The nucleocytoplasmic shuttling protein CIZ reduces adult bone mass by inhibiting bone morphogenetic protein–induced bone formation

X-ray examinations of the bone revealed that gross morphology of the femora in CIZ-KO mice was similar to that in wild type (Fig. 1, A and B). Body weight in CIZ-KO mice was ∼10% less at the 8-wk time point, whereas it caught up with the weight of wild-type mice by 50 wk (Fig. 1 C). X ray of the long bone revealed that trabecular spicules present in the distal end of femora were observed to be denser in CIZ-KO compared with wild type (Fig. 2 A). Similarly, femoral neck (Fig. 2 A) was more radiopaque in CIZ-KO compared with wild type. Radiopacity levels in the distal ends of the femur (condyle regions) were similar between the two genotypes (Fig. 2 A).

To further visualize the changes in the trabecular bone structures in the femora, two-dimensional microCT examination was conducted in the metaphyseal regions of the femora within a midsagittal plane. This demonstrated that more cancellous bone was present in the femora in CIZ-KO mice than wild-type mice (Fig. 2 B). Quantification of the trabecular bone on the microCT sections of the femora (Fig. 2 B, rectangle) indicated that BV/TV levels were increased by ∼40% in CIZ-KO mice (Fig. 2 C). Increase was not limited to long bones as microCT examination of the lumbar vertebrae indicated ∼15% increase in trabecular bone mass (Fig. 2, D and E).

To examine whether CIZ deficiency–induced increase in bone volume could be due to reduction in bone resorption, histomorphometric analysis was conducted to obtain bone resorption parameters. Osteoclast morphology (not depicted) as well as quantitative bone resorption parameters, including osteoclast number per bone surface (N.Oc/BS) and osteoclast surface per bone surface (Oc.S/BS) in the histological sections, were similar between wild-type and CIZ-deficient mice (Fig. 3, A and B). Thus, bone resorption would not be a major target of CIZ action. In contrast with the bone resorption parameters, histomorphometry on bone formation parameters in CIZ-deficient mice based on calcein double labeling indicated that the bone formation rate and mineral apposition rate were enhanced in CIZ-KO mice in comparison with those in wild-type mice (Fig. 4, A–C). Therefore, CIZ deficiency enhances specifically bone formation activity in vivo without altering bone resorption to increase trabecular bone volume.

We further investigated whether CIZ deficiency enhances osteoblastic activity in a cell-autonomous manner. For this purpose, bone marrow cells were obtained from the femora of wild-type and CIZ-deficient mice and were cultured in the presence of ascorbic acid and β-glycerophosphate. Alkaline phosphatase (ALP) activity was measured after 10 d in cultures. Bone marrow cells from CIZ-KO mice revealed higher basal levels ALP activity than the cells from wild-type mice (Fig. 5).

To examine whether such CIZ deficiency–induced enhancement of ALP activity could be translated into alteration in calcified tissue formation, nodule formation assays were conducted. As shown in Fig. 6 A, CIZ deficiency enhanced the formation of alizarin red–S-positive mineralized nodules in the bone marrow cells in culture. Quantification of the alizarin red-positive nodules revealed approximately threefold enhancement of the nodule formation in CIZ-deficient bone marrow cells (Fig. 6 B). These data indicate that CIZ deficiency enhances osteoblastic activity in a cell-autonomous manner. We also performed bone marrow cultures to examine the effect of CIZ deficiency on osteoclastogenesis. CIZ deficiency did not alter the levels of tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cell formation in culture, indicating that CIZ deficiency does not target cells in osteoclast lineage (Fig. 6 C).

To explore the molecular bases for CIZ deficiency–induced enhancement in bone formation in vivo, we examined expression profile of osteoblastic phenotype-related genes in the bone of these mice. Total RNA was extracted from whole humeri and was subjected to RT-PCR analysis. The mRNA expression levels of osteoblastic phenotype-related genes such as type I collagen (COL), ALP, and osteopontin (OPN), as well as upstream transcription factor gene, osterix (OSX), were enhanced in CIZ-KO mice (Fig. 7, A and B). These data reveal that the increase in bone-forming activity in the cells in CIZ-deficient mice in vivo at least in part is based on the enhancement of expression of the genes encoding osteoblastic phenotype-related proteins as well as upstream osteoblastic transcription factor.

