakae Tanaka, MD, PhD: University of Tokyo, Tokyo, Japan. Drs. Nakamura and Tanaka have received consulting fees, speaking fees, and/or honoraria from Eisai Pharmaceuticals (less than $10,000). Address correspondence to Sakae Tanaka, MD, PhD, Department of Orthopaedic Surgery, Faculty of Medicine, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan. E-mail: [email protected] Submitted for publication January 16, 2011; accepted in revised form August 18, 2011. 3908 DISTINGUISHING THE FUNCTIONS OF RISEDRONATE containing type and a non–nitrogen-containing type. The difference in the structures results in different bisphosphonate mechanisms of action for antiresorbing activity (1). Non–nitrogen-containing bisphosphonates are reported to act through the intracellular accumulation of nonhydrolyzable ATP analogs that exert cytotoxic effects on osteoclasts (5), while nitrogencontaining bisphosphonates inhibit the mevalonate pathway and prevent the posttranslational prenylation of small GTP-binding proteins such as Ras, Rho, Rac, and Cdc42 (6). Previous studies have demonstrated that small G proteins control the differentiation, apoptosis, cytoskeletal organization, and vesicular trafficking of osteoclasts (7–10). Therefore, the antiresorptive properties of nitrogen-containing bisphosphonates can be exhibited by inducing osteoclast apoptosis (11,12), altering the osteoclast cytoskeleton (3,13), or affecting osteoclast vesicular and membrane trafficking (14). Recently, it was reported that nitrogen-containing bisphosphonates suppress the bone-resorbing activity of osteoclasts independently of apoptosis (15). In addition, Weinstein et al demonstrated that bone biopsy specimens from patients with long-term alendronate administration exhibited an increased osteoclast number (OcN) in proportion to the accumulated amount of alendronate (16). These observations have led to the notion that the antiresorptive function of nitrogen-containing bisphosphonates is regulated through a signaling pathway different from that of their proapoptotic function. The present study aimed to distinguish the molecular mechanisms regulating the proapoptotic function and antiresorptive function of nitrogen-containing bisphosphonates. We previously reported that the proapoptotic Bcl-2 homology 3 (BH3)–only domain protein Bim plays a crucial role in the osteoclast apoptosis induced by cytokine deprivation (17). In the present study, we showed that risedronate-induced osteoclast apoptosis was markedly suppressed in Bim⫺/⫺ mice, while risedronate increased bone volume (BV) in the mice to a level similar to that in wild-type (WT) mice. We also demonstrated that the cell viability and boneresorbing function of osteoclasts were regulated by the ERK/Bim pathway and the Akt pathway, respectively. MATERIALS AND METHODS Animals. Newborn and 8-week-old male ddY and C57BL/6J mice were purchased from Sankyo Labo Service Co. Bim⫺/⫺ mice (on a C57BL/6 genetic background) were kindly provided by Dr. Philippe Bouillet and Dr. Andreas Strasser (Walter and Elisa Hall Institute, Melbourne, Victoria, Australia). The breeding and genotyping of Bim⫺/⫺ mice were 3909 performed as previously described (18). All animals were housed under specific pathogen–free conditions and treated with humane care under the approval of the Animal Care and Use Committee of the University of Tokyo. Generation of osteoclasts and survival/bone resorption assay. Osteoclasts were generated using a coculture system previously described (7). Briefly, when murine osteoblastic cells and bone marrow cells were cultured on collagen gel– coated dishes in the presence of 10 nM 1␣,25-dihydroxyvitamin D3 (1␣,25[OH]2D3) and 1 M prostaglandin E2 (PGE2), osteoclasts were differentiated on day 6 of culture. The cells were dispersed by treatment with 0.1% bacterial collagenase (Wako Pure Chemical) for 10 minutes and then used for the pit formation assay and survival assay as follows. For the bone resorption assay, the cells were resuspended in ␣-minimum essential medium (␣-MEM) containing 10% fetal bovine serum (FBS), replated on dentin slices, and cultured for 24 hours with or without the indicated concentrations of risedronate. After cells were removed by treating the dentin slices with 1M NH4OH, the resorption areas were visualized by staining with 1% toluidine blue. The resorption pit area was quantified using an image analyzing system (Microanalyzer; Japan Poladigital). For the survival assay, the resuspended cells were placed on culture dishes. After 5 hours of incubation, the cultures were treated with ␣-MEM containing 0.1% collagenase and 0.2% Dispase for 10 minutes to remove osteoblastic cells and purify osteoclasts, and the remaining osteoclasts were further cultured in ␣-MEM containing 10% FBS with or without the indicated concentrations of risedronate. For the survival assay using caspase inhibitors, the osteoclasts purified in the way described above were cultured in the presence or absence of 30 M risedronate with DMSO, 30 M caspase 8 inhibitor (Z-IETD-FMK), or 30 M caspase 9 inhibitor (Z-LEHD-FMK). Cell viability/survival rate was expressed as the percentage of morphologically intact tartrateresistant