Cells which exhibit high motility, such as osteoclasts, dendritic cells, and macrophages, have actin-rich structures called podosomes, which mediate cell attachment and migration. Podosomes are composed of dense F-actin columns (the actin core) and a diffuse meshwork of F-actin surrounding the actin core (the actin cloud).1 The osteoclast is a highly differentiated, multinucleated, bone-resorbing cell of hematopoietic origin. The podosomes in fully differentiated osteoclasts cultured on glass are observable at the cell periphery and form a ring like structure called the podosome belt.2 The sealing zone is another ring-like F-actin–rich structure that bone-resorbing osteoclasts form on mineralized matrix.3 The sealing zone serves as the part of the osteoclast that attaches to a target and provides an enclosed space known as a “resorption lacuna” between cells and bone. The membrane of the osteoclast circumscribed within this sealing zone is folded in order to form a ruffled border through which proton and matrix-degrading enzymes are released into the resorption lacunae. High-resolution electron microscopy revealed that the podosome belt and the sealing zone employ the same structural unit of F-actin and suggested that the difference between them is the result of the degree of compaction.4 However, the regulatory mechanism underlying this compaction is still unknown.
Microtubules, one of the major components of the cytoskeleton, are polymers of α-tubulin and β-tubulin. Microtubules have a fast-growing end (plus-end) and slow-growing end (minus-end). In most cells, the minus-end binds to the microtubule organizing center and the plus-end extends toward the periphery. The plus-end microtubules switch from the assembly to the disassembly phase, and this phenomenon, dynamic instability, plays a crucial role in microtubule attachment to targets.5 The plus-end possesses tracking proteins (+TIPs), a family of microtubule-associated proteins (MAPs) that accumulate and are involved in both the stabilization of the microtubule and the interaction between the microtubule and the cell cortex. Mammalian osteoclasts have few centrosomal microtubule organizing centers, and the microtubules originate from pericentriolar matrix proteins on the nuclei.6 In fully differentiated mammalian osteoclasts attached on glass, most of the nuclei are clustered at the cell periphery. From this location the microtubules radiate in a circular pattern.6
Recent studies have revealed that microtubules have an essential role in regulating the formation of podosome belts and sealing zones in osteoclasts.7, 8 In particular, the acetylated form of tubulin, which is an indication of stabilized tubulin, is considered to play an essential role in regulating osteoclast activity.2, 9 Hazama and colleagues10 reported that the deacetylation of tubulin in osteoclasts is associated with the release of lytic granules required for bone-resorption. Destaing and colleagues9 reported that the stabilization of the podosome belt and the sealing zone is associated with an increase in acetylated tubulin, which is in turn controlled by the Rho-mDia2-HDAC6 pathway, indicating the central role of HDAC6 in regulation of tubulin deacetylation. However, the specific mechanisms of the anchoring and stabilization of tubulin by MAPs in osteoclasts are not well understood.
Protein kinase B (Akt) is a major downstream effector of phosphatidylinositol 3-kinase (PI3K) and mediates cellular processes such as proliferation, migration, and survival. The Akt family members consist of Akt1/PKBα,Akt2/PKBβ, and Akt3/PKBγ. These isoforms are similar in structure and size, and are considered to be redundant to one another.11 Akt1 and Akt2 are ubiquitously expressed in mammals, whereas Akt3 is expressed dominantly in the limited tissues such as the testis and brain.11
This study aimed to clarify the regulatory effect on the bone-resorbing activity of osteoclasts by the Akt pathway. Akt was found to positively regulate the bone-resorbing activity of osteoclasts by stabilizing microtubules and enhancing sealing zone formation. The microtubule stabilization effect of Akt was regulated by the binding of MAPs with microtubules through glycogen synthase kinase-3β (GSK-3β). Osteoclast-specific Akt1 and Akt2 double-deficient mice were shown to exhibit mild osteosclerosis caused by a decreased bone-resorbing function of osteoclasts. These results suggest that Akt plays an essential role in bone resorption by regulating the stable interaction between microtubules and the cell cortex, as well as regulating dynamic microtubule formation.
