Bone 60 (2014) 112–121

Contents lists available at ScienceDirect

Bone journal homepage: www.elsevier.com/locate/bone

Original Full Length Article

Prostaglandin D2 induces apoptosis of human osteoclasts through ERK1/2 and Akt signaling pathways☆ Li Yue a,b, Sonia Haroun b, Jean-Luc Parent a,b, Artur J. de Brum-Fernandes a,b,⁎ a b

Department of Pharmacology, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, 3001 12e Avenue Nord, Sherbrooke, Quebec J1H 5N4, Canada Division of Rheumatology, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, 3001 12e Avenue Nord, Sherbrooke, Quebec J1H 5N4, Canada

a r t i c l e

i n f o

Article history: Received 11 July 2013 Revised 6 December 2013 Accepted 9 December 2013 Available online 15 December 2013 Edited by: Hong-Hee Kim Keywords: Osteoclasts Apoptosis Prostaglandin D2 Akt ERK1/2 β-Arrestin

a b s t r a c t In a recent study we have shown that prostaglandin D2 (PGD2) induces human osteoclast (OC) apoptosis through the activation of the chemoattractant receptor homologous molecule expressed on T-helper type 2 cell (CRTH2) receptor and the intrinsic apoptotic pathway. However, the molecular mechanisms underlying this response remain elusive. The objective of this study is to investigate the intracellular signaling pathways mediating PGD2-induced OC apoptosis. OCs were generated by in vitro differentiation of human peripheral blood mononuclear cells (PBMCs), and then treated with or without the selective inhibitors of mitogen-activated protein kinase-extracellular signal-regulated kinase (ERK) kinase, (MEK)-1/2, phosphatidylinositol3-kinase (PI3K) and NF-κB/IκB kinase-2 (IKK2) prior to the treatments of PGD2 as well as its agonists and antagonists. Fluorogenic substrate assay and immunoblotting were performed to determine the caspase-3 activity and key proteins involved in Akt, ERK1/2 and NF-κB signaling pathways. Treatments with both PGD2 and a CRTH2 agonist decreased ERK1/2 (Thr202/Tyr204) and Akt (Ser473) phosphorylation, whereas both treatments increased β-arrestin-1 phosphorylation (Ser412) in the presence of naproxen, which was used to eliminate endogenous prostaglandin production. In the absence of naproxen, treatment with a CRTH2 antagonist increased both ERK1/2 and Akt phosphorylations, and reduced the phosphorylation of β-arrestin-1. Treatment of OCs with a selective MEK-1/2 inhibitor increased caspase-3 activity and OC apoptosis induced by both PGD2 and a CRTH2 agonist. Moreover, a CRTH2 antagonist diminished the selective MEK-1/2 inhibitor-induced increase in caspase-3 activity in the presence of endogenous prostaglandins. In addition, treatment of OCs with a selective PI3K inhibitor decreased ERK1/ 2 (Thr202/Tyr204) phosphorylation caused by PGD2, whereas increased ERK1/2 (Thr202/Tyr204) phosphorylation by a CRTH2 antagonist was attenuated with a PI3K inhibitor treatment. The DP receptor was not implicated in any of the parameters evaluated. Treatment of OCs with PGD2 as well as its receptor agonists and antagonists did not alter the phosphorylation of RelA/p65 (Ser536). Moreover, the caspase-3 activity was not altered in OCs treated with a selective IKK2/NF-κB inhibitor. In conclusion, endogenous or exogenous PGD2 induces CRTH2dependent apoptosis in human differentiated OCs; β-arrestin-1, ERK1/2, and Akt, but not IKK2/NF-κB are probably implicated in the signaling pathways of this receptor in the model studied. © 2013 Elsevier Inc. All rights reserved.

Introduction

Abbreviations: CRTH2, chemoattractant receptor homologous molecule expressed on T-helper type 2 cells; DMSO, dimethyl sulfoxide; DP, D-Type prostanoid; ERK, extracellular signal-regulated kinase; FBS, fetal bovine serum; GPCR, G protein-coupled receptor; GRK5, GPCR kinase 5; IKK2, IκB kinase-2; IL, interleukin; M-CSF, macrophagecolony stimulating factor; MEK, mitogen-activated protein kinase-ERK kinase; OB, osteoblast; OC, osteoclast; PBMCs, peripheral blood mononuclear cells; PGD2, prostaglandin D2; PI3K, phosphatidylinositol3-kinase; RANKL, receptor activator for nuclear factor κB ligand; TRAP, tartrate-resistant acid phosphatase. ☆ This work was presented in part at the 34th ASBMR Annual Meeting (Minneapolis, MN, USA, Oct. 12–15, 2012). ⁎ Corresponding author at: Division of Rheumatology, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, 3001 12e Avenue Nord, local 3858, Sherbrooke, Quebec J1H 5N4, Canada. Fax: +1 819 564 5265. E-mail addresses: [email protected] (L. Yue), [email protected] (S. Haroun), [email protected] (J.-L. Parent), [email protected] (A.J. de Brum-Fernandes). 8756-3282/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.bone.2013.12.011

Bone homeostasis depends on balanced bone resorption by osteoclasts (OCs) and bone formation by osteoblasts (OBs) [1]. Excessive bone loss is a pathologic mechanism in several local and systemic diseases including osteoporosis, periodontitis, rheumatoid arthritis and bone metastatic tumor-driven osteolysis. These different diseases show an increase in bone resorption by OCs. This increased bone resorption may be due to increases in the differentiation and activation or to changes in cell death and apoptosis. OC formation, differentiation and activation are regulated by numerous cytokines, growth factors and hormones, such as macrophage-colony stimulating factor (M-CSF) [2], receptor activator for nuclear factor κB ligand (RANKL) [3] and tumor necrosis factor-α [4]. Among the factors essential for OC differentiation and activation, interleukin (IL)-1 and IL-6 directly [4,5] promote OC generation, whereas estrogen, transforming growth factor-β,

