The precise regulation of Ca2+ dynamics is crucial for proper differentiation and function of osteoclasts. Here we show the involvement of plasma membrane Ca2+ ATPase (PMCA) isoforms 1 and 4 in osteoclastogenesis. In immature/undifferentiated cells, PMCAs inhibited receptor activator of NF-κB ligand–induced Ca2+ oscillations and osteoclast differentiation in vitro. Interestingly, nuclear factor of activated T cell c1 (NFATc1) directly stimulated PMCA transcription, whereas the PMCA-mediated Ca2+ efflux prevented NFATc1 activation, forming a negative regulatory loop. PMCA4 also had an anti-osteoclastogenic effect by reducing NO, which facilitates preosteoclast fusion. In addition to their role in immature cells, increased expression of PMCAs in mature osteoclasts prevented osteoclast apoptosis both in vitro and in vivo. Mice heterozygous for PMCA1 or null for PMCA4 showed an osteopenic phenotype with more osteoclasts on bone surface. Furthermore, PMCA4 expression levels correlated with peak bone mass in premenopausal women. Thus, our results suggest that PMCAs play important roles for the regulation of bone homeostasis in both mice and humans by modulating Ca2+ signaling in osteoclasts.
Bone homeostasis is maintained by the concerted action of bone-forming osteoblasts and bone-resorbing osteoclasts. In the bone microenvironment, osteoclast differentiation is governed by macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor κB ligand (RANKL) provided by osteoblasts, stromal cells, and lymphocytes (Boyle et al., 2003; Walsh et al., 2006). M-CSF ensures the survival of osteoclast precursors, and RANKL stimulates signaling pathways required for osteoclastogenesis (Boyle et al., 2003; Walsh et al., 2006). Upon maturation, osteoclasts tightly seal off bone surfaces and secrete acids and proteases to digest bone matrices.
The sophisticated control of both extracellular and intracellular Ca2+ concentrations is pivotal to the proper development and function of osteoclasts (Lorget et al., 2000; Nowycky and Thomas, 2002). During osteoclast differentiation by RANKL, cytosolic Ca2+ concentrations show an oscillatory pattern (Takayanagi et al., 2002; Asagiri et al., 2005). For the initiation of RANKL-dependent Ca2+ oscillation, the activation of PLCγ and the engagement of inositol 1,4,5-triphosphate receptor type2 have been suggested to be essential (Kuroda et al., 2008; Shinohara et al., 2008; Yoon et al., 2009). RANKL-induced Ca2+ oscillation triggers calcineurin-dependent dephosphorylation and nuclear translocation of nuclear factor of activated T cell c1 (NFATc1; Asagiri et al., 2005; Yang and Li, 2007). Because the activity of NFATc1 controls the transcription of osteoclastogenic genes (Asagiri et al., 2005; Walsh et al., 2006), the molecular mechanisms by which Ca2+ oscillations are generated and regulated are of great importance in the understanding of mechanisms for osteoclast differentiation. In addition, mature osteoclasts absorb vast amounts of Ca2+ accompanied by organic bone degradation products during resorption. Because excessive Ca2+ ions are toxic to osteoclasts, osteoclast survival is ensured by the extrusion of Ca2+ into the surrounding extracellular space via transcytosis (Salo et al., 1997) or certain Ca2+ transporters including Na+/Ca2+ exchangers (Moonga et al., 2001; Li et al., 2007).
In proteomic and genomic screening experiments, we discovered that the expression of plasma membrane Ca2+ ATPase (PMCA) isoforms 1 and 4 was dramatically increased during the late phase of osteoclast differentiation. PMCA belongs to the P-type pump family that maintains intracellular Ca2+ homeostasis by exporting Ca2+ from the cytoplasm to the extracellular space (Di Leva et al., 2008; Brini, 2009). It was reported that PMCA1 knockout is lethal to mice during early embryonic development (Okunade et al., 2004). The major observed phenotype of PMCA4 knockout mice was male infertility as a result of reduced sperm motility (Schuh et al., 2004). Besides its role as a Ca2+ pump, PMCA has been suggested to function as a signaling molecule in recent studies (Buch et al., 2005; Cartwright et al., 2009). Of particular note, PMCA4 functionally interacts with nitric oxide synthase (NOS). Overexpression of PMCA4 dramatically down-regulated NO synthesis in the ambient Ca2+ concentration where NOS operates (Schuh et al., 2001).
Here, we report that PMCAs play dual roles in osteoclast differentiation and survival by regulating RANKL-induced Ca2+ oscillations in preosteoclasts and mediating Ca2+ extrusion in mature osteoclasts. Furthermore, PMCA deficiency induced a low bone mass phenotype in mice. In addition, high PMCA4 expression levels showed a positive correlation with peak bone mass in premenopausal women. These results suggest a novel role for PMCAs in osteoclast development and bone homeostasis.