acid phosphatase (TRAP)–positive multinucleated cells. TRAP staining was performed at pH 5.0 in the presence of L(⫹)-tartaric acid using naphthol-AS-MX phosphate in N,N-dimethylformamide as the substrate. The number of TRAP-positive osteoclasts remaining at the different time points is shown as a percentage of the cells at time zero. Sealing zone formation on dentin slices was analyzed as follows. The cells were cultured on dentin slices as described above. After 16 hours, the cells were fixed in phosphate buffered saline (PBS) containing 4% paraformaldehyde for 10 minutes and then stained by TRAP. Cells were then incubated for 30 minutes with rhodamine-conjugated phalloidin solution (Molecular Probes) to visualize F-actin. Sealing zone formation on dentin slices was observed using fluorescence microscopy (Biozero; Keyence). The sealing zone formation rate is represented as the proportion of the cells with an uninterrupted ring-like F-actin structure among the TRAP-positive multinucleated cells. Western blotting and enzyme-linked immunosorbent assay (ELISA). All extraction procedures were performed at 4oC or on ice. Cells were washed with ice-cold PBS, and proteins were extracted with Tris–NaCl–EDTA buffer (1% Nonidet P40, 10 mM Tris HCl [pH 7.8], 150 mM NaCl, 1 mM EDTA, 2 mM Na3VO4, 10 mM NaF, and 10 g/ml aprotinin). The lysates were clarified by centrifugation at 12,000g for 10 minutes. For Western blotting analysis, lysates were subjected 3910 to sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis with a 7.5–15% Tris glycine gradient gel or 15% Tris glycine gel and transferred onto nitrocellulose membranes (Bio-Rad). After blocking with 6% milk/Tris buffered saline– Tween, membranes were incubated with primary antibodies to cleaved caspase 3, caspase 8, cleaved caspase 9, Bcl-xL, Bcl-2, phospho-Akt at Ser473, Akt, phospho-ERK, ERK (all from Cell Signaling Technology), Bim (Santa Cruz Biotechnology), cytochrome c oxidase 4, cytochrome c (both from BD Biosciences), or ␤ -actin (Sigma-Aldrich), followed by horseradish peroxidase–conjugated goat anti-mouse IgG and goat antirabbit IgG (both from Promega). Immunoreactive bands were visualized with ECL Plus (Amersham) according to the manufacturer’s instructions. The blots were stripped by incubation for 20 minutes in stripping buffer (2% SDS, 100 mM 2-mercaptoethanol, and 62.5 mM Tris HCl [pH 6.7]) at 50°C and reprobed with additional antibodies. Separation of the mitochondrial fraction and cytosolic fraction was performed using an ApoAlert cell fractionation kit (Clontech) according to the manufacturer’s instructions. Phospho–Akt-1 and phospho–p44/42 MAPK were also quantified using a PathScan Phospho–Akt-1 (Ser473) Sandwich ELISA kit and PathScan Phospho–p44/42 MAPK (Thr202/ Tyr204) Sandwich ELISA kit (Cell Signaling Technology) according to the manufacturer’s instructions. Real-time polymerase chain reaction (PCR). Messenger RNA (mRNA) was isolated from osteoclasts using Isogen (Wako Pure Chemical), and an aliquot (1 g) was reverse-transcribed using a QuantiTect Reverse Transcription kit (Qiagen) to make single-stranded complementary DNA. PCR was performed using a ABI Prism 7000 Sequence Detection System (Applied Biosystems) using QuantiTect SYBR Green PCR Master Mix (Qiagen) according to the manufacturer’s protocol. All reactions were run in triplicate. After data collection, the relative mRNA copy number of a specific gene was calculated with a standard curve generated with serially diluted plasmids containing PCR amplicon sequences, and normalized to rodent total RNA with mouse ␤-actin serving as the internal control. Standard plasmids were synthesized with a TOPO TA cloning kit (Invitrogen) according to the manufacturer’s instructions. The primers we used to detect the common form of all Bim splice variants were 5⬘CTTCCATACGACAGTCTC-3⬘ and 5⬘-AACCATTTGAGGGTGGTCTTC-3⬘. Expression constructs and gene transduction. The adenovirus vectors used in the experiments, and the genes carried by the vectors, were as follows: AxGFP (green fluorescence protein [GFP] gene), AxMekCA (constitutively active MEK-1 gene), and AxAktCA (constitutively active Akt-1 gene). AxGFP was used as a control vector. AxMekCA and AxAktCA were provided by Dr. Hideki Katagiri (Tohoku University). Constitutively active MEK-1 was generated by an exchange of the 2 Raf-1 phosphorylation sites on Ser218 and Ser222 for Glu218 and Glu222, and constitutively active Akt-1 was generated by adding a myristoylation signal sequence to its N-terminus. Viral titers were determined by the end point dilution assay, and the viruses were used at 50 multiplicities of infection (MOI). Infection of osteoclasts by adenovirus vectors was carried out following the method previously described (19). Briefly, on day 4 of culture, when osteoclasts began to appear, mouse cocultures were incubated for 1 hour at 37°C with a small amount of ␣-MEM containing the recombinant adeno- MATSUMOTO ET AL viruses at the desired MOI. Cells were then washed twice with PBS and further incubated at 37°C in ␣-MEM containing 10% FBS, 10 nM 1␣,25(OH)2D3, and 1 M PGE2. Experiments were performed 36 hours after the infection. Retrovirus packaging