The p85α Subunit of Class I_A Phosphatidylinositol 3-Kinase Regulates the Expression of Multiple Genes Involved in Osteoclast Maturation and Migration

Abstract

Intracellular signals involved in the maturation and function of osteoclasts are poorly understood. Here, we demonstrate that osteoclasts express multiple regulatory subunits of class I_A phosphatidylinositol 3-kinase (PI3-K) although the expression of the full-length form of p85α is most abundant. In vivo, deficiency of p85α results in a significantly greater number of trabeculae and significantly lower spacing between trabeculae as well as increased bone mass in both males and females compared to their sex-matched wild-type controls. Consistently, p85α^−/− osteoclast progenitors show impaired growth and differentiation, which is associated with reduced activation of Akt and mitogen-activated protein kinase extracellular signal-regulated kinase 1 (Erk1)/Erk2 in vitro. Furthermore, a significant reduction in the ability of p85α^−/− osteoclasts to adhere to as well as to migrate via integrin αvβ3 was observed, which was associated with reduced bone resorption. Microarray as well as quantitative real-time PCR analysis of p85α^−/− osteoclasts revealed a significant reduction in the expression of several genes associated with the maturation and migration of osteoclasts, including microphathalmia-associated transcription factor, tartrate-resistant acid phosphatase, cathepsin K, and β3 integrin. Restoring the expression of the full-length form of p85α but not the version with a deletion of the Src homology-3 domain restored the maturation of p85α^−/− osteoclasts to wild-type levels. These results highlight the importance of the full-length version of the p85α subunit of class I_A PI3-K in controlling multiple aspects of osteoclast functions.

Osteoclasts (OCs) are derived from precursors of monocyte/macrophage lineage, whose growth and maturation are mainly dependent on two osteoblast/stromal cell-derived cytokines, including macrophage colony stimulating factor (M-CSF) and receptor activator of NF-κB ligand (RANKL) (22, 30, 35, 63). The critical role for these two cytokines in OC growth and differentiation has been further illustrated by studying mice lacking the expression of RANKL and M-CSF (30, 64). These mice show severe osteopetrosis and lack mature OCs. M-CSF and RANKL regulate OC progenitor (OCp) growth and function in part by regulating the expression of several OC genes, including tartrate-resistant acid phosphatase (TRAP), cathepsin K, calcitonin receptor, and integrin β3 (15, 29). Stimulation of OC precursors by RANKL and M-CSF results in the activation of a number of signaling molecules, including Gab2, Grb2, Vav, Src homology-2 (SH2)-containing inositol-5-phosphatase (SHIP), Dap12, JNK, p38, and phosphatidylinositol 3-kinase (PI3-K). Mice deficient in the expression of some of these molecules have demonstrated the precise involvement of each of these enzymes and/or adaptor proteins in osteoclastogenesis (16, 17, 47, 55). In addition to signaling molecules, several transcription factors have also been implicated in regulating osteoclastogenesis. Some of these factors regulate early growth and survival of OCs while others contribute to late stage maturation of OCs. Examples of these transcription factors include c-fos, NF-κB, PU.1, microphathalmia-associated transcription factor (MITF), and Jun dimerization protein 2 (JDP2) (27, 51, 52).

While several signaling molecules in OC development have been identified, the role of PI3-K appears to be particularly important. This is partly because PI3-K has been shown to be a critical downstream effector from at least three distinct cell surface receptors in OCs, including M-CSF receptor, αvβ3, and RANK (20). Importantly, all three molecules and their downstream substrates have been identified as candidate therapeutic targets for treatment of OC-related bone disorders. In OCs, once activated by M-CSF, PI3-K influences survival as well as actin remodeling (38, 39). Pharmacologic inhibitors of PI3-K, including LY294002 and wortmannin, show a dramatic reduction in the development of OCs in cultures treated with M-CSF and RANKL. Furthermore, treating OCs with pharmacologic inhibitors of PI3-K also impairs the resorptive activity of OCs, which is associated with impaired ruffling, actin ring formation, and reduced pit formation (36). Thus, these results clearly implicate the PI3-K pathway as an important pathway in regulating OC functions. However, these conclusions have been largely based on studies conducted with commercially available PI3-K inhibitors such as wortmannin and LY294002. These compounds target the ATP-binding site of all PI3-K family members and interfere with PI3-K-related kinases at elevated concentrations. Additionally, these inhibitors are broad spectrum and nonspecific and are associated with extreme toxicity. Therefore, to better understand and therapeutically manipulate the PI3-K pathway in OCs, quantitative as well as qualitative data evaluating the output of individual regulatory and/or catalytic subunit(s) of PI3-K are essential.

PI3-K belongs to a family of enzymes that are involved in phosphorylating PI lipids at the 3′ position (26). Based on sequence similarity and biochemical properties, PI3-Ks have been divided into three classes. Class I PI3-Ks have been shown to regulate receptor tyrosine kinase-mediated responses. Members of the class I_A PI3-K are heterodimers, which are composed of a catalytic subunit, p110 (α, β, or δ), and a regulatory subunit, p85 (α or β). The p85 regulatory subunit binds to phosphorylated tyrosine residues via their SH2 domains, resulting in the recruitment and activation of the p110 catalytic subunit at the plasma membrane. In addition to the presence of SH2 domains, p85 also consists of an SH3 domain, a proline-rich domain (PRD), and a domain homologous to the breakpoint cluster region (BCR) gene product (54). Although the precise role of these domains in p85α-regulated functions is poorly understood, it has been suggested that they might be involved in targeting p85α to distinct subcellular compartments and/or in the recruitment of additional signaling molecules. Interestingly, the shorter regulatory subunits of p85, namely p55α, p50α, and p55γ, that do not contain these additional domains (i.e., SH3, PRD, or BCR) appear to have distinct biologic activities in cells (42, 50).

In the current study, we demonstrate that OCs express multiple regulatory subunits of class I_A PI3-K although the expression of the full-length form of p85α is most abundant. In vivo, deficiency of p85α results in a significantly greater number of trabeculae and significantly lower spacing between trabeculae as well as increased bone mass in both males and females compared to their sex-matched wild-type controls. Consistently, p85α^−/− OCps show impaired growth and differentiation, which is associated with reduced activation of Akt and mitogen-activated protein (MAP) kinase extracellular signal-regulated kinase 1 (Erk1)/Erk2 in vitro. Furthermore, a significant reduction in the ability of p85α^−/− OCs to adhere to as well as to migrate via integrin αvβ3 was observed, which is associated with reduced bone resorption. Microarray as well as quantitative PCR analysis on p85α^−/− OCs revealed a significant reduction in the expression of several genes associated with the maturation and migration of OCs, including MITF, TRAP, cathepsin K, and β3 integrin. Restoring the expression of the full-length form of p85α but not the version with a deletion of the SH3 domain restored the maturation of p85α^−/− OCs to wild-type levels. These results suggest that the functional defects in p85α-deficient OCs are observed in spite of the continuous expression of p50α and p55α subunits and highlight the importance of the SH3, PRD, and the BCR domain of p85α in regulating OC functions.

MATERIALS AND METHODS

Mice.

p85α^+/− mice have been previously described (46, 50). p85α^−/− mice were obtained by mating of p85α^+/− mice. The genotype of the p85α^−/− mice was determined by PCR as previously described (46). The targeting strategy allowed selective disruption of p85α expression while leaving p55α and p50α isoforms intact (46). These mice were maintained under specific-pathogen-free conditions in the Indiana University Laboratory Animal Research Center, Indianapolis, IN.

Histochemistry and histomorphometry.

Female p85α^−/− and wild-type mice were sacrificed at 6 to 8 weeks of age. Prior to sacrifice, the mice were administered intraperitoneal injections of calcein (20 mg/kg) and alizarin complexone (25 mg/kg) 5 and 2 days before sacrifice, respectively. The distal third of each femur was dissected, fixed in 4% paraformaldehyde, dehydrated in graded alcohols, and embedded (undecalcificed) in methylmethacrylate. The polymerized blocks were sectioned at 6 μm in the frontal plane, mounted on charged microscope slides, and either left unstained (for fluorochrome histomorphometry) or reacted for TRAP activity using a commercially available reaction kit (Sigma-Aldrich). The TRAP-stained sections were counterstained with von Kossa to reveal mineralized tissue and analyzed on a Nikon Optiphot microscope equipped with the BioQuant histomorphometry software (R&M Biometrics). On each section, the numbers of TRAP-positive (TRAP⁺) cells adjacent to the trabecular bone matrix were counted in the secondary spongiosa, beginning 0.5 mm proximal to the growth plate. OC counts were standardized to the total trabecular surface.