PMCAs are induced by RANKL during osteoclastogenesis
In an effort to identify new molecular players associated with osteoclastogenesis, we analyzed gene expression changes during osteoclast differentiation from human peripheral blood mononuclear cells (PBMCs) using DNA microarrays (Chang et al., 2008a). Among the genes significantly increased, ATP2B4 (gene encoding PMCA4 Ca2+ pump) was present (Fig. 1 A). Interestingly, in a proteomic study involving cell-surface protein purification and subsequent liquid chromatography/mass spectrometry/mass spectrometry (LC/MS/MS) analyses, the protein level of PMCA1 was found to become higher in mouse osteoclast precursors upon stimulation with RANKL (Fig. 1 B; Lee et al., 2008). Among the four known isoforms of PMCA (Clapham, 2007; Brini, 2009), the mRNA levels of Atp2b1 (gene encoding PMCA1) and Atp2b4 prominently increased during osteoclastogenesis from mouse bone marrow–derived macrophages (BMMs), whereas Atp2b2 (PMCA2) and Atp2b3 (PMCA3) were not detected during the entire course of osteoclastogenesis (Fig. 1 C). In contrast, the expression of isoforms of Na+/Ca+ exchanger (NCX), a family of Ca2+ transporters expressed in mature osteoclasts (Moonga et al., 2002; Li et al., 2007), either decreased (Slc8a1 [gene encoding NCX1]) or was undetectable during osteoclast differentiation (Fig. 1 D). Quantitative real-time PCR experiments (Fig. 1 E) and Western blot analyses using a pan-PMCA antibody (Fig. 1 F) further corroborated the up-regulation of PMCA at both the mRNA and protein levels during osteoclastogenesis from BMMs. Confocal microscopy studies revealed that PMCA localized to the basolateral membranes of mature osteoclasts cultured on dentin slices (Fig. 1 G). Although the question of whether PMCAs are also expressed on the apical (resorbing) membranes could not be clearly answered by confocal analyses of cells on dentin disc because of autofluorescence of dentin, exclusive basolateral localization of PMCA was evident in osteoclasts plated on glass coverslips (unpublished data).
The transcription factor NFATc1 is a key regulator of osteoclastogenesis and its expression is induced by RANKL (Takayanagi et al., 2002; Yang and Li, 2007). Notably, we found using the wed-based prediction program PROMO that the promoter regions of Atp2b1 and Atp2b4 contain several putative NFATc1 binding sites. To test the possibility of involvement of NFATc1 in the regulation of PMCA expression, we overexpressed a constitutive-active form of NFATc1 (NFATc1-CA) in BMMs. As shown in Fig. 1 H, NFATc1-CA overexpression was sufficient to induce PMCA expression in the absence of RANKL stimulation (Fig. 1 H). In addition, chromatin immunoprecipitation (ChIP) experiments revealed that NFATc1 bound to the promoter regions of Atp2b1 and Atp2b4 in RANKL-treated BMMs (Fig. 1 I). These data indicate that PMCA expression is up-regulated during osteoclastogenesis in an NFATc1-dependent manner and suggest the possibility that PMCAs extrude Ca2+ across the basolateral membrane of osteoclasts.
PMCA deficiency enhances osteoclastogenesis
Because the expression of PMCA isoforms significantly increased during osteoclastogenesis, their role for osteoclast differentiation was tested by introducing siRNA into mouse BMM osteoclast precursors. The isoform-specific knockdown of PMCA1 and PMCA4 was confirmed by RT-PCR analysis (Fig. 2 A) and Western blotting (Fig. 2 B). The PMCA siRNA-transfected BMMs were further cultured in the presence of RANKL to allow the formation of mature osteoclasts. The knockdown of PMCAs remarkably increased the formation of tartrate-resistant acid phosphatase (TRAP)–positive multinuclear osteoclasts compared with control siRNA-transfected cells (Fig. 2 C, top). The increased osteoclastogenesis by PMCA knockdown resulted in a significantly enhanced resorption (Fig. 2 C, bottom), indicating the osteoclasts generated were functionally competent. Because PMCA knockdown increased osteoclast number and size (Fig. 2 C, top), the expression of osteoclast marker genes was analyzed by quantitative real-time PCR (Fig. 2 D). The knockdown of PMCA1 and PMCA4 significantly augmented the induction of fusion marker genes Tm7sf4 (DC-STAMP) and Atp6v0d2 (V-ATPase) as well as differentiation marker genes Ctsk (cathepsin K) and Acp5 (TRAP) by RANKL stimulation. BMMs obtained from Atp2b1+/− (PMCA1-heterozygous) and Atp2b4−/−(PMCA4-null) mice also exhibited significantly accelerated osteoclastogenesis in vitro, compared with wild-type cells (Fig. 2 E). Finally, the in vivo effect of PMCA knockdown was analyzed after injecting PMCA-targeting siRNA oligonucleotides onto mouse calvariae. A microcomputed tomography (μCT) analysis (Fig. 2 F, top) and TRAP staining (Fig. 2 F, bottom) of calvarial bones clearly revealed that bone resorption as well as osteoclast formation increased significantly in the absence of PMCA, with concomitant up-regulation of the TRAP gene expression (Fig. 2 G). Similarly, enhanced osteoclastogenesis was observed when PMCA was silenced in calvariae organ culture experiments (Fig. S1).
PMCAs regulate the Ca2+ oscillation–NFATc1 axis in preosteoclasts
It has been suggested that RANKL-induced Ca2+ oscillations are crucial for the maintenance of cytosolic Ca2+ concentration required for osteoclast differentiation (Takayanagi et al., 2002; Walsh et al., 2006; Shinohara et al., 2008). Eosin (tetrabromofluorescein) has been suggested as a potent inhibitor of PMCAs (IC50 = 1 µM; Gatto and Milanick, 1993; Gatto et al., 1995). Treatment of eosin not only strongly enhanced RANKL-induced Ca2+ oscillations in preosteoclasts in vitro but also augmented the formation of osteoclasts in vivo (Fig. S2). Consistently, the knockdown of PMCA1 or PMCA4 induced robust Ca2+ oscillations in preosteoclasts (Fig. 3 A). NFATc1 is a Ca2+-dependent key transcription factor for osteoclastogenesis (Kim et al., 2005; Sharma et al., 2007). The dephosphorylation of NFATc1 by calcineurin followed by nuclear translocation and binding to its own promoter up-regulates NFATc1 expression during osteoclast differentiation through a process called auto-amplification (Asagiri et al., 2005). In accordance with the increased Ca2+ oscillations, both PMCA1 and PMCA4 siRNA significantly increased the number of cells with nuclear NFATc1, whereas the addition of a calcineurin inhibitor cyclosporine A almost completely abolished NFATc1 nuclear localization (Fig. 3, B and C). Western blotting analyses further confirmed elevated NFATc1 levels in nuclear fractions of PMCA knockdown cells (Fig. 3 D). In sharp contrast to the PMCA knockdown, the overexpression of rat PMCA1 or mouse PMCA4 in BMMs (Fig. 3 F) strongly suppressed RANKL-dependent Ca2+ oscillations (Fig. 3 E) as well as osteoclastogenesis (Fig. 3, E and G). However, the silencing of PMCA1 and PMCA4 did not markedly alter other receptor activators of NF-κB signaling pathways including the phosphorylation of MAPKs, the expression of c-Jun and c-Fos, or the activation of PLCγ1 (Fig. S3).
