CILIA-LIKE STRUCTURES AND POLYCYSTIN-1 IN OSTEOBLASTS/OSTEOCYTES AND ASSOCIATED ABNORMALITIES IN SKELETOGENESIS AND RUNX2 EXPRESSION

Abstract

We examined the osteoblast/osteocyte expression and function of polycystin-1 (PC1), a transmembrane protein that is a component of the polycystin-2 (PC2)-ciliary mechanosensor complex in renal epithelial cells. We found that MC3T3-E1 osteoblasts and MLO-Y4 osteocytes express transcripts for PC1, PC2 and the ciliary proteins Tg737 and Kif3a. Immunohistochemical analysis detected cilia-like structures in MC3T3-E1 osteoblastic and MLO-Y4 osteocyte-like cell lines as well as primary osteocytes and osteoblasts from calvaria. Pkd1^m1Bei mice have inactivating missense mutations of Pkd1 gene that encodes PC1. Pkd1^m1Bei homozygous mutant mice demonstrated delayed endochondral and intramembranous bone formation, whereas heterozygous Pkd1^m1Bei mutant mice had osteopenia caused by reduced osteoblastic function. Heterozygous and homozygous Pkd1^m1Bei mutant mice displayed a gene dose-dependent decrease in the expression of Runx2 and osteoblast-related genes. In addition, overexpression of constitutively active PC1 C-terminal constructs in MC3T3-E1 osteoblasts resulted in an increase in Runx2 P1 promoter activity and endogenous Runx2 expression, as well as an increase in osteoblast differentiation markers. Conversely, osteoblasts derived from Pkd1^m1Bei homozygous mutant mice had significant reductions in endogenous Runx2 expression, osteoblastic markers and differentiation capacity ex vivo. Co-expression of constitutively active PC1 C-terminal construct into Pkd1^m1Bei homozygous osteoblasts was sufficient to normalize Runx2 P1 promoter activity. These findings are consistent with a possible functional role of cilia and PC1 in anabolic signaling in osteoblasts /osteocytes.

The study of the genetics of hereditary cystic diseases of the kidney led to the recognition of a novel mechanosensing mechanism involving cilia, PKD1 and PKD2 (1-5). Autosomal dominant polycystic kidney disease (ADPKD) is caused by inactivating mutations of PKD1 (which produces polycystin-1, PC1) or PKD2 (which produces polycystin-2, PC2) (1). Also, mutations of Tg737, kif3a, and cpk, which encode the ciliary proteins polaris, Kif3a, and cystin, causes autosomal recessive polycystic kidney disease (6–9). PC1 is a 4,303 amino acid cell-surface-expressed protein that has a multidomain extracellular region, an 11 transmembrane region (10), and a ~200 amino acid cytoplasmic C-terminal tail (3), whereas PC2 is a Ca²⁺-permeable cation channel belonging to the transient receptor potential (TRP) family (11–13). There is compelling evidence that PC1 forms heterodimers with PC2 to create a polycystin complex that localizes with Tg737, Kif3a and cpk in cilium and may be involved in cell-matrix interfaces and cell-cell contacts (1,5,9,14). The polycystin complex, either through its association with cilium or cell-to-cell or cell-to-matrix interactions, has been implicated as a mechanosensor in renal epithelial cells (15–17). Emerging data, however, favor a role of the polycystin complex as a flow sensor. In this regard, fluid shear increases intracellular Ca²⁺ in wild-type epithelial cells (16,18), whereas cells with mutations in PKD1 or in which PC2 has been inactivated with an antibody do not activate flow-induced Ca²⁺ signaling (16,18,19). Whether the Cilia-PC1/PC2 complex might play a more generalized role in mechanosensing other tissues, such as bone, has not been examined.

Mechanical stimulation of the skeleton by exercise is an important anabolic signal in bone, which leads to increased osteoblastic proliferation and matrix deposition (20–23), whereas the absence of mechanical stimulation, as occurs with immobilization, disuse and exposure to low gravity, causes bone loss (24,25). Osteocytes, lining cells, and potentially osteoblasts are the key mechanosensitive cells in bone that transduce force or load into anabolic cell signals leading to new bone formation. Mechanosensing in bone leads to the release of autocrine factors such as prostaglandin E2 (PGE₂), prostacyclin I2 (PGI₂), ATP, UTP, transforming growth factor beta (TGF-β), fibroblast growth factor 2 (FGF-2), insulin-like growth factors (IGFs), and nitric oxide (NO) (20–22,26). Runx2 appears to be a down-stream target, since this essential osteoblastic transcript-tion factor is upregulated by mechanical strain, extracellular nucleotides, and oxidative stress in osteoblasts (27–32), and bone loss due to unloading is exacerbated in heterozygous Runx2- deficient mice (33). Runx2 also regulates the expression of genes, such as cyclooxygenase-2 (Cox-2), osteopontin, osterix, osteocalcin, and α₁(I) procollagen, which are all induced by mechanical strain (20,34).

Several primary mechanosensing molecules in bone have been proposed, including several G-protein-coupled receptors (e.g., prostaglandin and purinergic receptors), integrin receptors, connexins/gap-junctions, and stretch-activated and purinergic channels as well as cell-to-cell and cell-to-matrix interactions (20–22,26). The initial mechanosesning event that stimulates the release of osteogenic factors and the activation of second messengers leading to the osteoinductive response, however, has not been established. The possibility that polycystin 1 may be involved in mechano-sensing in osteoblasts/ osteocytes, however, is suggested by the observation that Pkd1 null mice have abnormal skeletal development characterized by spina bifida occulta and osteochondro-dysplasia (35-37). Although cilia-like structures have been reported in bone cells, definitive proof of cilia in osteoblasts/osteocytes is lacking (38-40). Mechanical strain in bone can generate interstitial fluid flow that exerts fluid shear stress on bone cells and stretches tissues, however, that could potentially activate PC1 and cilia (23).

