P2X7 Nucleotide Receptors Mediate Blebbing in Osteoblasts through a Pathway Involving Lysophosphatidic Acid*

Extracellular nucleotides, released in response to mechanical or inflammatory stimuli, signal through P2 receptors in many cell types, including osteoblasts. P2X7 receptors are ATP-gated cation channels that can induce formation of large membrane pores. Disruption of the gene encoding the P2X7 receptor leads to decreased periosteal bone formation and insensitivity of the skeleton to mechanical stimulation. Our purpose was to investigate signaling pathways coupled to P2X7 activation in osteoblasts. Live cell imaging showed that ATP or 2 ′,3 ′-O-(4-benzoylbenzoyl)-ATP (BzATP), but not UTP, UDP, or 2-methylthio-ADP, induced dynamic membrane blebbing in calvarial osteoblasts. Blebbing was observed in calvarial cells from wildtype but not P2X7 knock-out mice. P2X7 receptors coupled to activation of phospholipase D and A2, inhibition of which suppressed BzATP-induced blebbing. Activation of these phospholipases leads to production of lysophosphatidic acid (LPA). LPA caused dynamic blebbing in osteoblasts from both wild-type and P2X7 knock-out mice, similar to that induced by BzATP in wildtype cells. However, LPA-induced blebbing was more rapid in onset and was not affected by inhibition of phospholipase D or A2. Blockade or desensitization of LPA receptors suppressed blebbing in response to LPA and BzATP, without affecting P2X7-stimulated pore formation. Thus, LPA functions downstream of P2X7 receptors to induce membrane blebbing. Furthermore, inhibition of Rho-associated kinase abolished blebbing induced by both BzATP and LPA. In summary, we propose a novel signaling axis that links P2X7 receptors through phospholipases to production of LPA and activation of Rho-associated kinase. This pathway may contribute to P2X7-stimulated osteogenesis during skeletal development and mechanotransduction.

In response to the appropriate cues, cells of mesenchymal origin differentiate to become mature osteoblasts that lay down the organic matrix of bone, which subsequently mineralizes (1). Some of these cells become osteocytes, trapped inside the matrix but still connected to each other by cell processes. This network of osteoblasts and osteocytes is suggested to be part of the mechanotransduction system in bone (2). There is growing evidence that extracellular nucleotides and their receptors on these cells mediate responses of the skeleton to mechanical loading.
In many cell types, nucleotides are released in response to a variety of stimuli, including stretch, hypoxia, osmotic swelling. and cell lysis (3). Fluid shear induces ATP release from osteoblast-like cells in vitro (4), and it has been proposed that osteoblasts and osteocytes release nucleotides such as ATP in response to mechanical loading in vivo (5). Extracellular nucleotides bind to cell surface P2 receptors, either P2X, which are ATP-gated nonselective cation channels, or P2Y, which are G protein-coupled receptors. There are seven subtypes of P2X (P2X1-7) and eight subtypes of P2Y (P2Y1, -2, -4, -6, and - [11][12][13][14] receptors currently identified in mammals. Osteoblasts express multiple subtypes of P2X and P2Y receptors (5,6). Interestingly, disruption of the gene encoding the P2X7 receptor results in an osteopenic phenotype involving decreased periosteal bone formation by osteoblasts and increased trabecular bone resorption by osteoclasts (7). Furthermore, stimulation of periosteal bone growth by mechanical loading is markedly attenuated in mice lacking the P2X7 receptor (8). However, the presence of functional P2X7 receptors in osteoblasts is controversial (7,9,10), and the signaling pathways that may be activated by these receptors in osteoblasts are not known.
Like other P2X receptors, P2X7 subunits consist of two transmembrane domains with an extracellular loop and intracellular N and C termini (11). Functional receptors are homomeric consisting of multiple P2X7 subunits. The P2X7 receptor behaves as a nonselective cation channel, and its activation can lead to the formation of pores permeable to hydrophilic molecules as large as 900 Da (12). It has remained controversial as to whether pore formation arises by dilation of the channel or activation of a separate pore complex (13), with recent evidence favoring the latter possibility (14).
In some cell types, activation of P2X7 receptors leads to the formation of membrane blebs (15)(16)(17)(18)(19). Zeiosis, the dynamic protrusion and retraction of blebs (20) can occur during mitosis, motility, and apoptosis in various cell types. The mechanism underlying membrane blebbing is thought to involve actomyosin contraction, but the initiating events are poorly understood. Interestingly, cell shape and tension generated by actomyosin have been shown to regulate osteoblast differentiation from human mesenchymal stem cells (21).