Subjects and Methods
Newborn and 5-week-old male C57BL/6J mice were purchased from Sankyo Labo Service Co. (Tokyo, Japan). The breeding and genotyping of the Akt1fl/fl and Akt2fl/fl mice (kindly provided by Morris J. Birnbaum, University of Pennsylvania) were performed as described.12, 13 To generate osteoclast-specific Akt1 and Akt2 double-conditional knockout (DKO) mice, we used cathepsin K-Cre mice (kindly provided by Shigeaki Kato, Soma Chuo Hospital), in which the Cre recombinase gene is knocked into the cathepsin K locus and specifically expressed in osteoclasts.14 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, as described.15 Briefly, when murine osteoblastic cells and bone marrow cells were cocultured on collagen gel–coated dishes in the presence of 10 nM 1α,25-dihydroxyvitamin D3 [1α,25(OH)2D3] and 1 µM 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, Tokyo, Japan) for 10 minutes and then used for the pit formation assay and the survival assay as follows. For the bone resorption assay, the cells were resuspended in α modified essential medium (α-MEM) containing 10% fetal bovine serum (FBS), replated on dentin slices, and cultured for 24 hours. After cells were removed by treating the dentin slices with 1 M NH4OH, the resorption areas were visualized by staining with 1% toluidine blue. The resorption pit area was quantified using an image analyzing system.
The sealing zone formation on dentin slices was analyzed as follows. The cells were cultured on dentin slice as described in the previous paragraph. After 16 hours, the cells were fixed in PBS containing 4% paraformaldehyde for 10 minutes and then stained with tartrate-resistant acid phosphatase (TRAP). TRAP staining was performed at pH 5.0 in the presence of L(+)-tartaric acid using naphthol AS-MX phosphate (Sigma-Aldrich, St. Louis, MO, USA) in N-N dimethyl formamide as the substrate. Cells were then incubated for 30 minutes with rhodamine-conjugated phalloidin solution (Molecular Probes, Eugene, OR, USA) to visualize F-actin. Sealing zone formation on dentin slices was observed under fluorescence microscopy (Biozero; KEYENCE, Woodcliff Lake, NJ, USA). The sealing zone formation rate is represented as the proportion of the cells having an uninterrupted ring-like F-actin structure among the TRAP-positive multinucleated cells.
Triton cytoskeleton extraction
Extraction of the triton-insoluble cytoskeletal fraction was performed as described.16 Briefly, osteoclast cultures were washed twice with ice-cold PBS, once with Buffer 1 (0.1 M PIPES [pH 6.9], 1 mM MgSO4, 2 M glycerol, 2 mM EGTA, 0.02 trypsin inhibitory units per mL aprotinin) and then incubated for 3 minutes with Buffer 2 (Buffer1 + 0.2% [wt/vol] Triton X-100). This fraction is referred to as the Triton X-100–solubule or “cytosolic” fraction. The Triton X-100–insoluble remainders were washed twice with Buffer 3 (1 M piperazine-N,N′-bis(2-ethanesulfonic acid) [PIPES] [pH 6.9], 1 mM MgSO4, 2 M glycerol, 0.02 trypsin inhibitory units per mL aprotinin), and dissolved with Ca2+-containing medium (0.1 M PIPES [pH 6.9], 1 mM MgSO4, 5 mM CaCl2, 0.02 trypsin inhibitory units per mL aprotinin). The lysate was centrifuged at 12,000 g for 10 minutes at 4°C, and the supernatant was collected. This fraction is referred to as the Triton-X-100–insoluble fraction or “Triton cytoskeleton” fraction. The lysates were used for immunoblotting analysis.
Western blotting and immunoprecipitation
All the extraction procedures were performed at 4°C or on ice. Cells were washed with ice-cold PBS, and proteins were extracted with TNE buffer (1% NP-40, 10 mM Tris-HCl [pH 7.8], 150 mM NaCl, 1 mM EDTA, 2 mM Na3VO4, 10 mM NaF, and 10 µg/mL of aprotinin). The lysates were clarified by centrifugation at 12,000g for 10 minutes. For Western blotting analysis, lysates were subjected to SDS-PAGE with 7.5% to 15% Tris-Glycin gradient gels or 15% Tris-Glycin gels, and transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). After blocking with 6% milk/TBS-T, membranes were incubated with primary antibodies to phospho-Akt (Ser473), Akt, phospho-GSK-3β (Ser9), GSK-3β, β-Catenin, Akt1, Akt2 (Cell Signaling Technology Inc., Beverly, MA, USA), APC, CLIP-170, Lis1 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), p150 Glued (BD Biosciences Inc., San Diego, CA, USA), EB1, acetylated tubulin, α-tubulin, β-actin (Sigma-Aldrich, St. Louis, MO, USA), or CLASP2 (ProteinTech Group, Chicago, IL, USA), followed by horseradish peroxidase (HRP)-conjugated goat anti-mouse immunoglobulin G (IgG), goat anti-rabbit IgG (Promega, Madison, WI, USA), and donkey anti-goat IgG (Santa Cruz Biotechnology). Immunoreactive bands were visualized with ECL Plus (Amersham, Pittsburgh, PA, USA), 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 then reprobed with additional antibodies.
