Inorganic Pyrophosphate Promotes Osteoclastogenic Commitment and Survival of Bone Marrow Derived Monocytes mediated by Egr-1 up-regulation and MITF phosphorylation

Several reports emphasized the importance of inorganic pyrophosphate (PPi) in hindering osteoblast differentiation and bone matrix mineralization. Its ubiquitous presence is thought to prevent “soft” tissue calcification, whereas its degradation to Pi in bones and teeth by alkaline phosphatase (ALP) may facilitate crystal growth. While the inhibiting role of PPi on osteoblast differentiation and function is largely understood, less is known about its effects on osteoclast determination and activity. In this study, we investigated the role of PPi in bone resorption using calverial organ cultures ex vivo. We present an evidence that PPi stimulated calvarial bone resorption marked by calcium (Ca2+) release in the condition media (CM). We then examined PPi effects on osteoclast differentiation using mouse bone marrow-derived monocytes (BMMs). Our results revealed that PPi enhanced osteoclast differentiation ex vivo, marked by increased number and size of TRAP-stained mature osteoclasts. Moreover, PPi stimulated osteoclastogenesis in BMMs co-cultured with osteoblasts. These data supported the increased osteoclast activity in bone resorption using functional osteo-assays. The finding of PU.1-Egr-1 dependent up-regulation of c-FMS and RANK receptors in BMMs supported the enhanced pre-osteoclast commitment and differentiation. Moreover, osteoclast survival was enhanced by activation of MITF-BCL-2 pathway that was mediated by MAPK-ERK1/2 signaling. Last, our data showed that PPi up-regulated ANK; PPi transporter, during osteoclast differentiation through ERK1/2 phosphorylation whereas mutation of ANK inhibited osteoclastogenesis. Collectively, our data suggest that PPi promotes osteoclast differentiation, survival, and function through PU.1 up-regulation and MITF phosphorylation whereas ANK loss-of-function inhibited osteoclastogenesis.


INTRODUCTION
Mineralization of the mammalian skeleton is a tightly regulated process, in which the key regulators play critical roles in defining the site, timing and the extent of mineral deposition. In bone, calcium (Ca 2+ ) and phosphate-based apatite crystal formation is initiated by membrane bound matrix vesicles, deposited to the type I collagen-rich organic matrix and propagate along the collagen fibrils (4,35). Unlike in the highly insoluble exoskeleton of numerous invertebrates, the calcium phosphate apatite crystal in mammalian skeleton is not a perfect crystalline with its stability subjected to a local pH change (46). A slight shift from neutral to acidic pH converts bone apatite crystals to soluble ions, and vice versa. The dynamic nature of vertebral skeleton is important not only to support homeostasis of serum mineral ions but also to facilitate bone remodeling, allowing enlargement, contouring and repair of bones (46).
Inorganic pyrophosphate (PPi), which consists of two molecules of inorganic phosphates connected by a high-energy ester bond, is best known for its ability to inhibit tissue calcification (49). In mammalian cells, PPi does not appear to be synthesized de novo. Instead, it is generated as an abundant byproduct or metabolic intermediate of numerous intracellular and extracellular biochemical reactions (55). Intracellular PPi is generated and stored largely in mitochondria, but it is also detected in endoplasmic reticulum and Golgi (27). The extracellular PPi concentration in the skeletal tissue is determined by several types of plasma membrane proteins: ectonucleotide pyrophosphatase (ENPP), which generates PPi from ATP (51), tissue non-specific alkaline phosphatase (TNAP), which hydrolyzes PPi into two inorganic phosphates (Pi) (58), and progressive ankyloses (ANKH) transporter; which is involved in PPi in-and efflux (8).
While the functional role of intracellular PPi in mammalian cells remains elusive, extracellular PPi has been extensively studied for its inhibitory role in tissue calcification (28). Extracellular PPi directly binds to the surface of basic calcium phosphate hydroxyapatites and interferes with the propagation of crystal formation, contributing to the formation of poorly ordered bone crystal structure (3). In addition, exogenous PPi at micro-molar concentrations stimulates the expression of osteopontin; a negative regulator of mineralization that also inhibits the PPi hydrolyzing activity of TNAP in MC3T3-E1 osteoblast cultures (3).
The physiological significance of extracellular PPi in tissue calcification is observed in various human pathologies and genetically modified mouse strains with dysregulated PPi metabolism, where a reduction of extracellular PPi is associated with ectopic calcification of arteries (48), joint tissues (65), renal stone formation (43), and hyperostosis of craniofacial bones (50). Under physiological concentrations, it is well documented that the mineral crystals formed of extracellular PPi are mainly the basic calcium phosphate crystals (29). Increased extracellular PPi concentrations, on the other hand, is associated with hypo-mineralization of the bone matrix as in Tnap -/mice or patients with deficient TNAP activities (20). Moreover, higher concentrations of PPi results in formation of calcium pyrophosphate dihydrate (CPPD) crystals with spontaneous deposition in the articular tissues leading to an arthritis condition called; chondrocalcinosis (29).
Another syndrome characterized by PPi metabolic disorder is the craniometaphyseal dysplasia (CMD) with genetic nonsense mutation of ankh that underlie increased intracellular and decreased extracellular PPi (11). The ankh protein is type II transmembrane with [10][11][12] helices, spanning the outer cell membrane, and is associated with PPi efflux. Most of the ANKH mutations are located in cytoplasmic domains close to the C-terminus (44). Previous study reported ANKH gene mutation impairs osteoclast differentiation and function that will disrupt bone remodeling (50). Reduced bone turnover can contribute to the bone thickening characteristic of CMD. Another study showed that ANKH gene mutation increases osteoblast activity rather than a defect in bone remodeling (53). Moreover, ANKH gene mutation reduces PPi transportation out of the cells. A shortage of extracellular PPi will increase bone mineralization, which also contribute to the bone overgrowth of CMD (53). Earlier studies of CMD point to the important regulatory role of PPi in bone modeling/remodeling process (6).
Given the expanding functional implication of PPi in bone remodeling, it is imperative to test the role of PPi in bone resorption. We hypothesize that extracellular PPi has a direct stimulating effect on osteoclasts, promoting differentiation of bone marrow-derived monocytes (BMMs) into mature osteoclasts that are active in bone resorption. In this study, murine calvarial organ cultures and BMMs have been essentially used to determine the potential role of exogenous PPi in pre-osteoclast differentiation, and function. Here, we report that PPi enhances commitment and survival of BMMs into mature osteoclasts in a PU.1 dependent up-regulation of EGR-1, and MAPK-dependent phosphorylation of MITF. Thus, our data suggest PPi works as a positive regulator of osteoclastogenesis and osteoclast-mediated bone resorption.