was performed by transfection of the pMx vectors into BOSC cells. Retrovirus construction of Bcl-2 and Bcl-xL and infection of the osteoclast precursors were carried out as previously described (20). In vivo risedronate treatment. For in vivo analysis, normal saline or 0.01 mg risedronate/kg body weight was subcutaneously injected daily for 14 days into 14-week-old Bim⫺/⫺ and 14-week-old WT mice (3 animals/group). The day after the final injection, blood samples were collected retroorbitally under anesthesia from Bim⫺/⫺ mice and WT mice with or without risedronate treatment immediately prior to killing. Sera were obtained using a capillary blood collection tube with serum separator (Becton Dickinson). The serum concentration of C-terminal crosslinking telopeptide of type I collagen (CTX-I) was measured by RatLaps ELISA (Nordic Bioscience). Plain radiographs were obtained using a soft x-ray apparatus (CMB-2; Softex), and the bone mineral density (BMD) was measured by dual-energy x-ray absorptiometry using a bone mineral analyzer (PIXImus Densitometer; GE Medical Systems). Histologic analyses. Tissues were fixed in 4% paraformaldehyde/PBS, decalcified in 10% EDTA for 2 weeks at 4oC, embedded in paraffin, and cut into 3 m–thick sections. Hematoxylin and eosin staining was performed according to the standard procedure. Histomorphometric analysis was performed in the primary and secondary spongiosa of the proximal tibia beginning from the lowest point of the growth plate to a point 1.0 mm distally. Osteoclasts were identified by TRAP staining. The total number of osteoclasts was expressed as the number per millimeter at the cancellous bone perimeter (BPm). Giant osteoclasts were defined as cells that had more than 8 nuclear profiles (2-dimensional images of a section) and that were detached from bone. Cells undergoing apoptosis were identified by means of the TUNEL method, which specifically labels the 3⬘-hydroxyl terminal DNA strand breaks. For the TUNEL procedure, all agents, including buffers, were part of a kit (In Situ Cell Death Detection kit, peroxidase; Roche Applied Science); the staining procedure was carried out according to the manufacturer’s recommendation. Apoptotic cells were defined as being TUNEL positive and having apoptotic morphologic features of chromatin condensation, nuclear fragmentation, and cytoplasmic contraction or fragmentation. Statistical analysis. The in vitro experiments were repeated at least 3 times, and the representative data from 1 experiment were used for statistical analysis. Results are expressed as the mean ⫾ SD of data obtained from 6 independent culture dishes in 1 typical experiment. In the in vivo experiments, the data obtained from 3 different animals were used. Statistical analyses were performed using Student’s unpaired 2-tailed t-test or analysis of variance. RESULTS Risedronate induces osteoclast apoptosis through the mitochondrial pathway. When the osteoclasts generated in vitro were treated with risedronate, the DISTINGUISHING THE FUNCTIONS OF RISEDRONATE Figure 1. A, Dose-dependent reduction by risedronate (RIS) of the number of tartrate-resistant acid phosphatase (TRAP)–positive osteoclasts (OCs), shown as a percentage of the cells at time zero. ⴱ ⫽ P ⬍ 0.01 versus untreated cells (control). B, Risedronate-induced expression of cleaved caspase 3, caspase 8, and cleaved caspase 9 in osteoclasts. Positive control (PC) cell lysates for cleaved caspase 8 immunoblotting (IB) were obtained from spleen cells stimulated with 400 ng/ml FasL for 16 hours. C, Risedronate-induced release of cytochrome c from the mitochondrial fraction (M) into the cytosolic fraction (C) in osteoclasts. Cytochrome c oxidase 4 (Cox4) is a marker of the mitochondrial fraction. D, Effects of caspase 8 inhibitor (Z-IETD-FMK) and caspase 9 inhibitor (Z-LEHD-FMK). Top, TRAP staining of representative cultures. Bars ⫽ 100 m. Bottom, Number of TRAP-positive osteoclasts shown as a percentage of the cells at time zero. ⴱ ⫽ P ⬍ 0.01 versus cultures without risedronate treatment. E, Effects of retroviral overexpression of Bcl-2 (RxBcl-2) and Bcl-xL (RxBcl-xL). Top, Overexpression of Bcl-2 and Bcl-xL confirmed by Western blotting. Bottom, Number of TRAP-positive osteoclasts shown as a percentage of the cells at time zero. ⴱ ⫽ P ⬍ 0.01 versus cultures without risedronate treatment. In A, D, and E, experiments were repeated at least 3 times, and results are expressed as the mean ⫾ SD of data obtained from 6 independent culture dishes in 1 typical experiment. number of osteoclasts was reduced in a dose-dependent manner, and risedronate at concentrations of ⱖ3 M induced a significant reduction in the OcN after 24 hours of treatment (Figure 1A). Osteoclasts treated with risedronate exhibited condensation and segmentation of the nuclei by Hoechst 33342 fluorescence microscopic ana- 3911 lysis, morphologic features which are reminiscent of apoptosis (Figure 2). The 17-kd and 19-kd forms of cleaved caspase 3, active fragments generated from procaspase 3, were increased after 16 hours of risedronate treatment, as determined by Western blot analysis (Figure 1B), further confirming that risedronate induces apoptosis in