Immunoprecipitation was performed by incubating 500 mg of total cell protein lysate with 2 µg of antibody for 2 hours on ice and then adding 20 µL of protein G-agarose. After incubation for 1 hour at 4°C with end-over-end mixing, the immune complex was recovered by centrifugation and washed five times with buffer containing 20 mM HEPES-NaOH, pH 7.4, 150 mM NaCl, 10 mM EDTA, and 0.02% NP-40. The immunoprecipitates were then subjected to SDS-PAGE and Western blot analysis.
Expression constructs and gene transduction
The adenovirus vectors used in the experiments and the genes carried by the vectors were as follows: AxTb-YFP (tubulin-yellow fluorescence fusion protein gene), AxEB1-YFP (the EB1-yellow fluorescence fusion protein gene), AxCre (Cre recombinase gene), and AxAktCA (the constitutively active Akt1 gene). AxTb-YFP and AxEB1-YFP were provided by Dr. Takao Nakata (Tokyo Medical and Dental University, Tokyo, Japan). AxAktCA was provided by Dr. Hideki Katagiri (Tohoku University, Miyagi, Japan). Constitutively active Akt1 was generated by adding a myristoylation signal sequence to its N terminus. The control virus (Ax1W1) was provided by Dr. Izumu Saito (The University of Tokyo, Tokyo, Japan). Viral titers were determined by the end point dilution assay, and the viruses were used at 50 multiplicity of infection (MOI). The infection of adenovirus vectors into osteoclasts was carried out following the method described.17 In short, on day 4 of the culture, when osteoclasts began to appear, mouse cocultures were incubated for 1 hour at 37°C with a small amount of α-MEM containing the recombinant adenoviruses at the desired MOI. Cells were then washed twice with PBS and further incubated at 37°C in α-MEM containing 10% FBS, 10 nM 1α,25(OH)2D3, and 1 µM PGE2. Experiments were performed 36 hours after the infection.
Osteoclasts infected with AxTb-YFP were purified by treating them with 1 µM nocodazole for 30 minutes. After 30 minutes, the cells were washed three times with PBS, and then fixed with 4% paraformaldehyde to detect nocodazole-resistant tubulin. Cells were fixed with 4% paraformaldehyde for 0, 30, 60, 90, and 120 minutes after the washout of nocodazole for observation of the microtubule reorganization.
Immunofluorescence and time-lapse imaging
For live imaging experiments, osteoclasts cultured on glass coverslips were infected with AxTb-YFP and incubated with α-MEM containing 5 µg/mL Hoechst 33258 for 30 minutes. Osteoclasts were treated with nocodazole as described in Nocodazole treatment above, and incubated with α-MEM, or with α-MEM containing 1 µM Akt inhibitor IV (Calbiochem, San Diego, CA, USA), or with α-MEM containing 20 mM LiCl. Osteoclasts were placed in a thermostat containing chamber for microscopic analysis, allowing the temperature to be maintained at 37°C throughout the experiment, and cells were imaged under fluorescence microscopy (BZ-8100; KEYENCE). Images were collected at 90-second intervals for 120 minutes.
Radiographic, micro–computed tomography, and histomorphometric analyses
Plain radiographs were taken using a soft X-ray apparatus (CMB-2; SOFTEX, Tokyo, Japan). Micro–computed tomography (µCT) scanning of the distal femur was performed using a ScanXmate-L090 Scanner (Comscantechno Co., Ltd., Yokohama, Japan). Three-dimensional microstructural image data were reconstructed and structural indices were calculated using TRI/3D-BON software (RATOC Systems, Osaka, Japan).