PMCAs regulate bone homeostasis in vivo
To examine the in vivo role of PMCAs in bone homeostasis, femurs from Atp2b1+/− (PMCA1-heterozygous) and Atp2b4−/− (PMCA4-null) mice were analyzed. The von Kossa staining of femurs from 6-wk-old male mice revealed a lower amount of mineralized bone in both Atp2b1+/− and Atp2b4−/− mice compared with respective wild-type mice (Fig. 4 A). Similarly, the μCT analyses of metaphyseal regions of femurs showed lower trabecular bone volume and bone mineral density in both PMCA-insufficient mice (Fig. 4 A). Finally, histological assessments confirmed a significant reduction in trabecular bone volumes (hematoxylin and eosin staining) and a marked increase in the osteoclast surface (TRAP staining) in Atp2b1+/− and Atp2b4−/− mice compared with wild-type mice. These effects of PMCA deficiency on bone homeostasis were osteoclast specific because the number of osteoblasts was similar in both PMCA-insufficient mice to that of corresponding wild-type mice (Fig. 4 A). Furthermore, the mineral apposition rate was not significantly different in Atp2b1+/− or Atp2b4−/− mice compared with wild-type mice (Fig. 4 B). In addition, the knockdown of PMCA expression or the inhibition of PMCA activity in calvarial cells did not affect osteoblast differentiation in vitro (Fig. S4).
PMCAs facilitate the survival of bone-resorbing mature osteoclasts
Intracellular Ca2+ overload is highly toxic to cells. The disposal of surplus intracellular Ca2+ is particularly important in mature osteoclasts because massive Ca2+ influx occurs when osteoclasts resorb the mineral components of bone (Salo et al., 1997). In this context, we hypothesized that increased PMCA expression in mature osteoclasts might facilitate the removal of intracellular Ca2+ during bone resorption, preventing Ca2+-induced apoptosis. To test this hypothesis, the TUNEL staining was performed with trabecular bone sections of femurs from Atp2b1+/− and Atp2b4−/− mice. Both PMCA-deficient mice exhibited conspicuously more TUNEL-positive osteoclasts compared with wild-type mice (Fig. 5, A and B), supporting the notion of an anti-apoptotic role for PMCAs in mature osteoclasts. When cultured on dentin slices in vitro, osteoclasts derived from Atp2b1+/− and Atp2b4−/− mice also showed more apoptotic nuclei compared with wild-type cells (Fig. 5, C and D). To further examine the role of PMCAs in bone-resorbing osteoclasts, isolated mature osteoclasts were transfected with PMCA siRNAs and cultured on dentin slices. The TUNEL staining of these osteoclasts showed a significantly higher number of TUNEL-positive cells in PMCA-silenced culture (Fig. 5, E and F). In addition, PMCA-silencing in osteoclasts promoted the cleavage of PARP and caspase-3 upon ionomycin treatment (Fig. 5 G), indicating that osteoclasts are more susceptible to Ca2+-induced apoptosis in the absence of PMCAs. The increase in apoptosis led to a marked reduction in resorbed areas on dentin slices when PMCAs were silenced in mature osteoclasts, although slightly deeper resorption pits were formed upon the knockdown of PMCAs (Fig. 5 H). Similar enhancement of apoptosis by PMCA siRNA was also observed when osteoclasts were grown on glass slides and challenged with ionomycin to mimic Ca2+ entry during bone resorption, whereas the overexpression of PMCAs strongly prevented Ca2+-induced apoptosis (Fig. 5, I and J).
PMCA4 but not PMCA1 is involved in NO-dependent osteoclast fusion
The knockdown of PMCA4 alone induced comparable acceleration of osteoclast differentiation, despite that the PMCA1 isoform was dominantly expressed over PMCA4 in osteoclast precursors (Figs. 1 and 2). This result may suggest a possibility of additional function of PMCA4 involved in the regulation of osteoclastogenesis. Interestingly, PMCA4 was suggested to inhibit NO production (Schuh et al., 2001), which is known to promote the fusion of preosteoclasts in the late phase of osteoclastogenesis (Nilforoushan et al., 2009). As shown in Fig. 6 A, the Ca2+-dependent neuronal NOS (nNOS) expression was significantly increased upon RANKL stimulation of BMMs. Furthermore, PMCA4 physically interacted with nNOS in preosteoclasts, whereas PMCA1 did not (Fig. 6 B). The knockdown of PMCA4, but not that of PMCA1, significantly enhanced NO production in preosteoclasts (Fig. 6 C). NOC-12, an NO donor, greatly augmented the size of osteoclasts, supporting the role of NO in preosteoclast fusion (Fig. 6 D). Although both PMCA1 and PMCA4 siRNA increased the formation of larger osteoclasts, only the PMCA4 siRNA-induced development of large osteoclasts was inhibited by the NOS inhibitor L-NMMA (Fig. 6, D and E). To further delineate the role of PMCA4 in nNOS-mediated osteoclasts fusion, siRNA oligonucleotides targeting nNOS were introduced into PMCA-silenced preosteoclasts, which were further cultured in the presence of RANKL to allow the formation of mature osteoclasts with the confirmation of specific knockdown by Western blotting (Fig. 6, F and G). The knockdown of nNOS only slightly reduced the osteoclast fusion index (Kaneda et al., 2000) in control and PMCA1-silenced cells (Fig. 6 H). However, nNOS siRNA dramatically reduced the osteoclasts fusion index in PMCA4-silenced osteoclast precursors. These data support the role of PMCA4 in regulating NO production and osteoclast fusion by inhibiting nNOS activity.
