In Vivo Targeted Deletion of Calpain Small Subunit, Capn4, in Cells of the Osteoblast Lineage Impairs Cell Proliferation, Differentiation, and Bone Formation*

Calpains are intracellular cysteine proteases, which include widely expressed μ- and m-calpains (1). Both μ-calpains and m-calpains are heterodimers consisting of a large catalytic subunit and a small regulatory subunit. The calpain small subunit encoded by the gene Capn4 directly binds to the intracellular C-terminal tail (C-tail) of the receptor for parathyroid hormone and parathyroid hormone-related peptide and modulates its cellular functions in osteoblasts in vitro (2). To investigate a potential role of the calpain small subunit in osteoblasts in vivo, we generated osteoblast-specific Capn4 knock-out mice using the Cre-LoxP system (3). Mutant mice had smaller bodies with shorter limbs, reduced trabecular bone with thinner cortices, and decreased osteoblast number. In vitro analysis confirmed that deletion of Capn4 in osteoblasts severely affected multiple osteoblast functions including proliferation, differentiation, and matrix mineralization. Collectively, our findings provide the first in vivo demonstration that the calpain small subunit is essential for proper osteoblast activity and bone remodeling.

Calpains are a family of Ca 2ϩ -dependent intracellular cysteine proteases that include ubiquitously expressedand m-calpains (1,4). Both -calpains and m-calpains form heterodimers consisting of a large catalytic subunit (80 kDa) encoded by the genes Capn1 and Capn2, respectively, and a small regulatory subunit (28 kDa) encoded by the gene Capn4 (1). Disruption of Capn4 eliminates both -calpains and m-calpain activities in embryonic fibroblasts (5), suggesting that the calpain small subunit is essential for maintenance of calpain stability and activity. Notably, genetic ablation of Capn4 results in early embryonic lethality, which demonstrates an essential role of the calpain small subunit during development (5,6).
Several lines of evidence suggest that calpains are necessary for proper osteoblast function in vitro (7)(8)(9)(10)(11). However, a physiological role of calpains in osteoblasts in vivo remains to be established. Chemical inhibition of calpain activity reduces osteoblast proliferation and differentiation in the MC3T3-E1 osteoblastic cell line (11). Moreover, we have previously reported that the calpain small subunit directly binds to the intracellular C-tail of the receptor for PTH 2 and PTH-related peptide and modulates its ligand-mediated cellular functions (2). Both ligands are known regulators of bone homeostasis in vivo through their direct actions on cells of the osteoblast lineage (12)(13)(14). Taken together, these findings suggest that the calpain small subunit could play a critical but yet unknown role in osteoblast biology in vivo.
To test this hypothesis, we have conditionally ablated Capn4 in cells of the osteoblast lineage in vivo by using the Cre-LoxP system. Lack of the calpain small subunit in osteoblasts caused a significant decrease of both trabecular and cortical bone, which was associated with a severe impairment of osteoblast proliferation and differentiation. These findings are the first in vivo evidence that the calpain small subunit plays a crucial role in osteoblast function and bone homeostasis.
Whole Mount Skeletal Staining-The whole mount skeletal staining was performed as described previously (20). Briefly, newborn mice were fixed in ethanol for 5 days and then in acetone for 2 days. Staining with Alizarin red S and Alcian blue was performed for 3 days at 37°C. After washing with distilled water, the skeleton was cleared in 1% KOH and taken through graded steps into 100% glycerol.
Histomorphometry-For dynamic histomorphometry, animals were injected intraperitoneally with fluorochromes, calcein, and demeclocycline (20 g/g of body weight, Sigma-Aldrich), 3 and 10 days before sacrifice, respectively. Bones were fixed and embedded in methyl methacrylate resin as described previously (21). Five-micrometer sections were stained with Masson Trichrome or coverslipped unstained, and histomorphometric analysis was performed with the Osteomeasure system (Osteometrics Inc., Atlanta, GA) using standard procedures. Tibial sections were measured in the proximal metaphysis beginning 340 m below the chondro-osseous junction.
