Increased Osteoclastogenesis in Mice Lacking the Carcinoembryonic Antigen-Related Cell Adhesion Molecule 1

Alterations in bone remodeling are a major public health issue, as therapeutic options for widespread bone disorders such as osteoporosis and tumor-induced osteolysis are still limited. Therefore, a detailed understanding of the regulatory mechanism governing bone cell differentiation in health and disease are of utmost clinical importance. Here we report a novel function of carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1), a member of the immunoglobulin superfamily involved in inflammation and tumorigenesis, in the physiologic regulation of bone remodeling. Assessing the expression of all members of the murine Ceacam family in bone tissue and marrow, we found CEACAM1 and CEACAM10 to be differentially expressed in both bone-forming osteoblasts and bone-resorbing osteoclasts. While Ceacam10-deficient mice displayed no alteration in structural bone parameters, static histomorphometry demonstrated a reduced trabecular bone mass in mice lacking CEACAM1. Furthermore, cellular and dynamic histomorphometry revealed an increased osteoclast formation in Ceacam1-deficient mice, while osteoblast parameters and the bone formation rate remained unchanged. In line with these findings, we detected accelerated osteoclastogenesis in Ceacam1-deficient bone marrow cells, while osteoblast differentiation, as determined by mineralization and alkaline phosphatase assays, was not affected. Therefore, our results provide in vivo and in vitro evidence for a physiologic role of CEACAM1 in the regulation of osteoclastogenesis.


Introduction
In the healthy organism, bone remodeling is performed by the balanced activity of bone-forming osteoblasts and bone-resorbing osteoclasts, assuring the constant renewal of bone tissue and maintenance of adequate bone stability [1,2]. In osteoporosis, the most prevalent bone disease worldwide, a relative increase of bone resorption over bone formation occurs, thereby resulting in bone loss and a subsequent increase in fracture risk [3]. As excessive osteoclastogenesis is detrimental not only in osteoporosis, but also tumor-induced osteolysis and Paget's disease of bone [4,5], the molecular understanding of the processes regulating osteoclast formation and function is of paramount clinical importance.
Osteoclasts represent highly specialized, multinuclear giant cells, which are formed by the fusion of hematopoietic precursor cells from the monocyte/ macrophages lineage. The process of osteoclast formation (osteoclastogenesis) depends on two essential cytokines, macrophage colony-stimulating factor (M-CSF) [6,7] and receptor activator of nuclear factor kappa-B ligand (RANKL) [8,9], which are produced by bone marrow cells and osteoblasts, respectively. While M-CSF is required for the early differentiation of monocytes and macrophages, RANKL is essential for the subsequent cellular fusion to yield mature and functional osteoclasts. This is best demonstrated by mice lacking RANKL which display osteopetrosis, a condition characterized by the absence of functional osteoclasts and resulting in a marked increase in bone mass with consecutive displacement of bone marrow [10,11]. Through binding to the receptor activator of nuclear factor kB (RANK), expressed on osteoclasts and their precursors, RANKL activates two key transcription factors, nuclear factor kappalight-chain-enhancer of activated B-cells (NF-kB) and cytoplasmic calcineurin/ nuclear factor of activated t cells (NFATC1), which have been demonstrated to be of crucial importance for osteoclastogenesis [12,13] Once fully differentiated, osteoclasts express Acp5 (Tartrate-resistant acid phosphatase) and Calcr (Calcitonin receptor) and attach to the bone matrix, which is subsequently resorbed by the secretion of hydrochloric acid and matrix-degrading peptidases [14].
While many systemic and local factors, including endocrine organs, the central nervous system, and mechanical load bearing, have been identified as pivotal regulators of bone turnover [15,16], recent research has unraveled an unanticipated role of cell adhesion molecules in the regulation of bone cell differentiation and function. For example, vascular cellular adhesion molecule 1, which is expressed on myeloma cells and interacts with integrins mediating osteoclast attachment to bone surface, was shown to tether osteoclast progenitors to accelerate their maturation, thus facilitating tumor-induced osteolysis [17,18]. Furthermore, it could be demonstrated that the intercellular adhesion molecule-1 provides a high affinity adhesion between osteoblast and osteoclast precursors, thereby enhancing the binding of Rank to membrane-bound Rankl on osteoblasts [19]. Another group of cell-to-cell adhesion molecules that has raised great scientific and clinical interest in recent years are carcinoembryonic antigen-related cell adhesion molecules (CEACAMs), representing a subdivision of the immunoglobulin-related glycoproteins. Apart from functioning as receptors for host-specific bacteria and viruses, CEACAMs have been shown to regulate tissue architecture, cell-to-cell recognition, tumor proliferation, neovascularization and metastasis [20]. However, despite the extensive characterization of CEACAMs in pathologic conditions such as inflammation and cancer, their role in bone remodeling remained unclear to date.
In the present study, we found Ceacam1 and Ceacam10 to be expressed in bone marrow and tissue, including osteoblasts and osteoclast precursors. While no alterations in bone remodeling were detected in Ceacam10-deficient mice, an osteoporotic bone phenotype due to increased osteoclastogenesis was observed in mice lacking Ceacam1. Ex vivo assays demonstrated an increased osteoclast formation in bone marrow cultures derived from Ceacam1-deficient mice, which was accompanied by an elevated expression of Nfatc1, the master transcription factor governing osteoclastogenesis. Taken together, these findings not only provide in vivo and in vitro evidence for a role of CEACAM1 in the regulation of bone remodeling, they also raise the possibility that pharmacologic targeting of CEACAM1 may be an alternative approach to treat skeletal disorders caused by excessive bone resorption.