High PMCA4 gene expression correlates with high peak bone mass in humans
With ample evidence for PMCA in osteoclast regulation in mice, we next sought to investigate the potential association between PMCA gene expression and bone homeostasis in humans. To this end, we analyzed a set of publicly available genomics data deposited in GEO (accession no. GSE7158) in which DNA microarray experiments were performed using blood monocytes obtained from 878 healthy Chinese women aged between 20 and 45. The microarray results were compared between 12 subjects with lowest peak bone mass and 14 subjects with highest peak bone mass. In this dataset, PMCA4 gene (ATP2B4) expression exhibited a significantly higher level in women with high peak bone mass (Fig. 7, A and B).
The precise regulation of Ca2+ dynamics is crucial for proper differentiation and function of osteoclasts. RANKL induces oscillations of intracellular Ca2+ concentrations that trigger NFATc1 activation, which is a prerequisite for osteoclast differentiation. For the initiation of RANKL-dependent Ca2+ oscillations in osteoclasts, PLCγ activation is essential (Shinohara et al., 2008; Yoon et al., 2009). Furthermore, Kuroda et al., (2008) reported that the inositol 1,4,5-triphosphate receptor was critical for RANKL-induced Ca2+ oscillations and osteoclastogenesis. However, mechanisms for the termination of Ca2+ oscillations remain unknown although Ca2+ oscillations gradually disappear in the late stages of osteoclastogenesis. In the present study, we showed that the expression of two major Ca2+ extrusion pumps, PMCA1 and PMCA4, increased dramatically during the late stages of osteoclast differentiation (Fig. 1). Notably, overexpression of PMCAs in osteoclast precursors dramatically diminished the amplitudes of Ca2+ oscillations, reducing the osteoclastogenic potential (Fig. 3). In contrast, the knockdown of PMCAs resulted in a marked increase in Ca2+ oscillations. Consistent with these in vitro data, radiological and histomorphometric analyses of Atp2b1+/− and Atp2b4−/− mice revealed increased osteoclastogenesis with concomitant reduction of bone volume without any difference in bone formation (Fig. 4). To our knowledge, this study is the first demonstration that the plasma membrane Ca2+ pump is directly involved in the regulation of RANKL-induced Ca2+ oscillations and osteoclast differentiation both in vitro and in vivo. Interestingly, the knockdown of both PMCA1 and PMCA4 isoforms slightly increased the expression SERCA2 and reduced that of TRPV5 (Fig. S5), both of which have been shown to regulate cytosolic Ca2+ concentrations in osteoclasts (van der Eerden et al., 2005; Yang et al., 2009). However, it is highly unlikely that the observed pro-osteoclastogenic effects induced by PMCA deficiency were mainly caused by the indirect regulation of those proteins because individual knockdown of PMCA1 or PMCA4 significantly increased Ca2+ oscillations and osteoclastogenesis without conspicuous alterations in the SERCA2 and TRPV5 expression. Nonetheless, detailed dissection on the cross talk between the Ca2+ regulators in osteoclasts such as PMCA, SERCA, and TRPV is required to fully understand the mechanism by which osteoclast Ca2+ oscillations are controlled.
The transcriptional induction of PMCA genes was dependent on NFATc1 activity (Fig. 1). The promoter region of Atp2b1 and Atp2b4 contained several putative NFATc1 binding sites that indeed associated with NFATc1 (Fig. 1 I). Combined with the observation that the NFATc1 nuclear translocation was significantly increased in PMCA-silenced cells (Fig. 3 B), it is likely that an NFATc1-PMCA negative regulatory loop is operating to fine-tune osteoclast differentiation. We have reported such feedback regulation of NFATc1 by dual specificity tyrosine kinase 1A and Nogo-A during osteoclastogenesis (Lee et al., 2009, 2012). Thus, these data add another line of evidence that NFATc1 activity is regulated by positive and negative feedback mechanisms.