In the current study, we examined whether osteoblasts/osteocytes express Pkd1 and Pkd2 and other cilia genes, determined whether mono cilia are present in these cells, assessed the skeletal phenotype in Pkd1^m1Bei mice that have an inactivating missense mutation in the Pkd1 gene (37), and evaluated the function of PC1 in osteoblasts ex vivo. We have found evidence for the presence of polycystin complexes and cilia in osteoblasts/osteocytes that are coupled to Runx2-dependent regulation of osteoblast/ osteoctye function.

EXPERIMENTAL PROCEDURES

Mice

Pkd1^m1Bei heterozygous mice, which has an inactivating point mutation in Pkd1 gene caused by ENU mutagenesis leading to substitution of an arginine for methionine in the first transmembrane domain of the PC1 protein (37), were obtained from mutant Mouse Regional Resource Center (UNC). These mice were bred and maintained on a C57BL/6J background. The mice were genotyped using SYBR® Green (Bio-Rad) real-time PCR reagents, extracted genomic DNA and the following primers: wild-type (T allele) forward primer: 5'- CTG GTG ACC TAT GTG GTC AT -3', mutant (G allele) forward primer: 5' - CTG GTG ACC TAT GTG GTC AG -3', and common reverse primer: 5' - AGC CGG TCT TAA CAA GTA TTT C -3'. The Pkd1 allele discrimination was determined by delayed signal on one mismatch. Animal experiments were performed following review and approval by University of Kansas Medical Center’s Animal Care and Use Committee.

Cell Cultures and Transient or Stable Transfections

MLO-Y4, osteocyte-like cells, were plated onto culture dishes coated with rat tail type I collagen, whereas the MC3T3-E1 cells were plated directly onto the plastic culture dish. For total RNA isolation, MC3T3-E1 cells were cultured in α-MEM containing 10% FBS supplemented with 5 mM β-glycerophosphate and 25 μg/ml of ascorbic acid for 4, 7, and 14 days, while MLO-Y4 cells were cultured in α-MEM supplemented with 5% fetal bovine serum and 5% calf serum for 4, 7, and 14 days. For transient transfections, a number of 1 x 10⁶ of MC3T3-E1 or MLO-Y4 cells were transfected with either control expression vector, gain-of-function, or loss-of-function PC1 C-tail constructs along with the Runx2-P1 luciferase reporter (p1.4Runx2-P1-Luc) construct by using the electroporation protocol from Amaxa’s Biosystems as described by the manufacturer (Amaxa Inc.). A total of 6.6 μg of plasmid DNA was used for each electroporation, with 3.6 μg of PC1 C-tail construct, 2.4μg of p1.4Runx2-P1-Luc reporter, and 0.6 μg of Renilla luciferase-null (RL-null) as internal control plasmid. Promoter activity was assessed by measuring luciferase activity 48 hours after transfection. The cells were then lysed in Passive lysis buffer (Promega), and 20 μl of cell lysate was used with the dual luciferase assay kit (Promega) using an EG&G Berthold 9507 Luminometer. Stable transfection of MC3T3-E1 was performed by a protocol that maintains the differentiation potential of these osteoblasts as previously described (41). Briefly, MC3T3-E1 cells were only transfected with 6.0 μg of PC1 C-tail fusion construct and selected by incubation in media containing 500 g /ml G418 (Life Technologies, Inc.) 48 hours after transfection.

Measurement of ALP Activity and Mineralization Assays in Immortalized Osteoblasts Cultures

Calvaria from Pkd1^m1Bei E15.5 embryos were used for the isolation of osteoblasts by sequential collagenase digestion as previously described (42,43). To engineer immortal osteoblast cell lines, isolated primary osteoblasts were infected using a retroviral vector carrying SV40 large and small T antigen. To induce differentiation, immortalized osteoblasts were plated at a density of 1x 10⁵ cells per well in a 6-well plate, and grown for period of up to 14 days in α-MEM containing 10% FBS supplemented with 5 mM β-glycerophophate and 25 μg/ml of ascorbic acid. Alkaline phosphatase activity and Alizarin red-S histochemical staining for mineralization were performed as previously described (42). Total DNA content was measured with a PicoGreen^® dsDNA quantitation reagent and kit (Molecular Probes, Eugene, OR).

RT-PCR Analysis

RT-PCR was done using the Titan^TM One tube RT-PCR kit from Roche Applied Science. DNase I-treated total RNA (1.0 μg) was reverse transcribed into cDNA with the reverse primer described before (44). The reverse transcription reaction was incubated at 50 °C for 30 min. PCR was performed with thermal cycling parameters of 94 °C for 30 s, 60 °C for 30 s and 68 °C for 30 s for 35 cycles followed by final extension at 68 °C for 7 min. To amplify mouse Pkd1 transcript, we used the forward primer (5'-CAT TGT ACC CCT GGA GGA GA-3') in combination with the reverse primer (5'-GAT GTC CAG GCT GTT TCG AT-3'). To amplify mouse Pkd2 transcript, we used the forward primer (5'-GGG GCT GCT ACA GTT TCT TG-3') with the reverse primer (5'-CCG GAG ACT CTC TGA GAT GG-3'). The mouse Tg737 transcript was RT-PCR amplified using the forward primer (5'-TCC AAC TGA TTC CCA AGC TC-3') and the reverse primer (5'-TGG CAC TCA GTC GTT CAC TC-3'). Mouse GAPDH was amplified as a control for the RT-PCR reactions.

Immunofluorescence

MC3T3-E1 osteoblasts and MLO-Y4 osteocytes were grown on collagen-coated coverslips and kept at confluence for at least 3 days. At the end of the culture, the cells were washed three times with PBS, then fixed with cold 4% paraformaldehyde/0.2% Triton for 10 minutes at room temperature and washed with PBS 3 times. The coverslips were incubated for 30 minutes in 1% BSA before incubation with the primary acetylated alpha-tubulin antibody (Sigma Aldrich, T6793) for 1 hour at room temperature. After washing three times in PBS they were treated with secondary Texas Red-labeled anti-mouse IgG (Jackson ImmunoResearch, 715-076-150) in 1% BSA for 1 hour at room temperature and washed three times in PBS before mounting with ProLong® Gold antifade reagent (Invitrogen, P36935). Nuclei were counter-stained with DAPI. Photographs were taken under a microscope with magnifications of 60x.