osteoclasts. Interestingly, cleaved caspase 3 was observed immediately after purification (time zero) and decreased 8 hours after purification in control osteoclasts. We think that this is because the purification maneuver using collagenase and Dispase may somehow be harmful and temporarily stimulate caspase 3. There are 2 distinct apoptosis pathways in mammals. One is the death receptor pathway, which is initiated by death receptors, such as the tumor necrosis factor receptor, which contain the intracellular death domain. Upon ligand binding, intracellular signaling is propagated through FADD adaptor protein–mediated activation of caspase 8. The other is the mitochondrial pathway, which is regulated by the pro- and antiapoptotic Bcl-2 family members and induces the release of cytochrome c from mitochondria, leading to Apaf-1 adaptor–mediated activation of caspase 9. After 16 hours of risedronate treatment, cytochrome c was released into the cytosolic fraction more abundantly in osteoclasts receiving risedronate treatment than in control osteoclasts (Figure 1C). The cleaved form of caspase 9 was clearly increased in osteoclasts, while the 18-kd active fragment of caspase 8 was hardly observed (Figure 1B). The risedronate-induced reduction of the number of osteoclasts was fully reversed by a caspase 9 inhibitor (Z-LEHD-FMK) but only partially by a caspase 8 inhibitor (Z-IETD-FMK) (Figure 1D). The reduced viability of osteoclasts receiving risedronate treatment was completely recovered by retrovirus vector–mediated overex- Figure 2. Risedronate-induced nuclear condensation and segmentation in osteoclasts. Osteoclasts generated in vitro were left untreated or were treated with 3 M risedronate for 24 hours, and nuclear condensation and segmentation were examined by Hoechst 33342 fluorescence microscopic analysis. Bars ⫽ 20 m. See Figure 1 for definitions. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN) 1529-0131. 3912 pression of the antiapoptotic Bcl-2 family members Bcl-2 or Bcl-xL (Figure 1E). These results indicate that risedronate mainly induces osteoclast apoptosis through the mitochondria-dependent pathway. Resistance of Bimⴚ/ⴚ mouse osteoclasts to risedronate-induced apoptosis. We then examined the effect of risedronate on the expression levels of pro- and antiapoptotic Bcl-2 family members. Immunoblot analysis revealed an increase in Bim expression within 8 hours of risedronate addition, while the expression levels of the antiapoptotic Bcl-2 family members, Bcl-2 and BclxL, remained unchanged (Figure 3A). Among the 3 isoforms of Bim generated by alternative splicing, BimEL was mainly induced by risedronate treatment in osteoclasts. BimL was also up-regulated in response to risedronate treatment (Figure 3A). The expression of other proapoptotic family members, such as Bax or Bad, was also not altered by risedronate treatment (Figure 3A). No significant difference in the Bim mRNA level was observed by real-time PCR between osteoclasts cultured with and those cultured without risedronate, suggesting that the changes in the Bim protein levels are due to posttranslational mechanisms (Figure 3B). We then examined the effect of risedronate on the osteoclasts generated from Bim⫺/⫺ mouse bone marrow cells. Risedronate markedly reduced the number of WT mouse osteoclasts, and ⬍20% of the cells remained after 24 hours of treatment, while risedronate had no significant effect on Bim⫺/⫺ mouse osteoclasts (P ⫽ 0.329) (Figure 3C). We previously reported that Bim⫺/⫺ mouse osteoclasts exhibit less bone-resorbing activity than WT mouse osteoclasts. Interestingly, in spite of the prolonged survival of Bim⫺/⫺ mouse osteoclasts compared to WT mouse osteoclasts, pit formation was similarly and almost completely inhibited by risedronate in both WT mouse osteoclasts and Bim⫺/⫺ mouse osteoclasts (Figure 3D). The mean ⫾ SD reduction rates were 85.3 ⫾ 5.2% in WT mouse osteoclasts and 90.2 ⫾ 5.1% in Bim⫺/⫺ mouse osteoclasts, with no significant difference (P ⫽ 0.15). Suppression of osteoclast apoptosis in Bimⴚ/ⴚ mice does not alter the antiresorptive effect of risedronate in vivo. To examine the impact of Bim deficiency in risedronate-induced osteoclast apoptosis in vivo, risedronate (0.01 mg/kg body weight) was administered to the Bim⫺/⫺ mice and WT mice once a day for 14 days. We previously reported that Bim⫺/⫺ mice have mild osteosclerosis due to the decreased bone-resorbing activity of osteoclasts, although the number of osteoclasts increased in Bim⫺/⫺ mice (17). Radiographic analysis showed that risedronate treatment increased the radio- MATSUMOTO ET AL Figure 3. A, Western blotting showing effects of risedronate on the expression of Bcl-2 family proteins. B, Transcriptional regulation of Bim in osteoclasts as determined by real-time polymerase chain reaction. Results are expressed as the mean ⫾ SD of 3 independent experiments. C, Resistance of Bim ⫺/⫺ mouse osteoclasts to risedronate-induced cell reduction. Osteoclasts generated from bone marrow cells of wild-type (WT) or Bim⫺/⫺ mice were cultured in the presence or absence of 30 mM risedronate. Top, TRAP staining of representative cultures. Bars ⫽ 100 m. Bottom, Number of