In our experiments, the increased expression of PMCAs persisted during osteoclastogenesis and reached a maximal level in mature osteoclasts (Fig. 1), leading to the hypothesis that these Ca2+ pumps might have additional roles in mature osteoclasts in addition to their role to diminish Ca2+ oscillations in developing osteoclasts. During bone resorption, a large amount of Ca2+ released from the bone enters osteoclasts via several Ca2+ transporters or channels localized on the apical membrane of osteoclasts including a plasma membrane ryanodine receptor (Moonga et al., 2002), TRPV5 (van der Eerden et al., 2005), and NCX (Li et al., 2007). Because high intracellular Ca2+ can be toxic to osteoclasts, excessive Ca2+ needs to be sequestered from the cytosol or expelled into the extracellular space. In this context, a continuous discharge of Ca2+ across the basolateral membrane has been observed in bone-resorbing osteoclasts (Berger et al., 2001). Although a Ca2+-ATPase activity was discovered in the purified plasma membranes in chicken osteoclasts over two decades ago (Bekker and Gay, 1990), studies on its identification and function in osteoclasts have not been thoroughly performed until the present. Here we suggest that PMCAs localized on the basolateral membrane of mature osteoclasts (Fig. 1) might operate to extrude superfluous intracellular Ca2+ across the basolateral membrane to prevent Ca2+-induced apoptosis of osteoclasts. In support of this hypothesis, we observed a significantly higher number of apoptotic osteoclasts in Atp2b1+/− mice and Atp2b4−/− mice compared with that of wild-type mice (Fig. 5). These results were further supported by in vitro experiments in which the knockdown of PMCAs resulted in a dramatically enhanced apoptosis in osteoclasts either cultured on bone slices or on glasses (Fig. 5, E and I). It should be noted that the plasma membrane NCX is another candidate considered for the removal of intracellular Ca2+ from osteoclasts. However, we observed that the mRNA expression of Ncx1 (Slc8a1) dramatically decreased during osteoclast differentiation, whereas the transcripts of Ncx2 and Ncx3 were not detected during osteoclastogenesis (Fig. 1 D). Furthermore, a recent publication by Li et al., (2007) predicted that NCX works in an influx mode on apical membranes promoting the entry of Ca2+ present in resorption lacunae into bone-resorbing osteoclasts. Combined with the reported anti-apoptotic roles of PMCA in HepG2, HeLa, mouse smooth muscle cells, and T cells (Chami et al., 2003; Okunade et al., 2004; Pellegrini and Scorrano, 2007), the up-regulation of PMCA expression in bone-resorbing osteoclasts might be beneficial in preventing apoptosis by extruding excessive Ca2+.
Although both isoforms were expressed in osteoclasts, PMCA1 and PMCA4 do not seem to be functionally redundant. The siRNA-mediated specific knockdown of either isoform of PMCA alone was sufficient to significantly enhance Ca2+ oscillations and osteoclast differentiation (Figs. 2 and 3). Furthermore, PMCA4 but not PMCA1 was involved in the regulation of NO, which is suggested to enhance osteoclast fusion. In RANKL-treated osteoclast precursors, only PMCA4 was coimmunoprecipitated with nNOS (Fig. 6 B), which is in accordance with previous studies that showed a physical interaction between PMCA4 and nNOS in smooth muscle cells (Schuh et al., 2003) and the inhibition of nNOS by PMCA4 caused by a local decrease in Ca2+ concentration (Schuh et al., 2001). The importance of PMCA–nNOS interaction in osteoclast fusion was clearly demonstrated by the dramatic decrease in the osteoclast fusion induced by PMCA4 knockdown in the absence of nNOS (Fig. 6 H). Thus, our data indicate that PMCA4 has a unique role of NO regulation during osteoclastogenesis, in addition to the modulation of Ca2+–NFATc1 axis. Notably, the expression level of PMCA4 rather than that of PMCA1 exhibited a higher correlation with peak bone mass in women (Fig. 7, A and B). The precise underlying mechanisms by which bone mass is regulated by PMCAs in humans, including the possible involvement of NO, need to be investigated in further studies.
To summarize, we propose dual roles of PMCAs during osteoclastogenesis (Fig. 7, C and D). During osteoclast differentiation (Fig. 7 C), NFATc1 increases PMCA expression in a RANKL-dependent manner. In late stages of osteoclast differentiation, high levels of PMCA cause efflux of intracellular Ca2+, reducing Ca2+ oscillations and limiting NFATc1 activities. This autoregulatory loop consisting of NFATc1, PMCA, and Ca2+ oscillations fine-tunes osteoclastogenesis. An additional mode of osteoclastogenesis regulation exists in which PMCA4 inhibits NO production that expedites osteoclast fusion. Therefore, PMCAs may limit excessive osteoclast formation by lowering intracellular Ca2+ and NO. In bone-resorbing osteoclasts (Fig. 7 D), maximal PMCA expression ensures osteoclast survival by the extrusion of excessive Ca2+ across the basolateral membrane in the face of massive Ca2+ entry upon bone resorption. An osteopenic bone phenotype was observed in both Atp2b1+/− and Atp2b4−/− mice, suggesting that in the absence of PMCA the advantageous conditions for osteoclast differentiation dominated over adverse effects on osteoclast survival in vivo. Because PMCA4 expression correlated with high peak bone mass in women, the modulation of PMCA4 expression or activity might serve as a novel strategy against bone erosive diseases such as osteoporosis.
Materials and methods
Recombinant human soluble RANKL and human M-CSF were purchased from PeproTech. Lipofectamine 2000 and Fura-2/AM were purchased from Invitrogen. Antibodies against PMCAs were purchased from Santa Cruz Biotechnology, Inc. (pan-PMCA) and Thermo Fisher Scientific (PMCA1 and PMCA4). Phospho-specific antobodies for p38, ERK1/2, JNK, CREB, and PLCγ1 were obtained from Cell Signaling Technology. Antibodies against p38, ERK1/2, JNK, c-Jun, PLCγ1, and cleaved caspases-3 were also purchased from Cell Signaling Technology. Antibodies for c-Fos, NFATc1, tubulin, PARP, and lamin B were obtained from Santa Cruz Biotechnology, Inc. Anti–β-actin, FITC-conjugated cholera toxin B subunit, ionomycin, eosin Y, the leukocyte acid phosphatase assay kit, and all other chemicals were obtained from Sigma-Aldrich.