From the unstained sections, trabecular bone turnover was assessed in the secondary spongiosa by measuring the extent of single-label and double-label perimeter (sL.Pm and dL.Pm, respectively, in the formula below) and the area of bone (dL.Ar) between the calcein and alizarin labels. Derived histomorphometric parameters included mineralizing surface (the percentage of mineralized surface/bone surface [MS/BS]), a measure of active bone-forming surface, calculated as follows: MS/BS = [1/2 (sL.Pm + dL.Pm)]/Tt.Pm × 100, where Tt.Pm is the total perimeter; the mineral apposition rate ([MAR] μm/day), a measure of the rate of radial expansion of new bone, was calculated as dL.Ar/dL.Pm/3 days; and the bone formation rate, an overall measure of bone formation that combines MS/BS and MAR, was calculated as MS/BS × MAR × 3.65.

μCT.

Trabecular bone mass and architectural properties in the femoral distal metaphysis were evaluated using a high-resolution desktop microcomputed tomography (μCT) imaging system (μCT-20; Scanco Medical AG, Basserdorf, Switzerland). Each femur (n = 7 males and 3 females per group) was scanned from the distal 70% to 90% of its total length. A microfocus X-ray tube with a focal spot of 10 μm was used as a source. For each slice, 600 projections were taken over 216° (180° plus half of the fan angle on either side). Approximately 270 microtomograph slices were acquired per bone using a slice increment of 17 μm. A standard convolution back-projection procedure with a Shepp and Logan filter was used to reconstruct the CT images in 1,024-by-1,024 pixel matrices. The trabecular area was partitioned manually from the cortical shell using the Scanco software. From the isolated three-dimensional trabecular networks, the following parameters were calculated (respective units are given): bone volume fraction (calculated as a percentage of the total volume), connectivity density (mm⁻³), trabecular number (mm⁻¹), trabecular thickness (μm), and trabecular separation (μm).

Generation of murine OCs.

Bone marrow cells from wild-type and p85α^−/− mice were cultured in OC culture medium (α-minimal essential medium; 1% penicillin-streptomycin and 10% fetal bovine serum) supplemented with 10 ng/ml of M-CSF overnight. The next day nonadherent cells were collected and cultured in the presence of 10 ng/ml of M-CSF and 100 ng/ml of RANKL for 6 days. Culture medium was changed every 2 days. Multinucleated OCs were identified by TRAP activity assay. Briefly, adherent cells were fixed with 10% formaldehyde for 10 min at room temperature and then with ethanol-acetone (50:50, vol/vol) for 1 min and washed with phosphate-buffered saline (PBS), and TRAP staining was performed using a commercially available kit (Sigma 387-A). TRAP⁺ cells with three or more nuclei were counted for quantification, and results are represented per field.

OC proliferation and survival.

A proliferation assay was conducted according to Faccio et al., with minor modifications (17). Briefly, 0.5 × 10⁵ OCps from wild-type and p85α^−/− mice were cultured in the presence of various concentrations of M-CSF or RANKL or both in 96-well tissue culture plates. After 2 days, 1.0 μCi of [³H]thymidine (Amersham) was added for 6 h before harvesting. Cells were harvested using an automated 96-well cell harvester (Brandel, Gaithersburg, MD), and thymidine incorporation was determined as counts per minute. Survival assay was conducted according to Yang et al. (61). Briefly, cells were starved of serum and cytokines for various time intervals, harvested by the addition of trypsin-EDTA, and resuspended in 1× binding buffer (BD Pharmingen). A total of 1 × 10⁶ cells were suspended in binding buffer containing annexin V-phycoerythrin (BD Pharmingen) and 7-aminoactinomycin D (7-AAD; BD Pharmingen) and incubated for 20 min at room temperature in the dark. Additional binding buffer was added before analysis by flow cytometry was performed.

OC migration and adhesion assay.

Migration of OCs was evaluated using a transwell assay as described previously, with minor modifications (44, 60, 61). To verify the number of cells loaded into each transwell, OCps previously cultured in M-CSF and RANKL for 5 days were lifted from the plates after the addition of 0.05% trypsin and 0.2% EDTA·4Na in Hanks balanced salt solution and scored to identify TRAP⁺ cells. Equivalent numbers of cells were loaded onto the upper chamber of an 8 μM polycarbonate transwell coated with vitronectin for 15 h in a humidified incubator at 37°C, and a lower chamber was added containing α-minimal essential medium, 0.1% bovine serum albumin, and M-CSF. After 4 h of incubation, cells that migrated to the bottom of the chamber were stained for TRAP, and the number of TRAP⁺ cells per field was then counted (Empire Imaging Systems, Plattsburgh, NY). For adhesion experiments, OCps (1 × 10⁵ cells/ml) were placed into 24-well plates coated with vitronectin (20 μg/ml) supplemented with M-CSF as described previously (62).

Bone resorption assay.

Single-cell suspensions of OCs were seeded onto dentine slices (47, 61) (ALPCO Diagnostic, Windham, NH) and incubated at 37°C in 5% CO₂ in the presence of M-CSF and RANKL. Following 7 days of culture, the slices were rinsed with PBS, left overnight in 1 M ammonium hydroxide, and stained with 1% toluidine blue in 0.5% sodium tetraborate solution. The number of resorptive areas or “pits” per low-power field on each bone slice was counted using reflective light microscopy. The area (mm²) of each pit was evaluated by measuring the width and length using QCapture Pro software (version 5.1) by an investigator who was blinded to the experimental groups.

Immunofluorescence microscopy.

To evaluate the cytoskeletal organization in p85α^−/− OCs, immunofluorescence microscopy was performed as previously described (40, 45). Briefly, OCs were grown on coverslips, washed with PBS, and permeabilized with 0.01% saponin in 80 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)], pH 6.8, 5 mM EGTA, and 1 mM MgCl₂ for 5 min at room temperature. Cells were fixed in 3% paraformaldehyde in PBS for 20 min and quenched with 50 mM NH₄Cl for 10 min. Cells were then washed in 2% bovine serum albumin with 0.01% saponin in PBS for 5 min to block nonspecific binding. Fluorescein isothiocyanate-conjugated phalloidin (Sigma, St. Louis, MO) was used to incubate the permeabilized cells for 1 h at room temperature. After three washes with the same buffer, nuclei were stained with 4′,6′-diamidino-2-phenylindole (DAPI). Slides were washed and then mounted with 80% glycerol in PBS. The cells were observed, and fluorescent images were taken with a Nikon TE 2000-5 fluorescent microscope.

Western blot analysis.

Wild-type or p85α^−/− OCps were starved of serum and growth factors and stimulated with M-CSF, RANKL, or both cytokines for the indicated times (see Fig. 9). The reaction was stopped by the addition of cold PBS, and cells were lysed as previously described (48). An equal amount of protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and separated protein was transferred onto a nitrocellulose membrane. Western blot analysis was performed using an anti-phospho-Akt and an anti-phospho-Erk antibody (Cell Signaling, Beverly, MA). Pan-p85 antibody and p85α (recognizes the SH3 domain of p85α)-specific antibody as well as the antihemagglutinin (anti-HA) antibody were purchased from Upstate Biotechnology, NY.

FIG. 9.

Reduced activation of PI3-K, Akt, and MAP kinase in p85α^−/− OCps in response to both M-CSF and RANKL. Wild-type (WT) and p85α^−/− OCps were starved and stimulated with a combination of M-CSF and RANKL (A), RANKL (100 ng/ml) (B), or M-CSF (10 ng/ml) (C) for 2 and 5 min. Equal amounts of cell lysates were subjected to a PI3-K lipid assay as described in the Materials and Methods section. Upper panels demonstrate quantitative reduction in the level of PI3-K activity in p85α^−/− OCps. Arrows in the bottom panels indicate the level of activation of PI3P and PIP2 in response to the indicated cytokines. (D and E) Reduced activation of Akt and Erk1/2 MAP kinase in p85α^−/− OCps. Wild-type and p85α^−/− OCps were starved and stimulated with M-CSF (10 ng/ml) or RANKL (100 ng/ml) or a combination of both. Equal amounts of cell lysates were subjected to Western blot analysis using an anti-phospho-Akt or an anti-phospho-Erk1/2 antibody. The amount of total Akt and Erk in each lane is indicated.