Animals
Ceacam1-and Ceacam10-deficient animals were generated and genotyped as described previously [21,22,23]. All animal experiments were approved by the local animal care committee.

Skeletal analysis
All mice received dual calcein injections for the determination of the bone formation rate at 9 and 7 days before sacrifice. The lumbar vertebrae were dehydrated and embedded non-decalcified into methyl methacrylate for sectioning. 4 mm-thin sections were stained with toluidine blue or von Kossa/van Gieson procedure as described [24]. Static and cellular histomorphometry was carried out using the OsteoMeasure system (Osteometrics, Decatur, USA) following the guidelines of the American Society of Bone and Mineral Research. Dynamic histomorphometry for the determination of the bone formation rate was performed on non-stained 12 mm-sections. The cortical thickness and the mean diameter of femora were quantified by mCT scanning using a mCT 40 (Scanco Medical).

Primary osteoblasts
Primary osteoblasts were isolated from bone marrow cells derived from 12 to 18week-old mice as described [25]. At 80% confluency, cells were differentiated by adding b-glycerophosphate and ascorbic acid. For staining with alizarin red, cells were fixed with 90% ethanol, washed twice with water and incubated with 40 mM alizarin red staining solution (pH 4.2). After washing, the stained cultures were incubated with 10% acetic acid to quantify matrix mineralization. After removing the cellular remnants by centrifugation, the supernatant was neutralized with ammonium hydroxide, and absorption was determined at 405 nm.

Primary osteoclasts
For the assessment of osteoclastogenesis, the bone marrow from 12 to 18-weekold mice was flushed out of the femora as described previously [25]. At day 2 of differentiation, M-CSF (Peprotech) was added, followed by RANKL (Peprotech) at day 4.
For TRAP-activity staining, cells were washed with PBS and then fixed in cold methanol. After two washing steps with distilled water, the fixed cells were dried before the TRAP-specific substrate naphthol AS-MX phosphate (Sigma-Aldrich) was added. For the quantification of osteoclast formation, the number of TRAPpositive, multinucleated (.3 nuclei/cell) cells per well was determined.

RAW264.7 cell line
RAW264.7 cells were obtained from ATCC (Wesel, Germany) and cultured in DMEM containing 10% fetal calf serum. For induction of osteoclastogenesis cells were cultured with RANKL for 5 days.