L. Yue et al. / Bone 60 (2014) 112–121

interferon-γ, IL-4, IL-12 alone or in synergy with IL-18 inhibit OC survival [6–10]. Differentiated OCs have a short life span, and several signaling pathways are implicated in the regulation of OC apoptosis, including Akt, extracellular signal-regulated kinase (ERK)1/2 and NF-κB [11–16], which can interactively or independently regulate OC apoptosis [12,15–19]. Generally, ERK1/2, Akt and NF-κB signaling pathways promote OC survival and prevent stress-induced OC apoptosis [16–19]. However, there is a report showing that ERK1/2 inhibition by a selective MEK-1 inhibitor reduces OC apoptosis [15] and that IKK2/NF-κB signaling is not required for OC survival [20]. Prostaglandins, a group of lipids derived from arachidonic acid, possess both pro- and anti-apoptotic functions depending on the differentiation stage and type of cells studied [21–24]. PGD2 plays its effects primarily by binding two specific receptors, D-type prostanoid receptor (DP) and chemoattractant receptor homologous molecule expressed on T-helper type 2 cells (CRTH2) [25,26]. PGD2 and its metabolites induce apoptosis of eosinophils [23], neuronal cells [27], non-small cell lung carcinoma cells [28] and human leukemia cells [29]. Interestingly, PGD2 can also exhibit anti-apoptotic function in eosinophils [30] and in human Th2 cells [24]. We have recently shown that PGD2 induces human OC apoptosis through activation of the CRTH2 receptor and the intrinsic apoptosis pathway [31]. However, the roles played by ERK1/ 2, Akt and NF-κB signaling pathways for PGD2/CRTH2-induced human OC apoptosis have not been studied. The objective of this study is to investigate molecular signaling implicated in PGD2-induced OC apoptosis. Materials and methods Materials Fetal bovine serum (FBS) and penicillin–streptomycin were purchased from Gibco (distributed by Invitrogen Canada, Inc., Burlington, ON, Canada). RANKL and M-CSF were obtained from PeproTech, Inc. (Rocky Hill, NJ, USA). TACS Blue Label Kit was purchased from R&D Systems (Minneapolis, MN, USA). Ficoll-Paque PLUS was purchased from Amersham Biosciences part of GE Healthcare (Bjorkgatan, Uppsala, Sweden). The following compounds were purchased from Cayman Chemical (Ann. Arbor, MI, USA): PGD2, CRTH2 agonist DK-PGD2, DP agonist BW 245C, CRTH2 antagonist CAY10471, and DP antagonist BW A868C; they were all diluted in dimethyl sulfoxide (DMSO). LY294002 and caspase-3 fluorogenic substrate were purchased from Calbiochem (Merck, Germany). PHA-408 was synthesized at Pfizer (St Louis, MO, USA). U0126, Akt, phospho-Akt (Ser473), RelA/p65 and phospho-RelA/ p65 (Ser536) antibodies as well as secondary antibodies (anti-rabbit and anti-mouse) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). ERK1 and p-ERK1/2 (Thr202/Tyr204) antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Anti-β-arrestin-1 (phospho Ser412) antibody was purchased from Abcam Inc. (Cambridge, MA, USA). All other reagents were purchased from Sigma-Aldrich Canada, Ltd. (Oakville, ON, Canada). Cell culture Osteoclasts were differentiated from human peripheral blood mononuclear cells (PBMCs) isolated from healthy donors blood as described previously [31–33]. The use of human PBMCs was approved by the Ethics Review Board of the Faculté de médecine et des sciences de la santé of the Université de Sherbrooke. All subjects provided an informed consent. Briefly, mononucleated cells were isolated by dextran sedimentation and Ficoll density gradient centrifugation. The cells were plated at a density of 1.5 × 106 cells/cm2 in 96-well plates for OC apoptosis detection using TACS assay kit, 48-well plates for bone resorption, or 12-well plates for immunoblotting at 37 °C and 5% CO2 in humidified atmosphere with medium changes twice a week. The cells were incubated in αMEM medium with 10% FBS, 1% penicillin–streptomycin, M-CSF (10 ng/ml) and RANKL (50 ng/ml) for 21 days.

113

Cell treatments After 21 days of differentiation, the OCs were starved for 24 h in 2% FBS-containing medium without M-CSF or RANKL, and then pre-treated with or without a selective PI3K inhibitor (LY294002, 1 μM), a selective inhibitor of MEK-1/2 (U0126, 1 μM) or a selective IKK-2 inhibitor (PHA408, 2 μM) for 30 min prior to the treatments for 24 h with PGD2, PGD2 agonists (DP agonist: BW 245; CRTH2 agonist: DK-PGD2) in the presence of naproxen (10 μM), or with PGD2 antagonists (DP antagonist: BW A868C; CRTH2 antagonist: CAY10471) in the absence of 10 μM naproxen. Naproxen, an inhibitor of both COX-1 and COX-2 enzymes, was used when OCs were treated with PGD2 or selective DP and CRTH2 agonists to eliminate endogenous production of prostaglandins. On the contrary, naproxen was not used when OCs were treated with selective PGD2 receptor antagonists as these cells generated enough PGD2 to induce cellular apoptosis, as shown in our previous study [31]. Tartrate-resistant acid phosphatase (TRAP) staining TRAP staining was performed to identify the human OCs using a commercial kit (catalog number 387A-1KT, Sigma-Aldrich) according to the manufacturer's instructions [31]. Differentiated OCs were washed with PBS, and incubated with a solution of Naphthol AS-BI phosphoric acid and freshly diazotized Fast Garnet GBC in the presence of tartrate at 37 °C for 35 min. Multinucleated (three or more nuclei) TRAPpositive cells (dark red/purple) were identified under light microscopy (OLYMPUS CK2, Markham, ON, Canada). Microphotography was done with Leica DMIRE2 inverted fluorescent microscope (Leica microsystems, Wetzlar, Germany), Retiga EX camera (QImaging, Burnaby, British Columbia, Canada) and OpenLab 5.5.0 Imaging software (OpenLab, Lexington, MA). Resorption assay To confirm the bone resorption ability of human differentiated OCs, PBMCs were cultured for 21 days on devitalized bovine cortical bone slices in 48-well plates under the same culture conditions as described above [34]. The differentiated cells were then incubated for 10 more days at 37 °C in 10% CO2. The cells were re-fed twice a week with medium supplemented with M-CSF (10 ng/ml) and RANKL (50 ng/ml). After the 31-day culture, each well was washed with PBS, and bone slices were then stained with 0.2% toluidine blue for 5 min to detect the formation of lacunar resorption pits. The resorption pits (blue to purple) were photographed and quantified with the image analysis program Simple PCI from Compix Inc., Imaging Systems (Cranberry Township, PA, USA). Determination of OC apoptosis rate using TACS Blue Labeling TACS Blue Label Kit was used to detect OC apoptosis in 96-well plates following the manufacturer's protocol [7,31]. The OCs were fixed, washed and permeabilized with an ethanol:acetic acid (2:1) solution, and then biotinylated nucleotides were incorporated by terminal deoxynucleotidyl transferase. The biotinylated nucleotides were examined using streptavidin-horseradish peroxidase conjugate followed by the substrate, TACS Blue Label. An insoluble blue precipitate is generated in nuclei where DNA fragments underwent double-stranded breaks. Blue multinucleated (three or more nuclei) cells were counted as apoptotic OCs, and pink ones as alive OCs. The stained cells were counted manually in five fields using a light microscope. The apoptosis rate was evaluated as the percentage of blue OCs with three or more nuclei. Caspase-3 activity assay Caspase-3 activity was determined in cultured OCs in vitro using the caspase-3 fluorogenic substrate assay with F-2500 FL Spectrophotometer [31,35]. After treatment, differentiated OCs were washed twice with

114

L. Yue et al. / Bone 60 (2014) 112–121

PBS before being lysed for 20 min on ice in RIPA buffer (150 mM NaCl, 50 mM Tris–HCl, PH 8, 1% Igepal, 0.5% deoxycholate, 10 mM Na4PP, 0.1% SDS, and 0.5 mM EDTA) containing a protease inhibitor cocktail. OC lysate proteins (30 μg) were incubated with 2 μl of Ac-DEVD-AMC (caspase-3-like substrate, 5 mM) in a reaction buffer (100 mM HEPES, 20% glycerol, and 5 mM DTT) for 2 h at 37 °C. Caspase-3-like activity was determined at an excitation/emission wavelength pair of 380 nm/ 405–500 nm.

three times, and the data shown are representative of all results obtained. Comparison between groups was performed with the one-way ANOVA test followed by Dunnett post-test or Mann–Whitney test using GraphPad PRISM 6 software.