PI3-K assay.

OCps were starved and stimulated with RANKL or M-CSF or a combination of both for 2 and 5 min at 37°C. Cells were lysed in 100 μl of PI3-K lysis buffer containing 125 mM Tris, pH 7.0, 25 mM MgCl₂, 5 mM EGTA, and protease inhibitors. Equal amounts of protein lysates were subjected to a PI3-K activity assay (BCA Protein Assay kit; Pierce, Rockford, IL). The activity of PI3-K in whole-cell lysates was measured by the amount of [γ-³²P]ATP incorporated into the lipid substrates, which were separated using thin-layer chromatography. The reaction mixture containing 1 μl of l-α-PI-4-monophosphate, 4 μl of l-α-PI-4,5-diphosphate, 0.5 μl of l-α-phosphatidyl-l-serine (Avanti Polar Lipids, Alabaster, AL), and [γ-³²P]ATP (10 μCi/reaction) was added to the lysates and incubated at 37°C for 10 min. The reaction was terminated by the addition of 105 μl of 1 N HCl, followed by the addition of 160 μl of chloroform-methanol (1:1, vol/vol). The samples were vortexed vigorously and centrifuged to separate the aqueous and organic phases. The lipid-containing organic phase was resolved on silica-coated thin-layer chromatography plates (Whatman, England) with a solvent mixer of 65% propanol-1% glacial acetic acid for 12 h, air dried on silica plate, and exposed to X-ray film (Kodak) with a Dupont intensifying screen at room temperature.

Measurement of TRAP 5b in mouse serum.

Solid-phase immunofixed enzyme activity assay for the determination of OC-derived TRAP form 5b (TRAP 5b) in mouse serum was assessed as described by the manufacturer (Immunodiagnostic Systems, Inc., Fountain Hills, AZ).

Microarray expression profiling and data analysis.

The mRNA samples were submitted to Johns Hopkins Hospital Microarray core facility, and expression profiling was performed using Affymetrix MOE430_2 chips. To estimate the gene expression signals, data analysis was conducted on the chips' CEL file probe signal values at the Affymetrix probe pair (perfect match probe and mismatch probe) level by employing the statistical technique of robust multiarray analysis using GC content for background adjustment (7, 58) with the Bioconductor package available at www.bioconductor.org. This probe-level data processing includes a normalization procedure utilizing quantile normalization (7) to reduce the obscuring variation between microarrays, which might be introduced during the processes of sample preparation, manufacture, fluorescence labeling, hybridization, and/or scanning. With the signal intensities estimated above, an empirical Bayes method with the gamma-gamma modeling, as implemented in the R package EBarrays, was used to estimate the posterior probabilities of the differential expression of genes between wild-type and p85α^−/− OCs. The criterion of a posterior probability of >0.5, which means that the posterior probability is greater than what might have been produced by chance, was used to produce the differentially expressed gene lists. All computation was performed in the R environment, and all Bioconductor packages are available at www.bioconductor.org.

Relative quantitation of mRNAs by real-time quantitative RT-PCR.

Total RNA was isolated from wild-type and p85α^−/− OCs using an RNeasy Mini kit (Qiagen). Reverse-transcriptase reactions were done using a SuperScript First-Strand cDNA Synthesis system (Invitrogen Life Technologies, Carlsbad, CA). Five nanograms of first-strand cDNA was used in reverse transcription-PCR (RT-PCR) to ensure linear amplification of sequences. Real-time PCR was performed on an ABI Prism 7500 Sequence Detection system using Sybr Green PCR Master mix (Applied Biosystems) and following the manufacturer's protocols. Primer sequences used in this study were (900 nM, each) the following: Acp-5/TRAP, CCCAATGCCCCATTCCA and CGGTTCTGGCGATCTCTTTG; calcitonin receptor, CGCATCCGCTTGAATGTG and TCTGTCTTTCCCCAGGAAATGA; cathepsin K, GGCTGTGGAGGCGGCTAT and AGAGTCAATGCCTCCGTTCTG; matrix metallopeptidase 9 (MMP-9), TATTTTTGTGTGGCGTCTGAGAA and GAGGTGGTTTAGCCGGTGAA; integrin β3, TGTGTGCCTGGTGCTCAGA and AGCAGGTTCTCCTTCAGGTTACA; JDP2, GCCATGCATTGCAAACACA and GGGAGGTGGATTGCAGTCTATG; and β-tubulin (housekeeping gene), CTGGGAGGTGATAAGCGATGA andCGCTGTCACCGTGGTAGGT. Each of these primer sets gave a unique product. PCR assays were performed in triplicate, and the data were pooled. Values obtained for levels of mRNAs were normalized to the levels of β-tubulin mRNA.

Construction and expression of p85α expressing retroviral vectors.

Whole spleen was used as source of total RNA for the synthesis of cDNA encoding p85α. Total RNA was extracted using TRI reagent (Molecular Research Center, Inc., Cincinnati, OH) following the manufacturer's protocol. Briefly, 1 ml of TRI reagent was added into 100 mg of tissue for cell lysis and then vigorously mixed with 0.2 ml of chloroform followed by centrifugation at 12,000 × g for 15 min. The aqueous phase was removed and mixed with 0.5 ml of isopropanol to allow RNA precipitation. The RNA was precipitated by centrifugation at 12,000 × g for 8 min and then washed with 1 ml of 75% ethanol. The dry RNA pellet was dissolved in RNase-free water. cDNA was synthesized using a Superscript First-Strand Synthesis system for RT-PCR (Invitrogen Life Technologies, Carlsbad, CA). Following synthesis of cDNA, the following primers were used for a p85α PCR: forward, 5′-GAATTCATGTACCCATACGATGTTCCAGATTACGCTATGAGTGCAGAGGGCTACCAG; reverse, 5′-CTCGAGTCATCGCCTCTGTTGTGCATATAC. Restriction sites used for cloning purposes have been underlined. The 5′ end of the primer contains an HA sequence to discriminate between exogenous and endogenous p85α regulatory subunits. PCR was performed using the following conditions: an initial denaturation step at 94°C for 2 min followed by 35 cycles of 94°C for 15 s, 60°C for 1 min, and 72°C for 2 min, with a final step of 72°C for 7 min. For the cloning of p85α with a deletion of SH3 (i.e., p85αΔSH3; amino acid residues 81 to 724 of p85α), the full-length version of p85α was used as a template and amplified using the following primers: forward, 5′-CCAGAATTCATGTACCCATACGATGTTCCAGATTAC GCTAGAATTTCACCCCCT ACTCCC; and reverse, 5′-CCACTCGAGTCATCGCCTCTGTTGTGCATATACTGG (restriction sites are underlined). The amplified cDNA was cloned into the EcoRI and XhoI sites upstream of an internal entry site and the enhanced green fluorescence (EGFP) protein containing bicistronic retroviral vector MIEG3 (57).

Retroviral supernatants for transduction of primary OCps were generated using the Phoenix ecotropic packaging cell line transfected with retroviral vector plasmids using a calcium phosphate transfection kit (Invitrogen, Carlsbad, CA). Supernatants were collected at 48 h posttransfection and filtered through 0.45-μm-pore-size membranes. For transductions using bone marrow-derived OCps, bone marrow cells were subjected to Histopaque-Ficol density gradient centrifugation. Low-density cells were collected, suspended in Iscove's modified Dulbecco's medium containing 20% fetal bovine serum and 1% penicillin-streptomycin, and prestimulated in non-tissue culture plates for 2 days prior to transduction on retronectin (Peprotech, Rocky Hill, NJ). Forty-eight hours after infection, cells expressing similar levels of EGFP in every group (vector alone and p85α) were sorted to homogeneity, and OCs were generated by growing them in the presence of M-CSF and RANKL. After 6 days of culture, multinucleated OCs were identified by a TRAP activity assay. Total numbers of cells were enumerated by counting TRAP⁺ cells in 24-well tissue culture plates.