Calvarial Organ Culture.
Intact cranial vaults that include frontal, parietal and interparietal bones were dissected from C57BL/6J mice, postnatal day 6. They were placed with the concave side down on a stainless steel grids in 12-well, and cultured in a-MEM containing 0.1% BSA, 50µg/ml ascorbic acid, and antibiotics-antimycotic (50 µg/ml penicillin, 50 µg/ml streptomycin, and 0.25 µg/ml amphotericin B) for 24 hours. Murine calvaria were washed with phosphate buffered saline solution (PBS) and cultured for additional three days in the fresh medium containing various combinations of chemical and biochemical reagents, as indicated in the Result section. The calvarial organs were then assessed for osteoclast bone resorption, as described below.

ANK-mutant Mice.
C3FeB6 A/A W-J -ANK ank /J: (ank/ank) mice, and their wild-type: C3FeB6 A/A W-J were obtained from the Jackson Laboratory, JAX stock #000200 (Bar Harbor, ME). Heterozygote breeders were used to generate and study ank/ank littermates, with genotypes analyzed by RT-PCR, as described previously (23). All mouse colonies were maintained in The Children Hospital of Philadelphia facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), under veterinary supervision and according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cell Culture.
Mouse macrophage cell line RAW 264.7 (ATCC, Manassas, VA) was maintained in DMEM containing 10% FBS supplemented with 1% nonessential amino acids, 1% L-glutamine and antibiotics-antimycotic. For osteoclast differentiation, the RAW cells were cultured for 7 days in the medium supplemented with 35 ng/ml RANKL. RAW 264.7/C4 cells are stably transfected cell line for MITF expression (RAW MITF ) was a kind gift from Dr. David Hume (University of Edinburgh). RAW MITF cells were maintained in a complete DMEM media as described above with geneticin (450 μg/ml), added for selection.
Mouse bone marrow-derived monocytes (BMMs) were isolated from 8-week old C57BL/6J, C3FeB6A/A W-J -ANK ank /J: (ank/ank) male mice, and their wild-type: C3FeB6A/A W-J , as described previously (2). Briefly, Mice were euthanized, and their tibiae and femurs were dissected using sterile technique, scraped off adherent soft tissue, and placed in ice-chilled complete media (a-MEM containing 10% FBS, and antibiotics-antimycotic). The ends of the bone were cut off, and the bone marrow was flushed out from the shaft, using PBS in syringe with a 26-gauge needle. The extruded bone marrow was triturated by repeatedly pipetting up and down, and centrifuged. The cell pellet was re-suspended in complete medium and cultured overnight. Next day, non-adherent BMMs were collected, passed through a 40 µm nylon mesh cell strainer, counted, and plated for future experiments.
For the experiments requiring NaPPi pretreatment, BMMs were plated at 2x10 5 cells/well in a 96-well and cultured for 48 hours in a complete medium supplemented with 25 ng/ml MCSF and PPi (100-1000 µM). The pretreated cells were washed with PBS and switched to a complete medium supplemented with 35 ng/ml RANKL in addition to 25 ng/ml MCSF. Cells were fed every two days by replacing 50% of the medium with a fresh complete medium containing 2X concentration of RANKL and MCSF. After 7-day culture period, cells were analyzed for TRAP activity and TRAP+ osteoclasts as described below.
For the co-culture experiments requiring NaPPi pretreatment, primary calvarial osteoblasts were isolated as described previously (1). Osteoblasts were plated at 1.

Enzyme Immunoassay (EIA) for Collagen Helical Peptide (CTX-1).
The MicroVue™ Helical Peptide EIA kit (Quidel, San Diego, CA) was used to measure the level of a helical peptide containing the residues 620-633 of the type I collagen a1 chain or more formally carboxy-terminal collagen crosslinks, known by the acronym (CTX-1) in conditioned media from cell and organ cultures, as instructed by the manufacturer. Each treatment group consisted of 3-replicates.