TRAPpositive osteoclasts shown as a percentage of the cells at time zero. D, Suppression of bone resorption by risedronate in both WT and Bim⫺/⫺ mouse osteoclasts. Top, Resorption pits of representative cultures. Bars ⫽ 100 m. Bottom, Pit area per cell. In C and D, experiments were repeated at least 3 times, and results are expressed as the mean ⫾ SD of data obtained from 6 independent culture dishes in 1 typical experiment. ⴱ ⫽ P ⬍ 0.01 versus cultures without risedronate. See Figure 1 for other definitions. opacity, especially in the distal femur, and increased BMD in the Bim⫺/⫺ mice at a level equivalent to that in the WT mice (Figures 4A and B). Risedronate treatment reduced the mean level of serum CTX-I, a measure of resorption breakdown products, to approximately half in WT and in Bim⫺/⫺ mice (Figure 4C). DISTINGUISHING THE FUNCTIONS OF RISEDRONATE 3913 Figure 4. A, Representative radiographic images of the distal femur of 14-week-old male Bim⫺/⫺ and wild-type (WT) mice that were left untreated or were treated with risedronate for 14 days. B, Bone mineral density of the distal femur in Bim⫺/⫺ and WT mice that were left untreated or were treated with risedronate. ⴱ ⫽ P ⬍ 0.05 versus untreated WT mice; ⴱⴱ ⫽ P ⬍ 0.05 versus untreated Bim⫺/⫺ mice. C, Serum concentration of C-terminal crosslinking telopeptide of type I collagen (CTX-I) in Bim⫺/⫺ and WT mice. ⴱ ⫽ P ⬍ 0.01 versus untreated mice of the same genotype. D, TRAP staining of the distal femur of WT and Bim⫺/⫺ mice. Bars ⫽ 100 m. E, Osteoclast number per bone perimeter (OcN/BPm) in Bim⫺/⫺ and WT mice. ⴱ ⫽ P ⬍ 0.01 versus untreated WT mice. F, Risedronate-induced osteoclast apoptosis in Bim⫺/⫺ and WT mice. ⴱ ⫽ P ⬍ 0.01 versus untreated mice of the same genotype. G, TRAP staining of normal osteoclasts in untreated WT mice (left) and giant osteoclasts in risedronate-treated WT mice (right). Bars ⫽ 10 m. H, Proportion of giant osteoclasts in Bim⫺/⫺ and WT mice. ⴱ ⫽ P ⬍ 0.01 versus untreated mice of the same genotype. In B, C, E, F, and H, results are expressed as the mean ⫾ SD of 3 samples obtained from different animals. See Figure 1 for other definitions. OcN/BPm was increased by risedronate treatment by ⬃50% in WT mice, but not in Bim⫺/⫺ mice (Figures 4D and E). The mean ratio of apoptotic osteoclasts to total osteoclasts was ⬃12% and ⬃2% in risedronate-treated WT mice and risedronate-treated Bim⫺/⫺ mice, respectively (Figure 4F). These results demonstrate that Bim⫺/⫺ mouse osteoclasts are resistant to risedronate-induced apoptosis in vivo as well as in vitro, but the suppression of osteoclast apoptosis does not affect the in vivo effect of risedronate increasing bone mass. Interestingly, risedronate treatment induced an ⬃5-fold increase of giant osteoclasts compared to untreated controls in both WT and Bim⫺/⫺ mice (Figures 4G and H). Effect of risedronate on ERK and Akt activity in osteoclasts. It was previously reported that the suppression of bone resorption by alendronate and risedronate is independent of their effects on apoptosis (15). As shown in Figure 5A, risedronate significantly suppressed pit formation at a concentration of 0.03 M, a much lower concentration than that required to induce osteoclast apoptosis (Figure 1A). To address the possible underlying molecular mechanisms, we examined the effect of risedronate on ERK and Akt activity in osteoclasts. Risedronate suppressed both Akt and ERK activity in osteoclasts in a time- and dose-dependent manner (Figures 5B and C). To compare the effect of risedronate on ERK and Akt activity in further detail, we performed ELISA analysis. Interestingly, risedronate suppressed ERK activity in osteoclasts at no less than 3 M, while it significantly suppressed Akt activity at 0.03 M (Figure 5D). Regulation of osteoclast apoptosis and activity by the ERK and Akt pathways, respectively. To analyze the role of ERK and Akt signals in osteoclasts, we separately activated these pathways in osteoclasts by infecting them with AxAktCA or AxMekCA. As shown in Figure 6A, these adenoviruses efficiently activated the Akt and ERK pathways in osteoclasts as shown by phospho-Akt or phospho-ERK immunoblotting. The mean ⫾ SD risedronate-induced reduction rate of OcN was 70.0 ⫾ 12.0% with empty vector, 40.2 ⫾ 8.5% with AxAktCA, and 8.4 ⫾ 12.1% with AxMekCA. Both AxAktCA and AxMekCA infection significantly reversed the effect of 3914 MATSUMOTO ET AL risedronate compared to empty vector infection (P ⬍ 0.05 for empty vector versus AxAktCA, P ⬍ 0.01 for empty vector versus AxMekCA), although risedronate still significantly reduced the number of osteoclasts infected with AxAktCA. The effect of AxMekCA was stronger than that of AxAktCA (8.4 ⫾ 12.1% versus 40.2 ⫾ 8.5%; P ⬍ 0.05) (Figure 6B). We then examined the effect of the Akt and Figure 5. A, Dose-dependent suppression of pit formation by risedronate. B and C, Time-dependent (B) and dose-dependent (C) effects of risedronate on Akt and ERK activity in osteoclasts shown by Western blotting with anti–phospho-ERK antibody and anti–phosphoAkt antibody. The blots were reprobed with anti-ERK antibody and anti-Akt antibody, respectively. D, Quantitative enzyme-linked immunosorbent assay for the