To examine the mechanism of how CIZ deficiency affects osteoblastic activities, effects of CIZ deficiency on bone morphogenetic protein (BMP) actions were examined in vitro and in vivo. First, CIZ deficiency effects on BMP actions were examined in the bone marrow cells in culture. ALP activity in wild-type bone marrow cells was enhanced by approximately threefold by BMP treatment (100 ng/ml recombinant human [rh]BMP-2 for 10 d; Fig. 8). Such BMP-enhanced ALP activity levels in wild-type cells were comparable to ALP activity levels in vehicle-treated CIZ-deficient cells, indicating that the absence of CIZ per se resulted in the enhancement of ALP levels as potently as BMP treatment. BMP treatment of CIZ-deficient cells further enhanced ALP levels up to threefold compared with those in BMP-treated wild-type cells (up to ninefold compared with vehicle-treated wild-type cells). Thus, for the same BMP treatment, presence versus absence of CIZ resulted in one- versus threefold increase in ALP activity levels, respectively, indicating that CIZ deficiency enhances BMP effects in vitro in a cell-autonomous manner.

Furthermore, we asked whether such in vitro CIZ deficiency effects on BMP actions can be translated into bone formation in vivo. For this purpose, bone marrow ablation was conducted because BMP expression was reported to be observed during bone regeneration in bone marrow ablation experiments (20). Bone marrow ablation was performed on the femora of wild-type and CIZ-KO mice, and the bones were examined using two-dimensional microCT 10 d later. Images of CIZ-deficient mice revealed denser newly formed trabecular bone patterns in the ablated area in the femora after bone marrow ablation compared with wild type (Fig. 9 A, rectangular area). Quantification of the newly formed trabecular bone (BV/TV) indicated that fractional bone volume in wild-type mice was increased by 3.0-fold relative to intact contralateral femora after ablation, whereas CIZ deficiency enhanced the increase in fractional bone volume by 3.7-fold relative to intact contralateral femora (Fig. 9 B, P < 0.05), even though the basal intact bone mass levels were higher than wild type. These observations indicate that CIZ deficiency enhances new trabecular bone formation in the ablation-induced bone regeneration model in vivo.

As the bone marrow ablation model does not specifically address whether CIZ deficiency–induced in vivo bone formation is through modulation of BMP actions alone or through bone formation in response to a mixture of injury-related cytokines, we tested whether CIZ deficiency enhances BMP-induced bone formation directly in vivo. 5 μg rhBMP-2 was injected onto the calvariae of wild-type and CIZ-KO mice to induce orthotopic bone formation. The rhBMP-2 injection–induced de novo bone formation was evaluated based on soft X-ray images using an image analyzer and was expressed as the area of the newly formed bone. Direct rhBMP-2 injection produced new bone, which was detected on soft X-ray images as shown in Fig. 10 A (arrow). Quantification of the area of the newly formed bone indicated that approximately twofold greater bone was formed by rhBMP-2 injection onto the calvariae of CIZ-deficient mice than that produced by rhBMP-2 injection onto the wild-type calvariae (Fig. 10 B). These data demonstrate that BMP-2–induced bone formation is enhanced in the absence of CIZ, indicating that CIZ suppresses BMP-signaling in vivo.

Our data revealed that deficiency in CIZ increased basal bone mass in adult mice. This phenotype was through the enhancement of bone formation in vivo, increased mineralized nodule formation in bone marrow cells, and elevated expression levels of the genes encoding osteoblastic phenotype-related marker proteins including COL1, ALP, OPN, and OSX in bone in vivo. CIZ deficiency also enhanced bone marrow ablation-induced new bone formation in vivo as well as BMP injection–induced de novo osteogenesis in adult mice. Thus, CIZ acts as an inhibitor of bone formation in adult mice in vivo at least in part through its suppression of BMP actions.

CIZ localizes at adhesion plaques, transfers into nuclear compartments, binds to consensus DNA sequences, and activates promoters of the genes encoding enzymes that degrade matrix proteins (15). As CIZ is expressed in cultured osteoblasts (17–19), this protein was thought to be a modulator of osteoblastic function. However, there was an ambiguity regarding CIZ actions in vitro because CIZ was reported to regulate expression of osteoblast-related genes positively or negatively depending on culture conditions (18, 19). Thus, in vitro overexpression experiments failed to determine CIZ function in osteoblasts. Our data in this paper conclusively indicated that CIZ deficiency enhanced adult bone volume levels in vivo. Therefore, CIZ is an inhibitory determinant for adult bone mass in vivo. CIZ deficiency enhances basal as well as BMP-induced alkaline phosphates activities in osteoblasts in culture. This implies that in vivo observations on the elevation of basal bone mass in CIZ-deficient mice as well as CIZ deficiency enhancement in BMP-induced bone formation on calvaria in vivo would be reflecting cell-autonomous actions of CIZ. Furthermore, we have observed previously in cultured osteoblasts that CIZ overexpression suppressed Smad-dependent activation of transcription (19). In the present work, CIZ deficiency enhanced bone formation and mineral apposition rates without altering osteoclast number and osteoclast surface in vivo. Therefore, CIZ suppresses adult bone mass levels by reduction in osteoblastic activity through the inhibition of BMP actions via interference of Smad activities.