Quantitative TRAP Activity Solution Assay and TRAP Staining.
BMM and osteoclasts cultures in a 96 well plate were fixed with 10% formalin for 15 minutes,

ex vivo Assays for Osteoclastic Bone Resorption.
BMMs were cultured for 48 hours in the OsteoAssay™ Human Bone Plate (Lonza, Walkersville, MD) coated with a thin layer of adherent human bone particles in complete medium containing 25 ng/ml MCSF and PPi (100-1000 µM). BMMs were then switched to RANKL and MCSF for additional 10 days as described above. Bone resorption was assessed by measuring Ca 2+ and type I collagen helical peptide 620-633 (CTX-1) in conditioned media harvested at the end of culture.
In parallel experiments, murine BMMs were cultured with MCSF and NaPPi for 48 hours and then differentiated with RANKL for additional 8 days on BD BioCoat™ Osteologic™ MultiTest Slides (BD Biosciences, Bedford, MA) coated with a synthetic calcium phosphate film. The cell culture condition was same as that for the OsteoAssay™ Human Bone Plate. At the end of culture, multi-well slides were treated with 1 M NH4OH to remove cells, washed 3 times with ddH2O, and processed for von Kossa staining and visualized using Olympus IX71 microscope.
The total number and area of resorption pits were captured and analyzed using image J program as described below.
In parallel experiments, co-culture of murine BMMs and osteoblasts were established on dentin discs (OsteoSite, Boldon, UK) in 96-well and treated as described above. On day 12, cells were removed with 1 M NH4OH and discs were stained with 1% toluidine blue. The total bone resorption area was captured and analyzed using image J program as described below.
Cell survival Assays -Cell survival of BMMs and RAW MITF cells were measured after 48 hours treatment with NaPPi and/or U0126 or SB203580 in complete or serum-free medium as described previously. Twelve mM of MTT substrate (Thiazolyl Blue Tetrazolium Bromide) (Sigma) was added to the 24-well plate and was incubated at 37 o C. Dimethyl sulfoxide (DMSO); (200 μl/well) was added to solubilize the formazan reaction and read for optical density at 570 nm.

Northern Blot Analysis.
Total RNA was isolated from cells using TRIzol® Reagent (Invitrogen) and confirmed for RNA integrity on an agarose gel. Five µg of RNA was separated on a 1% denaturing agarose gel, containing 2.2 M formaldehyde, and transferred onto a nylon membrane (Maximum Strength Nytran, Schleicher & Schuell, Keene, NH). The membrane was hybridized with cDNA probes for mouse ANK labeled with [a-32 P] dCTP using Rediprime TM II (Amersham Biosciences, Piscataway, NJ), washed, and exposed on X-ray film (Denville Scientific, Metuchen, NJ).

RT-PCR.
BMM cultures were treated with NaPPi at various concentrations (100-1000 µM) for 48 hours, as described above. Total RNA was isolated from cells using TRIzol® Reagent (Invitrogen) and confirmed for integrity on an agarose gel. Three μg of total RNA were reverse transcribed to cDNA in a volume of 10 μl containing the following: 150 ng Random Hexamers, 10 mM dNTP  Table 1.
Immunoblot Analysis -Protein was isolated in RIPA buffer (Sigma) and subjected to SDS-PAGE as described previously. Briefly, proteins isolated from primary cultures were mixed with Laemmli denature buffer (Bio-Rad, Hercules, CA) and heated at 100°C for 5 minutes. Samples were subjected to 10% SDS-PAGE in 1× TGS (Bio-Rad) for one and half hour with current settings at 30 mA per gel. Gel was then transferred to Hybond ECL nitrocellulose membrane (Amersham Bioscience, UK) by semi-dry transfer apparatus (Bio-Rad) for two hours with current settings at 50 mA per gel. The blot was incubated in blocking buffer (5% skim milk in TTBS) (Bio-Rad) for one hour. The blot was then incubated with the primary antibody diluted in fresh blocking buffer (1:1000) overnight at 4°C. The blot was washed in 1X TTBS for 3 times, 5 minutes each, and then incubated with HRP-conjugated secondary anti-mouse or anti-rabbit antibody diluted in fresh blocking buffer (1:5000), for one hour. The blot was washed in TTBS for 3 times and the signal was developed using Western Lightening Plus-ECL kit (PerkinElmer, Waltham, MA) and detected on HyBlot CL X-ray films (Denville).

Morphometric Image Analysis.
All images were captured using a Canon EOS 5D 12.8 MP digital SLR camera with EF 24-105 mm f/4 L IS USM Lens. All images were analyzed with Java-based image-processing program developed at the NIH and the Laboratory for Optical and Computational Instrumentation, Image J (v. 1.49o). The percent area fraction of TRAP positive staining was calculated for calvarial organ cultures as follow: dividing six randomly selected area fraction (The percentage of pixels in the selected image that have been highlighted in red), at or above the limit of threshold, by the same area fraction for total number of pixels through the entire field. Each treatment group consisted of three different murine calvaria.
Osteoclast activity of cells on synthetic calcium phosphate and dentin discs was assessed by counting resorption pits from six randomly selected fields per well, three wells per treatment group. The total resorption area was computed using the area measurement function in the image J program.

Statistical Analysis.
Statistical analyses were performed using Prism 4 (GraphPad Software, La Jolla, CA). one-way ANOVA was used to analyze quantitative PCR, semi-quantitative RT-PCR, Northern, and Western blot densitometry results for differences in analytes between groups. Then, post hoc analyses were carried out using the Bonferroni test for multiple comparisons and adjusted p values are reported. An adjusted p value of < 0.05 was considered significant for all analyses. A p-value of less than 0.05 was considered statistically significant. Data are presented as the mean + standard error (SE) of at least three independent assays.

