Type I Phosphotidylinosotol 4-Phosphate 5-Kinase γ Regulates Osteoclasts in a Bifunctional Manner*

Background: PI(4,5)P2, mainly synthesized by PIP5KIγ, is essential for normal cell function. Results: Deficiency or overexpression of PIP5KIγ, which results in decrease or increase of PI(4,5)P2, respectively, impairs osteoclast differentiation and function. Conclusion: Optimal PIP5KIγ and PI(4,5)P2 expression are important for osteoclast function. Significance: This is the first demonstration that PIP5KIγ regulates cells in a bifunctional manner. Type 1 phosphotidylinosotol-4 phosphate 5 kinase γ (PIP5KIγ) is central to generation of phosphotidylinosotol (4,5)P2 (PI(4,5)P2). PIP5KIγ also participates in cytoskeletal organization by delivering talin to integrins, thereby enhancing their ligand binding capacity. As the cytoskeleton is pivotal to osteoclast function, we hypothesized that absence of PIP5KIγ would compromise their resorptive capacity. Absence of the kinase diminishes PI(4,5) abundance and desensitizes precursors to RANK ligand-stimulated differentiation. Thus, PIP5KIγ−/− osteoclasts are reduced in number in vitro and confirm physiological relevance in vivo. Despite reduced numbers, PIP5KIγ−/− osteoclasts surprisingly have normal cytoskeletons and effectively resorb bone. PIP5KIγ overexpression, which increases PI(4,5)P2, also delays osteoclast differentiation and reduces cell number but in contrast to cells lacking the kinase, its excess disrupts the cytoskeleton. The cytoskeleton-disruptive effects of excess PIP5KIγ reflect its kinase activity and are independent of talin recognition. The combined arrested differentiation and disorganized cytoskeleton of PIP5KIγ-transduced osteoclasts compromises bone resorption. Thus, optimal PIP5KIγ and PI(4,5)P2 expression, by osteoclasts, are essential for skeletal homeostasis.

Bone resorption is the unique purview of the osteoclast. It involves cytoskeletal organization to polarize vesicles toward the bone-apposed plasma membrane into which they insert to deliver their cargo of matrix-degrading molecules to the extracellular, resorptive microenvironment (1). As phosphatidylinositols (PIs) 3 are key to cytoskeletal organization and function of all mammalian cells, it is likely that they participate in the bone resorptive process (2,3). PI(4,5)P 2 is among the most biologically relevant PIs as it modulates phenomena such as migration, apoptosis, and cell division (4,5). A key event in the biological activity of PI(4,5)P 2 is its metabolism by phospholipase C␥ into diacylglycerol and inositol-1,4,5-trisphosphate which, via ionized calcium, activates nuclear factor of activated T cells c1 (NFATc1) for osteoclast differentiation (4 -6). In consequence, arrest of PI(4,5)P 2 metabolism by deletion of phospholipase C␥2, the relevant isoform in the resorptive cell, inhibits osteoclastogenesis (7). PI(4,5)P 2 also regulates osteoclast cytoskeletal organization by interacting with gelsolin and WASP to control podosome signaling and formation of actin rings which isolate the resorptive milieu from the general extracellular space (3). However, phospholipase C␥2 deletion, which presumably increases PI(4,5)P 2 by preventing its metabolism, blunts the capacity of the cell to organize its cytoskeleton via the ␣v␤3 integrin and thus degrades bone (8,9). PI(4,5)P 2 is principally synthesized by phosphorylation of phosphatidylinositol-4 phosphate at the 5-position of the inositol ring via type I phosphatidylinositol 4-phosphate 5 kinase (PIP5KI). Three PIP5KI isoforms, ␣, ␤, and ␥, with conserved catalytic domains, are in hand. These isoforms enjoy discrete biological activities likely dictated by their non-overlapping distributions, reflecting, in turn, their distinct C-terminal regions.