Western blot analysis

Human PBMCs were cultured with αMEM medium containing M-CSF (10 ng/ml) and RANKL (50 ng/ml) for 21 days. To confirm the osteoclastic nature of the multinucleated cells, they were TRAP stained. As shown in Fig. 1A, TRAP positive (the dark red/purple) and multinucleated cells were found on day 21, and TRAP-positive cells containing three or more nuclei were counted as OCs. These cells were capable of bone resorption when cultured for 31 days on bovine cortical bone as shown by the formation of resorption lacunae (blue to purple) (Fig. 1B). OCs differentiated from PBMCs in 96-well plates were starved in 2% FBS-containing medium without M-CSF or RANKL for 24 h, and then the cells were treated with PGD2 (10 nM) for 24 h. Fig. 1C showed TACS assay on non-starved OCs, and that 87.75% cells are alive OCs, while Fig. 1D showed TACS assay on OCs treated with 10 nM of PGD2 for 24 h. Under these circumstances the percentages of apoptotic OCs were 12.25 ± 1.49% and 36.25 ± 4.79% in the control and PGD2treated groups, respectively (p b 0.01, N = 4).

After removing M-CSF and RANKL, as well as reducing FBS to 2% in αMEM medium for 24 h prior to the different treatments, the treated OCs were washed twice with ice-cold PBS and then lysed for 20 min on ice in RIPA lysis buffer [31]. Total proteins (20 μg) were subjected to SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA). The membrane was blocked with 5% BSA, incubated overnight at 4 °C with the primary antibodies against Akt phosphorylation at Ser473, total Akt, ERK1/2 phosphorylation at Thr202/Tyr204, total ERK1, RelA/p65 phosphorylation at Ser536, total RelA/p65 (1:1000 dilution) and β-arrestin-1 phosphorylation at Ser412 (1:5000 dilution). The membranes were incubated for 1 h at room temperature with a horseradish peroxidase-conjugated secondary anti-mouse or anti-rabbit antibody. After washing, bound antibodies were visualized by an enhanced chemiluminescence solution. All bands were measured by densitometry and normalized to total Akt, total ERK1, total RelA/p65 or actin using ImageJ software.

Results Functional identification of in vitro differentiated human OCs

Inhibition of the ERK1/2 signaling during PGD2-induced OC apoptosis Statistical analyses All data were expressed as means ± standard error of the mean (SEM) of samples obtained from different donors and considered significant when p b 0.05 (*). Each experiment was performed at least

Accumulating evidence shows the regulation of ERK1/2 signaling in OC apoptosis [15,16,36]. We first determined whether PGD2/CRTH2induced OC apoptosis was associated with ERK1/2 activation as indicated by phosphorylation at Thr202/Tyr204. OCs were treated with PGD2

Fig. 1. Representative light micrographs of identifications for human differentiated OCs. Mononuclear cells isolated from human PBMCs were cultured in αMEM medium with 10% FBS, 1% penicillin–streptomycin, M-CSF and RANKL until 21 days. A, TRAP staining analysis of human differentiated OCs generated after 21 days of culture in the presence of M-CSF and RANKL. TRAP-positive cells containing three or more nuclei were considered as OCs. White arrows indicate multinuclear OCs (dark red/purple). White scale bar = 100 μm. Images are the representative of four independent experiments. B, representative photomicrograph of bone slices. When PBMCs were cultured for 31 days on devitalized bovine cortical bone slices, they formed resorption pits (N = 4); black scale bar = 100 μm. C and D, TACS assay of human differentiated OCs. TACS Blue Labeling was used to quantify the OC apoptosis. Multinucleated cells with pink nuclei (black arrow) were counted as alive whereas multinucleated cells with blue nuclei (white arrow) were counted as apoptotic. At the end of PBMC cultures, the differentiated OCs were starved for 24 h in 2% FBS-containing medium without M-CSF or RANKL, and then the cells were treated with PGD2 (10 nM) for 24 h. Photomicrographs of TACS assay on non-starved OCs (C) and on OCs further stimulated with PGD2 (10 nM) for 24 h (D). Scale bar = 100 μm.

L. Yue et al. / Bone 60 (2014) 112–121

as well as its agonist and antagonist, and the phosphorylation of ERK1/ 2 at Thr202/Tyr204 was measured using Western blot. In the presence of naproxen, treatment with PGD2 (10 nM) or with the CRTH2 agonist DK-PGD2 (10 nM) attenuated ERK1/2 phosphorylation at Thr202/ Tyr204 in human OCs as compared to the vehicle-treated control (Figs. 2A and B). In the absence of naproxen, treatment with a selective CRTH2 antagonist CAY10471 (10 nM) caused an increase in ERK1/2 phosphorylation as compared to the control (Figs. 2C and D). In addition, reduced ERK1/2 phosphorylation was associated with increased activity of caspase-3 in OCs (Figs. 2E and F). Incubation with a DP agonist (BW 245C, 10 nM) or antagonist (BW A868C, 10 nM), in the presence and absence of naproxen, respectively, did not significantly change ERK1/2 phosphorylation (Figs. 2A–D). To further investigate the role of ERK1/2 signaling pathway in the PGD2/CRTH2-induced human OC apoptosis, differentiated OCs after 24 h starvation were treated with compound U0126, which inhibits the ERK1/2 upstream kinase MEK-1/2, for 30 min following by the different treatments. As expected, treatment with a selective MEK1/2 inhibitor U0126 (1 μM) attenuated ERK1/2 phosphorylation at Thr202/Tyr204 in OCs (Figs. 2A–D). Treatment with U0126 (1 μM) along with PGD2 or DK-PGD2 further reduced ERK1/2 phosphorylation at Thr202/Tyr204, whereas ERK1/2 phosphorylation

115

(Thr202/Tyr204) was increased in OCs treated with U0126 along with a CRTH2 antagonist CAY10471 as compared with U0126 alone (Figs. 2A–D). Finally, we determined the effect of U0126 on OC apoptosis in the presence or absence of CRTH2 ligands. Treatment with U0126 (1 μM) alone increased caspase-3-like activity in OCs as compared to the vehicle group (Fig. 2E). Similarly to our previous report [31], both PGD2 (10 nM) and the CRTH2 agonist DK-PGD2 (10 nM) increased caspase3 activity in human differentiated OCs (Fig. 2E). Interestingly, addition of U0126 to PGD2 or to the CRTH2 agonist DK-PGD2 further increased caspase-3 activity when compared with PGD2 or a CRTH2 agonist DKPGD2 alone, respectively (Fig. 2E). In addition, U0126 did not cause significant increase of caspase-3 activity in the presence of a CRTH2 antagonist CAY10471 (10 nM) (Fig. 2F). Neither the DP agonist BW 245C (10 nM) nor the DP antagonist BW A868C (10 nM) induced an increase of caspase-3 activity in the presence or absence of U0126 treatment as compared to the corresponding controls (Figs. 2E and F). Moreover, treatment with U0126 (0.1 nM–1 μM) for 30 min in the presence of exogenous PGD2 (10 nM) induced a dose-dependent increase in the percentage of apoptotic OCs (from 39.40 ± 5.62% to 58.00 ± 3.96%, N = 5, p b 0.05), with an EC50 of 7.91 ± 4.84 nM (Fig. 2G). Taken together, these results strongly suggest that PGD2-mediated OC apoptosis

Fig. 2. Participation of MEK-ERK1/2 pathway in apoptosis of OCs. After 21 days of differentiation, OCs were starved for 24 h in 2% FBS-containing medium and deprivation of RANKL and MCSF. The differentiated OCs were further pre-treated with or without a selective MEK-1/2 inhibitor (U0126, 1 μM) for 30 min prior to the stimulation for 24 h with 10 nM of PGD2, BW 245C (a DP agonist) or DK-PGD2 (a CRTH2 agonist) in the presence of 10 μM naproxen; with 10 nM of CAY10471 (a CRTH2 antagonist) or BW A868C (a DP antagonist) in the absence of naproxen for 24 h (presence of endogenous prostaglandins). A and C, the representative Western blot bands of total ERK1 and phosphorylated ERK1/2 at Thr202/Tyr204 in response to different treatments. B and D, relative density of phosphorylated ERK1/2 (Thr202/Tyr204) normalized to total ERK1 in the presence of different treatments (N = 3). E and F, caspase-3like activity on human OC apoptosis with different treatments (N = 3–5). In Figs. 3A–F, the group treated with naproxen and vehicle (DMSO) was used as a control when the OCs were stimulated with PGD2 or an agonist of its receptors in the presence of naproxen. Otherwise, the group treated with vehicle (DMSO) only was used as a control when the cells were incubated with an antagonist of PGD2 receptors in the absence of naproxen. G, concentration–response curve of MEK-1/2 inhibitor in the presence of 10 nM of PGD2 (N = 5); OCs treated with naproxen and PGD2 was used as a control. Data are means ± SEM, *p b 0.05, **p b 0.01, ***p b 0.001 vs. controls.