Pyrophosphate (PPi) stimulates bone resorption ex vivo. Previous reports investigated the
role of PPi in skeletal hypo-mineralization (21). In this study we asked whether exogenous PPi has a direct impact on bone resorption, using ex vivo murine calvarial bone dissected from C57BL/6 mice at postnatal day 6. The organ cultures were treated with sodium pyrophosphates (Na PPi) for 72 hours at final concentrations of (0-1000) µM and assayed for the Ca 2+ level in the conditioned media (CM). Lower doses of Na PPi <100 µM have no effect in calvarial bone resorption (data not shown), however higher doses of Na PPi ≥100 µM promoted osteolysis. As shown in Fig-1A, PPi treatment increased the Ca 2+ release from calvarial bone cultures into the media in a dose dependent manner (p< 0.01), compared to the untreated controls, suggesting that exogenous PPi has a calciotropic role. To verify the previous data, we measured CTX-1; an indicator for osteoclast activity, in conditioned media. PPi treatment increased the level of CTX-1 (p< 0.01), compared to the untreated controls (Fig-1B). These data were confirmed by the whole mount TRAP staining of calvarial organ cultures. PPi treatment increased the TRAP staining intensity (Fig-1C). Calculated percent area fraction of the stained calvaria confirmed the increased TRAP staining of PPi treated calvaria in a dose dependent manner (p< 0.01) (Fig-1D).
Moreover, calvarial organs treatment with PPi (1000 µM) increased size of TRAP positive multinucleated osteoclasts in the whole mount calvaria, compared to untreated control (Fig-1E). Quinazoline inhibited PPi-mediated bone resorption marked by reduced Ca 2+ levels in a dose dependent manner (p< 0.01), compared to calvarial treatment with PPi alone (Fig-1G). Taken together, these data provide an evidence that exogenous PPi stimulates osteoclast activity in bone resorption.
To investigate the calciotropic role of PPi on cytokines-mediated bone resorption, we measured Ca 2+ levels in conditioned media of calvaria treated with cytokines and/or PPi. Interestingly, PPi enhanced TNFα-mediated Ca 2+ release in a dose dependent manner (p< 0.01), compared to TNFα or PPi treated controls (Fig-2A). In contrast, the pyrophosphate analogue; disodium clodronates inhibited TNFα-mediated Ca 2+ release in a dose dependent manner (p< 0.01), compared to TNFα treated controls (Fig-2B).
To exclude the possibility of Pi action in PPi-stimulated bone resorption, calvarial bones were treated with sodium phosphates (Na Pi) for 72 hours and assayed for the Ca 2+ level in conditioned media. Interestingly, Na Pi treatment shifted the Ca 2+ off media in a dose dependent manner (p< 0.001), compared to controls (Fig-2E). This effect was reversed when PFA was added to Pi-treated calvarial bones (Fig-2F). These data restrain the possibility of PPi hydrolysis into Pi in the culture media, and also constrain the effect of Pi in PPi-mediated Ca 2+ release.
To further confirm the potential role of PPi in bone resorption, calvarial organ cultures were treated with probenecid; an anion transport inhibitor, that decrease PPi entrance intracellular. Interestingly, probenecid inhibited PPi-mediated Ca 2+ release in a dose dependent manner (p< 0.001), compared to PPi treated controls (Fig-2G). Taken together, these results unleash the authentic role of Na PPi in stimulation of bone resorption.

Pyrophosphate enhances osteoclast differentiation and function ex vivo.
In the following study, we investigated whether PPi has an effect on bone marrow-derived monocytes (BMMs) to differentiate into mature osteoclasts. Quantitative PCR analysis showed an increased TRAP expression in BMMs treated with exogenous PPi in a dose dependent manner (p< 0.01), compared to control (Fig-3A). Next, BMMs were treated with PPi (100-1000) µM and stained for TRAP. As shown in Fig-3B, BMMs treated with PPi demonstrated an increased number of TRAP positive mononuclear cells (osteoclast precursors; pre-osteoclasts) compared to control. Consistent with the previous data, TRAP activity was higher in BMMs treated with PPi (p< 0.001) (Fig-3C). Next, we examined osteoclast differentiation in BMMs treated with PPi.
Interestingly, BMMs pretreated with PPi showed a significant osteoclast differentiation, in response to RANKL, as indicated by larger TRAP-positive osteoclasts, compared to untreated counterparts (Fig-3D). These findings were supported with significant increase in TRAP activity, number and size of osteoclasts differentiated from PPi-treated BMMs (p < 0.001) compared to control (Fig-3E, F, G, and H). To further support our data, BMMs were treated with PPi and differentiated on glass slides then stained with Rhodamine for actin cytoskeleton. Consistently, the size of osteoclast was larger in BMMs pretreated with PPi then differentiated with RANKL compared to untreated controls (Fig-3I). Moreover, the acting ring was sharp and prominent in osteoclasts treated with PPi compared to control. These data suggest that PPi treatment enhance osteoclastogenesis.
Given that PPi promoted osteoclast differentiation, we explored the osteolytic activity of those mature osteoclasts. BMMs treated with PPi were differentiated into osteoclasts in response to RANKL, over human bone chips. Photomicrographs of bone chips showed large TRAP-positive osteoclasts with decrease of bone chips fragments in BMMs treated with PPi compared to untreated controls (Fig-4A). These findings were confirmed with marked increase in TRAP activity, Ca 2+ and CTX-1 levels in the CM of osteoclasts promoted from BMMs treated with PPi (p < 0.001) in a dose dependent manner (Fig-4B, C, and D).
These data were supported by testing the osteolytic activity of PPi promoted osteoclasts on bone discs. Photomicrographs of bone discs, after removal of osteoclasts, showed large pit area by PPi-promoted osteoclasts (Fig-4E). These findings were supported by increased number of resorption pits, total resorbed surface area, and the average pit size in PPi-promoted osteoclasts (p < 0.01) in a dose dependent manner (Fig-4F, G, and H). Collectively, these results show that PPi-promoted-osteoclasts have a robust activity in bone resorption. Taken together, these data provide a pivotal evidence on the regulatory role of PPi in osteoclast differentiation and function.