Talin promotes the capacity of osteoclasts to organize their cytoskeleton and thus resorb bone (15). Moreover, bone resorption is an exocytic process, which PIP5KI␥ regulates in other cells, and we find the kinase is expressed with differentiation of osteoclast precursors into the resorptive polykaryon (11, 16 -18). We find absence of PIP5KI␥ delays osteoclast differentiation in response to receptor activator of NFB ligand (RANKL) but once formed, the cells exhibit no cytoskeletal or functional abnormalities. In keeping with these in vitro observations, PIP5KI␥ Ϫ/Ϫ mice have arrested osteoclastogenesis in vivo. Surprisingly, excess PIP5KI␥ also delays osteoclast differentiation and in contrast to deficiency of the kinase, disrupts the cytoskeleton of the osteoclast in a talin-independent manner. Thus, reflecting abundance, PIP5KI␥ exerts varied effects on osteoclasts.

EXPERIMENTAL PROCEDURES
Mice-PIP5KI␥ ϩ/Ϫ (C57/BL6 background) mice were described previously (15). Mice were genotyped by PCR using 5Ј-GCCTCACAGAGATTTGACCTGTCA-3Ј (PIP5K␥ ϩ ) and 5Ј-CACCCCTTCCCAGCCTCTGA-3Ј (PIP5K␥ Ϫ ) as forward primers in combination with the 5Ј-CCTCACATCCTGCT-CACTCAGGACC-3Ј reverse primer. The products (250 and 650 bp for PIP5K␥ Ϫ and PIP5K␥ ϩ , respectively) were resolved on a 1% agarose gel. PIP5K␥ Ϫ/Ϫ and WT liver cells were obtained from postnatal day 0 pups and prepared as described (19). Animals were housed in the animal care unit of the Washington University School of Medicine and were maintained according to guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care. All animal experimentation was approved by the Animal Studies Committee of the Washington University School of Medicine.
Reagents-Recombinant M-CSF was obtained from R&D Systems (Minneapolis, MN). Glutathione S-transferase (GST)-RANKL was expressed in our laboratory as described (20) and unless specified used at 100 ng/ml. The sources of antibodies were as follows: rabbit anti-HA polyclonal antibody was from Covance (Emeryville, CA); antiphosphotyrosine mAb 4G10 (p-Tyr) was from Upstate (Charlottesville, VA); directed against the Src protein were gifts from Andrey Shaw (Department of Pathology, Washington University School of Medicine, St Louis, MO); mouse anti-NFATc1 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA); rabbit anti-␤3-integrin antibody, JNK, p-JNK, p38, p-p38, Akt, p-Akt, ERK, p-ERK, IB, and p-IB were from Cell Signaling (Beverly, MA); mouse anti-talin monoclonal and anti-actin antibodies were from Sigma; mouse anti-PIP5K␥ was from BD Biosciences; rabbit anti-PIP5K␥-pan and anti-PIP5K␥2 were for immunoblot (21); rabbit anti-PIP5K␥-pan WI155 for immunostaining was kindly provided by Dr. Anderson (14). All other chemicals were obtained from Sigma.
Macrophage Isolation and Osteoclast Culture-Primary liver and bone marrow macrophages were obtained and treated as described to generate osteoclasts (19,20). Cells were extracted with ␣-MEM and cultured in ␣-MEM containing 10% inactivated FBS, 100 international units/ml penicillin, and 100 g/ml streptomycin (␣-10 medium) with 1:10 CMG (conditioned medium supernatant containing recombinant M-CSF 1 g/ml) on Petri dishes (28). Cells were incubated at 37°C in 6% CO 2 for 4 days and then washed with PBS and lifted with 1ϫ trypsin/ EDTA (Invitrogen) in PBS. Unless specified differently, 1.5 ϫ 10 4 cells were cultured in 500 l ␣-MEM containing 10% heatinactivated FBS with 100 ng/ml GST-RANKL and 1:50 CMG in 48-well tissue culture plates, some of which contained a sterile bone slice. Cells were fixed and stained for tartrate-resistant acid phosphatase (TRAP) activity after 4 -5 days in culture using a commercial kit (387-A; Sigma-Aldrich). For preosteoclast generation, 1.5 ϫ 10 6 bone marrow macrophages (BMMs) were plated per 10-cm tissue culture dish and maintained in 1:50 CMG and 100 ng/ml GST-RANKL for 3 days. Preosteoclasts were lifted with 0.02% EDTA in PBS. For actin ring staining, cells were cultured on a bovine bone slice in the presence of M-CSF and RANKL for 6 days, at which time cells were fixed in 4% PFA, permeabilized in 0.1% Triton X-100, rinsed in PBS, and immunostained with Alexa Fluor 488 phalloidin (Invitrogen). To quantitate resorption lacunae, bone slices were incubated with peroxidase-conjugated wheat germ agglutinin (Sigma) for 1 h and stained with 3,3Ј-diaminobenzidine (Sigma).