Whole Mount Calvaria Immunostaining

For whole mount bone immunostaining for alpha tubulin, transgenic mice expressing the topaz variant of enhanced green fluorescent protein (eGFP-tpz) targeted to the osteocyte were used (45). These mice were a kindly provided by Dr. David Rowe (University of Connecticut Health Center, Farmington, CT 06030). The GFP reporter was expressed under control of a 7892bp fragment of the Dentin Matrix Protein-1 (Dmp1) promoter together with a 4439bp region containing the first exon, the first intron and part of exon 2. In these mice, GFP is expressed almost exclusively in osteocytes and preosteocytes. This transgenic mouse model therefore provides an extremely useful tool for imaging of osteocytes, either in situ within the bones or in cell cultures (45–47).

Whole calvaria from 12-day-old neonatal Dmp1-GFP mice were excised and the periostea carefully removed under a dissection microscope. The bones were fixed overnight at 4°C in the dark in neutral buffered formalin. The bones were then washed twice in PBS for 15 min, with shaking. Next, the bones were decalcified for one week in 10% EDTA, followed by washing three times in PBS at 4°C for 15 min. Background blocking was performed overnight at 4°C in PBS+1% normal horse serum + 0.05% sodium azide, followed by washing twice with PBS. As the primary antibody was a mouse monoclonal against alpha tubulin (Sigma Aldrich, T 6793), the vectastain M.O.M. kit was used according to manufacturer’s instructions to block endogenous mouse immunoglobulin. The bones were incubated for 4 hours at room temperature with primary antibody diluted in M.O.M. diluent. After four washes in PBS, the bones were then incubated for two hours with a Cy3-conjugated donkey anti-mouse second antibody, diluted 1:250 in M.O.M diluent (Jackson Immunoresearch Laboratories Inc., West Grove, PA). The bones were then washed four times in PBS and counterstained with DAPI nuclear stain (4μg/ml in PBS) for 30 seconds, followed by two final PBS washes. The whole bones were mounted under a coverslip in 2 drops of 9:1 glycerol:PBS containing 5% N-propyl gallate as an anti-fade reagent. The specimens were viewed on a Nikon E800 microscope under epifluorescent illumination. Digital photographs were acquired from multiple optical planes using an optronics CCD camera driven by the AnalySIS software (Olympus Soft Imaging Solutions Corp, Lakewood, CO). Extended focus images were generated using this software.

Scanning Electron Microscopy

Renal tubular MDCK and MLO-Y4 cells were cultured on Thermanox coverslips and at the end of the culture was washed with PBS and fixed with 10% formalin for 20 minutes, washed again with PBS, dehydrated in graded concentrations of ethanol and dried using Hexamethyl disilazone (HMDS) for 5 minutes. After dehydration the coverslips were mounted on a stub and sputtered with gold-palladium, and viewed with the FEI/Philips XL30 Field emission environmental SEM (48).

Whole Skeletal Mount Alizarin Red/Alcian Blue staining and Histological Preparation

This procedure provides information regarding developmental abnormalities. Embryos from 13.5 (E13.5) to 15.5 (E15.5) days of gestation were collected and fixed for more than 3 days in 95% ethanol. Samples were defatted for 2–3 days in acetone and stained sequentially with alcian blue and alizarin red S in 2% KOH. The stained skeleton preparations were cleared with 1%KOH/20% glycerol and stored in 50% ETOH/50% glycerol (42). Femurs from E15.5 embryos were decalcified at 4 °C in 12.5% EDTA and 2.5% paraformaldehyde in phosphate-buffered saline. Longitudinal sections were stained with hematoxylin and eosin to assess the histology of the growth plate and bone marrow cavity in femurs (42).

Micro-CT Analysis

The distal femoral metaphyses were scanned using a μCT 40 (Scanco Medical AG, Bassersdorf, Switzerland), and 167 slices of the metaphyses under the growth plate, constituting 1.0 mm in length, were selected. The three-dimensional (3D) images were generated using the following values for a gauss filter (sigma 0.8, support 1) and a threshold of 275. A 3D image analysis was done to determine bone volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp). Cortical bone was measured on the mid-shaft region of cortical bone. 50 slices of the diaphysis, constituting 0.3 mm in length, were selected. The mean cortical thickness (Ct.Th) was determined by distance measurements at 8 different points on the cortical slice (42).

Real-time RT-PCR and Western Blot

For quantitative real-time RT–PCR, 2.0 μg total RNA isolated from whole E15.5 embryos and femurs of 12-week-old mice was reverse transcribed as described (42). PCR reactions contained 100 ηg template (cDNA or RNA), 300 ηM each forward and reverse primer and 1X iQ^TM SYBR^® Green Supermix (Bio-Rad, Hercules, CA) in 50 μl. Samples were amplified for 40 cycles in an iCycler iQ^TM Real-Time PCR Detection System with an initial melt at 95 °C for 10 min followed by 40 cycles of 95 °C for 15 sec and 60 °C for 1 min. PCR product accumulation was monitored at multiple points during each cycle by measuring the increase in fluorescence caused by the binding of SybrGreen I to dsDNA. The threshold cycle (Ct) of tested-gene product from the indicated genotype was normalized to the Ct for 18s rRNA (49). Nuclear extracts were prepared using NE-PER^TM (Pierce Chemical Co., Rockford, IL). Protein concentrations were determined with a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). Equal quantities of protein were subjected to NuPAGE^TM 4–12% Bis-Tris Gel (Invitrogen) and were analyzed with standard Western blot protocols (HRP-conjugated secondary antibodies from Santa Cruz Biotechnology and ECL from Amersham). RUNX2 (sc-10758) and Lamin A/C (sc-6215) antibodies from Santa Cruz Biotechnology were used according to the manufacturer's instructions (44).