Pyrophosphate stimulates osteoclast differentiation and function in co-culture system.
As PPi stimulated bone resorption in murine calvarial organs in the data described above, we investigated the contribution of osteoblasts in PPi-enhanced pre-osteoclasts and osteoclastogenesis. Co-culture system of murine osteoblasts and BMMs was established and treated with vitamin D3 (positive control) or PPi (100-1000) µM and stained for TRAP.
Interestingly, BMMs co-cultured with osteoblasts and treated with PPi demonstrated obvious osteoclast differentiation indicated by large TRAP+ mature osteoclasts compared to untreated control (Fig-5A). These data were supported with marked increase in TRAP activity, number and size of osteoclasts differentiated from PPi-treated co-cultures (p < 0.001) compared to control (Fig-5B, C, and D). Surprisingly, the number and size of TRAP+ osteoclasts was not different (p > 0.05) in PPi-treated co-culture, compared to vitamin D3 positive control.
Given that PPi promoted osteoclast differentiation in co-culture system, we test the osteolytic activity of those mature osteoclasts. The co-culture system of murine Osteoblasts and BMMs was established on dentin discs and treated with vitamin D3 or PPi (1000) µM. Photomicrographs of dentin discs displayed an obvious bone resorption marked by large resorption pits and tracks in PPi-treated co-cultures compared to untreated controls (Fig-5E). The results were confirmed with marked increase (p < 0.001) in CTX-1 levels in the CM (Fig-5F) and the average total resorption area of dentin discs (Fig-5G) in PPi-treated co-cultures compared to untreated control.
Interestingly, the dentin discs displayed more complex bone resorption tracks in PPi-stimulated osteoclasts, compared to vitamin D3 positive control in the co-culture system (Fig-5E). These data suggest that PPi-stimulated osteoclastogenesis and bone resorption is mediated at least in part by osteoblasts.
To gain insight on the underlying mechanism of PPi-promoted osteoclasts in co-culture, we examined RANKL and OPG gene expressions in murine calvrial osteoblasts treated with PPi osteoblast CM compared to the untreated control (data not shown). Taken together, these data showed a clear evidence that PPi increased RANKL by osteoblasts that is responsible, at least in part, for PPi-stimulated osteoclastogenesis and bone resorption.

Activation of RANK-RANKL pathway in Pyrophosphate-enhanced pre-osteoclasts.
Next, we explored the down-stream signaling pathway that mediates enhanced osteoclastogenesis in PPi-treated BMMs. In this study, BMMs-treated PPi were analyzed for the early transcription factors responsible for BMM commitment, using semi-quantitative RT-PCR (RT-PCR) analysis. The transcription factors; PU.1 and MITF responsible for hematopoietic stem cells commitment into osteoclast progenitors (47), showed an up-regulation in PPi-treated BMMs in a dose dependent manner (Fig-6A). Similarly, the cFMS (MCSF receptor) and RANK receptor gene expressions were also increased in PPi-treated BMMs (Fig-6A). These data were confirmed by quantitative (q)RT-PCR that showed up-regulation (p < 0.01) of cFMS and RANK in PPi-treated BMMs (Fig-6B and C). Moreover, the transcription factors; c-Fos and c-Jun (AP-1 heterodimeric complex); downstream of MAPK signaling, also showed higher expression in PPitreated BMMs (Fig-6D and E). By examining the NF-кB mRNA, our findings showed a ~2-fold increase of NF-кB (P50/P65) expression in PPi-treated BMMs compared to untreated control (Fig-6F).
Next, we examined activation of MAP-Kinase (ERK1/2 and P38) which is downstream of RANK receptor in osteoclast differentiation (37). BMMs were treated with PPi and/or ERK1/2 inhibitor (U0126) and/or P38 inhibitor (SB203580) for up to 90 minutes. Western blot data showed increased phosphorylation of ERK1/2 (pERK) in PPi-treated BMMs reaching the maximum at 10 minutes, compared to untreated control (Fig-6G). On the other hand, U0126 inhibited PPi effect on ERK phosphorylation at 5 and 10 minutes (Fig-6G). Densitometric analysis confirmed ~3-fold inhibition of ERK phosphorylation in U0126-treated conditions, compared to PPi treatment alone (Fig-6H). Similarly, phosphorylation of P38 was increased significantly in PPi-treated BMMs, compared to untreated control, whereas SB203580 inhibited PPi effect on P38 phosphorylation (Fig-6I). Densitometry showed approximately ~2-fold decrease of P38 phosphorylation in SB203580-treated conditions, compared to PPi treatment alone (Fig-6J). Collectively, these data support the activation of downstream signaling cascade of RANK and MAPK in PPi-treated BMMs.

Pyrophosphate regulates PU.1 signaling pathway in BMMs.
As PU.1 is an early transcription factor in BMM pre-osteoclasts and has a pivotal role in regulation of cFMS; MCSF receptors (38), we investigated the role of PPi in the downstream signaling pathway of cFMS in PPi-treated BMMs. Interestingly, qRT-PCR showed that PU.1 is up-regulated (~2.5-fold) in PPi-treated BMMs in a dose dependent manner (Fig-7A). Next, we examined the transcription factors; Egr-1 and Egr-2 that have been reported as downstream mediators of PU.1 in BMMs (16). qRT-PCR analysis of Egr-1 gene expression was ~2.5-fold higher in PPi-treated BMMs in a dose dependent manner (Fig-7B). In contrast, Egr-2 expression was not different in PPi-treated BMMs and control (Fig-7C). To support our data and define the precise role of PPi on PU.1, we treated PUER cells with PPi and examined cFMS and RANK, Egr-1, and Egr-2 expressions. PUER cells are murine hematopoietic precursor cells that have been retro-virally transduced to express PU.1 fused to the estrogen receptor (ER) (57). PUER hi cells was used as appositive control to PUER lo cells. RT-PCR data showed marked up-regulation of cFMS and RANK expressions, when hydroxytamoxifen (OHT) is added to PUER hi cells, compared to untreated PUER hi cells (Fig-7D). On the other hand, PUER lo treated with PPi showed relative up-regulation of cFMS and RANK expressions compared to untreated PUER lo control by RT-PCR (Fig-7D). Densitometric analyses of RT-PCR data demonstrated ~2 to 6-fold increased levels of cFMS and RANK in PPi-treated PUER lo cells compared to untreated PUER lo control (Fig-7E, F). Interestingly, expression levels of cFMS and RANK were relatively similar (p > 0.05) in PPi-treated PUER lo cells and untreated PUER hi cells (Fig-7D, E, F).
Next, we characterized expression of Egr-1 and -2 in PUER lo and PUER hi cells.
Interestingly, Egr-1 expression was up-regulated in PPi-treated PUER lo cells, compared to untreated PUER lo control Fig-7D). These data were verified by densitometric analysis as ~3-fold increased level of Egr-1 in PPi-treated PUER lo cells (Fig-7G). However, Egr-2 expression no significant difference in PPi-treated PUER lo cells, untreated PUER lo cells, and OHT treated or untreated PUER hi cells (Fig-7D). These data were confirmed by densitometric analysis that showed relatively equal level of Egr-2 (p > 0.05) in all treated and untreated conditions (Fig-7H).
Taken together, we conclude that PPi stimulates PU.1 and its downstream signaling in osteoclast progenitors.