Plasmids and Retroviral Transduction-Wild-type constructs expressing mouse PIP5KI␥1, PIP5KI␥2, and kinase-inactive (KD) mutant were subcloned into the BamHI and XhoI sites of a pMX retroviral vector in which the puromycin resistance sequence was replaced with one coding for blastocidin resistance (18). cDNA was transfected transiently into Plat-E packaging cells using FuGENE 6 transfection reagent (Roche Applied Science). Virus was collected 48 h after transfection. Macrophages were infected with virus for 24 h in the presence of 1:10 CMG and 4 g/ml polybrene (Sigma). Cells were selected in the presence of CMG and 1 g/ml blasticidin (Calbiochem) for 3 days before use as osteoclast precursors.
RNA Extraction and Quantitative RT-PCR-RNA from cultured cells was isolated and purified using the RNeasy RNA purification kit (Qiagen, Valencia, CA); RLT lysis buffer was supplemented with ␤-mercaptoethanol (1%). Purified RNA was treated with DNase I (Invitrogen) before reverse transcription (RT). RT was performed using SuperScript III (Invitrogen), and quantitative PCR was performed using Power SYBR Green Master mix and gene-specific primers (Applied Biosystems, Foster City, CA). The quantitative PCR reaction was performed on an ABI Prism 7000 (Applied Biosystems).
Ca 2ϩ Imaging-Shortly before imaging, WT and PIP5KI␥deficient macrophages were incubated with 3 M Fura 4-AM (F14201, Invitrogen) for 30 min at 37°C in ␣-MEM containing 10 ng/ml M-CSF. The cells were washed and incubated in fresh ␣-MEM with 10 ng/ml M-CSF 30 min at 37°C. Ca 2ϩ imaging was performed with a 40ϫ objective of an inverted automated wide field epifluorescence DIC microscope (Leica DMIRE2, Leica Microsystems, Wetzlar, Germany). An objective lens heater was used to improve temperature homogeneity. Images (608 ϫ 512 pixels spatial and 12-bit intensity resolution) were recorded with a cooled Retiga 1300 camera (Qimaging, Burnaby, BC, Canada) every 10 s in 2 ϫ 2 binned acquisition mode, using 100 -300-ms exposures. Dynamic images were composed using NIH ImageJ software.
Phosphoinosotide Measurement (Sample Preparation)-Lipids were extracted as described previously by Pettitt et al. (22). Briefly, three ml of ice-cold CHCl 3 -MeOH (1:2; v/v) was added to the cell platelets of BMMs. After vigorous vortexing to ensure cell distribution, the samples were incubated on ice for 15 min, and then ice-cold CHCl 3 (1 ml) was added, followed by 1 ml of solution containing 1.76% KCl, 100 mM citric acid, 100 mM Na 2 HPO 4 , 5 mM EDTA, and 5 mM tetrabutyl ammonium hydrogen sulfate (pH 3.6). After mixing, the samples were incubated on ice for 5 min before centrifugation (200 ϫ g, 5 min) to complete the phase split. The lower organic phase was transferred into a clean silanized vial and evaporated to dryness under a stream of nitrogen.
Electrospray Ionization Mass Spectrometry (ESI-MS)-Mass spectrometric analyses of phosphatidylinositol phosphates (PIPs) were carried out in the negative-ion mode using a Thermo Scientific (San Jose, CA) LTQ Orbitrap Velos instrument with Xcalibur operating system. The instrument was operated in FT mode with resolution set at 100,000 (at m/z 400) and scan from 250 to 1000 m/z. The automatic gain control was set to 1 ϫ 10 6 ions with a maximal injection time of 100 ms, and one microscan was recorded. The lock mass option was enabled, and deprotonated stearic anion (C 17 H 35 CO 2 -; m/z 283.26425) was used for internal recalibration of the mass spectra. CID was done with a target value of 3000 in the linear ion trap, maximal injection time of 200 ms, collision energy of 20%, Q value of 0.25, and an activation time of 10 ms. The dry lipid extracts containing PIPs were redissolved in 50 l of 50% CH 3 OH/H 2 O, vortexed for 1 min, and centrifuged at 13,600 rpm for 3 min. The supernatant was transferred to a silanized tube and an aliquot (5 l) of the solution was mixed with 5 l of methanol (with 1% NH 4 OH) and loop injected onto the mass spectrometer via a built-in syringe pump, which maintained a continuous flow of 20 l/min with 80% methanol (with 0.5% NH 4 OH). The skimmer was set at ground potential, the electrospray needle was set at 3.0 kV, and temperature of the heated capillary was 250°C. Mass spectra were accumulated in the profile mode, and spectra were background subtracted.