In this study, we report that P2X7 receptors signal through phospholipase D (PLD) 4 and A 2 (PLA 2 ). The resulting lysophosphatidic acid (LPA) then acts on its receptors on osteoblasts to cause dynamic membrane blebbing via a pathway dependent on Rho-associated kinase. The P2X7-LPA axis may be of general significance for the formation of membrane blebs in other systems and for P2X7-mediated osteogenesis during skeletal development and mechanotransduction.
Animals-Generation of the P2rx7-deficient mouse by targeted disruption of the gene encoding this receptor was described (22). Mice were maintained in a mixed genetic background (129Ola ϫ C57Bl/6 ϫ DBA/2) by cross-breeding of P2rx7 ϩ/Ϫ mice (23). These studies were approved by the Council on Animal Care at the University of Western Ontario.
Osteoblast Isolation and Morphological Assessment-Calvarial cells were isolated from newborn rats and 5-7-day-old mice using sequential collagenase digestion as described (7) and cultured in ␣-MEM supplemented with 10% fetal bovine serum and 1% antibiotic solution. Cultures were characterized as osteoblast-enriched based on the following characteristics. Cells exhibited elevation of cAMP in response to parathyroid hormone and expressed transcripts for Runx2, Osterix, bone sialoprotein, and osteocalcin. Moreover, supplementation of cultures with ascorbic acid and ␤-glycerophosphate increased alkaline phosphatase activity and induced the formation of mineralized nodules (data not shown).
After 1 week, undifferentiated calvarial cell cultures were trypsinized, plated at density of 1.5 ϫ 10 4 cells/cm 2 , and further cultured for up to 3 days. To perform time-lapse recording, the culture medium was removed and replaced with nominally divalent cation-free buffer consisting of the following (in mM): NaCl 140, KCl 5, NaHEPES 20, and glucose 10, pH 7.35 Ϯ 0.02, 290 Ϯ 5 mosmol/liter. In some experiments, cells were bathed in low Ca 2ϩ buffer (containing 0.5 mM CaCl 2 ) or in HEPESbuffered MEM (containing 1.8 mM Ca 2ϩ and 0.8 mM Mg 2ϩ ). Cells were placed on a heater stage and maintained at ϳ35°C. Time-lapse experiments were carried out using either an inverted phase contrast microscope (Nikon Plan-Fluor ϫ20 objective, 0.45 numerical aperture) or a Hoffman modulation contrast microscope (Zeiss Plan-NeoFluor ϫ20 objective, 0.4 numerical aperture or ϫ40 objective, 0.75 numerical aperture). Brightness, contrast, and gamma adjustments were applied to the images shown in Figs. 1, 2, and 5 using Adobe Photoshop and CorelDraw. For quantitative analyses, the percentages of cells in each field exhibiting zeiosis before and after addition of test substances were determined. Fields obtained using ϫ20 objectives contained a total of 22 Ϯ 5 cells (mean Ϯ S.D.), and three fields were analyzed from each cell preparation. Experiments were performed on at least three independent preparations (total numbers of cells analyzed for each condition were 198 Ϯ 14, mean Ϯ S.D.).
FM4-64 Loading and Confocal Microscopy-Osteoblasts were trypsinized and plated on 35-mm glass bottom culture dishes for 1-3 days. Membranes were labeled by incubation with FM4-64 (2 g/ml) in medium for 10 min at 35-37°C. Medium was then replaced with nominally divalent cation-free buffer, and cells were observed by confocal microscopy at ϳ30°C (Zeiss model LSM 510, Jena, Germany). Images were acquired using Zeiss Plan-Apochromat ϫ63 objective (1.4 numerical aperture) and 488 nm Ar ϩ ion laser excitation. The emission was filtered at 560 nm long pass, and images were captured using time-lapse mode. In some experiments, cells were scanned in multiple focal planes parallel to the substrate to create a z-stack.