Tissues were fixed in 4% paraformaldehyde/PBS, decalcified in 10% EDTA, embedded in paraffin, and cut into 3-µm-thick sections. Hematoxylin and eosin (H&E) staining was performed according to the standard procedure. Osteoclasts were identified by TRAP staining. Histomorphometric analysis was performed in undecalcified sections from 0.2 mm below the growth plate to a point 1.125 mm distal of the primary spongiosa of the proximal tibia. For double-labeling, mice were injected subcutaneously with 16 mg/kg body weight of calcein on 4 days and 1 day before euthanasia.
Each series of experiments was repeated at least three times. The results are expressed as the mean ± SD. Statistical analyses were performed using a two-tailed unpaired Student's t test or ANOVA analysis.
An Akt inhibitor impairs the bone-resorbing activity and cytoskeletal organization of osteoclasts
We first examined the effect on osteoclasts of the Akt inhibitor IV, which was designed to target an adenosine triphosphate (ATP)-binding site of a kinase immediately upstream of Akt but downstream of PI3K. Treatment with 1 µM of the Akt inhibitor significantly reduced the pit formation without affecting the survival of osteoclasts (Fig. 1A, B). The proportion of osteoclasts containing sealing zones was significantly decreased by Akt inhibitor treatment (Fig. 1C, D). These results suggest that Akt regulates the bone-resorbing activity of osteoclasts by its effect on sealing zone formation.
Recent studies have demonstrated that the formation of both podosome belts and sealing zones in osteoclasts is associated with an increase in tubulin acetylation corresponding to microtubule stabilization.9 We therefore compared the structures of the podosome belts and acetylated tubulin between osteoclasts with and without the Akt inhibitor under fluorescent microscopy. The microtubules in mature osteoclasts on glass consist of two forms; one is a circular pattern of microtubules at the cell periphery, and the other is a pattern of microtubules radiating from the central cytoplasm to the cell periphery.18 Acetylated tubulin was distributed uniformly along these structures in mature osteoclasts, and the Akt inhibitor disrupted these orderly microtubule structures (Fig. 1E). The podosome belt positioning at the cell periphery in the untreated osteoclasts was rearranged in the cytoplasm after Akt inhibitor treatment (Fig. 1E). These results suggest that Akt critically regulates the structure and stability of microtubules in osteoclasts.
The forced activation of Akt results in the stabilization of microtubules and increased bone-resorbing activity
We then evaluated the effect of a gain-of-function of Akt by introducing the constitutively active Akt1 gene into osteoclasts using an adenovirus vector (AxAktCA). The forced activation of Akt by adenovirus infection increased the phosphorylation level of GSK-3β, a downstream effector of Akt, and also the amount of acetylated tubulin (Fig. 2A). The proportion of osteoclasts containing sealing zones were increased by AktCA expression, and their bone-resorbing activity was significantly enhanced (Fig. 2B, C). Immunofluorescence staining showed that a thick layer of acetylated tubulin was observed in the cell periphery of osteoclasts infected with AxAktCA (Fig. 2D). Nocodazole treatment disrupted microtubule organization in control osteoclasts, whereas the osteoclasts infected with AxAktCA contained a nocodazole-resistant pool of stable microtubules (Fig. 2D).
GSK3β inhibition affects the stabilization of microtubules
To confirm that the increased acetylated tubulin formation was regulated through the Akt/GSK-3β axis, we assessed the effect of LiCl and SB216763, which are specific inhibitors of GSK-3β. The osteoclasts were cultured for 5 hours in the presence of the indicated concentrations of LiCl or SB216763. LiCl and SB216763 dose-dependently increased the expression of the acetylated tubulin (Fig. 3A). Immunofluorescence analysis showed that the treatment with LiCl did not affect the structure of actin or tubulin, but rather induced the accumulation of the acetylated tubulin throughout the entire cytoplasm and increased the nocodazole-resistant pool of stable microtubules (Fig. 3B).