Western Blotting and Immunoprecipitation-Cultured cells were washed twice with ice-cold PBS and lysed in radioimmune precipitation assay buffer containing 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 mM NaF, and 1ϫ protease inhibitor mixture (Roche Applied Science). After incubation on ice for 10 min, cell lysates were clarified by centrifugation at 15,000 rpm for 10 min. Forty micrograms of total lysates were subjected to 8 -12% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto PVDF membranes. Filters were blocked in 0.1% casein in PBS for 1 h and incubated with primary antibodies at 4°C overnight followed by probing with fluorescence-labeled secondary antibodies (Jackson ImmunoResearch Laboratories). Proteins were detected with the Odyssey Infrared Imaging System (LI-COR Biosciences).
Actin Ring Staining and Bone Resorption Assay-Osteoclasts were generated on bone slices by exposure to 100 ng/ml RANKL and 20 ng/ml M-CSF. For actin ring staining, the cells were fixed in 4% paraformaldehyde and permeablized in 0.1% Triton X-100, rinsed in PBS, and immunostained with Alexa Fluor 488-phalloidin (Invitrogen). For bone resorption assay, osteoclasts were removed, and resorption pits were visualized by incubation of the specimen with 20 g/ml peroxidase-conjugated wheat germ agglutinin as described previously and stained with 3,3Ј-diaminobenzidine (Sigma).
Medium CTx-1 Assay-Macrophages were cultured on bone with 100 ng/ml RANKL and 1:50 CMG supernatant for 6 days. Medium (␣-10) was changed 1 day before harvesting. Medium CTx-1 concentration was determined using CrossLap for Culture ELISA kit (Nordic Bioscience Diagnosis A/S).
Immunofluorescence and Confocal Microscopy-Osteoclasts grown on glass slides in a 24-well plate were fixed with 4% paraformaldehyde in PBS for 20 min, followed by permeabilization and blocking in PBS/0.2% BSA for 30 min. The cells were then incubated with primary antibodies in 0.2% BSA/PBS overnight at 4°C. Primary antibody binding was visualized using fluorescent dye-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) in 0.2% BSA/PBS for 1 h. F-actin was stained with Alexa Fluor 488-phalloidin. Samples were mounted with 70% glycerol/PBS. Immunofluorescence-labeled cells were observed with a Carl Zeiss fluorescence microscope equipped with a CCD camera and analyzed using a Zeiss Spectral confocal laser scanning microscope equipped with Argon-Krypton lasers (LSM510 Zeiss Microsystems).
Dynamic Imaging of Osteoclasts-WT osteoclasts expressing PIP5K␥1, PIP5K␥2, and PIP5KI␥-KD, co-transduced with GFP-actin were maintained using standard culture conditions (37°C and 5% CO 2 , 95% air atmosphere) with RANKL and M-CSF, for 5 days, in a Bioptechs (non-liquid perfused) Delta T culture system, consisting of a heated, indium-tin-oxide-coated glass dish attached to a calibrated Bioptechs microperfusion peristaltic pump. All cultures were observed with a 20ϫ objective (numerical aperture of 0.4) of an inverted automated wide field epifluorescence DIC microscope (Leica DMIRE2, Leica Microsystems, Wetzlar, Germany). An objective lens heater was used to improve temperature homogeneity. Images (608 ϫ 512 pixels spatial and 12-bit intensity resolution) were recorded with a cooled Retiga 1300 camera (Qimaging, Burnaby, BC, Canada) every 2 min in 2 ϫ 2 binned acquisition mode, using 100 -300-ms exposures. Dynamic images were composed using NIH ImageJ software.
Histological Analysis of Mouse Pups-Embryonic bones were fixed in 10% formalin, and decalcified sections were stained for
Statistics-All data are expressed as mean Ϯ S.D., and statistical significance was calculated by analysis of variance.