Animals and in vitro osteoclastogenesis
The mutant PMCA1 (Atp2b1) and PMCA4 (Atp2b4) lines were prepared and maintained on the mixed (129Svj and Blackswiss) genetic background as previously described (Okunade et al., 2004). Loss of both copies of Atp2b1 caused embryonic lethality, but heterozygous mutants had no observable disease phenotype. Atp2b4−/− mutant mice showed no embryonic lethality and appeared externally normal. Although Atp2b1+/− mice and wild-type littermates were generated by breeding Atp2b1+/− and wild-type breeder-mates, Atp2b4−/− and wild-type littermates were generated by crossing Atp2b4+/− mice; genotypes were confirmed by PCR analysis using primers previously described (Okunade et al., 2004). All mice were maintained and procedures were performed as per guidelines by the National Institutes of Health (Guide for the Care and Use of Laboratory Animals). Animal experiment protocols were approved by the Committees on the Care and Use of Animals in Research at Seoul National University and University of Cincinnati. BMMs obtained from 5-wk-old ICR or PMCA mutant mice were used as osteoclast precursor cells for in vitro osteoclastogenesis experiments as described previously (Lee et al., 2008). In brief, mouse whole bone marrow cells, isolated by flushing the marrow space of femora and tibiae, were incubated overnight on culture dishes in α–modified Eagle medium (α-MEM) supplemented with 10% FBS. After discarding adherent cells, floating cells were further incubated with M-CSF (30 ng/ml) on Petri dishes. BMMs became adherent after a 3-d culture and were used as osteoclast precursor cells. Upon incubation of BMMs (3 × 104 cells/well in 48-well plates) with 30 ng/ml M-CSF and 100 ng/ml RANKL, >80% of the total cells became mononuclear TRAP-positive cells (preosteoclasts) after 2 d of culture. Fully mature multinucleated osteoclasts were formed after further incubation for 1 or 2 d. The osteoclast fusion index was defined as the number of nuclei per one multinucleated osteoclast (Kaneda et al., 2000). The exact time required for full differentiation varied slightly between experiments, and osteoclastogenesis was generally slower in transfected or virus-infected cells. To purify mature osteoclasts, BMMs (107 cells) were differentiated into osteoclast on collagen gel-coated 10-cm culture dishes in the presence of 30 ng/ml M-CSF and 100 ng/ml RANKL. Osteoclasts were detached by treating 0.2% collagenase (Invitrogen) at 37°C for 10 min, briefly centrifuged, and replated on culture dishes to allow reattachment for 1 h at 37°C. After the second round of collagenase treatment and gentle pipetting, only firmly attached mature osteoclasts remained whereas osteoclast precursors were removed.
Plasmid construction, transfection, and retroviral gene transfer
BMMs were transfected with pcDNA-rPMCA1 constructs (Bhargava et al., 2002) using Lipofectamin 2000. The entire coding region of mouse PMCA4 was PCR amplified from mouse osteoclast cDNA using the forward primer 5′-GGGCTCGAGCCACCATGACGAATCCACCAGGA-3′ and the reverse primer 5′-GGGGCGGCCGCTCAGACCGGTGTCTCCAG-3′. The amplified PCR product was cloned into a pMX-IG vector using XhoI and NotI sites. Retroviral packaging was performed by transfecting Plat-E cells with plasmids using Lipofectamin 2000. At 48 h after transfection, culture medium containing viral particles was collected and filtered through 0.45-µm syringe filters (Sartorius Stedim Biotech). For retroviral infection, BMMs were incubated in the virus-containing medium with 10 µg/ml polybrene and 30 ng/ml M-CSF for 24 h. The infection efficiency was >80% when measured for GFP fluorescence.
Gene knockdown by siRNA
The siRNA duplexes for PMCA1 (NM_026482_stealth_3610), PMCA4 (NM_213616_stealth_1507), and the negative universal control (medium GC content) were purchased from Invitrogen. Oligonucleotide siRNA duplexes were transfected into BMMs with Lipofectamin 2000 according to the manufacturer’s protocol.
The gene profiling of human PBMC-derived osteoclasts was described previously (Chang et al., 2008a). For osteoclast formation, hPBMCs were cultured in the presence of 30 ng/ml M-CSF and 100 ng/ml RANKL for 3 (preosteoclast) or 7 d (mature osteoclast). Total RNAs were extracted from hPBMCs, reverse transcribed, and transcribed in vitro into biotin-labeled cRNAs. These cRNAs were hybridized with the GeneChip Human Genome U133 Plus 2.0 Array (Affymetrix). The array chips were scanned with a GeneArray scanner (Affymetrix) and were analyzed by Microarray Suite 5.0 (Affymetrix). Publicly available gene expression datasets of human samples were downloaded from GEO (accession no. GSE7158) and the correlation between ATP2B1/ATP2B4 mRNA expression and peak bone mass was analyzed.
Cell-surface biotinylation and LC/MS/MS experiments
BMMs (4 × 105 cells/well in 6-well plates) were cultured with 30 ng/ml M-CSF alone or 200 ng/ml M-CSF plus RANKL for 36 h. A total of 2 × 107 cells were surface biotinylated using the Pinpoint cell-surface protein isolation kit (Thermo Fisher Scientific) following the manufacturer’s instructions. After cell lysis, biotinylated proteins were purified using a streptavidin column and subjected to SDS-PAGE. Protein bands showing differential expression were excised, digested with trypsin, and analyzed by nano-LC/MS/MS (Lee et al., 2008) equipped with an LTQ-FT mass spectrometer (Thermo Fisher Scientific) and an 1100 Series NanoLC pump (Agilent Technologies). Data from the mass spectrometer were analyzed using the Bioworks 3.2 software (Thermo Fisher Scientific). The search results were subsequently evaluated by PeptideProphet and home-built software.