116

L. Yue et al. / Bone 60 (2014) 112–121

is due to inhibition of ERK1/2 signaling pathway through the activation of the CRTH2 receptor.

Ser473 (Figs. 4A–D). These results demonstrate the reduction of Akt phosphorylation during PGD2/CRTH2-induced OC apoptosis.

Increased phosphorylation of β-arrestin-1 at Ser412 during PGD2-induced OC apoptosis

Akt signaling pathway is the upstream of ERK1/2 pathway during PGD2-induced OC apoptosis

Phosphorylation of β-arrestin-1 at Ser412 is associated with the inhibition of ERK1/2 activation [37]. Therefore, we measured the phosphorylation of β-arrestin-1 at Ser412 in OCs treated with CRTH2 ligands. As shown in Figs. 3A and B, treatment with PGD2 (10 nM) or the selective CRTH2 agonist (DK-PGD2, 10 nM) in the presence of naproxen increased the phosphorylation of β-arrestin-1 at Ser412 compared to the vehicle-treated control cells. Treatment with 10 nM of the CRTH2 antagonist CAY10471 in the absence of naproxen decreased phosphorylation of β-arrestin-1 at Ser412 in OCs, as compared with the vehicle-treated controls (Figs. 3C and D). Phosphorylation of β-arrestin-1 at Ser412 was not changed by treatment with neither a DP agonist (BW 245C, 10 nM) nor an antagonist (BW A868C, 10 nM) in the presence and in the absence of naproxen, respectively (Figs. 3A–D). These results show the increase in phosphorylation of β-arrestin-1 at Ser412 during PGD2/CRTH2-induced OC apoptosis.

In order to investigate the association of Akt and ERK1/2 signaling pathways in PGD2-induced OC apoptosis, the cells were treated with the PI3K inhibitor LY294002 (1 μM) for 30 min before analyzing ERK1/2 phosphorylation at Thr202/Tyr204 using Western blot. Similar to Fig. 2, stimulation with PGD2 (10 nM) decreased ERK1/2 phosphorylation at Thr202/Tyr204 in human OCs as compared to the vehicletreated control in the presence of naproxen to inhibit endogenous prostaglandin production (Figs. 5A and B). Treatment with the PI3K inhibitor LY294002 alone or along with exogenous PGD2 further decreased ERK1/2 phosphorylation at Thr202/Tyr204 as compared to the corresponding controls (Figs. 5A and B). In the absence of naproxen, treatment with the selective CRTH2 antagonist CAY10471 (10 nM) increased ERK phosphorylation as compared to the vehicle-treated control (Figs. 5C and D). Treatment with the PI3K inhibitor LY294002 attenuated CRTH2 antagonist compound CAY10471-induced increase in ERK1/2 phosphorylation at Thr202/Tyr204 in human OCs (Figs. 5C and D). These results suggest that ERK1/2 is a downstream of PI3K/Akt signaling during PGD2-induced OC apoptosis.

Reduced Akt phosphorylation on Ser473 during PGD2-induced OC apoptosis ERK1/2 and Akt can coordinately or independently regulate OC survival and apoptosis [16,36,38]. Thus, we determined the levels of Akt activation by studying its phosphorylation on Ser473 in OCs treated with different CRTH2 ligands by immunoblotting. In the presence of naproxen, treatment with PGD2 (10 nM) or the selective CRTH2 agonist (DK-PGD2, 10 nM) decreased the phosphorylation of Akt at Ser473 compared to the vehicle-treated control cells (Figs. 4A and B). Treatment with 10 nM of the CRTH2 antagonist compound CAY10471 in the absent of naproxen increased Akt phosphorylation in OCs, as compared with the vehicle-treated controls (Figs. 4C and D). However, treatment with either a DP agonist (BW 245C, 10 nM) or antagonist (BW A868C, 10 nM) did not have any effect on Akt phosphorylation at

IKK2/NF-κB signaling pathway is not implicated in the PGD2/CRTH2-induced human OC apoptosis Both ERK1/2 and NF-κB signals can independently influence or affect each other in the regulation of OC survival and apoptosis [12,16,36]. However, it remains unknown whether the NF-κB pathway is involved in PGD2-induced OC apoptosis. Cell treatment with 10 nM of PGD2, DP agonist BW 245C, CRTH2 agonist DK-PGD2, DP antagonist BW A868C, or CRTH2 antagonist CAY10471 did not affect the phosphorylation of RelA/p65 on Ser536 compared to their corresponding controls (Figs. 6A–D). Moreover, incubation with a selective IKK-2 inhibitor

Fig. 3. Modulation of β-arrestin-1 phosphorylation at Ser412 during human OC apoptosis. Differentiated OCs were starved for 24 h in 2% FBS-containing medium without RANKL or M-CSF, then the cells were treated with 10 nM of PGD2, BW 245C (a DP agonist) or DK-PGD2 (a CRTH2 agonist) in the presence of 10 μM naproxen; or with 10 nM of CAY10471 (a CRTH2 antagonist) or BW A868C (a DP antagonist) in the absence of naproxen for 24 h. A and C, the representative Western blot bands of phosphorylated β-arrestin-1 at Ser412 and actin with different treatments. B and D, relative density of phosphorylated β-arrestin-1 (Ser412) normalized to actin in response to different treatments (N = 3). The group treated with naproxen and vehicle (DMSO) was used as a control when the OCs were treated with PGD2 or an agonist of its receptors in the presence of naproxen. Otherwise, the group treated with vehicle (DMSO) only was used as a control when the cells were incubated with an antagonist of PGD2 receptors in the absence of naproxen. Data are means ± SEM, *p b 0.05 vs. controls.

L. Yue et al. / Bone 60 (2014) 112–121

117

Fig. 4. Involvement of Akt phosphorylation in OC apoptosis. At 21 days, the differentiated OCs were starved for 24 h in 2% FBS-containing medium without M-CSF and RANKL. The cells were further treated for 24 h with 10 nM of PGD2, BW 245C (a DP agonist) or DK-PGD2 (a CRTH2 agonist) in the presence of naproxen (10 μM); with 10 nM of CAY10471 (a CRTH2 antagonist) or BW A868C (a DP antagonist) in the absence of naproxen for 24 h. A and C, the representative Western blot bands of total Akt and phosphorylated Akt at Ser473 with different treatments. B and D, relative density of phosphorylated Akt (Ser473) normalized to total Akt in the presence of different treatments (N = 3). The group treated with naproxen and vehicle (DMSO) was used as a control when the OCs were stimulated with PGD2 or an agonist of its receptors in the presence of naproxen. The group treated with vehicle (DMSO) only was used as a control when the cells were incubated with an antagonist of PGD2 receptors in the absence of naproxen. Data are means ± SEM, *p b 0.05, **p b 0.01 vs. controls.