Pyrophosphate enhances differentiation and survival of pre-osteoclasts in MITF-dependent pathway.
Because we showed above that PPi up-regulated MITF expression in BMM differentiation, we investigated if MITF signaling pathway might be involved in BMM survival as well. MITF is a key transcription factor, responsible for BMM commitment into osteoclast progenitors (41), we verified MITF expression in PPi-treated BMMs. qRT-PCR data showed a dose-dependent increase of MITF expression (p < 0.01) in PPi-treated BMMs, compared to untreated control (Fig-8A). MITF has several isoforms where MITF-A and -E have been reported to regulate macrophage polarization and osteoclastogenesis, respectively (5). Semi-quantitative RT-PCR data showed relevant higher expressions of MITF-A and -E in PPi-treated BMMs in a dose dependent manner, compared to untreated control (Fig-8B). These data were confirmed by densitometry analysis that showed ~6 to 10-fold increase in MITF-A and MITF-E in PPi-treated BMMs compared to untreated control (Fig-8C, D). Next, we examined survival of osteoclast progenitors, where BMMs were primed with MCSF then treated with PPi in serum deprived conditions for 48 hours. Interestingly, BCL-2, a pro-survival transcription factor, showed a dosedependent increase (p < 0.01) in PPi-treated osteoclast progenitors compared to untreated control (Fig-8E). To study the downstream mechanism by which PPi stimulates BCL-2 expression, we examined BCL-2 in RAW 264.7 macrophage-cell lines, treated with PPi in comparison to RAW cells over-expressing MITF. Interestingly, PPi stimulated ~2-fold increase of BCL-2 in wild-type RAW cells compared to untreated control (Fig-8F). Moreover, BCL-2 expression in PPi-treated wild-type RAW cells was comparable to its expression in RAW cells over-expressing MITF.
Collectively, these data provide a supportive evidence that PPi stimulates survival of osteoclast progenitors, MITF and BCL-2, an effect mediated by MAPK signaling pathway.
Several studies affirmed the regulatory effects of MAP-Kinase on MITF in osteoclast progenitors (41). In this study, we examined if activation of MAPK controls survival of osteoclast progenitors, BMMs were primed with MCSF then treated with PPi and/or ERK-1 and -2 inhibitor (U0126), and/or P38 inhibitor (SB203580), in serum deprived conditions. Survival of osteoclast progenitor was enhanced (p < 0.01) when treated with PPi compared to untreated control (Fig-8G). However, treatment of osteoclast progenitors with PPi and U0126 decreased osteoclast survival significantly (p < 0.01) compared to PPi alone and comparable to untreated control (Fig-8G). On the other hand, SB203580 has no effect on PPi-enhanced survival of osteoclast progenitors, compared to PPi treatment alone. Moreover, U0126 has a dominant inhibitory effect (p < 0.05) on survival of osteoclast progenitors when treated with PPi and SB203580 (Fig-8G). To confirm that MAPK regulates MITF-mediated survival of osteoclast progenitors, we treated wild-type RAW cells and RAW cells overexpressing MITF with PPi and/or U0126, and/or SB203580 in serum deprived conditions. Consistent with the previous data, PPi treatment increased survival of wild-type RAW cells (p < 0.05) compared to untreated control and comparable to RAW overexpressing MITF (Fig-8H). Moreover, U0126 inhibited survival of PPi-treated wild-type RAW cells (p < 0.05) compared to PPi alone or RAW overexpressing MITF controls. However, U0126 didn't inhibit survival of RAW overexpressing MITF. Following, SB203580 had no effects neither on PPi-treated wild-type RAW cells nor on RAW overexpressing MITF (Fig-8H). Combined treatment with U0126 and SB203580 reduced survival of PPi-treated wild-type RAW cells (p < 0.05) compared to untreated control. However, U0126 and SB203580 didn't affect survival of RAW overexpressing MITF (Fig-8H). To support our conclusions, we tested MITF phosphorylation in osteoclast progenitors by Western blot.
Osteoclast progenitors were primed with MCSF then serum starved and treated with PPi and/or U0126 for the indicated time points. Interestingly, PPi treatment increased MITF phosphorylation on the Serine 307 residue at all-time points reaching the maximum at 90 minutes, compared to 0minute control (Fig-8I). In accordance with the previous data, U0126 inhibited PPi effects on MITF phosphorylation. Treatment of RAW cells with PPi and/or U0126 showed similar results (data not shown). Densitometric analysis of the Western blot demonstrated ~4-to 8-fold increase in MITF phosphorylation at 60 and 90 minutes in PPi-treated osteoclast progenitors compared to 0-minute control (Fig-8J). Moreover, U0126 inhibited MITF phosphorylation significantly compared to PPi alone at 60 and 90 minutes. Taken together, we conclude that PPi stimulates survival of osteoclast progenitors and MITF phosphorylation through ERK-1/2 signaling pathway.

