Kif3a Deficiency Reverses the Skeletal Abnormalities in Pkd1 Deficient Mice by Restoring the Balance Between Osteogenesis and Adipogenesis

Pkd1 localizes to primary cilia in osteoblasts and osteocytes. Targeted deletion of Pkd1 in osteoblasts results in osteopenia and abnormalities in Runx2-mediated osteoblast development. Kif3a, an intraflagellar transport protein required for cilia function, is also expressed in osteoblasts. To assess the relationship between Pkd1 and primary cilia function on bone development, we crossed heterozygous Pkd1- and Kif3a-deficient mice to create compound Pkd1 and Kif3a-deficient mice. Pkd1 haploinsufficiency (Pkd1 +/Δ) resulted in osteopenia, characterized by decreased bone mineral density, trabecular bone volume, and cortical thickness. In addition, deficiency of Pkd1 resulted in impaired osteoblastic differentiation and enhanced adipogenesis in both primary osteoblasts and/or bone marrow stromal cell cultures. These changes were associated with decreased Runx2 expression, increased PPARγ expression, and impaired hedgehog signaling as evidenced by decreased Gli2 expression in bone and osteoblast cultures. In contrast, heterozygous Kif3a +/Δ mice display no abnormalities in skeletal development or osteoblast function, but exhibited decreased adipogenic markers in bone and impaired adipogenesis in vitro in association with decreased PPARγ expression and upregulation of Gli2. Superimposed Kif3a deficiency onto Pkd1 +/Δ mice paradoxically corrected the effects of Pkd1 deficiency on bone mass, osteoblastic differentiation, and adipogenesis. In addition, Runx2, PPARγ and Gli2 expression in bone and osteoblasts were normalized in compound double Pkd1 +/Δ and Kif3a +/Δ heterozygous mice. The administration of sonic hedgehog, overexpression of Gli2, and the PC1 C-tail construct all increased Gli2 and Runx2-II expression, but decreased PPARγ2 gene expression in C3H10T1/2 cells. Our findings suggest a role for Pkd1 and Kif3a to counterbalance the regulation of osteogenesis and adipogenesis through differential regulation of Runx2 and PPARγ by Gli2.


Introduction
The primary cilium is a microtubule-based membrane protrusion that is assembled and maintained by the bidirectional intraflagellar transport (IFT) machinery [1] that is involved in the differentiation of mesenchymal stem cells into osteoblasts, chondrocytes and adipocytes. In this regard, disruption of primary cilia results in abnormal skeletal patterning, post-natal growth plate development, and skeletogenesis [2,3,4,5,6]. Conditional inactivation of kinesin family member 3A (Kif3a), a subunit of kinesin-2 motor complex, in mesenchymal stem cells results in severe patterning defects [3]. Conditional inactivation of Kif3a in chondrocytes results in post-natal dwarfism due to premature loss of the growth plate [2,4]. siRNA-mediated knock down of Kif3a in 3T3-L1 preadipocytes also leads to impaired adipocyte differentiation [7]. Primary cilia have also been identified in the osteoblast lineage and have been postulated to play a role in osteoblast differentiation [8,9].
Primary cilia and polycystins are co-expressed in cells within the osteoblast lineage [8] where they have been postulated to regulate skeletogenesis [4,9,25,26,27]. Although polycystin-1 (PC1), encoded by the Pkd1 gene, and Polycystin-2 (PC2), encoded by the Pkd2 gene, are mutated in autosomal dominant polycystic kidney disease [28,29,30,31], loss of polycystin function in mice also causes a severe skeletal phenotype. In this regard, homozygous loss of PC1 is associated with abnormal skeletal development through stimulation of the osteoblast-specific transcription factor Runx2-II [8,33]. Skeletal abnormalities are also observed in heterozygous Pkd1 mutant mice [8]. Moreover, Osteocalcin-Cre mediated conditional deletion of Pkd1 selectively in the osteoblast lineage results in osteopenia due to decreased osteoblast-mediated bone formation. Conditional deletion of Pkd1 in osteoblasts also results in increased adipogenesis in bone marrow stromal cell and impaired osteoblast differentiation, indicating that Pkd1 may also play a role in controlling a differentiation switch between the osteoblast and adipocyte lineages [34].