PHA-408 (2 μM) had no effect on caspase-3-like activity in the presence or absence of PGD2, DP agonist BW 245C, CRTH2 agonist DK-PGD2, DP antagonist BW A868C, or CRTH2 antagonist CAY10471 (Figs. 6E and F). These results indicate that the IKK2/NF-κB pathway is not involved in PGD2/CRTH2-induced human OC apoptosis. Discussion As cells primarily responsible for bone resorption, OCs play a crucial role in various bone diseases such as rheumatoid arthritis, osteoporosis and Paget's disease of bone, to name but a few. Induction of OC

apoptosis affects the rate of bone remodeling and turnover and leads to decreased bone resorption [39], being a potential therapeutic target for the treatment of such bone diseases. Tamoxifen and bisphosphonates, clinically used to decrease bone resorption, exert at least part of their beneficial effects by regulating OC apoptosis [40–42]. However, the physiological mechanisms regulating OC apoptosis are not fully known. Our previous publications suggest that PGD2 may be a physiologic anabolic agent for the bone: the production of PGD2 but not of PGE2 increases in humans in situations of high bone turnover such as fracture repair [43]. Human OBs produce PGD2 in response to biologically relevant stimuli [44,45] and express both the DP and CRTH2 receptors for

Fig. 5. Association of Akt and ERK1/2 signaling pathways in PGD2-induced OC apoptosis. The cells were starved for 24 h in 2% FBS-containing medium in the absence of RANKL and M-CSF, following pre-treatment with or without a selective PI3K inhibitor (LY294002, 1 μM) for 30 min prior to the stimulation for 24 h with 10 nM of PGD2 in the presence of 10 μM naproxen; with 10 nM of CRTH2 antagonist CAY10471 in the absence of naproxen for 24 h. The group treated with naproxen and vehicle (DMSO) was used as a control when the OCs were stimulated with PGD2 in the presence of naproxen. The group treated with vehicle (DMSO) only was used as a control when the cells were incubated with CRTH2 antagonist in the absence of naproxen. A and C, the representative western blot bands of total ERK1 and phosphorylated ERK1/2 at Thr202/Tyr204 in response to different treatments. B and D, relative density of phosphorylated ERK1/2 (Thr202/Tyr204) normalized to total ERK1 in the presence of different treatments (N = 3). Data are means ± SEM, *p b 0.05, **p b 0.01, ***p b 0.001 vs. controls.

118

L. Yue et al. / Bone 60 (2014) 112–121

Fig. 6. No effect on OC apoptosis through IKK2/NF-κB signaling pathway. After 21 days of differentiation, the cells were starved for 24 h in 2% FBS-containing medium without RANKL or M-CSF, then pre-treated with or without a selective IKK2 inhibitor (PHA-408, 2 μM) for 30 min prior to the stimulation for 24 h with 10 nM of PGD2, BW 245C (a DP agonist) or DK-PGD2 (a CRTH2 agonist) in the presence of 10 μM naproxen; with 10 nM of CAY10471 (a CRTH2 antagonist) or BW A868C (a DP antagonist) in the absence of naproxen for 24 h. The group treated with naproxen and vehicle (DMSO) was used as a control when the OCs were stimulated with PGD2 or an agonist of its receptors in the presence of naproxen. The group treated with vehicle (DMSO) only was used as a control when the cells were incubated with an antagonist of PGD2 receptors in the absence of naproxen. A and C, the representative Western blot bands of total RelA/p65 and phosphorylated RelA/p65 in response to different treatments. B and D, relative density of phosphorylated RelA/p65 normalized to total RelA/p65 in the presence of different treatments (N = 3). E and F, caspase-3-like activity on human OC apoptosis with different treatments (N = 3). Data are means ± SEM, *p b 0.05 vs. controls.

PGD2. Stimulation of DP decreases the expression of osteoprotegerin, while activation of CRTH2 decreases the expression of RANKL and is a potent chemotactic stimulus for OBs [45]. Human OCs also express DP and CRTH2 receptors: activation of the DP receptor strongly inhibits bone resorption by OCs, while activation of either DP or CRTH2 reduces osteoclastogenesis [32]. More recent results show that activation of the CRTH2 receptor induces human OC apoptosis through the intrinsic pathway [31]. The objective of the present study was to investigate the intracellular signaling pathways implicated in the induction of OC apoptosis by activation of the CRTH2 receptor. We used our previously described model of OCs differentiated in vitro from human PBMCs in the presence of RANKL and M-CSF [31–33]. This model generates relatively high number of multinuclear cells presenting the characteristic features and functions of OCs such as TRAP expression, actin ring formation, calcium response to calcitonin, and bone resorption, as shown in the present study and in our previous publications [31–33]. Since cells in this experimental model produce significant amounts of PGD2, as previously show [31], we used the following experimental approach: when testing the effects of PGD2, DP or CRTH2 agonists the endogenous synthesis of prostaglandins was inhibited by pre-incubation with the cyclooxygenases inhibitor naproxen, which was also used in the vehicle-treated controls;

this prevented competition of the endogenous PGD2 with the agonists used. On the other hand, when testing the effects of DP or CRTH2 antagonists, the endogenous production of prostaglandins was not inhibited, and naproxen was not added to the preparations. The participation of MEK-ERK1/2 cascade in the control of OC apoptosis remains contradictory, with reports showing its implication in the survival [11,16,46] or increased apoptosis of these cells [15]. These opposing results may depend on the experimental model, condition and species tested. The present study showed that a significant decrease in ERK1/2 phosphorylation was observed in OCs after treatment with PGD2 or a CRTH2 agonist, whereas treatment with a DP agonist or a DP antagonist did not have any effect on ERK1/2 phosphorylation. A CRTH2 agonist treatment further increased, whereas a CRTH2 antagonist diminished a MEK-1/2 inhibitor (U0126)-induced increase in caspase-3-like activity. These results suggest that PGD2 promotes CRTH2-dependent apoptosis in human OCs by inhibiting the ERK1/2 pathway. MEK-1/2 inhibitor treatment along with PGD2 and CRTH2 agonist led to more caspase-3 activity than the sum of the two individual responses. These results may indicate that the two drugs are synergistic or that another signaling pathway may be implicated in PGD2/CRTH2induced OC apoptosis. It is important to notice that the concentration of the selective MEK-1/2 inhibitor used in the present study (compound