Extraction of Lipids and Thin Layer Chromatography-Primary cultures of rat calvarial cells were trypsinized and plated in supplemented ␣-MEM at a density of 3 ϫ 10 4 cells/cm 2 . After 1 day, cultures were labeled with 1 Ci/ml [ 14 C]glycerol in serum-free ␣-MEM for 18 h. Medium containing unincorporated radiolabel was then replaced with nominally divalent cation-free buffer at 37°C containing inhibitors or vehicle. After 10 min, nucleotide was added. After a further 20 min, buffer was removed, and lipids were extracted from the cell layer by the method of Bligh and Dyer (24). Extracts were dried under a stream of N 2 and dissolved in chloroform together with unlabeled reference membrane lipids. Silica gel plates (LK4D, Whatman Inc., Florham, NJ) were sprayed with a solution of magnesium acetate (0.1 M) and oxalic acid (0.2 M). After the plates were dried, samples were loaded and separated by thin layer chromatography using the following solvent system (by volume) 34 of silica gel defined by reference lipids were collected, and radioactivity was quantified by liquid scintillation counting. Fluorescence Measurement of Cytosolic Free Calcium Concentration ([Ca 2ϩ ] i )-Cells were loaded by incubation with fura-2-AM (2 M) in culture medium without serum for 30 min at 37°C. Cultures then were washed, superfused continuously with nominally divalent cation-free buffer, and examined using a Nikon Diaphot inverted microscope (Nikon Fluor ϫ40 objective, 1.3 numerical aperture). The fluorescence emission at 510 nm with alternate excitation wavelengths of 345 and 380 nm was measured from individual cells using a Deltascan illumination system (Photon Technology International, London, Ontario, Canada). Fluorescence intensities at 345 and 380 nm were corrected for background by subtraction of values obtained from a region with no cells and the ratio used to calculate [Ca 2ϩ ] i . Test substances were dissolved in nominally divalent cation-free buffer and superfused onto single osteoblasts using pressure ejection from micropipettes positioned 30 -50 m from the cell (Picospritzer II, General Valve Corp., Fairfield, NJ).
Pore Formation Assay-To assess pore formation, we measured uptake of ethidium bromide as described (7). Calvarial cells were incubated in nominally divalent cation-free buffer for 10 min at 37°C with inhibitors or vehicle. BzATP or its vehicle was then added for 12 min together with ethidium bromide (20 g/ml). For some experiments, buffer was then replaced with fresh buffer containing Hoechst 33342 (5 g/ml) for 5 min to reveal nuclei. Cells were then washed and examined using an Axiovert S100 epifluorescence microscope (Zeiss). Other samples were washed and images acquired by confocal microscopy using a Zeiss Plan-Apochromat ϫ25 objective (0.8 numerical aperture) and 543 nm HeNe laser excitation. The emission was filtered at 560 nm long pass. Total numbers of cells from the same fields were counted using transmitted light images. Four fields were analyzed from each cell preparation, and experiments were performed on at least three independent preparations.
Statistical Analyses-Data are shown as means Ϯ S.D. or S.E., as indicated. Differences between two groups were assessed using t tests. Differences among three or more groups were evaluated by one-way or two-way analysis of variance (ANOVA) followed by Tukey or Bonferroni multiple comparison tests (respectively). Sigmoidal curves were fit by nonlinear regression using GraphPad Prism. Differences were accepted as statistically significant at p Ͻ 0.05.

The P2X7 Agonist BzATP Induces Dynamic and Reversible
Membrane Blebbing in Osteoblasts-Based on nucleotide-induced pore formation, we have shown previously that a subpopulation of calvarial cells expresses functional P2X7 receptors (7). Here we first monitored the morphology of murine osteoblasts using time-lapse Hoffman modulation contrast microscopy. Cultures were placed into a heater stage (ϳ35°C) and recorded using time-lapse for 10 min in nominally divalent cation-free buffer (Fig. 1, Before). Within this period, few cells exhibited membrane blebs (Fig. 1A, Before). Cultures were then treated with either 2Ј,3Ј-O-(4-benzoylbenzoyl)-ATP (BzATP, a relatively potent P2X7 agonist, 300 M) or vehicle (Control) for 20 min. BzATP caused dramatic changes in osteoblast morphology, including initial cellular retraction followed by the formation of dynamic plasma membrane blebs (zeiosis) with latency to onset of 4 -6 min (Fig. 1A, BzATP; see also supplemental Movie 1). Blebs enlarged and shrank in an asynchronous pattern with a mean lifetime of 1.2 Ϯ 0.3 min (n ϭ 8 blebs, three cells from three independent preparations). When BzATP was removed following 20 min, blebbing activity ceased within 5-10 min and cells respread (Fig. 1A, Wash), indicating reversibility.
We quantified the percentage of cells exhibiting zeiosis before and during the 20-min treatment interval. Although vehicle alone did not induce a significant change in the percentage of blebbing cells, BzATP treatment induced dynamic membrane blebbing in ϳ40% of cells in this series of experiments (Fig. 1B). Similar responses were observed in rat calvarial cells, indicating that blebbing is not restricted to murine osteoblasts.
Because membrane blebbing can be associated with apoptosis, we quantified the reversibility of BzATP-induced blebbing in murine osteoblasts. In this series of experiments, BzATP (300 M, 20 min) induced blebbing in 40% of cells. Within 60 min following removal of BzATP, only 9% of cells exhibited blebbing ( Fig. 1C), which was not significantly different from pretreatment values. Thus, blebbing induced in response to short term treatment with BzATP is fully reversible, ruling out the possibility that BzATP induces acute cell death.