measurement of phospho-Akt and phospho-ERK in osteoclasts treated with increasing concentrations of risedronate for 12 hours. Cultures stimulated with macrophage colonystimulating factor were used as the positive control. In A and D, experiments were repeated at least 3 times, and results are expressed as the mean ⫾ SD of data obtained from 6 independent culture wells in 1 typical experiment. ⴱ ⫽ P ⬍ 0.01 versus cultures without risedronate treatment. See Figure 1 for definitions. Figure 6. A, Western blotting shows adenovirus vector–mediated expression of constitutively active MEK-1 and Akt-1 in osteoclasts. B, Risedronate-induced reduction of osteoclast number was reversed by AxMekCA infection. AxAktCA only partially reversed the effect of risedronate. Left, TRAP staining of representative cultures. Bars ⫽ 100 m. Right, Number of TRAP-positive osteoclasts shown as a percentage of the cells at time zero. C, Risedronate-induced suppression of bone resorption was completely reversed by AxAktCA infection but not by AxMekCA infection. Left, Resorption pits of representative cultures. Bars ⫽ 100 m. Right, Pit area per cell. D, Risedronateinduced disruption of sealing zone formation was completely reversed by AxAktCA infection but not by AxMekCA infection. Left, Rhodamine-labeled phalloidin staining of representative cultures. Bars ⫽ 50 m. Right, Proportion of osteoclasts with sealing zone. In B–D, ⴱ ⫽ P ⬍ 0.01 versus the corresponding cultures without risedronate. Experiments were performed at least 3 times, and results are expressed as the mean ⫾ SD of data obtained from 6 independent culture dishes in 1 typical experiment. See Figure 1 for definitions. ERK pathways on bone resorption. The pit-forming activity of osteoclasts infected with the control vector and AxMekCA was almost completely suppressed by 30 M risedronate. In contrast, osteoclasts infected with DISTINGUISHING THE FUNCTIONS OF RISEDRONATE AxAktCA exhibited unaltered pit-forming activity in the presence of risedronate (Figure 6C). To obtain insight into the mechanism by which Akt activation cancelled the inhibitory effect of risedronate on bone resorption, we examined the sealing zone formation of osteoclasts cultured on dentin slices by rhodamine-labeled phalloidin staining. Sealing zone formation on the dentin slices was reduced to ⬃50% by risedronate treatment in osteoclasts infected with control vector and AxMekCA. In contrast, sealing zone formation was not affected by risedronate treatment in osteoclasts infected with AxAktCA (Figure 6D). These results suggest that the antiresorptive effect of risedronate is due to the disruption of cytoskeletal organization, which is caused by reduced Akt activity. DISCUSSION The nitrogen-containing bisphosphonates, including risedronate and alendronate, are among the most effective antiosteoporosis drugs, and clinical data have been accumulating in support of their effectiveness in reducing osteoporotic fractures. However, despite such clinical evidence, their exact mechanism of action remains to be elucidated. Although the nitrogencontaining bisphosphonates induce osteoclast apoptosis, recent studies indicate that they suppress the boneresorbing activity of osteoclasts at 10-fold lower doses than those required to induce osteoclast apoptosis (15), suggesting an antiresorptive mechanism of action other than osteoclast apoptosis. Nitrogen-containing bisphosphonates are known to inhibit the mevalonate pathway and reduce the prenylation of small GTP-binding proteins, such as Ras, Rho, Rac, and Cdc42 (6), and a critical role of the small GTP-binding proteins in osteoclast function has recently been reported (7–10). However, the upstream and downstream signaling pathways of small G proteins in osteoclasts have not been fully determined (21). In particular, it remains elusive whether nitrogencontaining bisphosphonate–induced osteoclast apoptosis and suppression of bone resorption are independently regulated or are controlled through the same mechanism. In this study, an effort was made to distinguish the molecular mechanisms underlying the proapoptotic and the antiresorptive effects of risedronate, and it was found that the ERK/Bim axis mainly regulates osteoclast apoptosis and that the Akt pathway regulates bone resorption. In addition, the in vivo effect of risedronate was elucidated using apoptosis-defective Bim⫺/⫺ mice. We first investigated the risedronate-induced os- 3915 teoclast apoptosis signaling pathway in vitro, and we found that the osteoclast apoptosis induced by risedronate occurs via the mitochondrial pathway, which is mediated by Bcl-2 family proteins. We previously reported that the proapoptotic BH3-only domain protein Bim plays a crucial role in mitochondria-mediated osteoclast apoptosis, and that the Bim expression level is posttranslationally regulated by the ERK-induced ubiquitin/proteasome pathway (17). Risedronate treatment increased Bim expression at the protein level, and Bim⫺/⫺ mouse osteoclasts were resistant to risedronateinduced apoptosis as compared to WT mouse osteoclasts, indicating an essential role of Bim in risedronate-induced osteoclast apoptosis. However, bone-resorbing activity was almost