L. Yue et al. / Bone 60 (2014) 112–121

U0126) did not completely inhibit ERK1/2 phosphorylation, higher concentrations inducing cell death. This incomplete inhibition may explain why both PGD2 and the CRTH2 agonist further reduced ERK1/2 phosphorylation and increased caspase-3 activity in the presence of U0126 treatment, although the participation of other pathways remains possible to explain these findings. CRTH2 is a G protein-coupled receptor (GPCR) expressed by Th2 lymphocytes, eosinophils and OCs, which mediates the functions of these cells in response to PGD2 [24,30,31]. Activation of CRTH2 by PGD2 leads to the Akt and ERK1/2 phosphorylation through G proteindependent pathway in human keratinocytes and Th2 lymphocytes [47,48]. This is corroborated by the findings that the activation of CRTH2 by PGD2 stimulates the PI3K signaling pathway, resulting in Akt phosphorylation to block the intrinsic apoptosis signaling pathway in human Th2 cells [24]. In addition, the signaling of CRTH2 receptor could be modulated by GPCR kinase 5 (GRK5) in HEK293 cells [49]. Recent studies have shown that GPCR can elicit signals through the interaction with scaffolding proteins, such as β-arrestins, which is independent of G protein coupling [50–52]. This arrestin-dependent pathway can synergize or oppose G protein-dependent signals, which may be due to the difference of agonist concentration, phosphorylation state of the receptor, alteration of receptor conformation, and availability of downstream effectors [53–56]. β-Arrestins can scaffold a number of kinases, such as PI3K, Akt and ERK1/2, leading to their activation or inactivation [52,56]. For example, β-arrestin-1 phosphorylation on Ser412 is required for GRK5-mediated inhibition of ERK1/2 signaling via G protein-independent pathway [57]. It has been shown that increased phosphorylation of β-arrestin-1 on Ser412 impairs its activity, leading to disruption of G protein-mediated ERK1/2 signal by insulin [37]. We observed that activation of CRTH2 reduced, whereas a CRTH2 antagonist increased ERK1/2 and Akt phosphorylation in OCs. This is in agreement with a study showing that the mutant of CRTH2 C-terminal tail increased ERK1/2 phosphorylation [58]. Importantly, both PGD2 and CRTH2 agonist augmented, whereas a CRTH2 antagonist decreased phosphorylation of β-arrestin-1 at Ser412 in OCs. Therefore, we speculate that in OCs, the activation of CRTH2 affects its C-terminal conformation and/or impairs β-arrestin-1 recruitment/ activity, thereby leading to inhibition of Akt and ERK1/2 as well as subsequent apoptosis. Further studies using β-arrestin-1 knockout OCs and co-immunoprecipitation techniques are needed to test this hypothesis. PI3K coordinately activates the MEK-ERK1/2 and NF-κB pathways to regulate OC survival and death [16,17]. It was reported that Akt isoforms – Akt1/2 – are abundantly expressed in both OBs and OCs, and regulate their differentiation and survival [59]. Previous findings revealed that double knockout mice of Akt1 and Akt2 affects bone development [60], and inhibition of PI3K/Akt and ERK1/2 causes apoptosis of mature OCs [46]. Xue et al. reported that CRTH2 mediated an inhibitory effect of PGD2 on the apoptosis of human Th2 cells through activation of PI3K pathway [24]. Similarly, we also found that the Akt phosphorylation was decreased during OC apoptosis. However, both PGD2 and CRTH2 agonist treatments increased the caspase-3-like activity in the presence of naproxen, whereas caspase-3-like activity was decreased by CRTH2 antagonist in OCs in the absence of naproxen. These observations suggest a cell-specific role of PGD2 and PI3K/Akt in regulating apoptosis. In general, PI3K activates MEK-ERK1/2 and NF-κB pathways [11,16,61,62]. On the contrary, Akt can also terminate ERK1/2 phosphorylation [63]. We found that treatment with a selective PI3K inhibitor reduced the phosphorylation of ERK1/2, which is associated with increased OC apoptosis induced by PGD2. As per previous studies [11,19,46], induction of OC apoptosis by both selective PI3K and MEK1/2 inhibitors is probably not due to their combined cytotoxicity. All these results suggest that PGD2 activates CRTH2, thereby inhibiting PI3K/Akt and subsequent ERK1/2 pathways leading to OC apoptosis. We have previously reported that PGD2/CRTH2 induced apoptosis of

119

human OCs uses the intrinsic apoptosis pathway [31]. Further experiments are needed to determine whether and how the PI3K/Akt and ERK1/2 pathways regulate OC apoptosis via the intrinsic apoptosis pathway. Two signaling pathways in the activation of NF-κB are known as the canonical and non-canonical pathways. Accumulating evidence shows that NF-κB is required for OC survival, and NF-κB inhibition/block causes OC apoptosis [12,13,64–66]. There is a conflicting report showing that the inhibition of NF-κB does not have any impact on OC survival or apoptosis [20]. Our study demonstrated that the treatment of OCs with PGD2, DP agonist (BW 245C), CRTH2 agonist (DK-PGD2), DP antagonist (BW A868C), or CRTH2 antagonist (CAY10471) did not alter the phosphorylation of RelA/p65 on Ser536. Furthermore, treatment with a selective IKK2/NF-κB inhibitor had no effect on OC apoptosis regardless of PGD2 treatment. These results suggest that IKK2/NF-κB pathway did not participate in the PGD2/CRTH2-induced human OC apoptosis, even though participation of non-canonical pathways in this effect cannot be ruled out. Conclusions In summary, PGD2 induced apoptosis via CRTH2 receptor, which was associated with impaired β-arrestin-1 activity as well as reduction of phosphorylation of ERK1/2 and Akt in human differentiated OCs. Treatment with the selective MEK-1/2 inhibitor further increased PGD2/ CRTH2-mediated OC apoptosis, whereas CRTH2 antagonist diminished MEK1/2 inhibitor-induced OC apoptosis. Moreover, PI3K inhibitor treatment reduced ERK1/2 phosphorylation in OCs treated with PGD2. The phosphorylation of RelA/p65 (ser536) was not altered in OCs treated with either PGD2 or its agonists/antagonists. In addition, treatment with a selective IKK2/NF-κB inhibitor had no effect on PGD2/ CRTH2-mediated OC apoptosis. The results presented in this study suggest that PGD2 induces human OC apoptosis through regulation of β-arrestin-1, ERK1/2 and Akt, but not the IKK2/NF-κB pathway. The development of selective inhibitors of PI3K and MEK-1/2 may provide an interventional avenue in halting the progression of bone diseases characterized by excessive bone loss through induction of OC apoptosis. Conflict of interest statement The authors have no conflict of interest to disclose. Funding Supported by the Canadian Institutes of Health Research (grant MOP89786). Acknowledgments This work was supported by the Canadian Institutes of Health Research (grant MOP89786). References [1] Singh A, Mehdi AA, Srivastava RN, Verma NS. Immunoregulation of bone remodelling. Int J Crit Illn Inj Sci 2012;2:75–81. [2] Nakanishi A, Hie M, Iitsuka N, Tsukamoto I. A crucial role for reactive oxygen species in macrophage colony-stimulating factor-induced RANK expression in osteoclastic differentiation. Int J Mol Med 2013;31:874–80. [3] Ikeda F, Matsubara T, Tsurukai T, Hata K, Nishimura R, Yoneda T. JNK/c-Jun signaling mediates an anti-apoptotic effect of RANKL in osteoclasts. J Bone Miner Res 2008;23:907–14. [4] Kudo O, Fujikawa Y, Itonaga I, Sabokbar A, Torisu T, Athanasou NA. Proinflammatory cytokine (TNFalpha/IL-1alpha) induction of human osteoclast formation. J Pathol 2002;198:220–7. [5] Kudo O, Sabokbar A, Pocock A, Itonaga I, Fujikawa Y, Athanasou NA. Interleukin-6 and interleukin-11 support human osteoclast formation by a RANKL-independent mechanism. Bone 2003;32:1–7.