Several signaling mechanisms have been reported to regulate adult bone mass in the last several years. Low-density lipoprotein receptor-related protein 5 (LRP5) acts as a coreceptor for Wingless-Int family member signaling, and loss of function mutation in LRP5 gene leads to osteoporosis. Similarly, other types of LRP5 mutation blocks the binding of dickkopf (DKK), an inhibitor of LRP5, and leads to increase in bone mass in humans. Another group of molecules is involved in negative regulation of adult bone mass via suppression of function; these include inhibitors of BMP such as Tob (21, 22). It is intriguing that multiple molecules such as Tob and CIZ are involved in suppression of osteoblastic functions. Bone is the storage site for BMP and, thus, the cells in bone could be subjected to continuous and possibly excessive exposure to BMP-driven stimuli. CIZ and Tob may partly block such excessive BMP signals and, thus, prevent abnormal or ectopic bone formation. Such negative regulation of BMP action by at least two distinct effectors, in this case CIZ and Tob, could be an important biological back-up pathway. This makes us speculate about the presence of strong potency in BMP action during the determination in adult bone mass. As deletion of even one of the two molecules CIZ and Tob results in exhibition of bone phenotype, simultaneous presence of both of the two is necessary for normal bone mass maintenance. Furthermore, we observed that both CIZ and Tob expression levels in osteoblasts in cultures are enhanced by BMP treatment. Thus, these two molecules form negative feedback circuits with BMP separately or in parallel. It can be observed in other biological systems that growth factors stimulate biological functions of cells while these molecules simultaneously enhance expression of inhibitors for their own activities (23–26). The balanced activities of both accelerators (stimulating cytokines) and decelerators (their inhibitors) would determine the normal balance in the biological functions of organs (in this case, bone). To our knowledge, CIZ is the first attachment- and transcription-related molecule that inhibits osteoblastic function in adult bone in vivo.

BMP is an important molecule in the body in various aspects. BMP is required for the maintenance of bone to support the body or to protect internal organs. BMP also regulates morphogenesis during embryonic development. In accordance with such a variety of BMP actions, numerous groups of BMP inhibitors have been identified. One group includes soluble BMP inhibitors such as noggin, chordin, SOST, and gremlin, which act outside the cells. These molecules inhibit BMP signaling via binding to BMP ligands to prevent their binding to their cognate receptors (27–29). BMP3 binds to BMP receptor and inhibits bone formation. Other groups consist of inhibitory Smad proteins, such as Smad6 and Smad7. These negative Smads inhibit intracellular BMP signaling by binding to BMP receptors or regulatory Smads (30, 31). A recently identified group contains proteins such as Tob, which acts as a Smad repressor to inhibit BMP signaling at the level of the nuclear transport or transcription (31). CIZ is special and distinct from other proteins because it localizes at adhesion plaques and simultaneously activates gene expression in addition to Smad modulation (19). It is also unique in terms of its localization at nuclear matrix (15). Thus, CIZ should be considered as a unique class of BMP inhibitor in vivo.

It is notable that CIZ deficiency did not result in major alterations in skeletal morphogenesis during embryonic stages. In the case of other BMP inhibitors, such as noggin, its deficiency causes hypertrophy of cartilage and severe fusion of joints. These knockout mice do not survive after birth (31). This may imply that CIZ deficiency could be backed up by the presence of other BMP inhibitors during the early period of life when embryonic development of the skeleton is taking place. Alternatively, CIZ-deficient mice may not have an obvious (major) skeletal phenotype in embryonic stage because of the rapid growth during that period of life. Our observations revealed that CIZ action is more important in bone metabolism in “adulthood” than in the embryonic stage, even when the target is BMP. In fact, we observed that high bone mass phenotype could be detectable even in mice of >50 wk old (unpublished data). These facts suggest that different BMP inhibitors would have diverse roles depending on time and location as well as cellular backgrounds. As most age-related diseases, including osteoporosis, affect individuals who grow and live normally up to adult stage, the roles of molecules, such as CIZ, could be critical for considering the pathogenesis of age-related bone diseases in human adults. CIZ-deficient mice also revealed defects in spermatogenesis (32). However, except for that, bone is the only organ to reveal a phenotype in these mice. Even for males, spermatogenesis may not be a major problem in an aged population; therefore, it would be of interest to contemplate development of new drugs to target against CIZ for osteoporosis in aged patients.