Akt regulates microtubule formation through GSK-3β
We next examined the involvement of Akt/GSK-3β in the dynamic formation of tubulin. The disrupted tubulin structure that had lost its circular pattern at the cell periphery as the result of nocodazole treatment was restored within 1 or 2 hours after the removal of nocodazole (Fig. 4A). Such recovery was not observed in osteoclasts treated with the Akt inhibitor and was promoted in the cells treated with LiCl. In accordance with the formation of the microtubules, the nuclei gathered in the center of the cell after nocodazole treatment were rearranged in the cell periphery in the control osteoclasts. This nuclear translocation was quantified by measuring the proportion of the nuclei placed at the cell periphery among all of the nuclei (peripheral nuclear translocation ratio, PNT ratio). The PNT ratio was significantly reduced by nocodazole treatment (67% ± 20% versus 9% ± 7%, p < 0.01), and recovered to the basal level within 60 minutes in the control osteoclasts and 30 minutes in the cells treated with LiCl. In contrast, the decreased PNT ratio was not fully recovered even 120 minutes after the removal of nocodazole in osteoclasts cultured in the presence of the Akt inhibitor (Fig. 4B). These results suggested that Akt correlates with the formation of microtubules and the accompanying nuclear translocation, which is mediated by GSK-3β.
The binding of MAPs with tubulin is regulated by Akt
Nuclear migration is known to be mediated by the activity of dynein/dynactin at the cell cortex.19 The dynactin complex is able to bind to both microtubules and dynein, and activates dynein's minus-end–directed motor activity. Many +TIPs, including dynactin, have important roles in anchoring and stabilizing microtubules. At the cell cortex, where microtubules are captured, dynein's minus-end–directed motor activity is converted to a force that pulls nuclei toward the cortex. Taken together, these findings indicate that the anchoring of microtubules to the cytoplasmic membrane is regulated by Akt in osteoclasts.
We therefore compared the binding of several representative MAPs to the Triton cytoskeleton between osteoclasts infected with control adenovirus vector or AxAktCA. Among the MAPs examined, the amount of EB1, APC, and p150Glued anchored to the Triton cytoskeleton was increased in osteoclasts infected with AxAktCA compared to control osteoclasts (Fig. 5A). p150Glued, a major component of dynactin, has sites that bind to EB1 and dynein.20 Mammalian EB1 and APC directly bind to each other and cooperate in the regulation of microtubule dynamics and microtubule-dependent processes under diverse conditions, including cell migration and mitosis.21, 22 We therefore examined whether Akt regulates the EB1-APC complex formation in osteoclasts. In control vector or AxAktCA-infected osteoclasts, APC was coimmunoprecipitated with EB1 (Fig. 5B). An association between EB1 and APC was observed in both the control and AxAktCA-infected osteoclasts, and the level of APC was substantially increased by AxAktCA infection (Fig. 5B).
Colocalization of EB1 and APC is lost in Akt1-deficient and Akt2-deficient osteoclasts
To further confirm the role of Akt in regulating MAPs, we generated osteoclast-specific Akt1 and Akt2 conditional knockout mice by crossing Akt1fl/fl, Akt2fl/fl, and cathepsin K-Cre knock-in mice.14 The cathepsin K-Cre+/–Akt1fl/flAkt2fl/fl mice (referred to herein as DKO mice) were born alive at the predicted Mendelian frequencies. Akt1 and Akt2 were markedly reduced in the osteoclasts from the DKO mice, whereas their expression in osteoblasts (Fig. 6A) and other tissues (data not shown) in DKO mice was comparable to that found in normal Akt1fl/flAkt2fl/fl littermates (referred to herein as DF mice representing double flox) (Fig. 6A). Osteoclasts were generated from bone marrow cells obtained from DKO or DF mice using receptor activator of NF-κB ligand (RANKL) and macrophage colony-stimulating factor (M-CSF). There was no apparent difference in RANKL-induced and M-CSF–induced osteoclast development or osteoclast-related gene expression between DKO and DF bone marrow cells (data not shown). Immunoprecipitation of EB1 from both DKO and DF osteoclasts showed that the association between EB1 and APC was substantially reduced by the deletion of Akt1 and Akt2 (Fig. 6B). Immunofluorescence analysis demonstrated that in the DF osteoclasts intense labeling of APC was detected in and around the nuclei and at the cell periphery, and APC colocalized well with EB1 except in the nuclei. In contrast, APC staining in DKO osteoclasts was detected potently in and around the nuclei and faintly at the cell periphery, and colocalized well with EB1 only around the nuclei (Fig. 6C). These results suggest that Akt has a role in regulating APC-EB1 complex formation in osteoclasts. We then compared the pit-forming activity and the sealing zone formation in the DF and DKO osteoclasts. The DKO osteoclasts exhibited a reduced bone-resorbing activity and sealing zone formation compared to the DF osteoclasts, which effect was recovered by the addition of LiCl (Fig. 6D–G).