Bone Densitometry and Bone Histomorphometric Analysis

Bone mineral density (BMD) of femurs was assessed at 12 weeks of age using a LUNARPIXIMUS bone densitometer (Lunar Corp., Madison, Wisconsin, USA) as previously described (50). Skeletons of mice were prelabeled with calcein (Sigma C-0875, 30 μg/g body weight) by intraperitoneal injection at days 1 and 5 in 12-week-old mice prior to collection of tibias. Tibias were removed from 12-week-old mice, fixed in 70% ethanol and processed for methyl methacrylate embedding by using Osteo-Bed Bone Embedding Kit (Polysciences, Inc.) (50). 10-μm cross sections of tibia at the joint of tibia and fibula were evaluated under fluorescent light as reported previously (51) by our laboratory. We determined mineral apposition rate (MAR) based on histomorphometry standards, endocortical and periosteal surfaces were analyzed separately.

Serum Biochemistry

Serum osteocalcin levels were measured using a mouse osteocalcin EIA kit (Biomedical Technologies Inc. Stoughton, MA, USA). Serum urea nitrogen (BUN) was determined using a BUN diagnostic kit from Pointe Scientific, Inc. Serum calcium (Ca) was measured by the colorimetric cresolphthalein binding method, and phosphorus (P) was measured by the phosphomolybdate–ascorbic acid method (Stanbio Laboratory, TX, USA). Serum osteoprotegerin (OPG) and free sRANK ligand (free sRANKL) were measured using mouse ELISA kits (Quantikine^®, R & D systems and Biomedical Medizinprodukte GmbH & Co KG), and serum TRAP was assayed with the ELISA-based SBA Sciences mouseTRAP^TM assay (Suomen Bioanalytiikka Oy).

Statistics

We evaluated differences between groups by one-way analysis of variance. All values are expressed as means ± SEM. All computations were performed using the Statgraphic statistical graphics system (STSC Inc).

RESULTS

Evidence for expression of Pkd1 and Pkd2 and the presence of cilia in osteoblasts/osteocytes

To investigate the possibility that Pkd1/Pkd2 and possibly cilia are present in bone cells, we examined if Pkd1, Pkd2 and the cilia-associated polaris (Tg737) and Kif3a transcripts are present in osteoblast and osteocyte cell lines. By RT-PCR we demonstrated that confluent cultures of MC3T3-E1 osteoblasts and MLO-Y4 osteocytes express Pkd1 and Pkd2 as well as polaris (Tg737) and Kif3a transcripts (Fig. 1A). To determine if osteoblasts and osteocytes express cilia-like structures, we examined MC3T3-E1 and MLO-Y4 cells as well as wild-type (Pkd1^+/+) and Pkd1 homozygous osteoblasts (Pkd1^m1Bei/m1Bei) by immuno-fluorescence (Fig 1 B) and scanning electron microscopy (SEM) (Fig 1C), and primary osteocytes by whole mount calvaria immunostaining (Fig 1D). As a control for our methods for demonstrating the presence of primary cilia, we used cilia expressing renal tubular derived MDCK cells as a positive control. These cells have been shown to express cilia by immunohistochemical analysis using an antibody to α-tubulin (52) and by SEM (48). Both confluent MC3T3-E1 osteoblast and MLO-Y4 osteoctye-like cell lines (Fig. 1B, upper two panels) displayed punctuate and linear staining with the α-tubulin antibody, consistent with the presence of single cilia-like structure per cell. We have also observed cilia in Pkd1^+/+ and Pkd1^m1Bei/m1Bei osteoblasts (Fig. 1B, lower two panels). The presence of cilia in Pkd1 homozygous osteoblasts is expected, since Pkd1 is known not to be required for ciliogenesis. A single cilia-like structure measuring 2–4 μm in length was also detected by SEM extending from the membrane surface in both MDCK (Fig. 1C, Left panel) and MLO-Y4 (Fig. 1C, Right panel) cells. In addition, the whole mount immunostaining of 12-day-old mouse calvaria showed α-tubulin positive structures resembling primary cilia in primary osteocytes (Fig. 1D, upper panel, GFP positive cells with dendrites) and primary osteoblasts Fig. 1D, lower panel, GFP negative). Collectively these data establish the presence of cilia-like structures in osteocytes and osteoblasts.

Fig. 1.

Evidence for expression of Pkd1 and Pkd2 and primary cilia in osteoblasts/osteocytes. A, RT-PCR analysis of Pkd1, Pkd2, Tg737 and Kif3a expressions in osteoblasts/osteocytes at various stages of maturation. The cilia gene polaris (Tg737) and Kif3a, as well as Pkd1 and Pkd2 were expressed in MC3T3-E1 osteoblasts (left panel) and MLO-Y4 osteocytes (right panel) as a function of culture duration. B, immunofluorescence of primary cilia in MC3T3-E1 osteoblastic and MLO-Y4 osteocyte-like cell line. Higher power magnification (60x) of imunostaining with α tubulin antibody demonstrated the presence of solitary binding (Red) with α-tubulin antibody, consistent with the presence of primary cilia-like structures in MC3T3-E1 osteoblasts and MLO-Y4 (upper two panels) osteocytes. We have also observed cilia in wild-type (Pkd1^+/+) and Pkd1 homozygous (Pkd1^m1Bei/m1Bei) osteoblasts (lower two panels). Second antibody without primary α-tubulin antibody displayed no such a structure in the immunostaining mentioned above (data not shown). Arrow heads indicate single primary cilia-like structures, Counterstaining with a nuclear marker (DAPI) is shown in blue. C, scanning electron micrograph of primary cilium. Arrows indicate primary cilia emerging from the MDCK (left panel) and MLO-Y4 osteocyte (right panel). D, whole calvaria immunostaining for primary cilia in osteocytes and osteoblasts from 12-day-old Dmp1-GFP mice. The whole calvaria are fixed and decalcified, then immunostained and mounted on a slide. The green cell is GFP positive osteocytes with dendrites. The red cell is negative for GFP. The nuclei are DAPI stained blue. Arrow heads indicate the single bright red, α-tubulin positive structures resembling cilia on the osteocyte-like and osteobalstic cells.