Expression analysis
Total RNA was isolated from indicated tissues, whole and marrow-flushed bones, or primary bone cells using the RNeasy Minikit (QIAGEN). Subsequently, DNase digestion was performed according to the manufacturer's instructions. Concentration and quality of extracted RNA was measured using an ND-1000 system (NanoDrop Technology). For cDNA synthesis, 0,5-1 mg of RNA was reverse transcribed using cAMV First-Strand Synthesis Kit (Invitrogen) according to the manufacturer's instructions. For semi-quantitative RT-PCR analysis, the following primers were used to amplify fragments of  [26]. Due to a considerable homology Ceacam1a and Ceacam2 were specifically amplified in addition to Ceacam1/2 in order to detect expression of Ceacam1. Quantitative RT-PCR expression analysis was performed using StepOnePlus predesigned TaqMan gene expression assays (Applied Biosystems). b-actin or Gapdh expression was used as an internal control, as indicated. For western blotting CEACAM1 protein was detected using the polyclonal rabbit anti-CEACAM1 antiserum ''P1'' after transfer of proteins from an SDS-page onto a PVDF membrane [27].

Immunofluorescence
For visualization of the actin cytoskeleton, cells were fixed in Formafix (Pathomed, Germany), followed by permeabilization with ice-cold acetone and rinses in phosphate buffered saline. Subsequently, cells were stained with fluorescein-labelled phalloidin and 49,6-diamidin-2-phenylindol (DAPI; both Molecular Probes, Germany) at room temperature in the dark (1:300 in phosphate buffered saline). After washes in phosphate buffered saline, slides were mounted with Vectashield mounting media (Vector Labs, United Kingdom). Pictures were taken on a Leica DM5000B fluorescent microscope (Leica, Germany).

Statistical analysis
All data were analyzed by two-tailed Student's t test using Excel software. All data are reported as mean ¡ SD. p,0.05 was considered statistically significant.