We also examined the ability of BzATP to induce blebbing in the presence of physiological concentrations of extracellular Ca 2ϩ and Mg 2ϩ . Under these conditions, BzATP induced retraction and blebbing of murine osteoblasts (Fig. 1D). A similar percentage of cells exhibited blebbing in the presence of divalent cations (Fig. 1E) as in nominally divalent cation-free buffer (Fig. 1B). However, it should be noted that a higher concentration of BzATP (1 mM) was required to induce this response in the presence of Ca 2ϩ and Mg 2ϩ . Subsequent studies were performed using nominally divalent cation-free or low Ca 2ϩ buffers.
Dynamic changes in morphology of individual cells were characterized using confocal microscopy. Murine osteoblasts were first loaded with the fluorescent dye FM4-64 to stain membranes. Cells were incubated in nominally divalent cationfree buffer and exposed to BzATP (300 M). BzATP induced the formation of multiple dynamic blebs (Fig. 1F). Blebs enlarged and shrunk in an asynchronous pattern with maximum diameters up to ϳ8 m. Observation of a series of optical sections parallel to the substratum after a 10-min stimulation with BzATP revealed that blebbing occurred over the entire cell surface and that blebs did not contain FM4-64-stained intracellular membranes (Fig. 1G).
P2X7 Activation Is Required for Nucleotide-induced Blebbing in Osteoblasts-BzATP can activate a number of P2 nucleotide receptors besides P2X7 (25). To elucidate which subtype(s) of P2 receptors cause blebbing in osteoblasts, we tested the effects of a variety of P2X and P2Y agonists in murine and rat osteoblasts. Before stimulation with nucleotides, few murine cells exhibited membrane blebs ( Fig. 2A, Before). UTP activates several P2Y receptors but does not activate any P2X receptors (3). FEBRUARY 2, 2007 • VOLUME 282 • NUMBER 5
UDP (100 M), which activates P2Y6 receptors, did not initiate blebbing (Fig. 2B). Similarly, 2-methylthio-ADP (a potent and selective P2Y1 receptor agonist) failed to induce blebbing, even at concentrations up to 3 mM (Fig. 2C). Concentrations of ATP (Յ100 M) that are sufficient to activate many P2X and P2Y receptors (but not P2X7) did not induce blebbing, whereas higher concentrations (Ն300 M) of ATP did. Consistent with the involvement of P2X7, BzATP was more potent than ATP ( Fig. 2C; EC 50 was 90 M for BzATP versus 310 M for ATP).
To determine unambiguously whether P2X7 receptors mediate membrane blebbing in osteoblasts, we compared responses of calvarial cells isolated from wild-type (WT) and P2X7 knock-out (KO) mice. There was no difference in the appearance of WT and P2X7 KO osteoblasts before stimulation with BzATP (Fig. 2D, Before). Notably, BzATP (300 M) induced robust membrane blebbing only in WT osteoblasts (Fig. 2D, BzATP). Approximately 40% of WT cells responded to BzATP with membrane blebbing, whereas there was no significant effect in P2X7 KO cells (Fig. 2E). Taken together, we conclude that nucleotide-induced blebbing in osteoblasts is induced specifically by activation of P2X7 receptors.
Effects of Phospholipase Inhibitors on P2X7-mediated Membrane Blebbing-It is poorly understood how activation of P2X7 receptors ultimately leads to membrane blebbing. P2X7 activation stimulates PLD in several cell types, including the human THP-1 monocytic cell line (26) and submandibular gland ductal cells (27). We investigated the possible role of PLD in P2X7-induced blebbing. Primary alcohols can substitute for water in the hydrolytic reaction catalyzed by mammalian PLDs. Thus, the primary alcohol 1-butanol inhibits the formation of phosphatidic acid, whereas the related tertiary alcohol tert-butanol does not affect PLD activity (28). In osteoblasts, 1-butanol (1% v/v) significantly suppressed blebbing induced by BzATP (Fig. 3A); in contrast, tert-butanol at the same concentration had no effect. These data suggest that P2X7 receptor-mediated blebbing is dependent on PLD activity.