Primary cilia and polycystins are functionally interconnected in many tissues. For example, loss of PC1 or primary cilia in the kidney results in same cystic phenotype. Indeed, polycystic disease can be caused in mouse models by homozygous loss-of-function mutations in proteins required for cilia formation or function, such as TG737, Kif3a, fibrocystin, and cystin [35,36,37,38]. Whether polycystins and primary cilia have interdependent functions in skeletogenesis is not known.
In the current study, we sought to examine if PC1 and primary cilia have interdependent functions in osteoblast and bone development. We crossed heterozygous Pkd1-deficient mice onto heterozygous Kif3a-deficient mice to create double heterozygous Pkd1 and Kif3a-deficient mice in an effort to impair both PC1 and primary cilia function. Unexpectedly, Kif3a deficiency upregulated Hh signaling and reversed the effect of mutant Pkd1 to impair osteoblastic differentiation and stimulate adipogenesis in vivo and in vitro. These effects on bone development occurred through crosstalk between Pkd1 and Hh pathways at the level of Gli2 expression in bone and osteoblasts. Thus, we have discovered a new interaction between Hh and Pkd1 components of primary cilia.

Confirmation of Pkd1 and Kif3a deficiency in vivo and in vitro
Since homozygous Pkd1 and Kif3a null mice are embryonic lethal [32,39,40], we examined compound heterozygous Pkd1 and Kif3a deficient mice to establish a potential functional link between Pkd1 and Kif3a. Crossing heterozygous Pkd1 +/D mice with heterozygous Kif3a +/D mice generated four genotypes that were born with the expected Mendelian frequency, including wild-type, heterozygous Pkd1 +/D , heterozygous Kif3a +/D , and double heterozygous (Pkd1 +/D ;Kif3a +/D ) mice (Fig. 1B). The overall survival and body weight of single heterozygous and double heterozygous Pkd1 +/D and Kif3a +/D mice was not different from wild-type littermates. There was no evidence of cyst formation in kidneys of either single or double heterozygous mice (data not shown).
Real-time RT-PCR analysis revealed that the expression of total Pkd1 or Kif3a transcripts from tibias of single heterozygous mice was reduced ,50% at six weeks of age compared to wild-type control (Fig. 1C), consistent with the deficiency of a single Pkd1 +/D or Kif3a +/D allele. The expression of Pkd1 or Kif3a transcripts in tibias of double heterozygous Pkd1 +/D ;Kif3a +/D was reduced to the same degree as their respective single heterozygous mice (Fig. 1C). We found that primary osteoblast cultures derived from heterozygous Kif3a mice had an approximately 50,60% reduction in mRNA expression levels of Kif3a transcripts (Fig. 1D). Although homozygous disruption of Kif3a disables anterograde IFT and leads to failure in the formation and maintenance of cilia [39], heterozygous disruption of Kif3a had no effect on ciliogenesis. Indeed, neither the reduction in Pkd1 or/and Kif3a altered the appearance of primary cilia in cultured osteoblasts ( Fig.  1E and 1F).

Lack of an effect of Kif3a deficiency on skeletogenesis
We found that heterozygous Kif3a +/D mice had no demonstrable bone abnormalities (Fig. 2). Indeed, the Kif3a +/D mice had both normal bone mineral density (BMD) and bone structure compared with age-matched wild-type control mice ( Fig. 2A and 2B). In addition, bone samples from single heterozygous Kif3a +/D mice had no detectable changes in markers of osteoblasts or osteoclasts (Table 1). Primary osteoblasts derived from Kif3a +/2 mice underwent an osteoblast differentiation and mineralization program similar to wild-type derived cells ( Fig. 3A-C). Kif3a deficiency, however, resulted in significant reductions in adipocyte-related markers, including adipocyte-specific fatty acid binding protein (aP2), lipoprotein lipase (Lpl), and Adiponectin in long bone samples (Fig. 3D).

Pkd1 deficiency induces osteopenia due to impaired osteoblast differentiation
We observed osteopenia in heterozygous Pkd1 +/D mice, similar to previously described heterozygous Pkd1 m1Bei mutant mice [8]. Indeed, haploinsufficiency of Pkd1 expression in bone tissues resulted in a ,10% reduction in femoral BMD in both male and female adult mice ( Fig. 2A). Micro-computed tomography (mCT) analysis revealed that the reduction in bone mass in single Pkd1 +/D mice was caused by a reduction in trabecular bone volume (BV/ TV, 29%) and cortical bone thickness (CtTh, 14%) (Fig. 2B). These reductions in bone volume and cortical thickness were associated with a significant decrease in mineral apposition rate in single Pkd1 +/D mice compared with age-matched control (Fig. 2C).