Role of ANK in Pyrophosphate-mediated osteoclastogenesis.
The data above described the roles of PPi in BMM survival, differentiation, and osteoclast function, so we pursued to examine the importance of ANK in regulation of PPi effects on osteoclastogenesis. ANK has been reported in various studies to transport inorganic PPi across the cell membrane (64). In the beginning, we examined ank mRNA during osteoclast differentiation by Northern blot. Expression of ank mRNA demonstrated a temporal increase during osteoclast differentiation, coming from BMMs and RAW 264.7 cells, respectively (Fig-9A). Densitometry showed a significant increase of ank expression during osteoclast differentiation reaching the maximum at 11 (p < 0.01) with BMMs and 6 days (p < 0.05) with RAW cells, respectively (Fig-9B). Western blot supported the significant increase in ANK levels in BMMs differentiated into osteoclasts (data not shown). Consistent with the previous data, ANK protein showed a temporal increase in osteoclasts differentiated from RAW cells reaching the highest level at day 6 (p < 0.01), compared to 0-undifferentiated control (Fig-9C, D). Next, we tested if PPi regulates ank in osteoclast progenitors. Interestingly, qPCR analysis of PPitreated BMMs showed a dose dependent increase in ank expression (p < 0.01), (Fig-9E).
To characterize the calciotropic role of PPi transporter; ANK, we treated calvarial bone organs with PPi and/or probenecid; an anion ANK inhibitor (12). PPi stimulated bone resorption marked by Ca 2+ release (p < 0.01), compared to untreated control (Fig-9F). However, probenecid inhibited PPi-mediated Ca 2+ release in a dose dependent manner (p < 0.001). To support these data, we treated BMMs with PPi and/or probenecid then osteoclast progenitors were differentiated with RANKL. TRAP activity was increased (p < 0.01) in osteoclast differentiated from PPi-treated BMMs, compared to RANK-L alone (Fig-9G). However, combined treatment of BMMs with PPi and probenecid decreased TRAP activity of osteoclasts, compared to PPi treatment alone (p < 0.05). To confirm the selective role of ANK in PPi transport during osteoclast differentiation, we used BMMs isolated from ank/ank mice and their wild-type littermates. ank/ank mice are characterized with nonsense mutation of ank gene leading to loss of function in active PPi transport (15). Treatment of ank/ank BMMs with PPi didn't have an effect, compared to upstroke increase (p < 0.01) in osteoclast differentiation in their wild-type PPitreated BMMs (Fig-9H). To investigate if ANK-PPi transport regulates ERK-1/2 phosphorylation, WT-ank BMMs were treated with PPi and/or probenecid in serum-deprived conditions. probenecid inhibited PPi-mediated ERK-1/2 phosphorylation in WT-ank BMMs, compared to PPi-treatment alone (Fig-9I). Densitometry showed a significant reduction (p < 0.05) of ERK-1/2 phosphorylation in probenecid/ PPi treated WT-ank BMMs at 5 and 10 minutes (Fig-9J). Taken together, these data provide a strong evidence that ANK transporter regulates PPi effects during osteoclast differentiation.