In conclusion, our observations demonstrate that CIZ deficiency enhances the levels of adult bone mass and bone formation. This was observed to be due to enhancement in the BMP signaling pathway in vivo and in osteoblasts in a cell-autonomous manner. Thus, CIZ would act as a critical suppressor for BMP signaling in osteoblastic function in adult in vivo.

CIZ-deficient mice were produced by Nakamoto et al. (32). Wild-type and CIZ-deficient mice in a DBA × C57BL/6 F2 background were derived from the original heterozygous cross. These mice were maintained as separate colonies. The mice were kept under controlled housing condition at 24°C on 12:12 light/dark cycles with the light cycle starting at 7:00 a.m., fed with standard laboratory chow (MF; Oriental Yeast), and given tap water. CIZ-deficient and wild-type female mice (8 wk old) were used in most of the experiments. For body weight data and some of the experiments, older mice were also examined. All experiments were conducted according to the institutional guidelines for animal welfare.

Trabecular bone structure and bone volume (bone volume/tissue volume [BV/TV]) were examined based on two-dimensional micro X-ray computed tomography (microCT) analysis, using a microCT apparatus (Musashi; Nittetsu-ELEX). The fractional bone volume (BV/TV) was measured in a square area of 0.8 mm² (2,000 μm × 400 μm) in the distal metaphyseal region of the femora or the fourth lumbar vertebrate and was quantified using Luzex-F Image analyzing system (Nireco). The threshold for the measurements was set at 110 for the analyses as determined previously based on the correlation to the values to those obtained in histomorphological measurements (33).

Bone marrow cells were obtained by flushing out the cells from the medullary cavity of femora with α-minimum Eagle's medium (α-MEM). The cells were rinsed, counted, and plated in 96-well plates (5 × 10⁴ cells/well) and were cultured in α-MEM supplemented with 10% FBS, 50 μg/ml ascorbic acid, 10 mM β-glycerophosphate, and antibiotics/antimycotics. In some of the cultures, the cells were treated with 100 ng/ml rhBMP-2. Medium was changed every 3 d, and the cells were cultured for 10 d. The cells were rinsed twice with ice-cold PBS and scraped into 10 mM Tris-HCl containing 2 mM MgCl₂ and 0.05% Triton X-100, pH 8.2. The cell lysates were briefly sonicated on ice after two cycles of freezing and thawing. Aliquots of supernatants were subjected to ALP activity measurement and protein assay according to Bradford's method. In brief, aliquots of cell lysates were mixed with assay buffer containing 10 mM p-nitrophenyl phosphate in 0.1 M sodium carbonate buffer, pH 10, supplemented with 1 mM MgCl₂ and followed by an incubation at 37°C for 30 min. After adding 1 M NaOH, the amounts of p-nitrophenol liberated in the assay mixtures were measured using a spectrophotometer.

Bone marrow cells were plated in 24-well plates (10⁶ cells/well). The cells were cultured in α-MEM supplemented with 10% FBS, 50 μg/ml ascorbic acid, 10 mM β-glycerophosphate, and antibiotics/antimycotics (100 U/ml penicillin G, 100 μg/ml streptomycin sulfate, and 0.25 μg/ml amphotericin B; GIBCO BRL). Medium was changed every 3 d, and the cells were cultured for 2 wk. At the end of the cultures, the cells were fixed with 10% formalin/saline and stained for calcium with alizarin red S to identify mineralized bone nodules.

The proximal and distal epiphyseal ends were removed from femora, and bone marrow cells were flushed out. The number of total bone marrow cells was counted, and the cells were plated in 48-well plates (0.8 cm² per well) at a density of 5 × 10⁵ cells/well. The cells were cultured in α-MEM supplemented with 10% FBS, 100 μg/ml antibiotics-antimycotics mixture, 10 nM 1,25(OH)₂ vitamin D₃, and 100 nM dexamethasone. The medium was changed on day 3. TRAP-positive osteoclast-like multinucleated cells (greater than two nuclei) were counted at the end of 5 d of culture.