We also found that low bone formation rates rather than increased bone resorption contributed to ostepenia in Pkd1 +/D mice. In this regard, bone derived from single heterozygous Pkd1 +/D mice had significant reductions in the osteoblast-lineage gene transcripts, including Runx2, Osteocalcin, Osterix, Osteoprotegerin (Opg), Rank ligand (RankL), and dentin matrix protein 1 (Dmp1) mRNA levels compared to wild-type control mice ( Table 1). The ratio of Opg/RankL that favors reduced osteoclastogenesis, and the bone expression of tartrate-resistant acid phosphatase (Trap), a marker of bone resorption, was reduced in heterozygous Pkd1 +/D mice (Table 1). In contrast, PPARc, Lpl, and Adiponectin but not aP2, markers of adipocyte differentiation, were significantly increased in the tibia of heterozygous Pkd1 +/D mice (Table 1). Serum levels of Osteocalcin and TRAP were also reduced in 6-week-old heterozygous Pkd1 +/D mice compared to wild-type littermates ( Table 2).
Rescue of osteopenia, abnormal mineralization and defective adipogenesis associated with Pkd1 deficiency in compound Kif3a and Pkd1 deficient mice Next, we assessed the effects of superimposed heterozygous deficiency of Kif3a on BMD and bone structure in Pkd1-deficient mice. Surprisingly superimposed Kif3a +/D deficiency onto heterozygous Pkd1 +/D mice corrected the skeletal phenotype observed in single mutant Pkd1 deficient mice (Fig. 2). Heterozygous deficiency of Kif3a in Pkd1 deficient mice corrects the effects of Pkd1 deficiency to lower bone mass ( Fig. 2A and 2B) and mineral apposition rate (MAR) (Fig. 2C). In addition, compound Pkd1 +/D ;Kif3a +/D heterozygous mice demonstrated normal expression of the osteoblastic, adipocytic and osteoclastic transcripts in whole bone (Table 1). Both heterozygous Kif3a +/D and compound Pkd1 +/D ;Kif3a +/D heterozygous mice had no significant alterations in serum markers of bone formation and resorption markers at 6 weeks of age compared with wild-type control ( Table 2). The apparent rescue of the Pkd1-mutant phenotype was not due to differences in genetic background, since both Pkd1 and Kif3a mutant mice had been crossed onto the C57BL/6J background for multiple generations. and deleted (Pkd1 D ) Pkd1 allele which has been removed the lox P cassette containing Exon 2-4 via Cre-mediated recombination (upper two panels), as well as wild-type (Kif3a + ) and deleted (Kif3a D ) Kif3a allele which has been excised the lox P cassette containing Exon 2 via Cre-mediated recombination (lower two panels). (B) Genotype PCR analysis of tail genomic DNA harvested from different individual mice. Four genotypes were generated in this breeding strategy. (C-D) Real-time RT-PCR analysis of total Pkd1 and Kif3a transcripts from the tibias of 6-week-old mice (C) and the cultured primary osteoblasts (D) by real-time RT-PCR. The level of Pkd1 or Kif3a transcripts exhibited almost ,50% decreases in long bone samples and primary cultured osteoblasts from single Pkd1 +/D or Kif3a +/D mice, and both Pkd1 and Kif3a transcripts retained the same reductions in compound Pkd1 +/D ;Kif3a +/D mice compared with wild-type control mice. The value of Pkd1 or Kif3a vs. cyclophilin A from the indicated genotype was a fold difference over wild-type. Data are expressed as the mean6SD from 5 to 6 individual mice. Values sharing the same superscript are not significantly different at P,0.05. (E-F) Immunofluorescence of primary cilia in cultured primary osteoblasts. Immunostaining of primary cilia (Red) was performed with acetylated a-tubulin antibody as described in the Materials and Methods. Counterstaining with a nuclear marker (DAPI blue) was used to calculate the percentage presence of primary cilia in cultured primary osteoblasts. There were no obvious number differences in the primary cilia among these four genotype osteoblasts. doi:10.1371/journal.pone.0015240.g001 Figure 2. Rescue of the skeletal phenotype in Pkd1-deficient mice by transfer onto Kif3a-deficient background. (A) Effects of Pkd1 or/ and Kif3a on BMD at 6 weeks of age. Compared with age-matched