DISCUSSION
The role of pyrophosphates (PPi) in tissue remodeling has been controversial. PPi preserve the differentiated phenotype of articular chondrocytes, whereas several studies have documented PPi generation, transport, and activity in vivo, most notably for chondrocytes and the tissues of joints (8). It is well established that pyrophosphates (PPi) inhibit tissue calcification in normal conditions (45,56). In bone, several studies reported PPi as a negative regulator of osteoblast differentiation and matrix mineralization (3,24). PPi was reported to inhibit mineralization by direct binding to hydroxyapatite crystals and inhibiting alkaline phosphatase activity (ALP) (3).
However, the significance of PPi in osteoclast differentiation and function was yet to be determined. This led us to further investigate the impact of PPi on osteoclastogenesis and so bone resorption.
In this study, we examined the direct effect of PPi on calvarial bone resorption and deficiency block pre-osteoclast differentiation and osteoclast function (9). In our study, upregulation of Egr-1 was evidenced in PPi-treated BMM cultures. Conversely, Egr-2 expression was not altered in PPi-treated BMMs. One study identified Egr-2 as a negative modulator of osteoclast differentiation using gain-of-function and loss-of-function approaches (31). On the other hand, Bradley and her group claimed the importance of Egr-2 in promotion of osteoclast survival through activation of MEK/ERK-dependent pathway (7). Our data suggest that stimulation of PU.1 dependent pathway in PPi-treated BMMs, resulted in osteoclast differentiation in response to RANKL. As enhanced OCL survival can possibly add to the increased number and size of osteoclasts, we investigated BMM survival programs in these PPi-treated BMMs. Our data showed that survival of BMMs is enhanced and mediated by MAPK-MITF phosphorylation. Expression analysis of overall MITF mRNA was also high in PPi-treated BMMs. There are 7 isoforms of MITF (40). Previous studies showed that MITF-A isoform is expressed in similar amounts in macrophages (BMMs) and osteoclasts, however MITF-E isoform is almost non-noticeable in BMMs, and its expression is considerably up-regulated during osteoclastogenesis (39). Moreover, MITF activation by serine phosphorylation at residue 307 was reported to recruit P38 MAPK and NFATc1 target genes during osteoclastogenesis (52).
Because enhanced osteoclast survival could possibly contribute to increased osteoclast number and size, we investigated osteoclast survival pathway in these PPi-treated conditions. Our data showed that PPi stimulated pre-osteoclast survival and BCL-2 mediated by MAPK signaling pathway. It is well supported the crucial role of BCL-2 as anti-apoptotic in osteoclastogenesis and osteoclast survival (42). Indeed, our data supported up-regulation of BCL-2 in PPi-treated pre-osteoclasts. Yamashita et al., reported the critical importance of BCL-2 in osteoclast survival using Bcl-2 −/− mice with increased bone mass due to at least in part declined osteoclastogenesis (61). Moreover, our study showed BCL-2 was downregulated in pre-osteoclasts treated with PPi and ERK inhibitor (U0126). In previous study, Subramanian and Shaha reported ERK phosphorylation up-regulates BCL-2 expression, associated with increased human macrophage survival (54). The overall cell survival of PPi treated pre-osteoclasts was decreased with ERK inhibitor, but not with P38 inhibitor. Interestingly, P38 was shown to induce the intrinsic apoptotic (Bim) molecule in human osteoclasts downstream of TGFb1-Smad-2 signaling pathway. Furthermore, TGFβ1-induced osteoclast apoptosis was declined by inhibiting the Smad pathway with abolished Bim up-regulation following TGF-β stimulation (25). Collectively, survival of PPi-treated pre-osteoclasts is enhanced, perhaps due to controlled up-regulation of MAPK-BCL-2 pathway. Next, our data showed an increased RANK signaling in PPi-treated BMMs resulted in activation of MAP kinase pathways through ERK (pERK) and P38 (pP38) phosphorylation.
Indeed, ERK inhibitor (U0126) and P38 inhibitor (SB280) decreased ERK and P38 phosphorylation in PPi-treated pre-osteoclasts. ERK1 was reported to stimulate osteoclast differentiation and their bone resorptive activity, as genetic disruption of ERK1 reduced numbers of osteoclast progenitors, and compromised pit formation (22). Moreover, it was reported that P38-MAPK activation is required for inducing osteoclast differentiation but not for their function (37).
Next, several reports well documented that RANK signaling activates MAP kinase, which in-turn activates AP-1 transcription factors (members of the Fra, Fos, and Jun) that activate downstream MITF, and thus, regulate osteoclastogenesis (36). Interestingly, our qPCR data showed up-regulation of c-Fos, c-Jun, and NFkB in PPi-treated pre-osteoclasts. A previous report has demonstrated that deletion of the c-FOS gene resulted in osteopetrosis by arresting osteoclast differentiation at the macrophage stage and this defect was completely rescued by expressing Fos protein (18). The role of PPi in activation of AP-1 warrant further investigation in future studies.
Next, it is well documented that PPi molecules are transported across the cell membrane through ank transporter (8). Specific rare genetic bone disorder known as craniometaphyseal dysplasia (CMD), characterized by progressive sclerosis and hyperostosis of the craniofacial bones has been linked to ANKH nonsense mutations resulting in increased intracellular and decreased extracellular PPi (19). Recent studies of CMD pointed to the role of PPi in regulation of bone modeling/remodeling process (26). Thus, a decrease in extracellular PPi may hinder normal bone remodeling, for instance, by inhibiting osteoclast differentiation or activity. In support of this notion, bone marrow-derived monocytes (BMMs) from a CMD knock-in mouse (ank ki /ank ki ) poorly differentiated to osteoclasts in cultures, compared to those from wild type mice (10). Consistent with the mouse data, the number of bone marrow-derived osteoclast-like cells from a CMD patient was only 40% of a normal individual, and they lacked osteoclastspecific vacuolar proton pump and the ability to absorb a dentin slice (60). Interestingly, our results showed temporal increase of ANK levels during osteoclast differentiation. Moreover, the ANK expression was escalated in PPi enhanced pre-osteoclasts that possibly indicate a positive feedback loop between PPi and ANK. In addition, probenecid; ANK inhibitor abrogated osteoclast differentiation and their function in PPi enhanced pre-osteoclasts. These data are supported by previous study showing PPi as a positive regulator of osteoclast differentiation events toward an osteoclast phenotype (32). Kim el al., reported the diminished osteoclast differentiation and activity of BMMs coming from ank/ank mutant mice (32). Since PPi stimulated temporal expression of ANK during osteoclast differentiation, and on the other hand probenecid limited PPi induced phosphorylation of ERK1/2. We conclude there's an increased possibility that PPi induces ANK expression through MAPK phosphorylation that will require further investigations. Interestingly, we showed that extracellular PPi was not able to rescue osteoclast differentiation in BMMs coming from ank/ank mutant mice compared to their wildtype littermates. These findings indicate that ANK is essential to transport PPi intracellular that up-regulate signaling pathways critical for osteoclast differentiation and thereby function. This notion is supported by the decreased osteoclastogenesis in knock-in ank ki /ank ki mice probably due to reduced pre-osteoclast fusion and migration capability; disrupted osteoclast actin ring formation; and abnormal osteoblast-osteoclast communication. In conclusion, our findings provide an evidence that PPi enhances pre-osteoclast commitment leading to increased osteoclast differentiation, survival, and function. The positive effects of PPi on pre-osteoclast commitment are mediated by PU.1-dependent up-regulation of c-FMS, RANK and MAPK-dependent phosphorylation of MITF signaling pathways (Fig-10). The PPi-enhanced preosteoclasts may be at least in part due to temporal increase of ANK expression during osteoclast differentiation, despite the fact that extracellular PPi did not rescue the osteoclast phenotype in ank/ank mutant mice. Generation of other genetically engineered mouse models, such as an osteoclast-specific knockout of ank, will help us to understand and identify the distinct signaling pathway(s) of ANK in osteoclast differentiation and thereby activity in bone resorption.