Homozygous mutant Pkd1^m1Bei mice have abnormal bone development associated with decreased Runx2

To determine if inactivating missense mutations in PC1 results in a bone phenotype, we examined skeletal development in Pkd1^m1Bei homozygous mouse (Pkd1^m1Bei/miBei) embryos. Pkd1^m1Bei mutant mice have a missense mutation (T to G at 9248bp) in the coding sequence of Pkd1 gene that encodes PC1 that results in an amino acid M to R substitution in the first transmembrane domain of the PC1 protein (37). Pkd1^m1Bei/miBei mice are embryonic lethal and have a kidney phenotype similar to that observed in Pkd1 knockout mice, whereas heterozygous Pkd1^+/miBei mice have no demonstrable kidney phenotype (35,36). In wild-type mice, we observed the anticipated development from an unmineralized cartilaginous bone template at E13.5 to a caudal progression of mineralization beginning with the skull, mandible and ribs by E14.5 and continuing with the spine and long bones by E15.5 (Fig. 2A). In contrast, in Pkd1^m1Bei/miBei mice, we observed abnormalities in endochondral bone formation. In E14.5 Pkd1^m1Bei/miBei embryos the cartilaginous skeleton was completely formed, but calcification was completely absent (Fig. 2A). However, at E15.5, Pkd1^m1Bei/miBei embryos displayed evidence of calcification, but the amount of mineralized bone was less than in wild-type controls (Fig. 2A). Defects were observed in both the calvaria and long bones, suggesting a delay in both intra-membranous and endochondral bone formation.

Fig. 2.

Defective skeletogenesis and decreased Runx2 expression in Pkd1^m1Bei homozygous mutant mice. A, alizarin red/Alcian blue staining of wild-type (Pkd1^+/+), heterozygous (Pkd1^+/m1Bei) and homozygous (Pkd1^m1Bei/m1Bei) embryos. Pkd1^m1Bei/m1Bei mice were smaller than Pkd1^+/m1Bei and Pkd1^+/+ mice. The unmineralized cartilage (indicated by blue staining) was not affected by mutant Pkd1. The skeletal calcification (indicated by red staining), however, was markedly delayed in E14.5–E15.5 Pkd1^m1Bei/m1Bei embryos as indicated by arrows. B, histological analysis (H&E) of endochondral ossification in longitudinal sections of decalcified femurs of Pkd1^+/+, Pkd1^+/m1Bei, Pkd1^m1Bei/m1Bei embryos. Low magnification (40x) of entire femur (left panel), higher power magnification (100x) of metaphyseal and diaphseal regions (right panel). A gene-dose dependent delay in endochondral bone formation was observed in Pkd1-mutant embryos as reflected by the reduced size of the marrow component in Pkd1^+/m1Bei and impaired vascular invasion resulting in incomplete remodeling of the cartilage anlage in Pkd1^m1Bei/m1Bei embryos. Arrow bar indicates bone marrow cavity (BMC) and arrows indicate bone collar (BC). C, Runx2 isoforms expression in embryos. A Pkd1 gene dose-dependent reduction in Runx2-II expression was observed in E15.5 Pkd1^m1Bei mutant embryos, while no difference was observed in Runx2-I isoform expression by real-time RT-PCR. D, osteoblasts and osteoclasts markers in embryos. A gene dose-dependent reduction in Runx2 downstream genes (osteocalcin,and osterix) were observed in E15.5 Pkd1^m1Bei mutant embryos, and a significant reduction in osteoblast (osteoprotegerin (OPG) and Rank ligand (RanKL) and osteoclast (tartrate resistant acid phosphatase, TRAP) markers are also observed in homozygous Pkd1^m1Bei/m1Bei embryos. Data are reported as the mean±SEM from at least three embryos. ^# and * respectively indicates significant difference from wild-type Pkd1^+/+ and heterozygous Pkd1^+/m1Bei mice at P<0.05 by one-way analysis of variance.

Histological analysis confirmed the delay in endochondral ossification in long bones of E15.5 Pkd1^m1Bei/miBei embryos (Fig. 2B). Homozygous Pkd1^m1Bei mice were characterized by delay in vascular invasion, resulting in the absence of a bone marrow cavity. These mice also exhibited a narrow bone collar. We also observed abnormalities bone in heterozygous Pkd1^m1Bei mice (Pkd1^+/m1Bei) that were not apparent in the whole skeletal preparations. In this regard, Pkd1^+/m1Bei displayed a reduced size of the bone marrow cavity, consistent with delayed conversion of the cartilaginous structure to primary spongiosa (Fig. 2B).

Runx2, a master gene regulating bone development, exists as two major gene products, Runx2-I derived from the P2 proximal promoter and Runx2-II (formerly called Osf2 or Cbfa1) derived from the P1 distal promoter (36, 37). Because the skeleton of Pkd1^m1Bei/miBei embryos resembled the skeletal phenotype in mice with selective Runx2-II deficiency (42), we determined if loss of polycystin-1 results in abnormalities of Runx2 expression. Runx2-I and Runx2-II mRNA levels in total RNA isolated from E15.5 embryos were quantified by real-time RT-PCR. A gene dose-dependent reduction in Runx2-II expression was observed, with levels being less in Pkd1^m1Bei homozygous embryos compared to Pkd1^m1Bei heterozygous embryos (Fig 2C). In contrast, no differences were observed in Runx2-I isoform expression in the different genotypes (Fig. 2C).

To investigate whether mutant Pkd1^m1Bei and the consequent reduction in Runx2-II expression resulted in concomitant defects in downstream gene expression, we examined by real-time RT-PCR the expression levels of osteoblast (osteocalcin, osterix, osteoprotegerin (OPG), and Rank ligand (RANKL)) and osteoclast (tartrate resistant acid phosphatase, TRAP) specific transcripts in wild-type, heterozygous Pkd1^+/m1Bei and homozygous Pkd1^m1Bei/m1Bei embryos. Similar to the reduction in Runx2-II, a gene dose-dependent reduction in osteocalcin and osterix expressions was observed in wild-type and Pkd1^m1Bei mutant embryos, and a significant reduction in others transcripts in homozygous Pkd1^m1Bei/m1Bei embryos was also found (Fig. 2D).