Results
To elucidate a potential role of CEACAMs in bone remodeling, we monitored the expression of all genes of the CEACAM family in whole bone, bone marrow and marrow-flushed bone. While the expression of most members was not detectable by semi-quantitative PCR, we exclusively found Ceacam1 and Ceacam10 expression in bone and bone marrow (S1 Figure). Comparing the expression in various tissues, Ceacam1 and Ceacam10 were expressed at comparable levels in bone samples (Fig. 1A). Interestingly, while the expression of Ceacam1 was highest in liver and femur, Ceacam10 was expressed at much higher levels in all bone specimens compared to non-skeletal tissues (Fig. 1A). To further characterize the expression of Ceacam1 and Ceacam10 in the course of osteoblast and osteoclast formation, we differentiated bone marrow cells into osteoblasts and osteoclasts and performed quantitative RT-PCR. Here we found that the expression of both Ceacam1 and Ceacam10 increased during the early stages of osteoblast differentiation, and decreased towards terminal osteoblast differentiation (Fig. 1B). In addition, although Ceacam1 and Ceacam10 were both expressed at high levels in undifferentiated bone marrow cultures, their expression markedly decreased during the course of osteoclastogenesis (Fig. 1C). This observation was confirmed in experiments using the macrophage cell line RAW264.7 where we found Ceacam1 to be differentially expressed during the course of osteoclastogenesis on mRNA (Fig. 1D) and protein level (Fig. 1E).
As this observation pointed towards a specific role of CEACAM1 and CEACAM10 in the regulation of bone remodeling, we next applied nondecalcified bone histology in mice lacking the respective genes. Von Kossa staining of spine sections from 6-month-old mice demonstrated a reduced bone mass in Ceacam1-deficient mice compared to WT controls, whereas no alteration could be detected in mice lacking CEACAM10 (Fig. 2A). These findings were confirmed by static histomorphometry, which demonstrated Ceacam1-deficient mice exhibit decreased trabecular bone volume accompanied by a reduction in trabecular number and an increase in trabecular separation (Fig. 2B). In contrast, none of these parameters were altered in mice lacking CEACAM10. In order to analyze whether the observed phenotype is also detectable in younger mice, we additionally performed static histomorphometry of 3-month-old mice. Spine sections of Ceacam1-deficient mice were characterized by a significantly reduced trabecular bone mass accompanied by reduced trabecular numbers (S2 Figure), similar to what was observed in 6-month-old animals. In addition, 3-month-old mutant animals displayed a reduced trabecular thickness and increased trabecular separation (S2 Figure). To address the question whether the lack of Ceacam1 not only affects skeletal architecture in vertebrae, but also long bones, we finally performed histomorphometry of non-decalcified tibia sections and mCT-scanning of femora derived from Ceacam1-deficient mice. While we found a decreased trabecular bone volume accompanied by a decrease in trabecular numbers in the tibiae of mutant animals (Fig. 3A), cortical thickness and mean diameter were unaltered in the femora of Ceacam1-deficient mice (Fig. 3B).
These findings suggested an important role of CEACAM1 in the regulation of trabecular bone remodeling, while CEACAM10 was found to have no overt effect. Therefore, we focused our further analyses on CEACAM1 and performed a full histomorphometric characterization of the spine sections derived from 6-monthold Ceacam1-deficient mice. Cellular histomorphometry using toluidine-blue stained spine sections revealed no alteration in the number of osteoblasts or osteoblast surface, indicating normal osteoblastogenesis in Ceacam1-deficient mice (Fig. 4A). Likewise, dynamic histomorphometry following dual calcein labeling demonstrated a normal bone formation rate, ruling out an impaired osteoblast function as the underlying cause of the observed phenotype. In contrast, Ceacam1-deficient mice were found to display an increased number of osteoclasts and osteoclast surface, suggesting increased bone resorption (Fig. 4B). Although an absolute increase in collagen degradation products was undetectable (Fig. 4C), normalization of serum crosslaps to the reduced bone mass revealed an elevated net bone resorption in mice lacking CEACAM1 (data not shown).
In order to analyze, whether the observed increase in osteoclastogenesis could be explained by gross alterations in OPG levels, we measured this cytokine in the serum of Ceacam1-deficient mice and controls by ELISA. While no changes were detectable in 3-month-old Ceacam1-deficient mice, we measured an increased concentration of OPG in 6-month-old mice, pointing towards an age-dependent compensatory regulation in vivo in the light of enhanced osteoclastogenesis. This assumption was confirmed in vitro, were we detected normal expression of Tnfsf11 and Tnfrsf11b, encoding RANKL and OPG, in primary osteoblasts at day 10 of differentiation (S3 Figure).
Investigating the possibility of cell-autonomous defects in these mice, we next differentiated WT and Ceacam1-deficient bone marrow cells into osteoblasts and osteoclasts ex vivo. Von Kossa staining at day 10 of differentiation revealed a normal formation of mineralized nodules in Ceacam1-deficient osteoblast cultures (Fig. 5A). Likewise, the assessment of alizarin red staining demonstrated regular extracellular matrix mineralization in osteoblasts derived from Ceacam1deficient mice. In addition, normal levels of intracellular alkaline phosphatase    activity were measured, ruling out a functional osteoblast defect caused by Ceacam1-deficiency (Fig. 5B). In contrast, TRAP-activity staining of Ceacam1deficient bone marrow cells cultured in the presence of M-CSF and RANKL displayed a marked increase in osteoclastogenesis at day 7 of differentiation ( Fig. 5C), while no alteration in the number of nuclei per osteoclast could be detected by immunofluorescence (Fig. 5D). To understand this effect on the molecular level, we finally screened for differentially expressed genes essential for osteoclast differentiation and function. On day 0 of differentiation, no significant differences could be measured among the tested genes (Fig. 5E). On day 3 of differentiation, only a tendency towards a reduced expression of Tm7sf4, encoding the fusion protein DC-STAMP, could be detected in Ceacam1-deficient cultures, whereas the expression of the late osteoclast marker gene Calcr was still undetectable in mutant and control cells. In contrast, on day 4 of differentiation we observed a trend towards enhanced expression of Acp5 and significantly increased expression of Nfatc1 (Fig. 5E). However, while the increased expression of Acp5 was only temporarily detectable, we found Nfatc1 to be overexpressed in Ceacam1-deficient osteoclasts again at day 7, pointing towards a crucial role of this key transcription factor in mediating the effects of CEACAM1 on osteoclastogenesis.