Osteoclast-specific Akt1-deficient and Akt2-deficient mice shows increased bone mass due to impaired absorption of osteoclasts
Although Akt DKO mice grew normally with no apparent morphological abnormalities, 12-week-old Akt DKO mice exhibited increased bone mass compared to their littermate DF mice on radiographic (Fig. 7A) and histological analyses (Fig. 7D). µCT analysis revealed a significant increase in bone volume per tissue volume and a significant decrease in trabecular separation compared with normal Akt DF littermates (Fig. 7B, C). There was no significant difference in trabecular thickness or trabecular number between the DKO mice and their normal littermates (Fig. 7C). Histomorphometric analysis of 12-week-old Akt DKO mice revealed a significant increase in the bone volume of the primary spongiosa (Fig. 7D). The number of osteoclasts per bone perimeter measurement and the extent of the eroded surface were decreased in DKO mice. A noteworthy finding in the DKO mice was the increase in the number of rounded osteoclasts detached from the bone surface (Fig. 7E), and the ratio of detached osteoclasts per total osteoclasts was significantly higher in the DKO mice compared with the DF mice (27.9% ± 4.6% versus 9.1% ± 2.3%, p < 0.01) (Fig. 7F,G). These mice also displayed an apparent, although not significant, decrease in the mineral apposition rate and displayed a significant decrease in the bone formation rate, indicating a low turnover of the bone metabolism. These findings suggest that the increased bone mass observed in the Akt DKO mice was caused by a decrease in the bone-resorbing activity of mature osteoclasts, rather than by increased bone formation per se.
Previous studies have revealed various roles of Akt in osteoclasts. M-CSF and RANKL promote osteoclast survival in part by activating the Akt pathway.23, 24 Sugatani and Hruska25 reported in a study using small interfering RNA (siRNA) that Akt1/Akt2 plays an essential role in the differentiation, but not in the survival of osteoclasts. Several other studies demonstrated the role of Akt in the survival of osteoclasts.23, 24 We previously reported that Akt regulates bone-resorbing activity through its organizing effect on the formation of sealing zones.26 In this study, we have presented in vitro and in vivo evidence that Akt positively regulates osteoclast activity by controlling microtubule stability.
There is accumulating evidence that the actin organization in osteoclasts is regulated by microtubules.7, 8 In particular, acetylated tubulin is reported to be important for changes in the podosome structure during osteoclast maturation.2, 9 Hazama and colleagues10 reported that deacetylation of acetylated tubulin is essential for the secretion of the lytic granules needed for bone resorption. We showed by immunofluorescence staining that an Akt inhibitor disturbed podosome belt formation in mature osteoclasts with a dysregulation of acetylated tubulin (Fig. 1E). The acetylated tubulin expression level was increased in osteoclasts expressing AktCA, suggesting that Akt positively regulates microtubule stability.
At the plus-end, microtubules continuously switch between assembly and disassembly, a phenomenon called “dynamic instability.” MAPs bind to the microtubules to regulate their stability, and some MAPs have a role in stabilizing microtubules and mediating their interaction with the cell cortex. The interaction between these MAPs and tubulin is lost upon phosphorylation by GSK-3β, which is negatively regulated by Akt.27 An Akt inhibitor blocked the peripheral translocation of the nuclei and microtubules to the cell periphery, whereas LiCl stimulated it, suggesting that these phenomena are positively controlled by the Akt/GSK-3β axis.