RESULTS

OCs express multiple regulatory subunits of class I_A PI3-K, and loss of the p85α regulatory subunit results in increased bone mass and bone density in vivo.

To determine which regulatory subunit of class I_A PI3-K is expressed in OCs, we performed Western blot analysis on early (3 days old) and late (6 days old) in vitro bone marrow-derived OCps from wild-type and p85α^−/− mice. As seen in Fig. 1, robust expression of p85α, p50α, and p55α isoforms of p85 in early (Fig. 1, lane 1) wild-type OCps as well as a modest but a significant reduction in the expression of the shorter isoforms (p50α and p55α) in p85α-deficient OCps was observed (lane 2). In contrast, during late stages of OC maturation (6-day-old OCs), the expression of the shorter isoforms of p85α (p50α and p55α) was significantly reduced in both wild-type as well as p85α-deficient OCps although a modest increase in the expression of the shorter isoform (p50α) was observed in p85α-deficient OCps compared to wild-type controls (lane 4). Overexpression of p50α in p85α-deficient adipocytes derived from these same mice has also been reported previously (50). The presence of the nonspecific band in the middle of the full-length and shorter isoforms has also been reported by other investigators using this same antibody (50). These results suggest that OCs express multiple regulatory subunits of class I_A PI3-K; however, the expression of the full-length p85α isoform is most abundant.

FIG. 1.

Expression of p85 (p85α, p50α, p55α, and p85β) in wild-type (WT) and p85α^−/− OCs. OCps were cultured in the presence of M-CSF (10 ng/ml) and RANKL (100 ng/ml) for either 3 days or 6 days, after which the cells were harvested and subjected to Western blot analysis using a pan-anti-p85 antibody (this antibody recognizes all regulatory subunits of class I_A PI3-K). Arrows in the top panel indicate the level of expression of p85 regulatory subunits in wild-type and p85α^−/− OCs. The bottom panel demonstrates total levels of β-actin in each lane.

Given the importance of the PI3-K pathway in regulating OC functions in vitro based on studies using pharmacologic inhibitors of this pathway, we examined if p85α is specifically required for bone development in vivo. To determine this, we characterized the skeletal defects in mice deficient in the expression of the full-length form of p85α. We measured the bone mass and architecture of 18-week-old wild-type control and p85α^−/− mice using μCT. Figure 2A and B demonstrate representativeμCT reconstructions from the distal third of the right femur in 18-week-old male and female wild-type and p85α^−/− mice. The anterior half of the bone has been digitally removed to reveal the trabecular bone compartment within the metaphysis. Numerous trabeculae and greater proximal encroachment of the trabecular network are readily apparent in p85α^−/− mice than in wild-type controls. Figure 2C shows a quantitative comparison of the trabecular number, trabecular spacing, bone volume fraction, and connectivity density between wild-type and p85α^−/− mice. Significantly greater trabecular number and significantly lower spacing between trabeculae were observed in both female p85α^−/− mice (Fig. 2A and C) and male p85α^−/− mice (Fig. 2B and C) than in their sex-matched wild-type littermates (n = 10 in each experiment), a common feature in high-bone-mass conditions. The bone volume fraction was also elevated in p85α^−/− mice. These data suggest that p85α^−/− mouse skeleton exhibits increased trabecular bone mass and architecture.

FIG. 2.

Deficiency of p85α in vivo results in increased bone mass. Representative μCT reconstructions from the distal third of the right femur in 18-week-old female (A) and male (B) wild-type (WT) and p85α^−/− mice. The anterior half of the bone has been digitally removed to reveal the trabecular bone compartment within the metaphysis. Note the more numerous trabeculae and the greater proximal encroachment of the trabecular network in the p85α^−/− mice. (C) μCT-derived measurements of the trabecular bone volume fraction (BV/TV), trabecular number (Tb.N), thickness (Tb.Th), and separation (Tb.Sp) revealed significantly more numerous trabeculae, with significantly less spacing between trabeculae, a common feature in high-bone-mass conditions (n = 20 mice including 10 wild-type [7 males and 3 females] and 10 p85α^−/− mice [7 males and 3 females]). Error bars represent ± 1 standard error of the mean. *, P < 0.01. (D) Quantitative analysis of the number of OC per bone volume in vivo in wild-type and p85α^−/− bone sections reacted for TRAP activity (n = 3). *, P < 0.05. TRAP 5b levels in the serum of 16- to 17-week-old (E) or 6- to 7-week-old (F) wild-type and p85α^−/− mice. Solid-phase immunofixed enzyme activity assay for the determination of OC-derived TRAP 5b in mouse serum was assessed as described in Materials and Methods. A significant increase in the serum TRAP 5b levels was noted in both old and young p85α^−/− mice compared to wild-type controls. For the 16- to 17-week age group, five wild-type and seven p85α^−/− mice were used. For the 6- to 7-week-old age group, 11 wild-type and 12 p85α^−/− mice were used. *, P < 0.01. (G to I) Trabecular bone turnover was assessed in the secondary spongiosa by measuring the extent of single label (sL.Pm) and double label (dL.Pm) perimeter and the area of bone (dL.Ar) between the calcein and alizarin labels. Derived histomorphometric parameters include mineralizing surface (MS/BS), a measure of active bone-forming surface; MAR, a measure of the rate of radial expansion of new bone; and the bone formation rate (BFR). Five wild-type and five p85α^−/− mice were used (P > 0.05).

Given the increase in bone volume in p85α^−/− mice, we next determined the number of OCs in p85α^−/− mice in vivo. The femurs of 7- to 8-week-old syngeneic p85α^−/− and wild-type mice were decalcified, and histological sections from the distal metaphysis were stained for the OC enzyme TRAP. Strikingly, there was a marked increase in the number of OCs per unit of trabecular surface (Fig. 2D) in p85α^−/− mice compared to wild-type controls. Recent studies have suggested that secreted TRAP 5b is an indicator of the number of OCs in vivo but not their activity (2-4, 11). To assess whether the increased numbers of OCs observed in p85α^−/− mice by TRAP staining were also observed using a surrogate assay, we enumerated TRAP 5b serum levels in both young and older wild-type and p85α^−/− mice. As seen in Fig. 2E and F, a significant increase in the level of serum TRAP 5b in both young (7 to 8 weeks old) and older (16 to 17 weeks old) p85α^−/− mice was observed compared to wild-type control mice. However, bone formation rates in p85α^−/− mice were indistinguishable from their wild-type counterparts (Fig. 2G to I), establishing that the increased skeletal mass of p85α^−/− mice does not reflect accelerated bone formation. Interestingly, enhanced numbers of OCs in vivo, along with elevated serum TRAP 5b levels, have been reported in mouse mutants lacking Vav guanine exchange factor as well as the adaptor protein Gab2 (16, 55). Both of these types of mutant mice demonstrate increased bone mass in vivo, which is associated with impaired OC function(s) in vitro. Taken together, our results demonstrate a requirement for p85α in the regulation of bone mass in vivo. Furthermore, our results suggest that, although an increase in OC number is observed in p85α-deficient mice in vivo, p85α-deficient OCs may be functionally impaired.

p85α is essential for OC growth and maturation.

To evaluate the effect of p85α in regulating OC growth, we next determined whether deficiency of p85α affects OCps in vitro. To study this, we stimulated mutant and wild-type OCps with increasing doses of M-CSF or increasing doses of RANKL or both. After 48 h of culture, cells were pulsed with [³H]thymidine for 6 h. A significant reduction in [³H]thymidine incorporation was observed in p85α^−/− OCps compared to wild-type controls at all doses of M-CSF (Fig. 3A) or RANKL (Fig. 3B) or in the presence of both M-CSF and RANKL (Fig. 3C). Given that PI3-K has been demonstrated to play an essential role in regulating the survival of cells in part by regulating the activation of Akt (9, 13, 43), we next examined survival/apoptosis in wild-type and p85α^−/− bone marrow-derived OCs. OCs were generated by culturing low-density mononuclear cells in the presence of M-CSF and RANKL as previously described (61). Purified OCs were cultured in M-CSF in the absence of serum, and apoptosis was assessed over a span of 18 h using annexin V staining. No significant difference in the rate of apoptosis was observed between wild-type and p85α^−/− OCs (Fig. 4A and B). Thus, deficiency of p85α in OCps confers on these cells hyporesponsiveness to the critical osteoclastogenic cytokines M-CSF and RANKL, which leads to reduction in OCp proliferation but not survival.