Discussion
The molecular understanding of bone remodeling represents an ongoing clinical challenge, as therapeutic options for skeletal disorders such as osteoporosis and tumor-induced osteolysis caused by excessive osteoclast function remain limited. While the biological actions of CEACAMs have been studied extensively in the context of pathological conditions including inflammation, stroke and cancer [20,22] a potential role in the regulation of bone turnover has not been sufficiently addressed to date. This is indeed surprising, given the growing body of evidence establishing a link between skeletal integrity and immune or cancer cells not only in basic, but also clinical research. For example, rheumatoid arthritis, a chronic autoimmune disease characterized by joint destruction, is considered to be driven by the secretion of pro-inflammatory cytokines including TNFa and IL-17 from lymphocytes, resulting in excessive activation of osteoclasts [28,29]. Likewise, expression of IL-1b and inhibitors of Wnt signaling, such as DKK-1, in tumor cells have been reported to be involved in malignant osteolysis through both an activation of osteoclastogenesis and an inhibition of bone formation (C) TRAP activity staining of terminally differentiated osteoclasts. The quantification of TRAP-positive, multinuclear osteoclasts is given on the right. (D) Immunofluorescence using DAPI (nucleus) and Phalloidin (actin) staining of osteoclast cultures at day 7 of differentiation. The quantification of the number of nuclei per osteoclast (Nuclei/Oc) is indicated on the right. (E) qRT-PCR expression analysis of the indicated genes (Acp5, Tartrate-resistant acid phosphatase; Tmf7sf4, DC-STAMP, dendrocyte-expressed seven transmembrane protein; NF-kB, nuclear factor kappa-light-chain-enhancer of activated Bcells; Nfat1c, nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 1, Tnfrsf11a, receptor activator of nuclear factor kB, Rank; c-Fos, FBJ osteosarcoma oncogene; and Calcr, Calcitonin receptor) at day 0, 3, 4 and 7 of osteoclast differentiation. All bars represent mean ¡ SD (n53 cultures per group). Asterisks indicate statistically significant differences (p,0.05). doi:10.1371/journal.pone.0114360.g005 Decreased Bone Mass in Ceacam1-Deficient Mice [30,31]. Based on their broad array of biological effects in the immune system and tumorigenesis, it was thus important to analyze the role of CEACAMs in the regulation of bone turnover.
Monitoring the expression of all genes encoding the members of the CEACAM family, we could exclusively detect the expression of Ceacam1 and Ceacam10 in bone tissue. Although this observation does not necessarily rule out a potential role of other Ceacams in bone remodeling, they pointed towards a specific function of CEACAM1 and CEACAM10 in regulating bone cell function. This notion was further supported by the high cDNA levels of the two encoding genes in bone tissue compared to other organs. Based on the specific expression dynamics of both Ceacam1 and Ceacam10 during osteoblast and osteoclast differentiation, we went on to analyze the bone phenotype of the respective animal models. While Ceacam10-deficiency was not associated with any alterations in structural bone parameters, static histomorphometry demonstrated a decreased trabecular bone volume in 3-and 6-month old Ceacam1-deficient mice. This was indeed an interesting finding, as Ceacam1-deficient mice, under basal conditions, have been reported to display no gross phenotypical abnormalities [20]. As mentioned above, while previous studies have primarily focused on the pathophysiologic functions of CEACAM1 in various in vivo and in vitro disease models, this observation pointed towards a physiologic role of CEACAM1 in the regulation of bone cell activity.