Analysis of Bromodeoxyuridine Incorporation-Two-weekold mice were injected intraperitoneally with 100 g of bromodeoxyuridine (BrdUrd) and 12 g of fluorodeoxyuridine per g of body weight 2 h before sacrifice (Sigma-Aldrich). After sacrifice, hind limbs were dissected, fixed, decalcified, and embedded in paraffin, and longitudinal sections across the tibia and femur were obtained. To identify actively proliferating cells, nuclei that had incorporated BrdUrd were detected using a Zymed BrdUrd immunostaining kit (Zymed Laboratories Inc., South San Francisco, CA). For each section, high power field images of the primary spongiosa and the proliferating chondrocyte layer were photographed. All BrdUrd-positive (black) and BrdUrd-negative (light blue) nuclei in these zones were counted separately. Four or five nonconsecutive sections were counted for each of two or three Capn4 flox/flox , Osx-Cre ϩ/Ϫ Capn4 flox/ϩ , and Osx-Cre ϩ/Ϫ Capn4 flox/flox littermate specimens at each time point.
Osteoblast Isolation and Culture-Osteoblasts were isolated from calvariae of newborn mice by serial digestion in ␣-minimum essential medium (␣MEM) (Invitrogen) containing 0.25 mg/ml type I and 0.75 mg/ml type II collagenases (Worthington, Lakewood, NJ). Calvariae were digested for 15 min at 37°C with constant agitation. The digestion solution was collected, washed with fresh medium, and digested five additional times. Digestions 3-6 were collected, washed with ␣MEM, and cultured in ␣MEM supplemented with 10% fetal bovine serum (HyClone, Logan, UT), 1% penicillin/streptomycin (Invitrogen Corp.) for 48 h. Medium was changed every other day.
Adenovirus Infection-Monolayer Capn4 flox/flox osteoblasts were infected with either control (adeno-lacZ) or Cre-recombinase adenovirus (adeno-Cre) at a multiplicity of infection of 100 for most experiments (26). Osteoblasts were harvested after 48 h. Genomic DNA and total RNA were extracted from osteoblasts for determination of efficiency of Cre-recombinase.
Bone Nodule Assay-Cells were cultured in ␣MEM with 10% fetal bovine serum and 1% penicillin/streptomycin supplemented with 50 g/ml L-ascorbic acid and 10 mM ␤-glycerophosphate (Sigma-Aldrich) (differentiation medium) for 21 and 28 days, respectively. Cells were then fixed with 70% ethanol, and mineralized calcium-phosphate deposits were stained with either Alizarin red S or von Kossa. Bone nodules were counted using a dissecting microscope. In some experiments, Alizarin red S dye was eluted and quantified as described elsewhere (27).
Osteoblast Apoptosis Assay in Vitro-Forty-eight h after adenoviral infection, osteoblasts were replated in 24-well plates at a density of 2.5 ϫ 10 4 cells/cm 2 . Cells were stained with annexin V-PE and 7-aminoactinomycin D using Guava PCA Nexin kit and analyzed by Guava personal cytometer (Guava Technology Inc., Hayward, CA).
Gene Silencing by siRNA-UMR106-01 (UMR) rat osteosarcoma cells were plated at a density of 5 ϫ 10 4 cells/cm 2 in 24-well plates. Capn4 and control nonfunctional siRNA were transfected into UMR cells using Oligofectamine as recommended by the manufacture (Invitrogen). Cells were harvested 48 h after transfection.
Statistics-Data were calculated from 3 to 5 independent experiments and expressed as the mean Ϯ S.E. of either duplicate or triplicate determinations. Statistical analysis was performed using the analysis of variance or the unpaired Student's t test. Statistical significance was determined using Fisher's projected least significant difference, and p values less than 0.05 were accepted as significant.