Nuclear migration is an important event associated with many cellular events such as migration and division, in which dynein, a minus-end–directed motor, is known to be involved.19, 28, 29 The dynein anchored to the cell cortex has been proposed to reel in the microtubules and the nuclei attached to these microtubules.19 The observation that nuclear migration is promoted by the activation of Akt or inhibition of GSK-3β raises the possibility that the proteins involved in the anchoring of dynein to the cell cortex are regulated by the Akt/GSK-3β pathway. Evidence has been accumulating that dynamics and organization of the microtubules are regulated by PI3K and its downstream kinases Akt and GSK-3β. For example, centrosome separation and mitotic spindle orientation in Drosophila melanogaster are facilitated by the stabilization that takes place between the microtubules and cortex, which is mediated by Akt and Zeste-white 3 (Zw3; the D. melanogaster homologue of GSK-3).30 In migrating fibroblasts, microtubule stabilization is regulated at the leading edge, which is regulated by the PI3K/Akt pathway.31 To date, several classes of proteins have been shown to be associated with the microtubule plus-ends and to interact with each other.32 Our study revealed that the forced activation of Akt increased the association of p150Glued, EB1, and APC with the triton-insoluble cytoskeleton fraction. EB1 associates with the newly growing microtubule plus-end, coupled with a rapid dissociation from the older part.33 EB1 is considered to play an essential role in microtubule dynamics by regulating the connection between microtubules and various microtubule trapping proteins on the cell cortex.20 APC is classified as a tumor suppressor gene, and its mutations are known to cause familial adenomatous polyposis.34 APC can stabilize microtubules directly by interacting with microtubules with its microtubule-binding region located between amino acids 2219 and 2580,35, 36 or indirectly by interacting with EB1.21 The interaction between APC and EB1 is reported to be important for the formation of stable microtubules in migrating fibroblasts37 and the carrying out of mitosis.38 Phosphorylation by GSK-3β decreases the ability of APC to bind to microtubules, which results in decreased microtubule stability.39 Akt facilitates a stable interaction between the cell cortex and microtubules by phosphorylating GSK-3β and decreasing its activity.27 p150Glued, one of the proteins that accumulate in the Triton-insoluble cytoskeletal fraction in osteoclasts infected with AxAktCA, is a major component of the dynactin complex which mediates dynein-driven activity. Because dynein directly associates with EB1,40 it is speculated that Akt promotes the motility of microtubules and nuclei by connecting dynein to the cell cortex through the APC-EB1 complex. In fact, immunofluorescence staining of DKO osteoclasts revealed a decreased colocalization of APC and EB1 at the cell periphery. The decreased bone-resorbing activity and inhibition of sealing zone formation in Akt DKO osteoclasts were both recovered by an inhibition of GSK-3β activity with LiCl.
Mice lacking the Akt1 gene are viable but their size is small compared to wild-type littermates.41 Kawamura and colleagues42 revealed that conventional Akt1 knockout mice exhibit low-turnover osteoporosis as a result of the impeded differentiation and survival of osteoblasts and osteoclasts. They speculated that the impaired differentiation of osteoclasts in Akt1 knockout mice is due to a reduced expression of RANKL in osteoblasts and the direct effect of this on osteoclasts. Akt2 knockout mice exhibit insulin resistance, with elevated blood glucose, mild growth deficiency, and an age-dependent loss of adipose tissue.43, 44 Mice lacking both the Akt1 gene and the Akt2 gene die shortly after birth, exhibiting severe growth deficiency, muscle atrophy, and impeded adipogenesis.45 To elucidate the physiological role of Akt in osteoclasts, we generated osteoclast-specific Akt1 and Akt2 double-knockout mice. We found that the mice exhibited mild osteosclerosis as a result of the decreased bone-resorbing activity of osteoclasts, which is consistent with the in vitro results using osteoclasts generated from DKO mouse bone marrow cells. The lack of Akt1/2 did not completely inhibit bone resorption either in vivo or in vitro. This may be because the deletion efficiency of Akt1 and Akt2 in DKO was not complete enough, as shown in Fig. 6A, or there may be other molecule(s) that substitute the function of Akt1/2.
In summary, our results provide both in vivo and in vitro evidence that the signaling pathways regulated by the Akt/GSK-3β axis play an essential role in normal osteoclast function by regulating cytoskeletal organization. Further investigation of these pathways in osteoclasts will provide further insights into the molecular mechanisms regulating cytoskeletal organization and function.
All authors state that they have no conflicts of interest.
This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and Health Science research grants from the Ministry of Health, Labor, and Welfare of Japan (to ST and YK). We thank Reiko Yamaguchi and Hajime Kawahara (Department of Orthopaedic Surgery, The University of Tokyo), who provided expert technical assistance. Pacific Edit reviewed the manuscript prior to submission.
Authors' roles: Study design and conception: TM, ST. Study conduct and data collection: TM, YN, JH, NT, TY. Data analysis: TM, YN, YK, ST. Data interpretation: TM, YK, KU, TK, KN, ST. Drafting manuscript: TM, ST. Revising manuscript content: All authors. Approving final version of manuscript: All authors. TM and ST take responsibility for the integrity of the data analysis.