120

L. Yue et al. / Bone 60 (2014) 112–121

[6] Krum SA, Miranda-Carboni GA, Hauschka PV, Carroll JS, Lane TF, Freedman LP, et al. Estrogen protects bone by inducing Fas ligand in osteoblasts to regulate osteoclast survival. EMBO J 2008;27:535–45. [7] Houde N, Chamoux E, Bisson M, Roux S. Transforming growth factor-beta1 (TGFbeta1) induces human osteoclast apoptosis by up-regulating Bim. J Biol Chem 2009;284:23397–404. [8] Huang W, O'Keefe RJ, Schwarz EM. Exposure to receptor–activator of NFkappaB ligand renders pre-osteoclasts resistant to IFN-gamma by inducing terminal differentiation. Arthritis Res Ther 2003;5:R49–59. [9] Bendixen AC, Shevde NK, Dienger KM, Willson TM, Funk CD, Pike JW. IL-4 inhibits osteoclast formation through a direct action on osteoclast precursors via peroxisome proliferator-activated receptor gamma 1. Proc Natl Acad Sci U S A 2001;98:2443–8. [10] Horwood NJ, Elliott J, Martin TJ, Gillespie MT. IL-12 alone and in synergy with IL-18 inhibits osteoclast formation in vitro. J Immunol 2001;166:4915–21. [11] Lee SE, Chung WJ, Kwak HB, Chung CH, Kwack KB, Lee ZH, et al. Tumor necrosis factor-alpha supports the survival of osteoclasts through the activation of Akt and ERK. J Biol Chem 2001;276:49343–9. [12] Abbas S, Abu-Amer Y. Dominant-negative IkappaB facilitates apoptosis of osteoclasts by tumor necrosis factor-alpha. J Biol Chem 2003;278:20077–82. [13] Penolazzi L, Lambertini E, Borgatti M, Piva R, Cozzani M, Giovannini I, et al. Decoy oligodeoxynucleotides targeting NF-kappaB transcription factors: induction of apoptosis in human primary osteoclasts. Biochem Pharmacol 2003;66:1189–98. [14] Xing L, Venegas AM, Chen A, Garrett-Beal L, Boyce BF, Varmus HE, et al. Genetic evidence for a role for Src family kinases in TNF family receptor signaling and cell survival. Genes Dev 2001;15:241–53. [15] Recchia I, Rucci N, Funari A, Migliaccio S, Taranta A, Longo M, et al. Reduction of c-Src activity by substituted 5,7-diphenyl-pyrrolo[2,3-d]-pyrimidines induces osteoclast apoptosis in vivo and in vitro. Involvement of ERK1/2 pathway. Bone 2004;34:65–79. [16] Gingery A, Bradley E, Shaw A, Oursler MJ. Phosphatidylinositol 3-kinase coordinately activates the MEK/ERK and AKT/NFkappaB pathways to maintain osteoclast survival. J Cell Biochem 2003;89:165–79. [17] Gingery A, Bradley EW, Pederson L, Ruan M, Horwood NJ, Oursler MJ. TGF-beta coordinately activates TAK1/MEK/AKT/NFkB and SMAD pathways to promote osteoclast survival. Exp Cell Res 2008;314:2725–38. [18] Fukuda A, Hikita A, Wakeyama H, Akiyama T, Oda H, Nakamura K, et al. Regulation of osteoclast apoptosis and motility by small GTPase binding protein Rac1. J Bone Miner Res 2005;20:2245–53. [19] Nakamura H, Hirata A, Tsuji T, Yamamoto T. Role of osteoclast extracellular signalregulated kinase (ERK) in cell survival and maintenance of cell polarity. J Bone Miner Res 2003;18:1198–205. [20] Miyazaki T, Katagiri H, Kanegae Y, Takayanagi H, Sawada Y, Yamamoto A, et al. Reciprocal role of ERK and NF-kappaB pathways in survival and activation of osteoclasts. J Cell Biol 2000;148:333–42. [21] Erl W, Weber C, Zernecke A, Neuzil J, Vosseler CA, Kim HJ, et al. Cyclopentenone prostaglandins induce endothelial cell apoptosis independent of the peroxisome proliferator-activated receptor-gamma. Eur J Immunol 2004;34:241–50. [22] Papadimitriou A, King AJ, Jones PM, Persaud SJ. Anti-apoptotic effects of arachidonic acid and prostaglandin E2 in pancreatic beta-cells. Cell Physiol Biochem 2007;20:607–16. [23] Ward C, Dransfield I, Murray J, Farrow SN, Haslett C, Rossi AG. Prostaglandin D2 and its metabolites induce caspase-dependent granulocyte apoptosis that is mediated via inhibition of I kappa B alpha degradation using a peroxisome proliferator-activated receptor-gamma-independent mechanism. J Immunol 2002;168:6232–43. [24] Xue L, Barrow A, Pettipher R. Novel function of CRTH2 in preventing apoptosis of human Th2 cells through activation of the phosphatidylinositol 3-kinase pathway. J Immunol 2009;182:7580–6. [25] Chiba T, Ueki S, Ito W, Kato H, Kamada R, Takeda M, et al. The opposing role of two prostaglandin D2 receptors, DP and CRTH2, in human eosinophil migration. Ann Allergy Asthma Immunol 2011;106:511–7. [26] Arima M, Fukuda T. Prostaglandin D2 receptors DP and CRTH2 in the pathogenesis of asthma. Curr Mol Med 2008;8:365–75. [27] Ragolia L, Palaia T, Frese L, Fishbane S, Maesaka JK. Prostaglandin D2 synthase induces apoptosis in PC12 neuronal cells. Neuroreport 2001;12:2623–8. [28] Wang JJ, Mak OT. Induction of apoptosis in non-small cell lung carcinoma A549 cells by PGD(2) metabolite, 15d-PGJ(2). Cell Biol Int 2011;35:1089–96. [29] Chen YC, Shen SC, Tsai SH. Prostaglandin D(2) and J(2) induce apoptosis in human leukemia cells via activation of the caspase 3 cascade and production of reactive oxygen species. Biochim Biophys Acta 2005;1743:291–304. [30] Gervais FG, Cruz RP, Chateauneuf A, Gale S, Sawyer N, Nantel F, et al. Selective modulation of chemokinesis, degranulation, and apoptosis in eosinophils through the PGD2 receptors CRTH2 and DP. J Allergy Clin Immunol 2001;108:982–8. [31] Yue L, Durand M, Lebeau Jacob MC, Hogan P, McManus S, Roux S, et al. Prostaglandin D2 induces apoptosis of human osteoclasts by activating the CRTH2 receptor and the intrinsic apoptosis pathway. Bone 2012;51:338–46. [32] Durand M, Gallant MA, de Brum-Fernandes AJ. Prostaglandin D2 receptors control osteoclastogenesis and the activity of human osteoclasts. J Bone Miner Res 2008;23:1097–105. [33] Hackett JA, Allard-Chamard H, Sarrazin P, de Fatima Lucena M, Gallant MA, Fortier I, et al. Prostaglandin production by human osteoclasts in culture. J Rheumatol 2006;33:1320–8. [34] Durand M, Boire G, Komarova SV, Dixon SJ, Sims SM, Harrison RE, et al. The increased in vitro osteoclastogenesis in patients with rheumatoid arthritis is due to increased percentage of precursors and decreased apoptosis—the In Vitro Osteoclast Differentiation in Arthritis (IODA) study. Bone 2011;48:588–96.