We next assessed the possible role of PLA 2 in P2X7-mediated blebbing. Two isoforms of cytosolic PLA 2 , calcium- In some experiments, BzATP was washed out after 20 min, and cells were allowed to recover in supplemented culture medium (without HCO 3 Ϫ ) for an additional 60 min, during which time blebbing ceased and cells respread (Wash, right panel). B, percentage of cells exhibiting dynamic blebbing was determined from time-lapse movies obtained before and during 20 min of treatment with vehicle (Control) or BzATP (300 M). * indicates significant difference compared with Before (p Ͻ 0.05). Data are means Ϯ S.E. (n ϭ 4 independent preparations, analyzed by two-way ANOVA). C, reversibility was quantified in a separate series of experiments. Cells were treated with BzATP (300 M) for 20 min and then BzATP was washed out as described above. The percentage of cells exhibiting dynamic blebbing was assessed during 20 min of treatment with BzATP and 40 -60 min following its removal. * indicates significant difference for BzATP compared with Wash (p Ͻ 0.05). Data are means Ϯ S.E. (n ϭ 3 independent preparations, analyzed by t test). D, to assess blebbing in the presence of physiological concentrations of divalent cations, osteoblasts were trypsinized, replated in serum-free medium containing 1.8 mM Ca 2ϩ and 0.8 mM Mg 2ϩ , and allowed to recover for 1 h. Cells were then monitored by time-lapse phase contrast microscopy at ϳ35°C for 10 min (Before). Addition of BzATP (1 mM) caused retraction and membrane blebbing (BzATP). Arrows indicate blebs. E, the percentage of cells exhibiting dynamic blebbing in the presence of divalent cations was determined from time-lapse movies obtained before and during 20 min of treatment with BzATP (1 mM). * indicates significant difference compared with Before (p Ͻ 0.05). Data are means Ϯ S.E. (n ϭ 3 independent preparations, analyzed by t test). F, to examine bleb morphology, osteoblasts were loaded with FM4-64 to stain membranes and observed using confocal microscopy. Cells were bathed in nominally divalent cation-free buffer at ϳ30°C. BzATP (300 M) was added at time 0. Images show 1-m-thick optical sections 10 m above the substratum through a single cell at 7-min intervals (n ϭ nucleus; blue arrowhead ϭ enlarging bleb; yellow arrowhead ϭ shrinking bleb). Images are representative of seven cells demonstrating dynamic membrane blebbing from four independent preparations. G, another osteoblast stained with FM4-64 was scanned in multiple focal planes parallel to the substrate following 10 min of stimulation with BzATP. Sections are 1 m thick, and distances above the substratum are indicated. Images are representative of four cells from three preparations. Scale bars are 10 m for all panels.
Effect of BzATP on Lipid Metabolism in Osteoblasts-To assay PLD activity, we exploited the unique ability of this enzyme to catalyze a transphosphatidylation reaction that, in the presence of glycerol, yields phosphatidylglycerol (31). Rat calvarial cell cultures were labeled with [ 14 C]glycerol and then challenged with BzATP (300 M) or its vehicle (Control) for 20 min. BzATP induced a significant increase in levels of phosphatidylglycerol (Fig. 4A). 1-Butanol (1% v/v) completely inhibited the production of phosphatidylglycerol induced by BzATP, whereas the PLA 2 inhibitor BEL (10 M) had no significant effect. Taken together, these data establish that BzATP activates PLD in these cells.
Interestingly, BzATP also stimulated production of lysophosphatidylglycerol (Fig. 4B). This effect was abolished by 1-butanol (1% v/v) as well as by BEL (10 M). These data indicate that BzATP activates both PLD and PLA 2 , which in turn leads to the production of phosphatidic acid and LPA.
LPA Induces Dynamic Membrane Blebbing in Osteoblasts-We next considered whether LPA itself could induce mem-  brane blebbing. LPA binds to specific G protein-coupled receptors in a variety of cell types, including rat and human osteoblasts (32,33). In both WT and P2X7 KO cells, LPA (10 M) induced zeiosis similar to that caused by BzATP in WT cells ( Fig. 5A; see also supplemental Movie 2). Blebs enlarged and shrunk in an asynchronous pattern with a mean lifetime of 1.7 Ϯ 0.3 min (n ϭ 8 blebs, three cells from three independent preparations). Approximately 35% of WT and P2X7 KO cells exhibited blebbing during the 20-min treatment (Fig. 5, B-D). Blebbing of P2X7 KO osteoblasts excludes the possibility that zeiosis arises from LPA-induced release of endogenous ATP leading to activation of P2X7 receptors.
It is possible that BzATP and LPA induce blebbing by acting through parallel pathways, both of which are dependent on PLD and PLA 2 activity. Therefore, we investigated whether LPA-induced blebbing was sensitive to inhibition of these phospholipases. Although 1-butanol and BEL inhibited BzATP-induced blebbing, they had no significant effect on blebbing caused by LPA (Fig. 5C). To further examine the relationship between P2X7 and LPA receptor signaling, we investigated the time course for onset of blebbing induced by LPA and BzATP in WT osteoblasts. The time to half-maximal blebbing was 3.0 Ϯ 0.4 min following addition of LPA (10 M) but at 6.5 Ϯ 0.3 min following BzATP (300 M) (Fig. 5D). Although a significantly greater percentage of cells exhibited blebbing in response to LPA than BzATP at 4 and 6 min, there was no significant difference in the maximal response (at 8 -10 min, when ϳ45% of cells exhibited blebbing). Taken together, these data support the view that LPA acts downstream of P2X7 receptor activation.