PIP5KI␥ Deficiency Inhibits Osteoclast Formation but Not
Function-Our first exercise in exploring the role of PIP5KI␥ in osteoclast formation and function was to determine the expression of the enzyme with precursor differentiation. PIP5KI␥ mRNA, measured by quantitative PCR, was detectable in naïve BMMs and increased ϳ5-fold with exposure to the osteoclastogenic cytokines, whereas PIP5KI␣ mRNA did not change (Fig. 1A). M-CSF/RANKL-induced protein expression was as profound, enhancing total PIP5KI␥ and PIP5KI␥2, 3.8-and 4.2fold, respectively (Fig. 1B).
We next assessed the impact of PIP5KI␥ on osteoclast formation. Because deletion of the kinase results in perinatal lethality, we utilized liver as the source of PIP5KI␥ Ϫ/Ϫ and WT osteoclast precursors.
To generate osteoclasts, we exposed liver-derived macrophages to M-CSF and increasing amounts of RANKL for 4 days and stained them for TRAP activity. Osteoclasts of both genotypes increased progressively with RANKL dose (Fig. 2, A and  B). In all circumstances, however, the number of polykaryons was significantly less in cultures containing PIP5KI␥ Ϫ/Ϫ cells. This difference in osteoclastogenesis was most apparent at relatively low concentrations of the cytokine. The decreased number of mutant osteoclasts was reflected by diminished expression of the specific differentiation markers, ␤3 integrin subunit, NFATc1, and c-Src (Fig. 2C).
To determine the physiological relevance of the decreased osteoclastogenesis attending PIP5KI␥ deficiency, in vitro, we asked whether the same obtains in vivo. Thus, we prepared histological sections of long bones of PIP5KI␥ Ϫ/Ϫ and WT newborn mice and stained them for TRAP activity. In keeping with our in vitro observations, PIP5KI␥-deficient mice have a striking paucity of osteoclasts (Fig. 3).
These experiments indicate that PIP5KI␥ is essential for generation of normal numbers of osteoclasts, in vitro and in vivo. In both circumstances, however, the kinase-deficient TRAP-expressing polykaryons were well spread and morphologically identical to their WT counterparts, suggesting normal cytoskeletal organization and function. In keeping with this posture, actin ring formation by mutant or WT prefusion osteoclasts generated on bone was indistinguishable (supplemental Fig.  1A). To directly assess resorptive capacity of individual cells, prefusion osteoclasts were generated on plastic by 3 days exposure to RANKL and M-CSF, at which time proliferation is largely arrested. The cells were lifted and equal numbers placed on bone slices. Resorption pits, quantified 3 days later, were excavated in equal numbers by WT and PIP5KI␥ Ϫ/Ϫ osteoclasts (supplemental Fig. 1, B and C).
PIP5KI␥ Regulates RANKL Signaling-Because M-CSF and RANKL are necessary and sufficient for osteoclastogenesis, we asked which of these cytokines is impacted by absence of  PIP5KI␥. Thus, we treated WT and mutant liver-derived osteoclast precursors with either M-CSF or RANKL and assessed activation of established effector molecules, which mediate osteoclast formation. Absence of PIP5KI␥ had no effect on M-CSF-induced phosphorylation of ERK, JNK, or AKT (Fig.  4A). The same cells exposed to RANKL, however, were incapable of normally activating AKT, JNK, or p38 (Fig. 4B). Because these data indicate RANKL mediates the osteoclastogenic properties of PIP5KI␥, we determined whether the cytokine activates the kinase? Thus, we transduced macrophages derived from PIP5KI␥ Ϫ/Ϫ liver with HA-PIP5KI␥ and exposed the cells to either of the two cytokines for as long as 15 min. HA immunoprecipitates were immunoblotted for tyrosine phosphorylation, a marker of the activation of the kinase (Fig. 4C). Consistent with induction of its effectors, RANKL, but not M-CSF, phosphorylates PIP5KI␥. The RANKL-induced calcium signal is essential for osteoclast differentiation. Because PIP5KI␥ modulates calcium transport, which in turn, activates the key osteoclastogenic transcription factor, NFATc1, we asked if absence of the kinase disrupts calcium signaling in osteoclast precursors. Consistent with impaired osteoclastogenesis, RANKL-induced calcium signaling is diminished in PIP5KI␥ Ϫ/Ϫ macrophages (Fig. 4, D and E). Thus, PIP5KI␥ promotes osteoclast differentiation by sensitizing its precursors to RANKL. The results are consistent with the decreased levels of PIP 2 in the brain of the PIP5KI␥ Ϫ/Ϫ mice (5,17). Furthermore, given the specificity of PIP5KI␥ in generating PI(4,5)P 2 , these observations suggest it is the major PIP 2 in osteoclast precursors.