FIG. 3.

Defective proliferation of p85α^−/− OCps. Wild-type (WT) and p85α^−/− OCps were cultured in the presence of increasing amounts of M-CSF (A) or RANKL (B) or both (C). After 2 days, proliferation was evaluated by a [³H]thymidine incorporation assay. Bars represent the mean [³H]thymidine incorporation in OCps (cpm ± standard deviation) from one representative experiment performed in triplicate. Similar results were observed in two independent experiments. *, P < 0.05 for wild-type versus p85α^−/− OCs.

FIG. 4.

Deficiency of p85α in OCps does not alter the survival of these cells. Serum- and cytokine-depleted OCps were stained with annexin V-phycoerythrin and 7-AAD and analyzed by flow cytometry. (A) Representative dot blot showing the percent survival of OCps as determined by the lack of staining of cells by either annexin V and/or 7-AAD (i.e., lower-left coordinate). (B) Bar graph demonstrating percentage of annexin V- and 7-AAD-negative cells at various time points in the absence of serum and cytokines.

Defective morphology and differentiation of p85α^−/− OCs.

It has been previously reported in mice that genetic deletion of SHIP, a negative regulator for PI3-K, results in enhanced differentiation of OCs (i.e., gain of function) (47). To determine how loss of the p85α regulatory subunit of class I_A PI3-K impacts maturation of OCs, we evaluated OC differentiation following OC culturing and subjected the cells to TRAP staining to identify multinucleated OCs (61). Wild-type and p85α^−/− OCs were generated by culturing OCps for 6 days with RANKL (100 ng/ml) and increasing concentrations of M-CSF (10, 30, and 100 ng/ml). As seen in Fig. 5A, osteoclastogenesis was significantly impaired in p85α^−/− cultures, as reflected by a significant decrease in the number of multinucleated TRAP⁺ cells at three concentrations of M-CSF together with RANKL (100 ng/ml) compared to wild-type cultures (Fig. 5A). Representative photographs (Fig. 5A) and a quantitative assessment of TRAP⁺ wild-type control and p85α^−/− OCs are shown in Fig. 5B and C. To examine whether deletion of p85α leads to a delay in OC differentiation, we cultured p85α^−/− OCs for an additional 2 days (total, 8 days) followed by TRAP staining. Culturing wild-type OCps for an additional 2 days induced apoptosis in a majority of very large multinucleated cells (data not shown); however, no improvement in the differentiation potential of p85α^−/− OCs was observed by maintaining these cells for an additional 2 days in culture (data not shown). These results suggest that loss of p85α impairs OC maturation and proliferation in response to M-CSF and RANKL stimulation.

FIG. 5.

Defective morphology and differentiation of p85α^−/− OCps. (A) Representative photomicrographs of OCps of the indicated genotypes generated in vitro following culture in M-CSF (10 to 100 ng/ml) and RANKL (100 ng/ml). OCps were identified by staining for TRAP activity. A representative field is shown. (B) Quantitative reduction in the number of multinucleated (>3 nuclei/field) TRAP⁺ OCs is shown. Bars represent the mean numbers of multinucleated cells (mean ± standard deviation) of one representative experiment performed in replicates of three (10 fields were counted per replicate). *, P < 0.05 for wild-type versus p85α^−/−. (C) Relative number of TRAP⁺ cells from five independent experiments is summarized in a line chart. *, P < 0.05 for wild-type versus p85α^−/− OCs. OD, optical density.

Expression of p85α in OCs is essential for αvβ3 and OPN-mediated migration, adhesion, and bone resorption in vitro.

Bone resorption by OCs is greatly dependent on their ability to adhere and migrate on the bone surface in part via integrin αvβ3 (33). To determine whether the loss of p85α alters the ability of these cells to migrate on extracellular matrix via integrins, equivalent numbers of wild-type and p85α^−/− OCs were placed in the upper chamber of a transwell coated with αvβ3 or osteopontin (OPN), and migration was assessed in response to M-CSF. A significant reduction in the migration of p85α^−/− OCs toward vitronectin and OPN was observed in comparison to wild-type controls (Fig. 6A). Further, deficiency of p85α in OCs also affected their ability to adhere to OPN (Fig. 6B).

FIG. 6.

Impaired migration and adhesion in p85α^−/− OCs. (A) OC migration via αvβ3 integrin or OPN in the presence of M-CSF. (B) OC adhesion via OPN in the presence of M-CSF. Results represent mean ± standard error of the mean of four independent experiments. *, P < 0.01; **, P < 0.001 (comparing p85α^−/− versus wild-type OCs). WT, wild type.

One of the functions of OCs is to form specialized cell-extracellular matrix to induce the degradation of bone matrix by releasing proteinases (8). We next asked whether deletion of p85α influences OC resorption activity. This process was assessed in vitro by culturing OCs on bone slices (dentin) and then evaluating the number and area of bone “pits” that were resorbed. Wild-type and p85α^−/− OCs (10⁵/dish) were cultured onto dentine slices for 7 days in the presence of M-CSF and RANKL. Following culture, dentine slices were stained with toluidine blue, and pit formation was evaluated. As seen in representative photomicrographs from two independent experiments, deficiency of p85α in OCs significantly impaired the ability of these cells to form pits (Fig. 7A). Quantitative assessment of the percent resorbing area from four independent experiments is shown in Fig. 7B. These results suggest that the bone resorption potential of p85α^−/− OCs is significantly reduced compared to the resorption potential of wild-type OCs.

FIG. 7.

Impaired bone resorption by p85α^−/− OCs. Bone resorptive activity was measured by a pit formation assay. (A) Representative photomicrographs of a bone resorption assay from two independent experiments following culture of OCps on dentine slices. The resorbed bone is stained dark blue. The number and area of resorbed regions, referred to as pits, are quantitated in panel B. Data are the mean from four independent experiments. *, P < 0.01, wild-type versus p85α^−/−. WT, wild type.

p85α regulates actin organization.

The OC bone-resorbing capacity involves the organization of the actin cytoskeleton to form a specialized matrix that initiates degradation of bone matrix by releasing proteases (8). This process involves several small actin-based adhesion structures called podosomes that are organized into complex structures identified as clusters, rings, and ultimately belts to form a functional sealing zone. Given the role of PI3-K in regulating cytoskeleton functions, we next evaluated the impact of p85α deficiency on modulating the actin cytoskeleton using previously reported criteria (12). Wild-type cultures demonstrated markedly larger multinucleated OCs than p85α^−/− cultures following phalloidin staining, as observed in the low-power field (×40 magnification) (Fig. 8) and a higher-power field (×100) (Fig. 8). The yellow arrowheads in Fig. 8, frame 3, indicate representative belt formation in the OCs while red arrows represent clusters. p85α^−/− OCs demonstrate significantly fewer numbers of belt structures than the wild-type (Fig. 8, frame 3). These data indicate that p85α plays an essential role in functional F-actin organization.

FIG. 8.

Altered actin organization in p85α^−/− OCs. Representative photomicrograph of OCps following staining with fluorescein isothiocyanate-conjugated phalloidin at magnifications of ×40 (frames 1 and 2) and ×100 (frames 3 and 4). Arrows in frames 1 and 2 indicate clusters. Arrows in frame 3 indicate belt-forming cells. The experiment was conducted on three independent occasions.

p85α^−/− OCps have reduced PI3-K, Akt, and Erk activation.