Effect of LPA Receptor Antagonist on P2X7 Receptor-induced Membrane Blebbing-Next, we directly addressed whether LPA receptor activation functions downstream of P2X7 receptors. We used VPC-32183, a selective antagonist of the LPA1 and LPA3 receptors (34). As expected, VPC-32183 (100 nM) blocked the ability of LPA to induce blebbing (Fig. 6A). Moreover, VPC-32183 suppressed blebbing induced by activation of the P2X7 receptor, reinforcing the involvement of LPA in this process. To address the possibility that VPC-32183 blocks P2X7 receptor function, we monitored BzATP-induced pore formation. Cultures were incubated in nominally divalent cation-free buffer containing ethidium bromide (a normally impermeant fluorescent probe) and BzATP (300 M). There was no difference in the percentage of cells that accumulated ethidium bromide (an index of pore formation) in the presence and absence of VPC-32183 (Fig. 6B). These observations rule out the possibility that VPC-32183 directly affects P2X7 receptor function.

LPA Receptor Desensitization Suppresses P2X7 Receptor-induced Membrane
Blebbing-To further address the involvement of LPA in P2X7 receptor-induced blebbing, we exploited the behavior of LPA receptors that desensitize, like many other G protein-coupled receptors, in response to prolonged application of agonist (35). Cultures of WT osteoblasts were incubated with LPA (100 M) for 15-18 h (control cells were incubated for the same period with vehicle). To confirm desensitization, we monitored changes in [Ca 2ϩ ] i in response to a brief application of LPA (10 M). LPA caused transient elevation of [Ca 2ϩ ] i in control cells (Fig. 6C, left), but prolonged incubation with LPA virtually abolished this response (Fig. 6C, right). LPA-induced Ca 2ϩ responses were quantified as the area under the [Ca 2ϩ ] i time trace (above base line). This analysis revealed that Ca 2ϩ responses in cells preincubated with LPA were reduced by 85% (Fig. 6D), indicating profound desensitization.
As expected, LPA failed to induce blebbing in desensitized cells (Fig. 6E). Importantly, P2X7 receptor-induced blebbing was also suppressed in desensitized cells. To address the possibility that LPA pretreatment causes heterologous desensitization of P2X7 receptors, we monitored BzATP-induced pore formation. There was no difference in the accumulation of ethidium bromide in desensitized and control cells (Fig. 6F). Taken together, the data presented in Figs. 4 -6 establish that P2X7 receptor activation induces downstream signaling through the LPA receptor that in turn leads to membrane blebbing.

DISCUSSION
The P2X7 receptor has a critical role in the regulation of bone formation and skeletal mechanotransduction (7,8). In this study, we examined the signaling events triggered by activation   FEBRUARY 2, 2007 • VOLUME 282 • NUMBER 5 of endogenous P2X7 receptors in osteoblasts. We show for the first time the involvement of PLD, PLA 2 , and the potent bioactive lipid LPA in P2X7-induced blebbing (Fig. 8). The P2X7-LPA axis may underlie the formation of membrane blebs in other cell types. Moreover, in osteoblasts, this pathway may enhance periosteal bone formation and mediate responses to mechanical stimulation.

P2X7-induced Blebbing in Osteoblasts
P2X7 Receptors Mediate Membrane Blebbing in Osteoblasts-ATP or BzATP induce blebbing in several cell types, including hepatocytes (39), thymocytes (40), and macrophage cell lines (16,18), presumably mediated by P2X7 receptors. However, ATP or BzATP can activate a number of P2 receptors besides P2X7 (25). Several characteristics of the blebbing response in osteoblasts are consistent with the involvement of P2X7 receptors. These include the relatively high concentration of ATP required to induce the response, the greater potency of BzATP compared with ATP, and the sensitivity of the response to divalent cations. Gain-of-function approaches involving heterologous expression of P2X7 receptors in HEK-293 cells have provided direct evidence for their role in blebbing (17,19). We are the first to use a loss-of-function approach to assess the role of P2X7 receptors in membrane blebbing.
Previous reports have provided conflicting data on the expression of P2X7 receptors in cells of the osteoblast lineage. Using immunocytochemical techniques, P2X7 receptors were not detected in sections of undecalcified, unfixed bone (9). However, P2X7 receptors have been identified in subpopulations of human and murine osteoblasts based on RT-PCR and nucleotide-induced pore formation (7,10). This study provides compelling functional evidence for expression of P2X7 receptors in cells of the osteoblast lineage.