Real-time PCR and RT-PCR analyses
For real-time PCR analysis, 1.5 µg of total RNA was reverse transcribed and PCR amplified with SYBR green master mix (Applied Biosystems) for 40 cycles of 15-s denaturation at 95°C and 1-min amplification at 60°C in ABI Prism 7500 System (Applied Biosystems). Relative mRNA expression levels were presented by normalizing against Hprt1 mRNA levels. For RT-PCR analysis, total RNAs were isolated with TRIzol reagent (Invitrogen) and 2 µg of RNAs were reverse transcribed with Superscript II (Invitrogen) according to the manufacturer’s instructions. The primer sets used in real-time PCR and RT-PCR are listed in Table S1.
ChIP assays were performed based on the protocol provided by the manufacturer (EZ ChIP kit; EMD Millipore) and the previously published method with slight modifications (Ha et al., 2010). In brief, BMMs were cultured with 30 ng/ml M-CSF and 100 ng/ml RANKL for 48 h to induce NFATc1 expression before cross-linking using formaldehyde. After sonication, the chromatin was immunoprecipitated with 5 µg each of control IgG antibody (Santa Cruz Biotechnology, Inc.) or NFATc1 antibody (Santa Cruz Biotechnology, Inc.). The eluted DNA fragments were analyzed by PCR using specific primers flanking the NFATc1 binding sites located within ∼1.5 kb upstream of Atp2b1 or Atp2b4 transcription initiation sites. Putative NFATc1 binding sites were identified using the web-based prediction program PROMO. Two putative NFATc1 binding sites of Atp2b1 promoter region (−1,103 to −1,092 and −52 to −44) and four putative NFATc1 binding sites of Atp2b4 promoter region (−1,292 to −1,283, −1,161 to −1,153, −1,047 to −1,039, and −51 to −43) were identified. PCR primer sequences are listed in Table S2. Input samples were also subjected to PCR with the same primers.
Western blotting and Immunoprecipitation
Western blotting and immunoprecipitation experiments were performed as previously described (Ryu et al., 2006; Kim et al., 2007). Cells were disrupted in lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 1% Triton X-100, 10 mM NaF, 1 mM Na3VO4, and complete protease inhibitor cocktail) and 30 or 45 µg of cell lysates were resolved by 8–10% SDS-PAGE. Separated proteins were transferred to a polyvinylidene difluoride membrane (GE Healthcare) and the membrane was blocked with 5% skim milk and probed with appropriate primary antibodies. After 1-h incubation with HRP-conjugated secondary antibodies, the immunoreactivity was detected using chemiluminescence. For coimmunoprecipitation experiments, 1 mg of cell lysates was immunoprecipitated with 2 µg of anti-PMCA1 or anti-PMCA4 antibodies and immunoblotted using nNOS or pan-PMCA antibodies. Control immunoprecipitation was performed using a mouse IgG isotype control antibody. Nuclear fractions were prepared by lysing cells in a hypotonic lysis buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, and 0.5 mM PMSF) and solubilizing nuclear pellets with the sequential addition of 15 µl of high salt buffer (20 mM Hepes, pH 7.9, 420 mM NaCl, 25% glycerol, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM PMSF, and 0.5 mM DTT) and 75 µl of storage buffer (20 mM Hepes, pH 7.9, 100 mM NaCl, 20% glycerol, 0.2 mM EDTA, 0.5 mM PMSF, and 0.5 mM DTT). Protein concentration was determined using a detergent-compatible protein assay kit (Bio-Rad Laboratories).
To detect the NFATc1 localization, preosteoclasts cultured on glass coverslips were fixed with 3.7% formaldehyde and permeabilized with 0.1% triton X-100. After blocking in PBS containing 1% BSA, coverslips were incubated with anti-laminB or anti-NFATc1 antibodies diluted (1:50, 1 h) in PBS containing 1% BSA for 2 h. Subsequently, cells were washed and stained with DAPI or Cy3-conjugated secondary antibodies (1:300, 1 h). For the measurement of PMCA localization, osteoclasts cultured on dentin discs were fixed and stained with FITC-conjugated cholera toxin (for GM1-containing plasma membrane labeling, 2.5 µg/ml, 20 min) plus anti–pan-PMCA antibodies. After washing, cells on dentin discs were mounted and images were obtained using a confocal microscope (FV300; Olympus).
Measurement of intracellular Ca2+ concentrations
BMMs on glass coverslips were cultured with 30 ng/ml M-CSF and 100 ng/ml RANKL for 48 h. For the measurement of Ca2+ oscillations in individual osteoclast precursor, cells were loaded with 5 µM Fura-2/AM and 0.05% pluronic F127 for 40 min at room temperature. After washing three times with Hank’s balanced salt solution (Gibco), the fluorescence was recorded at every 500 ms with 340/380 nm excitations and 510 nm emission at 37°C using a digital imaging system (Cascade 650; Photometrics) and Metafluor image analysis software (Universal Imaging).
Calcein-xylenol orange double labeling
To evaluate the mineral apposition rate in vivo, Atp2b1+/−, Atp2b4−/−, and wild-type littermate mice were sequentially injected with 25 mg/kg calcein and 90 mg/kg xylenol orange intraperitoneally with an interval of 6 d (Lee et al., 2009). At 3 d after the last injection, mice were killed and dissected tibiae were embedded in methyl methacrylate resins. Tissue sections were observed under a laser-scanning microscope (LSM5 PASCAL; Carl Zeiss).
Calvarial cells were isolated from 1-d-old mice as described previously (Lee et al., 2009). Osteoblast differentiation was induced by culturing cells in osteogenic medium (α-MEM containing 10 mM β-glycerophosphate and 100 µg/ml ascorbic acid) for 7 d, which was confirmed by AP staining.
The cytotoxicity of eosin on BMMs was assessed using CCK-8 reagents (Dojindo Laboratories) that produce formazan dye in the presence of live cells. The optical density was measured at 450 nm.