Osx-Cre ϩ/Ϫ Capn4 flox/flox Mice Have Decreased Trabecular Bone and Reduced Expression of Osteoblast-specific Markers-Detailed histological and in situ hybridization analyses were then performed on hind limbs isolated from 2-, 4-, and 9-weekold male mice, respectively, to investigate a putative role of Capn4 in bone modeling and remodeling. Tibiae of 2-week-old Osx-Cre ϩ/Ϫ Capn4 flox/flox mice had markedly reduced trabecular bone when compared with both Osx-Cre ϩ/Ϫ Capn4 flox/ϩ and Capn4 flox/flox control littermates ( Fig. 2A). Moreover, expression of osteoblast markers such as MMP13, Col.1, and OC was severely impaired in mutant bones ( Fig. 2A and data not shown). A similar phenotype was also detectable in 4-and 9-week-old mutant mice of either sex (Fig. 2, B and C, and data not shown). Surprisingly, at these ages, Osx-Cre ϩ/Ϫ Capn4 flox/flox/ϩ mice displayed a loss of trabecular bone, which was virtually indistinguishable from that of Osx-Cre ϩ/Ϫ Capn4 flox/flox mutants, as shown by routine histology and detection of Col.1 mRNA by in situ hybridization analysis (Fig.  2, B and C). Conversely, Osx-Cre ϩ/Ϫ Capn4 ϩ/ϩ mice did not present any detectable bone abnormality at any prenatal or postnatal age. CT analysis showed no significant differences in 12-week-old male and female Osx-Cre ϩ/Ϫ Capn4 ϩ/ϩ versus Capn4 flox/flox when compared with parameters such as bone volume per total volume (16.4 Ϯ 2.0 versus 15.3 Ϯ 4.0%), trabecular number (5.1 Ϯ 0.2 versus 4.6 Ϯ 0.8/mm), and trabecular spacing (187.3 Ϯ 10.2 versus 216.85 Ϯ 44.5 m) (data for female not shown). These results, thus, exclude any adverse effect of the Cre transgene per se on bone homeostasis. It is, therefore, likely that lack of a single Capn4 allele in cells of the osteoblast lineage is sufficient to cause bone loss, which, however, is detectable exclusively in adult mice.

Lack of Capn4 in Cells of the Osteoblast Lineage Reduces Bone Cortical Thickness and Bone
Architecture-Cortical bone morphology was examined by CT analysis at the mid-femoral diaphysis. Total and cortical bone areas and cortical thickness were significantly reduced in Osx-Cre ϩ/Ϫ Capn4 flox/flox mice versus controls, suggesting that the bone is smaller and that periosteal apposition is inhibited in Osx-Cre ϩ/Ϫ Capn4 flox/flox mice (Fig. 3B). In addition, the polar moment of inertia, an index of torsional strength, was significantly lower in mutants than controls (Fig. 3B). Thus, deletion of Capn4 in cells of the osteoblast lineage causes bone loss in both trabecular and cortical compartments and severely affects parameters that predict bone strength.
Lack of Capn4 Impairs Proliferation of Cells of the Osteoblast Lineage-The reduced osteoblast number observed in Osx-Cre ϩ/Ϫ Capn4 flox/flox mice could be a consequence of either decreased proliferation or increased apoptosis. To distinguish between these two possibilities, proliferation of cells of the osteoblast lineage was examined by BrdUrd incorporation in 1and 2-week-old mice. The number of BrdUrd-positive cells detected in the primary spongiosa was significantly reduced in Osx-Cre ϩ/Ϫ Capn4 flox/flox when compared with Osx-Cre ϩ/Ϫ Capn4 flox/ϩ and Capn4 flox/flox littermates (Fig. 4B, panel 1), whereas chondrocyte proliferation, which served as an internal control for proper BrdUrd uptake and distribution, was not affected (Fig. 4B, panel 2). No significant difference in the number of apoptotic cells determined by terminal deoxynucleotidyl transferase-mediated nick-end labeling staining was detected in 2-week-old mutant bones versus controls (data not shown). Collectively, these data indicate that the reduced osteoblast Lack of Capn4 Impairs Osteoblast Differentiation and Mineralization in Vitro-Osteoblast functions were next investigated in vitro. For this purpose, primary calvarial osteoblasts were harvested from Osx-Cre ϩ/Ϫ Capn4 flox/ϩ and Osx-Cre ϩ/Ϫ Capn4 flox/flox newborn mice and seeded at a density of 5.0 ϫ 10 3 cells/cm 2 . Levels of Capn4 mRNA expression were reduced by ϳ90% in Osx-Cre ϩ/Ϫ Capn4 flox/flox osteoblasts in comparison with Osx-Cre ϩ/Ϫ Capn4 flox/ϩ . Cell growth was indistinguishable between Osx-Cre ϩ/Ϫ Capn4 flox/flox and Osx-Cre ϩ/Ϫ Capn4 flox/ϩ calvarial cells until day 10 and became significantly lower in Osx-Cre ϩ/Ϫ Capn4 flox/flox cells than Osx-Cre ϩ/Ϫ Capn4 flox/ϩ cells on day 14. Osteoblast differentiation and mineralization, assessed by ALP activity and mRNA expression of various osteoblast markers, and number of bone nodules, normalized by cell number at the time of assay, respectively, were all significantly lower in Osx-Cre ϩ/Ϫ Capn4 flox/flox than in Osx-Cre ϩ/Ϫ Capn4 flox/ϩ calvarial cells (Fig. 5).