[35] Seong GJ, Park C, Kim CY, Hong YJ, So HS, Kim SD, et al. Mitomycin-C induces the apoptosis of human Tenon's capsule fibroblast by activation of c-Jun N-terminal kinase 1 and caspase-3 protease. Invest Ophthalmol Vis Sci 2005;46:3545–52. [36] Fong D, Bisson M, Laberge G, McManus S, Grenier G, Faucheux N, et al. Bone morphogenetic protein-9 activates Smad and ERK pathways and supports human osteoclast function and survival in vitro. Cell Signal 2013;25:717–28. [37] Hupfeld CJ, Resnik JL, Ugi S, Olefsky JM. Insulin-induced beta-arrestin1 Ser-412 phosphorylation is a mechanism for desensitization of ERK activation by Galphaicoupled receptors. J Biol Chem 2005;280:1016–23. [38] Kwak HB, Lee MS, Kim HS, Cho HJ, Kim JW, Lee ZH, et al. Proteasome inhibitors induce osteoclast survival by activating the Akt pathway. Biochem Biophys Res Commun 2008;377:1–6. [39] Piva R, Penolazzi L, Borgatti M, Lampronti I, Lambertini E, Torreggiani E, et al. Apoptosis of human primary osteoclasts treated with molecules targeting nuclear factor-kappaB. Ann N Y Acad Sci 2009;1171:448–56. [40] Weinstein RS, Manolagas SC. Apoptosis and osteoporosis. Am J Med 2000;108:153–64. [41] Benford HL, McGowan NW, Helfrich MH, Nuttall ME, Rogers MJ. Visualization of bisphosphonate-induced caspase-3 activity in apoptotic osteoclasts in vitro. Bone 2001;28:465–73. [42] Wu X, Ahn EY, McKenna MA, Yeo H, McDonald JM. Fas binding to calmodulin regulates apoptosis in osteoclasts. J Biol Chem 2005;280:29964–70. [43] Gallant MA, Chamoux E, Bisson M, Wolsen C, Parent JL, Roux S, et al. Increased concentrations of prostaglandin D2 during post-fracture bone remodeling. J Rheumatol 2010;37:644–9. [44] Mathurin K, Gallant MA, Germain P, Allard-Chamard H, Brisson J, Iorio-Morin C, et al. An interaction between L-prostaglandin D synthase and arrestin increases PGD2 production. J Biol Chem 2011;286:2696–706. [45] Gallant MA, Samadfam R, Hackett JA, Antoniou J, Parent JL, de Brum-Fernandes AJ. Production of prostaglandin D(2) by human osteoblasts and modulation of osteoprotegerin, RANKL, and cellular migration by DP and CRTH2 receptors. J Bone Miner Res 2005;20:672–81. [46] Lee ZH, Lee SE, Kim CW, Lee SH, Kim SW, Kwack K, et al. IL-1alpha stimulation of osteoclast survival through the PI 3-kinase/Akt and ERK pathways. J Biochem 2002;131:161–6. [47] Kanda N, Ishikawa T, Watanabe S. Prostaglandin D2 induces the production of human beta-defensin-3 in human keratinocytes. Biochem Pharmacol 2010;79:982–9. [48] Xue L, Gyles SL, Barrow A, Pettipher R. Inhibition of PI3K and calcineurin suppresses chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2)dependent responses of Th2 lymphocytes to prostaglandin D(2). Biochem Pharmacol 2007;73:843–53. [49] Gallant MA, Slipetz D, Hamelin E, Rochdi MD, Talbot S, de Brum-Fernandes AJ, et al. Differential regulation of the signaling and trafficking of the two prostaglandin D 2 receptors, prostanoid DP receptor and CRTH2. Eur J Pharmacol 2007;557:115–23. [50] Defea K. Beta-arrestins and heterotrimeric G-proteins: collaborators and competitors in signal transduction. Br J Pharmacol 2008;153(Suppl. 1):S298–309. [51] Mathiesen JM, Ulven T, Martini L, Gerlach LO, Heinemann A, Kostenis E. Identification of indole derivatives exclusively interfering with a G protein-independent signaling pathway of the prostaglandin D2 receptor CRTH2. Mol Pharmacol 2005;68:393–402. [52] Luttrell LM, Lefkowitz RJ. The role of beta-arrestins in the termination and transduction of G-protein-coupled receptor signals. J Cell Sci 2002;115:455–65. [53] Sun Y, Huang J, Xiang Y, Bastepe M, Juppner H, Kobilka BK, et al. Dosage-dependent switch from G protein-coupled to G protein-independent signaling by a GPCR. EMBO J 2007;26:53–64. [54] Azzi M, Charest PG, Angers S, Rousseau G, Kohout T, Bouvier M, et al. Betaarrestin-mediated activation of MAPK by inverse agonists reveals distinct active conformations for G protein-coupled receptors. Proc Natl Acad Sci U S A 2003;100:11406–11. [55] Yee DK, Suzuki A, Luo L, Fluharty SJ. Identification of structural determinants for G protein-independent activation of mitogen-activated protein kinases in the seventh transmembrane domain of the angiotensin II type 1 receptor. Mol Endocrinol 2006;20:1924–34. [56] Wang P, DeFea KA. Protease-activated receptor-2 simultaneously directs betaarrestin-1-dependent inhibition and Galphaq-dependent activation of phosphatidylinositol 3-kinase. Biochemistry 2006;45:9374–85. [57] Barthet G, Carrat G, Cassier E, Barker B, Gaven F, Pillot M, et al. Beta-arrestin1 phosphorylation by GRK5 regulates G protein-independent 5-HT4 receptor signalling. EMBO J 2009;28:2706–18. [58] Schroder R, Merten N, Mathiesen JM, Martini L, Kruljac-Letunic A, Krop F, et al. The C-terminal tail of CRTH2 is a key molecular determinant that constrains Galphai and downstream signaling cascade activation. J Biol Chem 2009;284:1324–36. [59] Kawamura N, Kugimiya F, Oshima Y, Ohba S, Ikeda T, Saito T, et al. Akt1 in osteoblasts and osteoclasts controls bone remodeling. PLoS One 2007;2:e1058. [60] Peng XD, Xu PZ, Chen ML, Hahn-Windgassen A, Skeen J, Jacobs J, et al. Dwarfism, impaired skin development, skeletal muscle atrophy, delayed bone development, and impeded adipogenesis in mice lacking Akt1 and Akt2. Genes Dev 2003;17:1352–65. [61] Agazie Y, Ischenko I, Hayman M. Concomitant activation of the PI3K-Akt and the Ras-ERK signaling pathways is essential for transformation by the V-SEA tyrosine kinase oncogene. Oncogene 2002;21:697–707. [62] Burow ME, Weldon CB, Melnik LI, Duong BN, Collins-Burow BM, Beckman BS, et al. PI3-K/AKT regulation of NF-kappaB signaling events in suppression of TNF-induced apoptosis. Biochem Biophys Res Commun 2000;271:342–5.

L. Yue et al. / Bone 60 (2014) 112–121 [63] Reusch HP, Zimmermann S, Schaefer M, Paul M, Moelling K. Regulation of Raf by Akt controls growth and differentiation in vascular smooth muscle cells. J Biol Chem 2001;276:33630–7. [64] Soysa NS, Alles N, Shimokawa H, Jimi E, Aoki K, Ohya K. Inhibition of the classical NFkappaB pathway prevents osteoclast bone-resorbing activity. J Bone Miner Metab 2009;27:131–9.

121

[65] Ozaki K, Takeda H, Iwahashi H, Kitano S, Hanazawa S. NF-kappaB inhibitors stimulate apoptosis of rabbit mature osteoclasts and inhibit bone resorption by these cells. FEBS Lett 1997;410:297–300. [66] Wu M, Wang Y, Deng L, Chen W, Li YP. TRAF family member-associated NF-kappaB activator (TANK) induced by RANKL negatively regulates osteoclasts survival and function. Int J Biol Sci 2012;8:1398–407.