Apoptosis is accompanied by zeiosis together with cell shrinkage, nuclear fragmentation, and formation of apoptotic bodies (36). Membrane blebbing has been reported in human osteoblasts exposed to BzATP for 40 min, a process that was suggested to be associated with apoptotic cell death (10). However, we found that P2X7-mediated blebbing was fully reversible upon removal of BzATP, and cells respread with extension of pseudopods and peripheral ruffling. These observations distinguish P2X7-induced blebbing in osteoblasts from apoptotic blebbing.
Blebbing of authentic osteoblasts induced by activation of endogenous P2X7 receptors resembles that described in HEK-293 cells heterologously expressing P2X7 receptors (17,19). In osteoblasts, half-maximal blebbing was observed 6 -7 min following addition of BzATP. This is considerably slower than in HEK-293 cells stably expressing P2X7 receptors, where halfmaximal activation is observed in less than 30 s (19). The timeto-initial onset of blebbing in HEK-293 cells was strongly dependent on agonist concentration, consistent with it being determined by the number of activated receptors. Thus, discrepancy in the kinetics of onset may reflect differences in P2X7 expression levels in native cells compared with transfected HEK-293 cells.
Novel Roles for Phospholipases in Mediating P2X7 Receptorinduced Membrane Blebbing-Early studies revealed that BzATP acts through P2X7 receptors in a macrophage-like cell line to rapidly activate PLD (41). In these cells, PLD is activated through a pathway that is not dependent on extracellular Ca 2ϩ , and PLD cannot be activated simply by elevation of [Ca 2ϩ ] i . Activation of P2X7 receptors has also been shown to stimulate PLD activity in lymphocytes (42), astrocytic cells (43), submandibular gland ductal cells (27), and thymocytes (44), although dependence on divalent cations is variable. In this study, we report that BzATP activates PLD in osteoblasts and that inhibition of PLD suppresses P2X7 receptor-induced blebbing, implicating lipid signaling in zeiosis.
Compared with PLD, less is known about the activation of PLA 2 by P2X7 receptors. High concentrations of ATP or BzATP stimulate both cPLA 2 and iPLA 2 in ductal cells of rat submandibular gland (29). BEL (a selective inhibitor of iPLA 2 ) blocked both PLA 2 activation and blebbing induced by BzATP in osteoblasts. In this regard, BzATP-induced blebbing was still observed when osteoblasts were bathed in the absence of extracellular Ca 2ϩ or when cytosolic Ca 2ϩ was chelated (data not shown). Taken together, these findings are consistent with the involvement of iPLA 2 in the pathway leading to membrane blebbing in osteoblasts.
LPA can be produced by the actions of PLD and PLA 2 (Fig. 8). Alternatively, LPA can be synthesized by PLA 2 acting on phosphatidylcholine to produce lysophosphatidylcholine, which is then hydrolyzed by the extracellular enzyme autotaxin (lysophospholipase D) (35). However, because autotaxin is not inhibited by 1-butanol (45), it is unlikely that it plays a significant role in BzATP-induced production of LPA by osteoblasts. In this study, 1-butanol abolished the synthesis of both lysophosphatidylglycerol and LPA triggered by BzATP (Fig. 4B and data not shown).
Role for LPA in Mediating P2X7 Receptor-induced Membrane Blebbing-We explored the dependence of P2X7-induced blebbing on LPA signaling using two approaches. Blockade or desensitization of LPA receptors suppressed the ability of both LPA and BzATP to induce blebbing, without affecting P2X7stimulated pore formation. 5 Thus, BzATP-induced blebbing is FIGURE 8. Proposed mechanism of P2X7-induced blebbing in osteoblasts. ATP or BzATP acts through the P2X7 receptor on osteoblasts to stimulate PLD and PLA 2 . These phospholipases catalyze the conversion of glycerophospholipid (PL) to phosphatidic acid (PA) and LPA. LPA acts on its receptor (LPAR) to stimulate Rho-associated kinase, which in turn causes blebbing.

P2X7-induced Blebbing in Osteoblasts
mediated by LPA acting through its endogenous receptors on osteoblasts. There are four mammalian LPA receptors (LPA1-4), all of which couple to heterotrimeric G proteins. LPA1 is widely distributed, LPA2 and LPA3 have more restricted distributions, and LPA4 is less well characterized and expressed at low levels (46). Several lines of evidence indicate that LPA-induced blebbing in osteoblasts arises from activation of LPA1. RT-PCR has revealed expression of LPA1 in rat osteoblasts (32) and both LPA1 and LPA2 in primary human osteoblastic cells (33). Moreover, LPA1 is the predominant receptor subtype present in murine MC3T3-E1 osteoblastic cells, with LPA3 transcripts barely detectable by real time RT-PCR (47).