Osteopenia in heterozygous Pkd1^m1Bei mice suggests a postnatal role for Pkd1 in bone

The homozygous Pkd1^m1Bei/m1Bei mice are embryonic lethal, thereby precluding assessment of the bone phenotype postnatally. To determine the impact of loss of one functional Pkd1 allele in the adult, we assessed bone mineral density (BMD) in 12-week-old male and female wild-type and heterozygous Pkd1^+/m1Bei mice, which have normal survival and no apparent histological abnormalities of the kidney (data not shown). As shown in Fig. 3A, Pkd1^+/m1Bei heterozygous mice have a 9% reduction in BMD. μCT analysis revealed that the reduction in bone mass in Pkd1^+/m1Bei heterozygous mice was caused by a reduction in trabecular bone volume (33%) and cortical bone thickness (6%) (Fig.3B). In addition, histological analysis of bone identified a significant decrease in mineral apposition rate in 12-week-old Pkd1^+/m1Bei mice compared with age-matched wild-type Pkd1^+/+ mice (Fig. 3C), suggesting impaired osteoblast-mediated bone formation. Further evidence for osteoblast dysfunction includes a reduction in Runx2-II and osterix expression (Fig. 3D) as well as osteocalcin, osteoprotegerin and Rank ligand in bone and serum from heterozygous Pkd1^+/m1Bei mice at 12 weeks of age (Fig. 3D and Table I). Serum levels and bone expression of TRAP, a marker of bone resorption, were also reduced in heterozygous Pkd1^+/m1Bei compared to wild-type littermates (Fig. 3D and Table I). These findings suggest that low bone formation rates rather than increased resorption contribute to the low BMD (Fig. 3, A-C) and bone volume of femurs observed in heterozygous Pkd1^+/m1Bei mice (Fig. 3B).

Fig. 3.

Skeletal abnormalities in 12-week-old Pkd1^+/m1Bei mice. A, bone mineral density (BMD) of femurs. There was significant reduction of femoral BMD in both male and female Pkd1^+/m1Bei heterozygous mice at 12 weeks compared to wild-type littermates by the PIXImus^TM mouse densitometer. B, 3-dimensional μCT images of femurs. 12-week-old Pkd1^+/m1Bei mice have diminished trabecular bone volume and reduced cortical thickness by μCT 3D analysis. C, bone formation. Mineral apposition rates in the tibia were significantly reduced in 12-week-old Pkd1^+/m1Bei mice. D, Runx2 isoforms, osteoblast and osteoclast marker expression. Runx2-II expression as well as osteoblast (osteocalcin, osterix, OPG and RanKL) and osteoclast (TRAP) markers in bone was significantly reduced in mRNA derived from whole femurs of 12 weeks Pkd1^+/m1Bei mice by real-time RT-PCR. Data represent the mean±SEM from at least three individual samples. ^# indicates significant difference from wild-type Pkd1^+/+ mice at P<0.05 by one-way analysis of variance.

TABLE I.

Biochemistry analysis of serum in 12-week wild-type and Pkd1^m1Bei mutant mice

Genotype	Pkd1+/+	Pkd1+/m1Bei
BUN(mg/dl)	18.4±0.97	18.9±0.98
Ca (mg/dl)	9.4±0.13	9.4±0.14
P (mg/dl)	8.7±0.42	8.8±0.43
Osteocalcin (ηg/ml)	534±41	405±38#
OPG (ηg/ml)	2.4±0.06	2.1±0.05#
Free sRANKL (pmol/l)	3.5±0.4	1.9±0.3#
TRAP (U/L)	16±0.6	12± 0.5 #

Data are mean ±S.E.M. from 4-5 individual mice.

^#indicates significant difference from wild-type Pkd1^+/+ mice at P<0.05 by one-way analysis of variance. Osteocalcin, OPG, and free sRANK-L are produced by osteoblasts, and TRAP is produced by osteoclasts.

Osteoblasts from E15.5 Pkd1^m1Bei/m1Bei mutant mice exhibit defective differentiation ex vivo

To confirm the presence of a primary effect of PC1 on osteoblasts, we characterized the differentiation potential of osteoblasts derived from homozygous Pkd1^m1Bei/m1Bei embryos. Consistent with the reduction in Runx2-II expression in bone of Pkd1^m1Bei/m1Bei homozygous mice, we found that osteoblasts derived from Pkd1^m1Bei/m1Bei homozygous mice also displayed significant reductions in both Runx2 mRNA and protein levels (Fig. 4A) as well as reduced P1 Runx2 promoter activity (see Fig 5C below). We also found that osteoblastic gene makers, including osteocalcin, osteopontin, osterix, and α₁(I) procollagen, were significantly decreased in Pkd1^m1Bei/m1Bei homozygous osteoblasts (data not shown). In addition, Pkd1^m1Bei/m1Bei osteoblasts displayed time-dependent increments in DNA content during the first 10 days in differentiation medium, but the DNA content was significantly lower at different time points compared with Pkd1^+/+ osteoblasts (Fig. 4B), indicating that there was a proliferation defect in Pkd1^m1Bei/m1Bei osteoblasts under differentiating conditions. We also found that mutant PC1 impairs osteoblastic differentiation and maturation, as evidenced by lower ALP activity and mineralization of extracellular matrix as assessed by Alizarin red S staining in Pkd1^m1Bei/m1Bei compared to wild-type osteoblast cultures grown for up to 14 days under differentiating conditions (Fig. 4, C and D).

Fig. 4.