Osteoclast number per bone surface (Oc.N/BS, N/mm) and the percentage of osteoclast surface to bone surface (Oc.S/BS, %) were obtained by counting TRAP-positive cells and by measuring the area in serial 7-μm-thick sagittal histological sections of femora of wild-type (n = 8) and CIZ-deficient (n = 6) mice (8 wk old) at 200-fold magnification. TRAP-positive cells forming resorption lacunae on the surface of the trabeculae and containing one or more nuclei were identified as osteoclasts. For in vivo fluorescent labeling, calcein (1.6 mg/kg body weight) was intraperitoneally injected on days 0 and 4. Wild-type (n = 8) and CIZ-deficient (n = 6) mice (8 wk old) were killed on day 6. Femora were fixed in 99.5% ethanol and embedded in methylmethacrylate without decalcification. Sagittal histological sections (7 μm thick) were prepared, and the calcein images were visualized by fluorescent light microscopy. Histomorphometry of bone sections was performed using at least six optical fields within the secondary spongiosa. The abbreviations for histomorphometric parameters were according to the recommendation by the American Society of Bone and Mineral Research Histomorphometry Nomenclature Committee (34).

Total RNA of whole humerus was extracted according to acid guanidium thiocyanate-phenol-chloroform method (35). First-strand cDNA was synthesized using 1 μg of the total RNA and Molony murine leukemia virus reverse transcriptase. Primers were synthesized on the basis of the reported mouse cDNA sequences for GAPDH (GenBank/EMBL/DDBJ accession no. NM_008084), OPN (GenBank/EMBL/DDBJ accession no. AF515708), ALP (GenBank/EMBL/DDBJ accession no. AF285233), COL (GenBank/EMBL/DDBJ accession no. U38544), and OSX (GenBank/EMBL/DDBJ accession no. AF184902). Primers for PCR reactions were as follows: GAPDH, forward, 5′-ACCACAGTCCATGCCATCAC-3′ and reverse, 5′-TCCACCACCCTGTTGCTGTA-3′; OPN, forward, 5′-CGACGATGATGACGATGATGAT-3′ and reverse, 5′-CTGGCTTTGGAACTTGCTTGAC-3′; ALP, forward, 5′-ATTGCCCTGAAACTCCAAAACC-3′ and reverse, 5′-CCTCTGGTGGCATCTCGTTATC-3′; COL, forward, 5′-TTTGTGGACCTCCGGCTC-3′ and reverse, 5′-AAGCAGAGCACTCGCCCT-3′; and OSX, forward, 5′-CTGGGGAAAGGAGGCACAAAGAAG-3′ and reverse, 5′-GGGTTAAGGGGAGCAAAGTCA-GAT-3′.

Amplification was performed at 21–25 cycles within a linear range. Each cycle was set at 94°C for 20 s; 60°C for 30 s; and 72°C for 40 s in a 25-μl reaction mixture containing 0.5 μl of each cDNA, 200 mM of each primer, 0.2 mM of dNTP, and 1 U Taq DNA polymerase (Takara). After amplification, 10 μl of each reaction mixture was subjected to electrophoresis to be analyzed on 1% agarose gel. The bands were visualized by ethidium bromide staining.

Bone mineral density (g/cm²) of the entire femora was measured based on dual-energy X-ray absorptiometry (DEXA) using an apparatus specifically designed for small animals (PIXI; GE Lunar) equipped with a laptop computer. Calibration was performed using a standard block according to the manufacturer's instruction.

Direct injection of BMP-2 onto the calvariae was conducted as described previously (36). In brief, rhBMP-2 (5 μg in 20 μl saline) was injected onto the parietal bones of wild-type (n = 8) and CIZ-deficient (n = 6) mice (8 wk old). rhBMP-2 was obtained from the Genetics Institute. Injection was performed every other day for 10 d, and the animals were killed 1 d after the last injection. Calvariae were fixed in 4% paraformaldehyde in PBS, pH 7.4, for 24 h, removed from the skull, and subjected to soft X-ray examination. X-ray images were recorded and quantified by using a Luzex-F automated image analysis system (Nireco).

The results were presented as mean ± SEM, and statistical evaluation was performed based on Student's t test. p-values <0.05 were considered to be significantly different.

This research was supported by the Grants-in-Aid received from the Japanese Ministry of Education (21st Century Center of Excellence Program, Molecular Destruction and Reconstitution of Tooth and Bone, nos. 14207056, 16659405, 16027215, and 16022221), grants from Japan Space forum, NASDA, and Japan Society for Promotion of Science (Integrated Action Initiative in Core to Core Program, Research for the Future Program, Genome Science).

The authors have no conflicting financial interests.