PIP5KI␥ Excess Disrupts Osteoclast Formation and Function-
With the expectation that PIP5KI␥1 or PIP5KI␥2 would rescue the osteoclastogenic failure of PIP5KI␥ deficiency, we transduced WT constructs, tagged to HA, into mutant liver cells that were then cultured in M-CSF and RANKL for 4 days. Unexpectedly, osteoclast number of either PIP5KI␥ Ϫ/Ϫ transductant did not increase and, in fact, decreased further relative to those bearing vector (Fig. 5, A and B). These observations raised the possibility that excess PI(4,5)P 2 synthesis may, similar to its deficiency, arrest osteoclastogenesis. This hypothesis is in keeping with the anti-osteoclastogenic effects of a likely cause of PI(4,5)P 2 accumulation, namely absence of phospholipase C␥2, which also disrupts organization of the cytoskeleton of the cell (7,8). In fact, while the magnitude of the 18:0/20:4-PIP 2 peak in empty vector-transduced BMMs is similar to that of WT, overexpressed PIP5KI␥1 and PIP5KI␥2 increases it by ϳ2 orders of magnitude (supplemental Table 1) and markedly alters the facade of PIP5KI␥ Ϫ/Ϫ osteoclasts. Both transductants exhibit features common to osteoclasts incapable of normal cytoskeletal organization in that they fail to spread and have a "crenated" appearance (1). Confirming this phenomenon is talin-independent, PIP5IK␥2, but not PIP5KI␥1, binds to the integrin-activating protein in osteoclasts (Fig. 5C).
To determine the effect of increased PIP5KI␥ on normal osteoclastogenesis, we transduced WT BMMs with PIP5KI␥1, PIP5KI␥2, or empty vector. Both splice variants delay osteoclast differentiation and disfigure the cell (Fig. 6, A-C). The magnitude of inhibited osteoclastogenesis and the distorted appearance of TRAP-expressing cells reflects the quantity of transduced PIP5KI␥ (Fig. 6, D and E). Furthermore, in control WT osteoclasts, PIP5KI␥ localizes with actin in podosome belts (Fig. 6F). Overexpressed PIP5KI␥2 and PIP5KI␥1 also co-localize with actin but not in podosome belts as these structures fail to form, thereby confirming cytoskeletal impairment (supplemental Movie 1). Thus, a surfeit of PIP5KI␥ disrupts formation of osteoclasts and organization of their cytoskeleton.
The multifunctional properties of PIP5KI␥ reflect the fact that it serves as both a kinase and adaptor protein. We, therefore, asked whether the inhibitory properties of the molecule require its enzymatic activity. Transduction of a kinase-inactive PIP5KI␥ mutant (PIP5KI␥-KD), incapable of catalyzing PI(4,5)P 2 synthesis, yielded normal appearing WT osteoclasts with podosome belts and did not suppress their number (Fig. 6,  A and B, supplemental Movie 2). This observation is fortified by

PI-4 Phosphate 5-Kinase ␥ Regulates Osteoclasts
failure of PIP5KI␥-KD to dampen expression of differentiation markers of WT osteoclast precursors cultured with RANKL or M-CSF or to reduce RANKL-stimulated NFATc1 nuclear translocation (Fig. 6, G and H).
To explore the functional implications of excess PIP5KI␥, we assessed the bone-degrading capacity of osteoclasts transduced with either isoform as well as that lacking kinase activity. Overexpressed PIP5KI␥2 or particularly PIP5KI␥1 substantially reduces resorptive pit formation and CTx mobilization, which does not occur absent kinase activity (Fig. 7, A-C). To determine whether the diminished resorption reflects impaired capacity of individual cells as well as a reduction in their num-ber, we cultured transduced BMMs in M-CSF and RANKL for 5 days on Petri dish at which time they are committed to the osteoclast phenotype. The cells were lifted, and equal numbers placed on bone slices for 2 days. Again, excess PIP5KI␥1 or PIP5KI␥2 impairs pit formation (Fig. 7D). Because the lifted cells are no longer replicative, this experiment establishes that dysfunction of the individual osteoclast contributes to the attenuated resorption induced by PIP5KI␥ excess.