Next, to further investigate a role of Capn4 in osteoblast differentiation, calvarial cells isolated from Capn4 flox/flox mice were infected with adenovirus expressing either Crerecombinase (adeno-Cre) or ␤-galactosidase (adeno-lacZ). The efficiency of Cre-recombinase was assessed at both genomic DNA and RNA levels using real-time qPCR. Approximately 85% of Capn4 flox alleles were successfully excised in adeno-Cre-infected cells, and their levels of Capn4 mRNA expression were reduced to less than 10% of those observed in adeno-lacZ-infected control cells. Forty-eight h after adenoviral infection, cells were replated at the density of 1.3 ϫ 10 5 cells/cm 2 (higher density) to allow them to reach an indistinguishable cell number at the time of assay and, thus, minimize the confounding effect of impaired proliferation on the differentiation process. Notably, cell number of both mutant and control cells similarly increased and reached a plateau around day 10, when cells were seeded at the high density (data not shown). When the number of the apoptotic events was determined by annexin V staining, no significant difference was detected in Capn4-null calvarial cells when compared with the controls (ratio of annexin V-positive early and late apoptotic cells per total cells; control, 14.3 Ϯ 1.8%; Capn4-null, 15.2 Ϯ 1.3%). Consistent with the in vivo data, calvarial cells lacking Capn4 showed significantly reduced ALP activity (Fig. 6A) as well as impaired expression of several genes including NaPi3, PGK, and Runx2 that are early markers of osteoblast differentiation (Fig. 6B) (28). Expression of late markers of osteoblast differentiation such as ALP, Col.1, and OC was also significantly down-regulated in Capn4-null cells versus controls (Fig. 6C). Lastly, Capn4-null cells grown in differentiation medium for 21 and 28 days, respectively, formed only 33% of the bone nodules observed in control cells (Fig. 6, D and E). Taken together, these in vitro results confirm that lack of Capn4 impairs both proliferation and differentiation of cells of the osteoblast lineage as indicated by the in vivo model.
Lack of Capn4 in Cells of the Osteoblast Lineage Reduced PTH-stimulated c-fos mRNA Expression in Vitro-To further examine an effect of Capn4 ablation on gene transcription in cells of the osteoblast lineage, primary calvarial cells were treated with PTH for 0, 1, and 2 h, and levels of c-fos mRNA expression were assessed using real-time qPCR. c-fos is a key downstream target gene of PTH, and its expression increases rapidly and transiently upon PTH stimulation (31)(32)(33). In adeno-lacZ-infected control calvarial cells, PTH treatment for 1 and 2 h increased c-fos mRNA levels 17-and 9-fold above basal levels, respectively. This effect was reduced by ϳ35% in adeno-Cre-infected Capn4-null calvarial cells (Fig. 7A). We also silenced the calpain small subunit in UMR cells by using siRNA technology. siRNAs that specifically targeted Capn4 (Capn4-1 and Capn4-2) reduced calpain activity to 25 and 55%, respectively, when compared with control siRNA (Fig. 7B, panel  1). Capn4-knock-out in UMR cells resulted in a significant impairment of PTH-mediated increase of c-fos expression (Fig. 7B, panel  2). Cell surface expression of the receptor for PTH and PTH-re-lated peptide determined by ligand binding assay was indistinguishable between control and Capn4-null cells (data not shown). Collectively, these results suggest that Capn4 could regulate osteoblast function, at least in part, by modulating PTH activity in these cells.

DISCUSSION
Here we report the novel finding that Capn4 is critically important in bone development and remodeling in vivo. Given the early lethality of the universal Capn4 knock-out embryos, Osx-Cre ϩ/Ϫ Capn4 flox/flox was an appropriate animal model to investigate the role of the calpain small subunit in cells of the osteoblast lineage (5,6).