completely abolished by risedronate treatment even in Bim⫺/⫺ mouse osteoclasts as well as in WT mouse osteoclasts. In addition, mandatory activation of ERK pathways by adenovirus vector–mediated overexpression of constitutively active MEK-1 significantly reversed risedronate-induced reduction of OcN even in the presence of risedronate, but was unable to recover the bone-resorbing activity of osteoclasts suppressed by risedronate. These results clearly demonstrate that the ERK/Bim axis critically regulates the proapoptotic but not the antiresorptive effect of risedronate. This was further confirmed in vivo by analyzing the effect of risedronate on Bim⫺/⫺ mice. Bim⫺/⫺ mouse osteoclasts were resistant to risedronate-induced apoptosis in vivo as well; nonetheless, risedronate suppressed bone resorption in Bim⫺/⫺ mice as efficiently as in WT mice, as determined by the decrease in serum CTX-I and the increase in BV. These results indicate that risedronate’s in vivo effect of increasing bone mass is mainly caused by suppressing the bone-resorbing activity of osteoclasts and not by inducing osteoclast apoptosis. Weinstein et al recently reported an increase in the number of osteoclasts and the appearance of giant osteoclasts in human bone biopsy samples after longterm oral alendronate treatment (16). They attributed the presence of the giant osteoclasts to the enhanced survival of osteoclasts, which enabled them to fuse with other mononuclear cells. In our risedronate treatment experiments, giant osteoclasts comprised ⬃5% of the total OcN both in the risedronate-treated WT and Bim⫺/⫺ mice as compared to 1% in the untreated mice. This suggests that nitrogen-containing bisphosphonate– induced giant osteoclast formation is independent of ERK/Bim-induced osteoclast apoptosis. Formation of giant osteoclasts by nitrogen-containing bisphospho- 3916 nates may need changes other than the enhancement of survival, such as the detachment from bone surface, the structure change of cytoskeleton, and the expression of cell adhesion factors. Further study will be required to clarify the exact mechanism of giant osteoclast formation. We found that risedronate suppressed the boneresorbing activity of osteoclasts at concentrations equivalent to those required to suppress Akt activity, while the induction of apoptosis and ERK suppression required much higher concentrations. These observations led us to the hypothesis that the antiresorptive and proapoptotic effects of risedronate are mediated by the Akt and ERK pathways, respectively. This hypothesis was confirmed by the fact that the risedronate-induced suppression of bone resorption was completely reversed by the activation of the Akt pathway by AxAktCA infection, while bone resorption was markedly suppressed in control vector– or AxMekCA-infected osteoclasts. After 16-hour incubation on dentin slices, a sealing zone was formed in 70–80% of the total osteoclasts infected with the control vector, AxAkt CA , and AxMekCA under the untreated condition. Risedronate treatment reduced the intact sealing zone formation to less than half in the osteoclasts infected with the control vector and AxMekCA, but had no effect on the osteoclasts infected with AxAktCA. These results suggest that risedronate-induced disruption of the sealing zone in osteoclasts was caused by the suppression of the Akt activity. Our findings suggest that the ERK pathway mainly regulates osteoclast survival and that the Akt pathway regulates osteoclast activity. However, overexpression of AktCA in fact increased OcN at the basal level and partially reversed risedronate-induced reduction of OcN (Figure 6B), indicating that the Akt pathway also regulates osteoclast survival. This was consistent with our previous observation that the overexpression of dominant-negative Rac1, which suppressed macrophage colony-stimulating factor–dependent activation of the Akt pathway but not the ERK pathway, partially inhibited osteoclast survival (9). Further study is required to segregate the roles of the ERK and Akt pathways in osteoclast survival. Akt is a downstream effector of phosphatidylinositol 3-kinase (PI3K) (22). It was reported that wortmannin, a specific inhibitor of PI3K, disrupted actin ring formation in osteoclasts (23). In addition, a recent study showed that deficiency of p85␣, a regulatory subunit of class IA PI3K, resulted in reduced Akt activity and MATSUMOTO ET AL defective actin ring formation in osteoclasts (24). These reports support our hypothesis that Akt critically regulates the cytoskeletal organization in osteoclasts. Further study is required to fully understand the molecular mechanisms underlying the Akt-mediated regulation of the osteoclast cytoskeleton. It should also be clarified in the future whether nitrogen-containing bisphosphonates other than risedronate exert their effect through the same mechanism. In addition, it remains unclear whether risedronate in fact inhibits bone resorption by suppressing osteoclast Akt activity in vivo. Further understanding of the functional mechanism of risedronate and other nitrogen-containing bisphosphonates would enable their optimal use in the clinic. ACKNOWLEDGMENTS The authors thank Reiko Yamaguchi and Hajime Kawahara (Department of Orthopaedic Surgery, University of Tokyo), who provided expert technical assistance. Pacific Edit reviewed the manuscript prior to submission. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Tanaka had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Matsumoto, Nagase, Tanaka. Acquisition of data. Matsumoto, Nagase, Iwasawa, Yasui, Masuda. Analysis and interpretation of data. Matsumoto, Kadono, Nakamura, Tanaka. REFERENCES 1. Russell RG, Watts NB, Ebetino FH, Rogers MJ. Mechanisms of action of bisphosphonates: similarities and differences and their potential influence on clinical efficacy. Osteoporos Int 2008;19: 733–59. 2. Ezra A, Golomb G. Administration routes and delivery systems of bisphosphonates for the treatment of bone resorption. Adv Drug Deliv Rev 2000;42:175–95. 3. Sato M, Grasser W, Endo N, Akins R, Simmons H, Thompson DD, et al. Bisphosphonate action: alendronate localization in rat bone and effects on osteoclast ultrastructure. J Clin Invest 1991; 88:2095–105. 4. Thompson K, Rogers MJ, Coxon FP, Crockett JC. Cytosolic entry of bisphosphonate drugs requires acidification of vesicles after fluid-phase endocytosis. Mol Pharmacol 2006;69:1624–32. 5. Rogers MJ, Russell RG, Blackburn GM, Williamson MP, Watts DJ. Metabolism of halogenated bisphosphonates by the cellular slime mould Dictyostelium discoideum. Biochem Biophys Res Commun 1992;189:414–23. 6. Luckman SP, Hughes DE, Coxon FP, Graham R, Russell G, Rogers MJ. Nitrogen-containing bisphosphonates inhibit the mevalonate pathway and prevent post-translational prenylation of GTP-binding proteins, including Ras. J Bone Miner Res 1998;13: 581–9. DISTINGUISHING THE FUNCTIONS OF RISEDRONATE 7. Miyazaki T, Katagiri H, Kanegae Y, Takayanagi H, Sawada Y, Yamamoto A, et al. Reciprocal role of ERK and NF-B pathways in survival and activation of osteoclasts. J Cell Biol 2000;148: 333–42. 8. Zhang D, Udagawa N, Nakamura I, Murakami H, Saito S, Yamasaki K, et al. The small GTP-binding protein, rho p21, is involved in bone resorption by regulating cytoskeletal organization in osteoclasts. J Cell Sci 1995;108:2285–92. 9. Fukuda A, Hikita A, Wakeyama H, Akiyama T, Oda H, Nakamura K, et al. Regulation of osteoclast apoptosis and motility by small GTPase binding protein Rac1. J Bone Miner Res 2005;20: 2245–53. 10. Razzouk S, Lieberherr M, Cournot G. Rac-GTPase, osteoclast cytoskeleton and bone resorption. Eur J Cell Biol 1999;78:249–55. 11. Hughes DE, Wright KR, Uy HL, Sasaki A, Yoneda T, Roodman GD, et al. Bisphosphonates promote apoptosis in murine osteoclasts in vitro and in vivo. J Bone Miner Res 1995;10:1478–87. 12. Reszka AA, Halasy-Nagy JM, Masarachia PJ, Rodan GA. Bisphosphonates act directly on the osteoclast to induce caspase cleavage of Mst1 kinase during apoptosis: a link between inhibition of the mevalonate pathway and regulation of an apoptosis-promoting kinase. J Biol Chem 1999;274:34967–73. 13. Selander K, Lehenkari P, Vaananen HK. The effects of bisphosphonates on the resorption cycle of isolated osteoclasts. Calcif Tissue Int 1994;55:368–75. 14. Alakangas A, Selander K, Mulari M, Halleen J, Lehenkari P, Monkkonen J, et al. Alendronate disturbs vesicular trafficking in osteoclasts. Calcif Tissue Int 2002;70:40–7. 15. Halasy-Nagy JM, Rodan GA, Reszka AA. Inhibition of bone resorption by alendronate and risedronate does not require osteoclast apoptosis. Bone 2001;29:553–9. 16. Weinstein RS, Roberson PK, Manolagas SC. Giant osteoclast 3917 17. 18. 19. 20. 21. 22. 23. 24. formation and long-term oral bisphosphonate therapy. N Engl J Med 2009;360:53–62. Akiyama T, Bouillet P, Miyazaki T, Kadono Y, Chikuda H, Chung UI, et al. Regulation of osteoclast apoptosis by ubiquitylation of proapoptotic BH3-only Bcl-2 family member Bim. EMBO J 2003;22:6653–64. Bouillet P, Metcalf D, Huang DC, Tarlinton DM, Kay TW, Kontgen F, et al. Proapoptotic Bcl-2 relative Bim required for certain apoptotic responses, leukocyte homeostasis, and to preclude autoimmunity. Science 1999;286:1735–8. Tanaka S, Takahashi T, Takayanagi H, Miyazaki T, Oda H, Nakamura K, et al. Modulation of osteoclast function by adenovirus vector-induced epidermal growth factor receptor. J Bone Miner Res 1998;13:1714–20. Nagase Y, Iwasawa M, Akiyama T, Kadono Y, Nakamura M, Oshima Y, et al. The anti-apoptotic molecule Bcl-2 regulates the differentiation, activation and survival of both osteoblasts and osteoclasts. J Biol Chem 2009;284:36659–69. Coxon FP, Rogers MJ. The role of prenylated small GTP-binding proteins in the regulation of osteoclast function. Calcif Tissue Int 2003;72:80–4. Scheid MP, Woodgett JR. PKB/AKT: functional insights from genetic models. Nat Rev Mol Cell Biol 2001;2:760–8. Nakamura I, Takahashi N, Sasaki T, Tanaka S, Udagawa N, Murakami H, et al. Wortmannin, a specific inhibitor of phosphatidylinositol-3 kinase, blocks osteoclastic bone resorption. FEBS Lett 1995;361:79–84. Munugalavadla V, Vemula S, Sims EC, Krishnan S, Chen S, Yan J, et al. The p85␣ subunit of class IA phosphatidylinositol 3-kinase regulates the expression of multiple genes involved in osteoclast maturation and migration. Mol Cell Biol 2008;28:7182–98.