DISCUSSION
PI(4,5)P 2 comprises a minor component of membrane phospholipids but profoundly influences the function of all cells. It serves not only as precursor to second messengers, but the intact PI is an important signaling molecule in its own right. Osteoclasts increase in number in mice lacking the 5Ј lipid phosphatase, SHIP1, due to cell cycle activation of their precursors (23). SHIP1 Ϫ/Ϫ BMMs are also hypersensitive to RANKL and M-CSF and, in consequence, generate large, active osteoclasts resulting in severe osteoporosis (24). Additionally, PI(4,5)P 2 and PIP 3 interact with gelsolin, thereby regulating podosome assembly and signaling in the bone resorptive cell (3,25). Finally, phospholipase C␥2, which metabolizes PI(4,5)P 2 to diacylglycerol and PIP 3 , modulates osteoclast formation and function (7,8).
The PIP5KI family, which accounts for the majority of PI(4,5)P 2 synthesis, consists of ␣, ␤ and ␥ isoforms which distribute in distinct intracellular locations and thus appear to regulate individual biological events. Because osteoclast function is attended by dramatic cytoskeletal reorganization, which in other cells is regulated by PIP5KI␥, we reasoned that deletion of this isoform is most likely to impact the resorptive cell (9,11,17,18). Our finding that PIP5KI␥2 increases in abundance as macrophages commit to the osteoclast phenotype fortifies this hypothesis. To determine whether the enzyme is necessary for osteoclast formation or function, we utilized mice globally deleted of all splice variants of PIP5KI␥. Because these animals die perinatally, we generated osteoclasts by culturing macrophage-containing hepatic aspirates in M-CSF and RANKL.
In keeping with the role of PIs in cell maturation, absence of PIP5KI␥ delays osteoclast differentiation, particularly that induced by relatively modest amounts of RANKL. Importantly, the requirement for PIP5KI␥ in skeletal metabolism is established by the paucity of osteoclasts in mice lacking the kinase.
As in the case of phospholipase C␥2 deficiency, osteoclastic cells lacking PIP5KI␥ respond normally to M-CSF but not RANKL (7). In both conditions, there is dampened activation and expression of osteoclast differentiation markers, including NFAT, which in other cells is regulated by diacylglycerol and PIP 3 (11). The blunted osteoclastogenesis of PIP5KI␥-and phospholipase C␥2-deficient osteoclasts is consistent with the expected reduction of PI(4,5)P 2 metabolites in both circumstances. Furthermore, as SHIP1 deficiency, which is attended by increased PIP 3 expression (26), is characterized by enhanced osteoclastogenesis, the reduced numbers of PIP5KI␥ Ϫ/Ϫ osteoclasts may reflect arrest of a second metabolic event whereby PI(4,5)P 2 is phosphorylated to PIP 3 by PI3K (27).
In contrast, PIP5KI␥ Ϫ/Ϫ osteoclasts exhibit no apparent cytoskeletal abnormalities. The cells spread as effectively as their WT counterparts, they form actin rings, and the individual bone resorbing capacities of both genotypes are indistinguishable. This observation is surprising as PI(4,5)P 2 regulates many aspects of cytoskeletal function in other cells and is purported to do so in the osteoclast (3). Furthermore, PIP5KI␥2 resides within focal adhesions and interacts with structural proteins such as talin, facilitating its association with integrins. Talin/integrin binding is a necessary event whereby these heterodimers organize the cytoskeleton, including that of osteoclasts (13,14,28). As most cells possess more than one isoform of PIP5KI, which may in some circumstances be redundant, a reasonable, albeit speculative, hypothesis holds that PIP5K␣ and/or PIP5K␤ prevent osteoclast dysfunction in face of PIP5K␥ deficiency (11,18). This posture is challenged, however, by the absence of MS-detectable PIP 2 in PIP5K␥ Ϫ/Ϫ osteoclast precursors in the form of BMMs.