Calpains belong to a family of intracellular cysteine proteases that have been shown to cleave numerous and diverse substrates (1). Our study provided direct evidence that lack of Capn4 severely affects proliferation of cells of the osteoblast lineage both in vivo and in vitro. Consistent with these findings, it was previously reported that a cell-permeable calpain inhibitor attenuates proliferation in the MC3T3-E1 cell line (10,11). The role of calpain in the cell cycle has been controversial. Capn4-null embryonic fibroblasts with no calpain activity proliferate normally (5). However, several studies have suggested that calpain plays a role in the cell cycle, mainly at G 1 to S transition, by regulating proteolysis of various proteins such as cyclin D and A, cyclin-dependent kinase 2, and p53 (31)(32)(33). Our novel animal model demonstrates that regulation of cell proliferation is indeed a crucial biological function of the calpain small subunit, at least in osteoblasts.
Lack of Capn4 in osteoblasts significantly reduces ALP activity and expression of differentiation markers such as ALP, Col.1, OC, and Runx2 mRNAs. Consistent with our findings, a cellpermeable calpain inhibitor was reported to attenuate osteoblast differentiation in the MC3T3-E1 cell line (11). The finding constitutes the first in vivo evidence that Capn4 is an essential modulator of cell terminal differentiation in osteoblasts. Notably, immunostaining forand m-calpains performed in the MC3T3-E1 cell line revealed that m-calpain is a cytoplasmic protein, whereas -calpain is both nuclear and cytoplasmic, which implies that calpains may control gene expression directly at the transcriptional level (11).
It has been reported that PTH stimulates expression of the immediate early gene, c-fos, which has a critical role in bone biology, by regulating phosphorylation of the cAMP response element binding protein at serine 133 by protein kinase A (34 -38). Moreover, we have previously reported that MC3T3-E1 osteoblastic cells stably expressing calpastatin show markedly reduced PTHmediated cAMP accumulation (2). Lastly, our present study indi-   JULY 25, 2008 • VOLUME 283 • NUMBER 30

Capn4 Is Essential for Proper Osteoblast Function
cates that lack of Capn4 reduces PTH-stimulated expression of c-fos in both primary calvarial and UMR cells. Collectively, these data thus suggest that Capn4 may regulate osteoblast functions, at least in part, by modulating PTH activity in these cells.
A role of calpains in bone homeostasis has been previously proposed by the analysis of universal -calpain knock-out mice, which develop osteopenia secondary to an increased osteoclast activity. No detectable osteoblast phenotype was reported in these mutant mice (39). Thus, their bone phenotype clearly differs from that of the osteoblast-specific Capn4 knockout mice, which showed both impaired osteoblast number and function and reduced osteoclast number and activity. Altogether, the phenotype of the Capn4 knockout mice may reflect either loss of m-calpain function or loss of both m-calpains and -calpains function in osteoblasts.
Several lines of evidence indicated that lack of a single Capn4 allele in cells of the osteoblast lineage causes a detectable bone phenotype. Osx-Cre ϩ/Ϫ Capn4 flox/ϩ mice started to lose body weight around 6 weeks of age, and their mean body weight became indistinguishable from that of Osx-Cre ϩ/Ϫ Capn4 flox/flox mice at 10 weeks of age. Although body weight only serves as an indirect indicator of skeletal growth, the data suggest impaired skeletal growth in Osx-Cre ϩ/Ϫ Capn4 flox/ϩ mice. Moreover, histological analysis showed that Osx-Cre ϩ/Ϫ Capn4 flox/ϩ mice rapidly develop an osteoporotic phenotype between 2 and 4 weeks of age. Lastly, proliferation of cells of the osteoblast lineage as assessed by BrdUrd incorporation at 2 weeks was significantly reduced in Osx-Cre ϩ/Ϫ Capn4 flox/ϩ mice when compared with Capn4 flox/flox mice, although it was still ϳ2-fold higher than in Osx-Cre ϩ/Ϫ Capn4 flox/flox specimens. Importantly, Osx-Cre ϩ/Ϫ Capn4 ϩ/ϩ were, virtually and as confirmed by CT analysis, indistinguishable from wild-type controls, which indicates that the bone phenotype of Osx-Cre ϩ/Ϫ Capn4 flox/ϩ was not the mere consequence of a nonspecific effect secondary to the presence of the Cre transgene in cells of the osteoblast lineage. These results are suggestive of a possible haploinsufficiency phenotype in mice lacking one allele of Capn4 in cells of the osteoblast lineage.
In summary, lack of Capn4 in cell of the osteoblast lineage negatively affects both trabecular and cortical bone by impairing proliferation and differentiation of osteoblastic cells. Our  data thus indicate that calpains could be important molecules in pathological conditions such as osteoporosis.