Binding of M-CSF to its receptor results in the autophosphorylation of the M-CSF receptor on several intracellular tyrosine residues, thereby creating docking sites for various SH2- and SH3-containing signaling molecules. Several of these tyrosine sites have been mapped in the M-CSF receptor. The p85α regulatory subunit of PI3-K binds to Tyr 721 (41), thereby bringing PI3-K into proximity with its substrates in the plasma membrane. To determine the in vivo consequences of loss of p85α in OCs on the activation of total PI3-K activity in wild-type and p85α^−/− OCps stimulated with either M-CSF alone, RANKL alone, or with a combination of both M-CSF and RANKL, OCps cultured in the presence of M-CSF and RANKL for 2 to 3 days were starved for 6 h of growth factors and stimulated with either M-CSF alone, RANKL alone, or a combination of both cytokines for 2 and 5 min. The lysates were subsequently subjected to a PI3-K activity assay as described in the Materials and Methods section. As seen in Fig. 9A to C, a significant reduction in the overall PI3-K activity was observed in p85α-deficient OCps treated with M-CSF, RANKL, or a combination of both cytokines. These results suggest that up to 40% of the M-CSF- and RANKL-induced PI3-K activity in OC progenitors is regulated via the full-length form of p85α.

Previous studies have shown that PI3-K can regulate the activation of both Akt and Erk MAP kinase (5). To determine the extent to which the deficiency of p85α modulates the activation of Erk MAP kinase and Akt in response to M-CSF, RANKL, or both of these cytokines, we performed Western blot analysis on OCps that were starved and then stimulated with these cytokines using phosphospecific antibodies that recognize the activated forms of Akt and Erk MAP kinase. As seen in Fig. 9D and E, deficiency of p85α in OCps results in a significant decrease in the activation of both Akt and Erk MAP kinase. Under these culture conditions stimulation of OCps with RANKL alone did not induce the activation of Erk MAP kinase although a significant increase in the activation of Erk MAP kinase was observed in wild-type OCps stimulated with both M-CSF and RANKL together. In contrast, activation of Akt in response to RANKL stimulation was robust and appeared to be further augmented in the presence of M-CSF (Fig. 9D, lanes 6 and 8). These results suggest that activation of both Akt as well as Erk MAP kinase is significantly modulated as a result of reduced over all PI3-K activity due to p85α deficiency in OCps.

Deficiency of p85α alters the expression of multiple OC-specific genes.

Previous studies have demonstrated an essential role for PI3-K in regulating multiple aspects of OC biology, including growth, differentiation, and actin-based functions. Since most of these results have been derived from experiments performed using pharmacologic inhibitors of PI3-K, which inhibit all classes of PI3-K, we sought to determine the nature of genes affected by the loss of a specific subunit of PI3-K, namely, p85α. We used Affymetrix GeneChips (MG-U74Av2) to profile mRNA expression in wild-type and p85α^−/− OCs. Wild-type and p85α^−/− OCs grown in the presence of M-CSF and RANKL for 6 days were utilized for isolating mRNA. We chose this time point to define alterations in gene expression related to OC growth and differentiation. Previous studies have shown that by 6 days of M-CSF and RANKL treatment, fully differentiated TRAP⁺ OCs are typically observed. Numerous known OC-specific genes were observed in OC samples derived from wild-type mice (Table 1). Importantly, a 13.9-fold reduction in the expression of the p85α subunit of class I_A PI3-K was observed in samples derived from p85α^−/− mice relative to the wild type, thus validating our experimental system. Several of the genes affected by the loss of p85α expression were related to OC growth and maturation, including macrophage-stimulating 1 receptor, OC-associated receptor, calcitonin receptor, TRAP, MMP-9, integrin αv, integrin β3, cathepsin K and transcription factors such as MITF and JDP2 (Table 1). In addition, the expression of several other genes was also affected due to lack of p85α. These include genes related to transcription, cell adhesion, signaling, cell growth, chemokines, proteases/inhibitors, transporters, extracellular protein, cytoskeletal protein as well as proteins involved in metabolism, electron transport, blood coagulation and metal ion binding (Table 1).

TABLE 1.

Microarray expression data

Category	Gene description	Fold change
Transcription	Microphthalmia-associated transcription factor	−1.73
	Early growth response 2	−1.8
	JDP2	−2.08
	Histone 2, H2aa1	−2.04
	Transcription factor AP-2, gamma	−1.62
	TNF-α-induced protein 3	−2.16
Cell adhesion	Integrin beta 3	−1.97
	Integrin alpha X	−1.87
	Epidermal growth factor-like repeats and discordin I-like domains 3	−2.9
	Sialic acid binding immunoglobulin-like lectin 5	−2.92
	Protocadherin 7	−2.82
	Milk fat globule-epidermal growth factor 8 protein	−1.9
	Integrin alpha v	−1.84
	Killer cell lectin-like receptor subfamily B member 1A	−3.0
	CD72 antigen	−1.93
Signaling	Regulator of G-protein signaling 1	−2.67
	PI3-K, regulatory subunit, polypeptide 1 (p85 alpha)	−13.97
	Down syndrome critical region homolog 1 (human)	−1.6
	Guanine nucleotide binding protein (G protein), gamma transducing activity polypeptide 2	−1.7
	Rab38, member of RAS oncogene family	−2.2
	Megakaryocyte-associated tyrosine kinase	−3.17
	PI-4-phosphate 5-kinase, type 1 alpha	−1.9
	PCTAIRE-motif protein kinase 3	−1.9
	SLAM family member 8	−3.0
	Related RAS viral (r-ras) oncogene homolog 2	−1.85
	Phosphoinositide-3-kinase adaptor protein 1	−1.7
	G protein-coupled receptor 68	−1.7
	MAP kinase kinase kinase kinase 1	−1.76
	Phosphodiesterase 1C	−1.7
	Ras homolog gene family, member J	−1.9
	Met proto-oncogene	−2.8
Cell growth	Growth differentiation factor 3	−1.8
	Platelet-derived growth factor, B polypeptide	−1.94
	Exostoses (multiple) 1	−1.92
	Ras homolog gene family, member C	−1.93
Chemokines	Chemokine (C-C motif) ligand 5	−2.2
	Chemokine (C-C motif) ligand 22	−1.63
Proteases/inhibitors	Neurolysin (metallopeptidase M3 family)	−2.64
	Carboxypeptidase E	−2.05
	Cathepsin K	−1.94
	Matrix metalloproteinase 9	−2.07
	Matrix metalloproteinase 19	−1.94
	Serine (or cysteine) proteinase inhibitor, clade D, member 1	−1.64
	Serine (or cysteine) proteinase inhibitor, clade B, member 9	−1.7
	Serine (or cysteine) proteinase inhibitor, clade B, member 6b	−4.2
	Serine (or cysteine) proteinase inhibitor, clade B, member 9b	−3.74
	Serine (or cysteine) proteinase inhibitor, clade B, member 1b	−2.17
	Serine (or cysteine) proteinase inhibitor, clade E, member 2	−2.44
Transporter/receptor	Histocompatibility 2, T region locus 24	−1.63
	Macrophage stimulating 1 receptor (c-met-related tyrosine kinase)	−2.12
	Protein C receptor, endothelial	−2.46
	Ryanodine receptor 1, skeletal muscle	−1.88
	Solute carrier family 4 (anion exchanger), member 2	−1.99
	Solute carrier family 37 (glycerol-3-phosphate transporter), member 2	−1.8
	Solute carrier family 6 (neurotransmitter transporter, betaine/GABA), member 12	−1.73
	Solute carrier family 6 (neurotransmitter transporter, serotonin), member 4	−1.64
	Potassium intermediate/small conductance calcium-activated channel, subfamily N, member 4	−2.2
	Potassium inwardly-rectifying channel, subfamily J, member 2	−2.02
	Growth hormone receptor	−1.9
	Calcitonin receptor	−2.65
	Transmembrane 7 superfamily member 4	−5.19
	Epithelial membrane protein 2	−2.13
	Solute carrier family 18 (vesicular monoamine), member 1	−1.68
	OC-associated receptor	−2.42
	Potassium intermediate/small conductance calcium-activated channel, subfamily N, member 4	−2.2
	ATPase, H+ transporting, lysosomal V0 subunit a isoform 1	−1.8
	Killer cell lectin-like receptor, subfamily A, member 17	−1.61
	Fibroblast growth factor receptor 1	−2.8
	Megalencephalic leukoencephalopathy with subcortical cysts 1 homolog (human)	−1.7
Extracellular protein	Ephrin B2	−2.48
	Glypican 1	−2.59
	MAM domain containing 2	−2.65
Cytoskeletal/structural protein	Myosin IE	−2.2
	Scinderin	−2.67
	Procollagen, type IV, alpha 5	−2.83
	Plakophillin	−2.62
Metabolism	Acetyl-coenzyme A synthetase 2 (ADP forming)	−1.7
	Enolase 2, gamma neuronal	−1.83
	Lipoprotein lipase	−2.13
	Phosphatidic acid phosphatase 2a	−2.0
	Carbonic anhydrase 2	−2.7
	Wolfram syndrome 1 homolog (human)	−2.04
	Earbohydrate sulfotransferase 11	−2.66
	3-Hydroxybutyrate dehydrogenase (heart, mitochondrial)	−2.48
	UDP-N-acetyl-alpha-d-galactosamine:polypeptide N-acetylgalactosaminyltransferase 9	−2.3
	Stearoyl-coenzyme A desaturase 2	−2.02
	Aldo-keto reductase family 1, member C18	−2.47
	2′-5′ Oligoadenylate synthetase 3	−2.32
	Stearoyl-coenzyme A desaturase 2	−2.02
	Acid phosphatase 5, tartrate resistant	−2.12
Electron transport	Cytochrome b-561	−2.0
Blood coagulation	Coagulation factor III	−4.08
	Coagulation factor II (thrombin) receptor	−2.36
Metal ion binding	Metallothionein 3	−4.6