In contrast to the normal appearance of individual bone resorptive cells lacking PIP5KI␥, its overexpression in mutant or WT cells inhibits spreading and yields a "pyknotic" appearance common to osteoclasts with disorganized cytoskeletons (1). The abnormal cytoskeleton attending excess PIP5K␥I is not unique to osteoclasts as focal adhesions are disrupted in other cells overexpressing the kinase (9,29). The mechanism of inhibited cytoskeletal organization, in these circumstances, is unknown but postulated to reflect local disruption of PI balance and removal of talin from focal adhesions by PIP5KI␥2 (9,13,14).
Our results indicate that the mechanisms of cytoskeletal dysfunction in PIP5KI-overexpressing osteoclasts may differ from those previously reported in other cells. Specifically, we find that cytoskeletal distortion occurs with equal efficiency in those bearing either PIP5KI␥2, which binds talin and removes it from focal adhesions, or PIP5KI␥1, which does not. Transduction of either splice variant fails to rescue the anti-osteoclastogenic effects of PIP5KI depletion, but this likely reflects abundance of the overexpressed proteins. Excess PIP5KI␥1 or PIP5KI␥2 mirrors absence of the parent gene, which arrests osteoclast formation due to delayed differentiation, but the overexpressed proteins also disrupt the cytoskeleton. The reduced number of osteoclasts in PIP5KI␥-transduced WT cells, in combination with the disruptive cytoskeletal effects, substantially dampens resorption. Similar to cytoskeletal organization, the diminished osteoclast abundance attending PIP5KI␥ excess reflects intact enzymatic activity as it does not occur in cells bearing a kinasearrested mutant.  . PIP5KI␥1 or PIP5KI␥2 overexpression arrests WT osteoclast formation and cytoskeletal organization in a kinase-dependent manner. A, WT BMMs transduced with vector, PIP5KI␥1, PIP5KI␥2, or kinase-inactive PIP5KI␥ (PIP5KI␥-KD) were cultured with RANKL and M-CSF for 4 days. The cells were stained for TRAP activity. B, appearance of cells as described in A. Scale bar, 250 m. C, WT BMMs transduced with HA-PIP5KI␥1, HA-PIP5KI␥2, or vector were exposed to M-CSF and RANKL. Markers of osteoclast differentiation, HA, and PIP5KI␥ were immunoblotted. ␤-Actin serves as loading control. D, WT BMMs were transduced with increasing amounts of PIP5KI␥2 plasmid, and expression was determined by immunoblot. ␤-Actin serves as loading control. E, WT BMMs transduced with increasing amounts of PIP5KI␥2 plasmid were cultured with RANKL and M-CSF for 4 days. The cells were stained for TRAP activity. Scale bar, 250 m. F, WT osteoclasts transduced with PIP5KI␥1, PIP5KI␥2, or vector (Vect) were stained with anti-PIP5KI␥ mAb (green) or phalloidin (red) to identify fibrillar actin. The kinase localizes with actin in peripheral podosome belts in vector-transduced cells (yellow). No podosome belts form in PIP5KI␥1-and PIP5KI␥2transduced osteoclasts, although kinase/actin co-localization is evident in other locations. Scale bar, 25 m. G, WT BMMs, transduced with vector or PIP5KI␥-KD, were cultured in M-CSF and RANKL. Markers of osteoclast differentiation were immunoblotted with time. ␤-Actin serves as loading control. H, WT preosteoclasts transduced with vector, PIP5KI␥1, PIP5KI␥2, or PIP5KI␥-KD were treated with RANKL for 20 min. Nuclear proteins were extracted and immunoblotted by NFATc1. The nuclear protein, lamin B1, serves as loading control.
As overexpressed PIP5KI␥ enhances PI(4,5)P 2 generation, our data raise the possibility that this PI, in excess, attenuates osteoclastogenesis and disrupts the cytoskeleton of the cell. These observations are in keeping with impairment of osteoclast formation and cytoskeletal organization attending deficient phospholipase C␥2 activity (7), another state of PI(4,5)P 2 abundance (9). Thus, both insufficient and excess PIP5KI␥ impact the osteoclast. Absence of the enzyme compromises its differentiation but not its function, whereas excess kinase disrupts both. These observations indicate that optimal amounts of PIP5KI␥ and its metabolites are essential for skeletal homeostasis.