Although Ceacam1 was differentially expressed during osteoblastogenesis, the cellular and dynamic histomorphometry failed to detect defective osteoblastogenesis or osteoblast function. Surprisingly, increased osteoclast parameters were found in trabecular bone, indicating accelerated osteoclastogenesis in Ceacam1deficient mice. To further characterize this effect at the cellular level, we differentiated bone marrow cells into osteoblasts and osteoclasts. In line with our in vivo observations, primary osteoblasts derived from Ceacam1-deficient mice displayed normal matrix mineralization and alkaline phosphatase activity in vitro. In contrast, although no alteration in the numbers of nuclei per osteoclasts could be detected, Ceacam1-deficient bone marrow cells demonstrated an increased osteoclastogenesis when cultured with M-CSF and RANKL. Therefore, it is now possible to conclude that CEACAM1 functions as a negative regulator of osteoclastogenesis in vivo and in vitro. The fact that we could detect increased levels of serum OPG in 6-month old animals is interesting, however does not explain the observed bone phenotype and increased osteoclastogenesis associated with the lack of CEACAM1. This is supported by the finding that 3-month-old mutant animals displayed a low bone mass phenotype despite normal OPG levels. Furthermore, as we failed to detect differences in the expression of Tnfsf11 and Tnfrsf11b in primary osteoblasts derived from Ceacam1-deficient mice, this particular phenomenon is most likely explained by an age-dependent counter regulatory mechanism rather than an intrinsic osteoblast defect.
On the molecular level, we could detect differential expression of Ceacam1 not only in bone marrow derived osteoclast progenitors, but also in the pure macrophage cell line RAW264.7, providing a potential explanation for the increased osteoclast formation in Ceacam1-deficient mice. Since the formation of mature osteoclasts primarily depends on RANKL-induced activation of key transcription factors and the subsequent expression of several osteoclast marker genes, we monitored the expression of NF-kb, Nfatc1, Acp5, Tmf7sf4, Tnfrsf11a, c-Fos, and Calcr during osteoclastogenesis in bone marrow cells derived from WT and Ceacam1-deficient mice. While indicators of mature osteoclasts, including Calcr and Acp5, were found to be expressed at similar or only temporarily elevated levels compared to WT controls, respectively, increased expression of Nfatc1 in Ceacam1-deficient cells was found at day 4 and 7 of osteoclast differentiation. As this coincided with the induction of monocyte/macrophage fusion into mature osteoclasts by the addition of RANKL to these cultures, it appears tempting to speculate that CEACAM1 may modulate osteoclastogenesis through the regulation of NFATC1, the master transcription factor required for osteoclast differentiation. Although a previous study reported a stimulatory effect of a short splice variant of CEACAM1 on the expression of Nfatc1 in CD4 + T cells [32], further research is necessary to elucidate the influence of CEACAM1 on the regulation of this transcription factor specifically in osteoclasts and myeloidderived cells. This is of particular importance, as the liver-specific inactivation of Ceacam1, characterized by hyperinsulinemia and glucose intolerance, results in decreased osteoclastogenesis [33]. Thus, it appears that the impact of global Ceacam1 deletion, including cells of the osteoclast lineage, is dominant over the effects of hepatic Ceacam1 expression on bone metabolism.
Taken together, our study reports a novel function of CEACAM1 in bone remodeling. Using in vivo and in vitro assays, we demonstrate that deficiency of CEACAM1 is associated with a reduced bone mass due to increased osteoclastogenesis, at least in mice. Given its previously reported function in regulating tumor cell differentiation and the immune system, future studies investigating the role of CEACAM1 in pathologic bone conditions, such as tumor-induced osteolysis and inflammation-induced bone loss, will be of crucial importance.