Pkd1^m1Bei/m1Bei osteoblasts have a developmental defect ex vivo. A, Runx2 expresssion in immortalized Pkd1^m1Bei/m1Bei osteoblasts. A, significant reduction of Runx2 expression was observed in immortalized Pkd1^m1Bei/m1Bei osteoblasts derived from E15.5 calvaria by real-time PCR (upper panel) and Western blot (lower panel). The 64 and 60 kDa bands in the Western blot represent Runx2 isoforms, the nuclear Lamin A/C (70 kDa) serves as a control (lower panel). B, DNA content. Pkd1^m1Bei/m1Bei osteoblasts displayed time-dependent increments in DNA content during 10 days of culture in differentiation medium, but the DNA content was significantly lower at different time points compared with Pkd1^+/+ osteoblasts. C, quantification of mineralization. Alizarin Red-S stain was extracted with 10% cetylpyridinium chloride and quantified as described in Methods. Pkd1^m1Bei/m1Bei osteoblasts had time-dependent increments in Alizarin Red-S accumulation during 14 days of culture, but the accumulation was significantly lower at different time points compared with Pkd1^+/+ osteoblasts. D, ALP activity. Pkd1^m1Bei/m1Bei osteoblasts displayed time-dependent increments in ALP activities during 14 days of culture, but the ALP activity was significantly lower at different time points compared with Pkd1^+/+ osteoblasts. Data are mean±SEM from at least three independent experiments. * indicates significant difference from wild-type Pkd1^+/+ at P<0.05 by one-way analysis of variance. Values sharing the same letter superscript are not significantly different at P<0.05.

Fig. 5.

Effects of Pkd1 (PC1) C-tail (gain-of-function) constructs on Runx2 expression in osteoblasts/osteocytes. A, schematic of PC1 C-tail fusion constructs and their coupling to signaling pathways. Chimeric constructs consist of the full-length or various deletion mutants lacking either the G-protein binding site or the coiled-coil domain linked to constructs encoding the human IgG and CD7 transmembrane domain in the pcDNA3.1 vector. B, effects of PC1 C-tail constructs on Runx2 P1 promoter activity in MC3T3-E1 osteoblasts. The C-tail fusion constructs containing the coiled-coiled (PC1-LT, PC1-HT and PC1-AT) domain activated Runx2 P1 promoter activity in MC3T3-E1 osteoblasts. Data are mean±S.E.M. from at least three independent experiments. Values sharing the same letter superscript are not significantly different at P<0.05. C, effects of PC1 C-tail constructs on Runx2 P1 promoter activity in Pkd1^m1Bei/m1Bei osteoblasts. Trasient co-transfection with PC1-AT construct normalized Runx2 P1 promoter activity in Pkd1^m1Bei/m1Bei osteoblasts to levels observed in wild-type Pkd1^+/+ osteoblasts. D, effects of PC1 C-tail constructs on endogenous Runx2 expresssion in osteoblastic cell lines. Increased Runx2 expression by real-time RT-PCR (upper panel) and Western blot (lower panel) in MC3T3-E1 stably transfected with PC1-AT are observed. The 64 and 60 kDa bands in the Western blot represent Runx2 isoforms, the nuclear Lamin A/C (70 kDa) serves as a control (lower panel). E, effects of PC1-AT overexpression on osteocalcin, osterix, osteoprotegerin, and Rank ligand message expression in MC3T3-E1 osteoblasts. Data are mean±SEM. from at least three independent experiments. * indicates significant difference from the sIgØ control at P<0.05 by one-way analysis of variance.

Pkd1 (PC1) C-tail (gain-of-function) constructs regulate the Runx2 expression in osteoblasts/ osteocytes

Activation of PC1 signaling can be achieved by overexpression of the C-terminus of PC1 (53). To confirm that the defect in Runx2 expression results from mutant PC1, we explored the function of PC1 in osteoblasts/ osteocytes by overexpressing PC1 C-tail (gain-of-function) constructs (Fig. 5A). The PC1-LT construct is comprised of human IgG CH2-CH3 region, TM region and entire C-tail containing the regions for G-protein activation, and the coiled-coil region that is required for coupling to PC2; PC1-HT contains a limited G-protein signaling domain; PC1-LS contains only the G-protein signaling domain; and PC1-AT contains only the coiled-coil region (Figure 5A). Transient co-transfection of all constructs along with the Runx2 P1 promoter-reporter construct into MC3T3-E1 osteoblasts significantly increased Runx2 P1 promoter activity (Fig 5B). The response was greater for the PC1-HT and PC1-AT constructs containing the coiled-coil domain. The activation of the Runx2 P1 promoter was significantly less with the PC1-LS domain containing only the site for G-protein coupling (Fig. 5B). Identical results were also observed in MLO-Y4 osteocytes (data not shown). We also examined the effects of overexpression of the PC1 C-tail constructs on Runx2 P1 promoter activity in Pkd1^m1Bei/m1Bei osteoblasts (Fig 5C). Transient cotransfection of the PC1-AT and P1 promoter constructs restored Runx2 P1 promoter activity in Pkd1^m1Bei/m1Bei osteoblasts to levels observed in wild-type Pkd1^+/+ osteoblasts (Fig. 5C).

Next, we examined whether the PC1 C-tail construct PC1-AT regulates endogenous Runx2 and bone-related gene expression in MC3T3-E1 osteoblasts. Consistent with the stimulation of Runx2 P1 promoter activity, we also observed significant increases of endogenous Runx2 message expression (Fig. 5D) as well as its downstream genes osteocalcin, osterix, osteoprotegerin and Rank ligand (Fig. 5E) in MC3T3-E1 stably transfected with the PC1-AT construct containing the coiled-coil domain. These findings are consistent with Pkd1 coupling to Runx2 expression and osteoblast differentiation.

DISCUSSION

The PC1/PC2 complex and cilia are an established paradigm for sensing mechanical strain and possibly other stimuli and regulating differentiation in kidney epithelial cells (16,17,19). In spite of two studies reporting skeletal abnormalities in Pkd1 null mice (35-37) and unconfirmed reports of cilia-like structures in osteoblasts/osteocytes by electron microscopy (38), the expression of the PC1/PC2 complex and cilia and their functional role in osteoblasts/ osteocytes has not been investigated in detail. In the current work, we provide evidence that the Pkd1 and Pkd2 gene productions, PC1 and PC2, are expressed along with cilia-related transcripts and cilia-like structures in osteoblasts/osteocytes. More importantly, we demonstrate that mutant PC1 is associated with a gene-dose dependent defect in bone development, decreased expression of the osteogenic Runx2 transcription factor and impaired osteoblast-mediated bone formation in vivo and ex vivo, consistent with an anabolic function of the Pkd1 gene product in osteoblasts/osteocytes.