To confirm the data generated using microarray experiments using another experimental approach, we performed real time quantitative RT-PCR analysis on mRNA samples from wild-type and p85α^−/− OCs generated in a manner similar to those described for the microarray analysis. We focused our analysis on genes known to play a critical role in OC growth and maturation and shown to be significantly downregulated in microarray analysis, including, macrophage-stimulating 1 receptor, OC-associated receptor, calcitonin receptor, TRAP, MMP-9, integrin αv, integrin β3, cathepsin K, transcription factor MITF, and JDP2. Consistent with the microarray analysis, two independent quantitative real-time RT-PCR experiments also demonstrated a two- to threefold reduction in the expression of these genes in p85α^−/− OCs compared to wild-type OCs (Fig. 10 and data not shown).

FIG. 10.

Quantitative RT-PCR analysis of OC-specific genes in wild-type (WT) and p85α^−/− OCps. Wild-type and p85α^−/− OCps were generated from bone marrow cultured for 6 days with RANKL (100 ng/ml) and M-CSF (10 ng/ml). Total mRNA was extracted as described in Materials and Methods. Expression of mRNA for TRAP, JDP2, cathepsin K (Ctsk), calcitonin receptor (Calcr), MMP-9, and integrin β3 (Itgβ3) was analyzed by real-time RT-PCR using β-tubulin mRNA as an endogenous control. Bars represent the mean ± standard deviation of one independent experiment performed in replicates of three. Similar results were seen in two independent experiments. *, P < 0.01, wild-type versus p85α^−/−.

In summary, microarray and quantitative RT-PCR analysis of wild-type and p85α^−/− OCs verified the induction of numerous target genes involved in OC maturation and growth and several genes involved in OC function. These data suggest that the transcriptional activation regulated by p85α in response to M-CSF/RANKL stimulation may be important for OC growth and differentiation.

Reconstituting the expression of p85α in p85α^−/− OCps rescues OC maturation defects.

Having established that p85α has a specific role in modulating OC functions in vivo and in vitro, we next examined whether the defect in OC differentiation in p85α^−/− mice and p85α^−/− OCs is directly related to the loss of p85α protein and not due to other changes in these cells. We cloned the p85α cDNA into a retroviral vector and then transduced wild-type or p85α^−/− OCps either with the empty vector or with a virus expressing the p85α cDNA. Transduction efficiency was monitored on the basis of EGFP expression. EGFP-positive cells were sorted to homogeneity and differentiated into OCs (Fig. 11). p85α^−/− OCps expressing the empty vector demonstrated a significant reduction in the generation of TRAP⁺ multinucleated cells (Fig. 11C). In contrast, reconstituting p85α^−/− OCps with a virus expressing the full-length p85α cDNA rescued the maturation (as observed by large multinucleated OCs) of p85α^−/− OCps to wild-type levels.

FIG. 11.

Reexpression of p85α into p85α^−/− OCps restores normal OC formation in vitro. (A) Flow cytometry profiles demonstrating the percentage of transduced cells (as determined by EGFP expression; x axis) expressing either the empty vector (control) or HA-tagged full-length form of p85α (p85α-HA) or an HA-tagged version of p85α with a deletion of the SH3 domain (p85αΔSH3). (B) Wild-type and p85α^−/− cells expressing various versions of p85α indicated in panel A were sorted to homogeneity and subjected to Western blot (WB) analysis. Arrows in the top panel indicate the expression of various p85α constructs as determined by an anti-HA antibody (ab). The middle panel demonstrates a comparison of the level of expression of the HA-tagged full-length version of p85α in p85α^−/− cells with that of endogenous levels of p85α in wild-type cells using an antibody that recognizes only the SH3 domain of p85α. The bottom panel demonstrates the level of protein in each lane as determined by the expression of β-actin. (C) Representative photomicrographs of OC culture following TRAP staining. Wild-type and p85α^−/− OCps expressing the indicated constructs were cultured in the presence of M-CSF and RANKL for 6 days and subjected to TRAP staining. Shown is a representative field. Similar results were observed in two independent experiments. WT, wild type; α, anti.

Since the maturation defects associated with p85α^−/− OCps were observed in spite of the presence of the shorter p50α and p55α isoforms, we hypothesized that the amino-terminal SH3 domain of p85α must contribute to the maturation of OCps downstream from M-CSF and RANKL. To test this, we constructed a version of p85α with a deletion of the SH3 domain and expressed it in p85α^−/− OCps. Figure 11A demonstrates the transduction efficiency of various constructs as assessed by EGFP expression shown on the x axis. Figure 11B demonstrates the relative expression of various constructs as assessed by Western blot analysis of the HA-tagged version of either the full-length form of p85α or the version with the SH3 domain deletion (Fig. 11B, top panel). Figure 11B (middle panel) demonstrates total p85α protein as assessed by an antibody against the SH3 domain of p85α. Note the undetectable expression of p85α in Fig. 11B (lane 3, middle panel). As seen in Fig. 11C, expression of p85α with the SH3 deletion in p85α^−/− OCps only marginally rescued OC maturation. Taken together, these results confirm the importance of the amino-terminal SH3 domain of p85α in regulating RANKL- and M-CSF-induced OC growth and maturation.

DISCUSSION

A large body of evidence has demonstrated that the PI3-K pathway is intimately associated with different phases of OC growth and development; however, the exact role this enzyme plays in OC function(s) in vivo and the role of specific isoforms of PI3-K in OC growth and development in vivo or in vitro remain poorly understood. The conclusions from most studies concerning the role of PI3-K in OCs have been based on pharmacologic inhibitors of PI3-K (wortmannin and LY294002). Although informative, these data are limited due to the nonspecific inhibition of all classes of PI3-Ks by pharmacologic inhibitors, thereby underscoring the need to examine the relevance of specific regulatory as well as catalytic subunits of PI3-K in OC development. Identifying the function of specific subunits is crucial as recent studies have begun to demonstrate specificity with respect to the function of distinct catalytic and regulatory subunits of PI3-K (59). Furthermore, if the PI3-K pathway is to be appropriately targeted for the treatment of diseases involving OCs, a better understanding of the relative contribution of various subunits of PI3-K in regulating OC functions is necessary. Our results in primary OCs demonstrate an essential and nonredundant role for p85α regulatory subunit of class I_A PI3-K in controlling multiple aspects of OC biology. Importantly, these defects are observed in spite of the continued presence of other subunits of class I_A PI3-K in these cells.

p85α and p85β subunits of class I_A PI3-K share near identity in the known functional domains in the carboxy terminus, including the amino-SH2 and the carboxy-SH2 domains, which are critical for mediating association with other SH2-containing proteins as well as binding to the p110 catalytic subunit. Interestingly, the domains in the carboxy terminus of p85α and p85β are also shared by the p50α and p55α subunits of class I_A PI3-K. Therefore, the basis for any differences in the interactions with the SH2 domains of other proteins is likely not to be derived from the carboxy terminus sequences of these regulatory subunits. Thus, the amino-terminal domain of p85α must play a unique role in regulating M-CSF/RANKL-induced growth, differentiation, and gene expression in OCs. Consistent with this notion, our in vitro results demonstrating a complete rescue in the maturation of p85α-deficient OCs by reconstituting the expression of the full-length form of p85α but not the version with the deletion of the SH3 domain suggests that the signals emanating from the SH3 domain of p85α are likely to play a critical role in OC maturation. Previous studies in heterologous cell lines have suggested that the SH3 domain of p85α can interact with proteins that consist of proline-rich regions, such as Sos and Cbl (25, 53). Studies are ongoing to identify proteins that preferentially bind to the amino-terminal end of p85α in OCs.