In keeping with a role for LPA1 in regulating osteoblast function, mice deficient in LPA1 exhibit craniofacial dysmorphism attributed to abnormal development of the facial bones (48). In contrast, LPA2 KO mice displayed no obvious phenotypic abnormalities (49). Thus, LPA1 appears to be the major LPA receptor in osteoblasts. Consistent with this possibility, we have shown that membrane blebbing induced by LPA and BzATP was suppressed by the selective LPA1/3 receptor antagonist VPC-32183. All LPA receptor subtypes are coupled through G q and/or G i to activation of phospholipase C and elevation of [Ca 2ϩ ] i . LPA1, but not LPA3, also couples through G 12/13 to Rho activation, which induces morphological changes in many cell types (35,46). Taken together, the evidence suggests that LPA1 mediates membrane blebbing in osteoblasts.
We found that blebbing induced by both BzATP and LPA is sensitive to Y-27632, an inhibitor of Rho-associated kinase. It has been shown previously that, at the concentration used in our study, Y-27632 inhibits Rho-associated kinase with little effect on other kinases, including protein kinase C, cAMP-dependent protein kinase, and myosin light chain kinase (38,50). P2X7 receptors induce zeiosis through a pathway dependent on the activity of Rho-associated kinase in other cell types (16 -19). Thus, it is conceivable that LPA signaling through Rho-associated kinase mediates blebbing induced by P2X7 activation in cell types in addition to osteoblasts. Rho-associated kinase has been shown in other systems to cause phosphorylation of myosin regulatory light chain, which in turn controls actomyosin filament assembly and contraction (36,37). It has been suggested that contraction of cortical actomyosin filaments causes their focal detachment from the plasma membrane and an increase in cytoplasmic pressure, both of which lead to the formation of blebs.
Possible Roles of P2X7-activated Signaling in Osteoblast Regulation and Function-ATP is released in response to mechanical strain, inflammatory stimuli, and cell damage (51). Because cytosolic concentrations of ATP are ϳ5 mM, it is conceivable that ATP release into confined extracellular compartments can result in levels sufficient to activate P2X7 receptors. Furthermore, an alternative pathway for P2X7 activation has been reported. This pathway involves a cell surface ADP-ribosyltransferase, which (upon exposure to NAD ϩ ) ADP-ribosylates the P2X7 receptor, leading to its irreversible activation (52). Regardless of its mechanism of activation, the P2X7 receptor plays important roles in vivo as revealed by deficits in P2X7 KO mice. These deficits include impaired inflammatory responses (53), decreased periosteal bone formation (7), and insensitivity of the skeleton to mechanical stimulation (8).
The phenotype of the P2X7 KO mouse indicates an important role for this receptor in the regulation and function of osteoblasts. Osteoblast blebs have been observed during the healing of experimental bone defects in vivo (54); however, the function of this phenomenon is poorly understood. Blebbing reflects vigorous actomyosin contraction that may have a role in cell migration or the assembly of extracellular matrix by osteoblasts.
In this study, we used blebbing as an end point to identify a subset of signaling pathways activated by P2X7 receptors in osteoblasts. LPA has been shown previously to have mitogenic and anti-apoptotic effects on rat osteoblasts (32,55). Mitogenic effects of LPA have been confirmed in human osteoblastic cells (33), and it has been reported recently that LPA enhances the motility of a murine osteoblast-like cell line (47). Moreover, Rho and Rho-associated kinase (downstream effectors of LPA1) have been shown to drive mesenchymal stem cells to differentiate into osteoblasts (21). Thus, P2X7-LPA axis may serve a critical role in the regulation of bone formation.
LPA can also act as a mitogen and motility factor for certain tumor cells (46). In this regard, LPA has been implicated recently in the metastasis of breast and ovarian cancers to bone (56). Thus, local production of LPA by osteoblasts may contribute to tumor metastasis and progression in the skeleton.
In addition to its contribution to the production of LPA, PLA 2 produces arachidonic acid, which in turn serves as a substrate for cyclooxygenases leading to the production of eicosanoids. Several eicosanoids, including prostaglandin E 2 , stimulate bone formation (57) and have been implicated in skeletal mechanotransduction (58). In this regard, fluid shear stress increases PGE 2 release by calvarial osteoblasts from WT, but not P2X7 KO mice (8). Interestingly, inhibition of prostaglandin synthesis using the cyclooxygenase inhibitor indomethacin prevents bone adaptation in response to mechanical strain in vivo (58).
In summary, we propose that mechanical stimuli induce release of ATP that then acts through P2X7 receptors on osteoblasts, leading to production of prostaglandins and LPA. These lipid mediators then act in an autocrine or paracrine manner to enhance bone formation, explaining the role of P2X7 receptors in skeletal development and mechanotransduction.