PC1 is known to be widely expressed and developmentally regulated in many tissues and cell types (54,55). Prior studies regarding bone, however, have been limited to embryonic tissues (35,36), and we are aware of no studies that have examined the expression of PC1 or PC2 in the osteoblast lineage or assessed the function of polycystins postnatally. Our studies are the first to demonstrate the presence of Pkd1 and its associated Pkd2 transcripts in bone mechanosen-sing osteoblasts/osteocytes. Indeed, we demonstrated that Pkd1 and Pkd2 transcripts and transcripts associated with cilia are expressed in osteoblasts/osteocytes cell lines (Fig.1). Primary cilia-like structures were also identified on the membrane surface of osteoblastic and osteocyte-like cell lines as well as primary osteocytes and osteoblasts in situ by immunostaining with anti-α-tubulin antibody (Fig. 1). Federman and Nichols reported the presence of primary cilia in rat calvarial osteocytes over 30 years ago (38,39), but this observation has not been widely accepted. Our findings clearly show the presence of a single cilium in postnatal primary osteocytes/osteoblasts in bone and in osteoblastic and osteocyte-like cell lines (Fig 1B-D).

We also provide additional information regarding PC1 function in bone through the characterization of the Pkd1^m1Bei mutant mouse, whose skeletal phenotype has not been previously described. We found a delay in endochondral bone formation during embryogenesis in the Pkd1^m1Bei homozygous mice (Fig. 2), similar to reported skeletal developmental abnormalities described in other Pkd1-deficient mouse models. The complete absence of a functional Pkd1 gene resulted in a delay in endochondral and intramembranous bone formation in vivo and an intrinsic abnormality of osteoblast function ex vivo, as evidenced by the defective maturation of osteoblasts derived from calvaria of E15.5 Pkd1^m1Bei/m1Bei embryos. Transcripts for markers of osteoblast function, including Runx2, osterix, and osteocalcin, were also reduced in Pkd1^m1Bei/m1Bei homozygous mouse embryos (Figs. 2 and 4).

We also discovered a postnatal reduction in osteoblast-mediated bone formation in heterozygous Pkd1^+/m1Bei mice, which has not been previously reported (35–37) (Fig. 3). The phenotype of heterozygous mutant mice was less severe than Pkd1^m1Bei/m1Bei homozygous mice, consisting of reduced BMD and trabecular bone volume due to diminished osteoblast-mediated bone formation, as evidenced by low mineral apposition rates, diminished osteoblast markers and absence of increased bone resorption (Fig. 3 and Table I). A gene-dose dependent effect of Pkd1 on osteoblast-mediated bone formation implicates PC1 as a potential mediator of osteoblast function. Abnormalities of skeletogenesis in Pkd1^m1Bei mutant mice could be solely due to the demonstrable abnormalities in the osteoblasts lineage or also involve defects in the function of chondrocytes, which also express cilia (56,57).

We also show that PC1 is an important regulator of osteoblast function through regulation of Runx2-dependent signaling both in vivo and in vitro. In this regard, we found evidence that Runx2-II, an essential transcriptional factor in controlling osteoblast-mediated bone formation, is downstream of Pkd1. Interestingly, the diminished metaphyseal bone volume and lower mineral apposition rate of Pkd1-deficient mice resembled the phenotype of selective Runx2-II-deficient mice (50). Moreover, we found evidence for a selective and gene-dose dependent reduction in the Runx2-II message levels in both embryos and bone of Pkd1^m1Bei mutant adult mice. In contrast, the Runx2-I isoform was not altered in either embryos or adult long bones. Finally, overexpression of constitutively active C-terminal PC1 chimeric constructs (53) into MC3T3-E1 and MLO-Y4 osteocytes resulted in stimulation of Runx2-II P1 promoter activity, endogenous Runx2-II transcripts and osteoblast differentiation markers (Fig. 5).

Although we did not directly study PC2 function in these studies, the observation that the PC1 C-terminal constructs containing the coil-coiled domain, which is known to couple PC1 with PC2 (11,58), are more potent stimulators of Runx2 expression, indirectly implicates a role of PC2. In addition, we have evidence for lower intracellular calcium levels in Pkd1^m1Bei/m1Bei homozygous osteoblasts, consistent with abnormal coupling to PC2 (data not shown). Further studies are needed to investigate the potential role of PC2-dependent regulation of intracellular calcium in mediating PC1 stimulation of Runx2 expression.

Our studies also did not establish a functional link between cilia and polycystin function or their roles as mechanosensors in bone. While the functional association of polycystin and cilia in renal epithelia cells implies a similar function in osteoblasts/osteocytes, polycystins might also have function independent of cilia in bone. In this regard, extracellular domain of PC1 may serve as a connection between cells and mediate stretch-sensitive cell-cell interactions (20–22). Alternatively, perturbations of the polycystin complex on the dendrite membrane could lead to changes in the cytoskeleton (59). In addition, cilia may have functions other than mechanosensing and actions that are independent of polycystins (57,60). Further studies are needed to establish a physical association between PC1, PC2 and cilia in osteoblasts/osteocytes, to establish that selective loss of PC1, PC2 and cilia in osteoblasts/ osteocytes results in defective mechanosensing in bone, and to elucidate the precise downstream signal transduction mechanism linking PC1 to activation of Runx2.

In conclusion, our data show that osteoblasts/osteocytes have cilia and express the cilia-associated Pkd1 and Pkd2 gene products, PC1 and PC2. Based on the phenotype changes in bone and ex vivo function of osteoblasts/osteocytes from Pkd1 mutant mice; we suggest that polycystin-1 is important for normal bone development and osteoblast/osteocyte-mediated bone formation. Cilia-polycystins are also a candidate for the elusive primary “mechanosenor” in bone. Additional studies are warranted, however, to determine if the cilia-PC1/PC2 complex is part of a skeletal mechanosensing mechanism that allows osteoblasts/osteocytes to sense and transduce mechanical forces into anabolic cell signals leading to new bone formation.