Our results in p85α^−/− OCs corroborate recent findings observed in SHIP-deficient (SHIP^−/−) OCs (47). SHIP is known to downregulate PI3-K-initiated signals by dephosphorylating PI-3,4,5-triphosphate (PIP3). Deficiency of SHIP results in enhanced survival of OC precursors, including hypersensitivity to M-CSF and RANKL (47). The size of SHIP^−/− OCs is significantly larger than that of wild-type OCs, and they contain significantly greater numbers of nuclei. Functionally, these cells exhibit greater resorptive activity than wild-type controls. Consistent with higher resorptive activity, trabecular thickness and the trabecular volume fraction are reduced in SHIP^−/− mice. Consistently, we show that deficiency of p85α, a positive regulator of PIP3, results in reduced sensitivity to M-CSF/RANKL, reduced OC size, and reduced resorptive activity leading to greater bone mass than in wild-type controls. These defects are likely due in part to reduced Akt and Erk MAP kinase activation in p85α^−/− OCs.

Although our results demonstrate that the deficiency of p85α impacts the growth, maturation, and actin-based functions in OCs, how precisely this occurs is likely to be complicated. To this end, our microarray and quantitative PCR studies demonstrate a significant reduction in the expression of multiple genes involved in the regulation of OC functions, including MITF as well as MITF target genes such as TRAP and cathepsin K. M-CSF- and RANKL-induced regulation of the transcription factor MITF is critical for normal OC development (21, 49, 56). MITF plays an essential role in multinucleation of OCs. Mutations in the MITF gene are associated with osteopetrosis due to impaired OC development in multiple species. OCs derived from cells bearing naturally occurring mutants of MITF tend to be smaller and mononuclear, rather than multinuclear, and often show lower TRAP activity than control OCs. The lack of normal numbers of nuclei in MITF mutant OCs has led to the idea that MITF may play an essential role in the fusion of OCs.

Mutations in the Ctsk gene cause the human disease pyknodysostosis (19). Pyknodysostosis is an autosomal recessive osteosclerotic disorder that is manifested in the form of short stature, skeletal dysplasia, and bone fragility (19). Deficiency of cathepsin K in mice results in mild osteopetrosis, elevated numbers of OCs, and increased bone mass (32). A finding similar to the one reported here in p85α^−/− mice. Since M-CSF has been shown to regulate cathepsin K levels, our results suggest that M-CSF-mediated p85α-induced PIP3 levels must play an essential role in this process.

In addition to reduced expression of MITF and MITF target genes such as cathepsin K in p85α^−/− OCs, the expression of the transcription factor JDP2 was also reduced twofold. Overexpression of JDP2 leads to activation of both TRAP and cathepsin K gene promoters as well as the formation of TRAP-positive multinuclear OCs (27). Antisense oligonucleotide to JDP2 strongly suppresses OC formation (27). The fact that p85α^−/− OCs demonstrate a twofold reduction in the expression of JDP2 along with reduced TRAP activity and cathepsin K expression supports the notion that perhaps JDP2 and MITF collaborate to regulate normal OC maturation, which is significantly impaired as a result of p85α deficiency in p85α^−/− OCs. Although MITF and JDP2 are implicated in regulating the expression of cathepsin K and TRAP, our results demonstrate that deficiency of p85α also affects the expression of other proteins involved in OC maturation, including MMP-9 and calcitonin receptor. Therefore, it is likely that the overall reduction in the expression of multiple OC-specific genes contributes to the p85α^−/− OC phenotype in vitro.

Bone resorption by OCs is partly dependent on cell adhesion and migration. In OCs, Rho family GTPases such as Rho and Rac downstream from the M-CSF receptor and αvβ3 integrin play an essential role in this process. Mice deficient in the expression of β3 or Vav3 (the guanine exchange factor for Rho and Rac) demonstrate defective bone resorption in part due to defects in cytoskeletal organization. Surprisingly, in spite of defects in bone resorption in β3^−/− and Vav3^−/− mice, these mutant mice consist of substantially greater numbers of OCs in vivo compared to wild-type controls (16, 33). Interestingly, deficiency of p85α also results in increased numbers of OCs in vivo, which is associated with defects in actin-based functions including adhesion and migration via αvβ3 and M-CSF. We have recently demonstrated that the stimulation of bone marrow-derived macrophages from p85α^−/− mice with M-CSF also results in reduced activation of Rho GTPase Rac (34). Thus, it is likely that the enhanced bone mass, along with defective resorptive capacity of p85α^−/− OCs, is in part contributed by altered Rac activation via αvβ3 and M-CSF.

Although the number of OCs is greater in p85α^−/− mice than in wild-type controls, these cells appear to be defective in their functional capabilities, including adhesion, migration, and bone resorption. These data suggest that perhaps impairing the ability of OCs to efficiently remove bone in vivo triggers a compensatory accumulation of functionally impaired OCs in p85α^−/− mice. Although the mechanism(s) underlying this response is unclear, the phenomenon of increased numbers of OCs in vivo due to an apparent deficiency in a signaling molecule in OCs does not seem to be unique to p85α-deficient mice. Others have reported a similar accumulation of OCs in vivo, which is associated with decreased net resorption after extended treatment with bisphosphonate. In addition, several mutant mice lacking signaling components of the M-CSF and/or RANKL as well as the αvβ3 signaling cascade have been described that also demonstrate higher numbers of TRAP 5b⁺ OCs in vivo (similar to those seen in p85α^−/− mice) but possess enhanced bone volume as a result of defective OC function in vitro, including bone resorption. These include mice that are deficient in the expression of the guanine exchange factor for Rac, Vav3 (16), β3 integrin (33), cathepsin K (32), NIK (37), and p62 (14) as well as mice lacking the membrane adaptor protein DAP12 (17, 23). A common feature of these mice with respect to osteoclastogenesis is the presence of severe and profound defects in OC growth and differentiation as well as other functions, including multinucleation in vitro in response to M-CSF and RANKL despite normal or elevated numbers of OCs in vivo. Although the exact reason behind the in vivo and in vitro disparity in these mutant mice is not clear, some hypotheses have been put forth to explain these differences (37). To this end, it has been suggested that in vitro culture conditions are likely to reflect activated or stress-induced osteoclastogenesis rather than basal osteoclast development (37). Alternatively, it has been proposed that a compensatory mechanism(s) may mask the in vivo bone phenotypes in some of these mutant mice (37). There are significant data to support both these possibilities. Studies have shown that several of the above-listed mouse mutants do not reveal substantial bone phenotypes under steady-state conditions but exhibit defects under conditions of stress. Likewise, although M-CSF and RANKL play an essential role in the growth and differentiation of osteoclasts in vitro, it is conceivable that in the absence of normal signaling via these cytokines in vivo, other cytokines within the bone marrow microenvironment, such as tumor necrosis factor alpha (TNF-α) and/or transforming growth factor β might compensate (1, 6, 10, 18, 24, 28, 31). Previous studies have shown that a combination of RANKL, TNF-α, and transforming growth factor β can rescue the OC defects observed in the absence of NIK in vitro (37). Thus, it is likely that the expression of these additional growth factors in the bone marrow environment allows normal OCs development in the absence of p85α in vivo, which would be reflective of basal but not activated or stressed-induced osteoclastogenesis. These possibilities are currently being investigated.

In summary, our results demonstrate that genetic disruption of the p85α subunit of class I_A PI3-K dampens cytokine- and integrin-based functions in OCs. We further show that these defects are associated with alteration in the expression of critical OC-specific genes previously implicated in regulating both OC maturation and actin-based functions. Thus, p85α plays an essential role in integrating signals downstream from cytokines and integrins in regulating osteoclastogenesis.