Lipopolysaccharide-induced bone resorption in vivo
Mice were injected with 5 mg/kg LPS (from Escherichia coli 0111:B4; Sigma-Aldrich) intraperitoneally twice with a 4-d interval as described previously (Chang et al., 2008b). 30 µl of 5 mM eosin were injected at the proximal ends of tibiae twice with a 4-d interval. At 7 d after the first injection, mice were killed and tibiae sections were stained for TRAP activity.
Calvarial bone resorption assay in vivo
Control or PMCA siRNA oligonucleotides (20 µM, 30 µl) were mixed with 10 µl Lipofectamine 2000 and injected into the subcutaneous space of calvariae of 5-wk-old ICR mice three times with 2-d intervals. As a positive control for bone resorption, collagen matrices soaked with RANKL (10 µg) were implanted on the periosteal surface of calvariae. At 6 d after final siRNA injection, mice were killed and calvariae were subjected to μCT or processed for histological analyses.
Bone histomorphometric analyses were performed on paraffin-embedded sections as described previously (Chang et al., 2008b). In brief, calvariae or tibiae were fixed in 4% paraformaldehyde, decalcified in 12% EDTA for 4 wk, and embedded in paraffin. 5-µm-thick tissue sections were subjected to TRAP staining or hematoxylin/eosin staining according to standard procedures. For the measurements of mineralized bones, tibiae were fixed in 4% paraformaldehyde, dehydrated in graded ethanol, and embedded in methyl methacrylate resin. Sections of 5 µm thickness were subjected to von Kossa’s silver nitrate staining followed by van Gieson’s counterstaining.
Femurs of 6-wk-old male Atp2b1+/−, Atp2b4−/−, and wild-type littermate mice were analyzed by μCT using the SkyScan 1072 system (SkyScan). Trabecular bone volume was measured in the 1-mm region in length, 1 mm below the distal growth plate of femurs. A total of 350–400 tomographic slices were acquired and 3D analyses were performed with CT volume software (ver 1.11; SkyScan).
Detection of apoptosis in vivo and in vitro by TUNEL assay
TUNEL assay using In situ Cell Death Detection kit (Roche) coupled to an AP-conjugated antobody was performed on bone tissue sections of 6-wk-old male Atp2b1+/−, Atp2b4−/−, and wild-type littermate mice to evaluate in vivo osteoclast apoptosis. For the detection of in vitro osteoclast apoptosis, mature osteoclasts were purified from the co-cultures of mouse bone marrow cells and calvarial osteoblasts (Ha et al., 2004). After transferring onto dentine discs (Immunodiagnostic Systems), mature osteoclasts (2 × 102 cells/disc) were transfected with PMCA siRNA oligonucleotides and further cultured with 30 ng/ml M-CSF and 100 ng/ml RANKL for 2 d. TUNEL assay with In situ Cell Death Detection kit coupled to an FITC-conjugated antibody was performed followed by Cy3-conjugated phalloidin (Invitrogen) and DAPI (Invitrogen) staining.
BMMs on glass coverslips were transfected with control or isoform-specific PMCA siRNAs and were further cultured with 30 ng/ml M-CSF and 100 ng/ml RANKL for 2 d. Cells were loaded with 5 µM of the cell-permeable fluorescent NO indicator DAF-2 DA (EMD Millipore) for 30 min. NO-dependent fluorescence was observed under a laser-scanning microscope. To confirm the specificity of DAF-2 DA fluorescence, cells were cultured overnight with 100 ng/ml LPS in the absence or presence of a NO synthase inhibitor, L-NMMA (10 µM; EMD Millipore), for 2 h before DAF-2 DA loading.
The Student’s t test was used to determine the significance of differences between two groups. Comparison of multiple results was performed by one-way analysis of variance followed by Student Knewman-Keuls post hoc tests. Differences with P < 0.05 were regarded as significant. The gene expression data in Fig. 7 were analyzed by Mann-Whitney U test.
Online supplemental material
Fig. S1 shows the effect of PMCA knockdown on osteoclastogenesis in calvariae organ culture models. Fig. S2 shows the enhanced RANKL-induced Ca2+ oscillations and osteoclast differentiation upon eosin treatment in vitro and in vivo. Fig. S3 shows the changes in P38, ERK1/2, JNK1/2, and PLCγ1 phosphorylation as well as c-Fos expression after RANKL stimulation in PMCA knocked down cells by Western blotting. Fig. S4 shows no change in osteoblast differentiation after PMCA knockdown or inhibition in vitro. Fig. S5 shows the expression levels of SERCA2 and TRPV5 in PMCA knocked down cells by Western blotting. Table S1 lists primer sets used in real-time PCR and RT-PCR experiments. Table S2 shows PCR primer sequences used in ChIP experiments.http://www.jcb.org/cgi/content/full/jcb.201204067/DC1
This work was supported by grants from the Science Research Center (20120000490) and the Ministry of Health and Welfare (A111787) to H.-H. Kim, and the Korea Research Foundation grants funded by the Korean government (KRF-2008-313-E00439 and NRF-2010-359-E00016 to Y. Lee and NRF-2010-860-20100096 to H.J. Kim). V. Prasad was supported by American Heart Association 11BGIA7720005 and National Institutes of Health grant HL061974.
The authors declare no conflict of interest.
bone marrow–derived macrophage
liquid chromatography/mass spectrometry/mass spectrometry
macrophage colony-stimulating factor
modified Eagle medium
nuclear factor of activated T cell c1
nitric oxide synthase
peripheral blood mononuclear cell
plasma membrane Ca2+ ATPase
receptor activator of NF-κB ligand
tartrate-resistant acid phosphatase