wild-type control, single heterozygous Pkd1 +/D mice had ,10% reduction in femoral BMD in both male and female adult mice. In contrast, there were no significant changes in the BMDs in single heterozygous Kif3a +/D mice, but the bone mass totally become normal in double heterozygous Pkd1 +/D ; Kif3a +/D mice, indicating Kif3a deficiency predominates over effects of Pkd1 deficiency on bone mass. The following data only represents male mice. (B) Effects of Pkd1 or/and Kif3a on bone structure of femoral metaphyses and midshaft diaphyses. mCT analysis revealed that single heterozygous Kif3a +/D mice exhibited a significant decrease in both BV/TV and CtTh, single heterozygous Kif3a +/D mice had no obvious changes in femoral bone structure, and Kif3a deficiency corrected the effects of Pkd1 To examine the cellular basis for these alterations in bone, we examined the impact of combined Pkd1 and Kif3a deficiency on cell proliferation, osteoblastic differentiation, and gene expression profiles in cells isolated from calvaria of newborn wild-type, single heterozygous and compound heterozygous Pkd1 and Kif3a deficient mice (Fig. 3). Cells isolated from calvaria are a mixed population capable of differentiation into both osteogenic and adipogenic pathways in culture [41,42]. We observed no significant changes in 5-bromo-2-deoxyuridine (BrdU) incorporation of calvarial derived cells between the four genotypes (data not shown), indicating neither Pkd1 nor Kif3a deficiency affected cell proliferation. Concordant with defects in both osteoblast and adipocyte developmental markers observed whole bone in vivo (Table 1), we found that cultured calvarial cells derived from single heterozygous Pkd1 +/D mice displayed lower alkaline phophatase (ALP) activity ( Fig. 3A) and diminished calcium deposition in extracellular matrix (Fig. 3B), reduced osteoblastic gene expression markers, such as Runx2 and Osteocalcin (Fig. 3C), and increased adipogenic gene expression, including PPARc and aP2 (Fig. 3D) compared to wild-type controls. Calvarial culture derived from heterozygous Pkd1 +/D also demonstrated decreased phosphorylation of the Akt (Fig. 3E), consistent with our prior results showing coupling of Pkd1 to the PI3K/Akt pathway in osteoblasts [43].
In contrast, single heterozygous Kif3a +/D osteoblasts had no abnormalities in ALP activity, mineralization of extracellular matrix, or osteogenic gene expressions profiles ( Fig. 3C-3E). Consistent with the decreased adipogenic markers in whole bone of Kif3a +/D mice, however, calvarial derived cells from Kif3a +/D mice expressed significantly lower levels of PPARc and aP2 compared to wild-type controls.
The defective osteoblastic differentiation, mineral deposition, and osteogenic gene expression in calvarial cultures derived Pkd1deficient mice was completely reversed by superimposing Kif3a deficiency. Indeed, calvarial cells derived compound heterozygous Pkd1 +/D ;Kif3a +/D mice had alkaline phosphatase activity, mineralization and expression of Runx2 and osteocalcin transcripts that did not differ from wild-type control cultures ( Fig. 3C-3F). Moreover, adipogenic gene expression (i.e., PPARc and aP2) that was increased in Pkd1 +/D derived calvarial cells was suppressed in compound heterozygous Pkd1 +/D ;Kif3a +/D cultures to levels not significantly different from Kif3a +/D -derived cells (Fig. 3F), consistent with a predominant role of Kif3a deficiency to regulate adipogenesis.
To further investigate the effects of Pkd1 or/and Kif3a deficiency on the adipogenesis potential, we investigated the adipogenic potential of bone marrow-derived mesenchymal stem cells (BMSC) derived from wild-type, heterozygous Pkd1 +/D , heterozygous Kif3a +/D , and double heterozygous (Pkd1 +/D ;Kif3a +/D ) mice grown in the presence of rosiglitazone [44]. Under these conditions, BMSC derived from wild-type mice cultures exhibited a low density of Oil Red O staining adipocytes. Single heterozygous Pkd1 +/D mice exhibited a marked increase in adipogenic cells as evidenced by cells with lipid droplet formation ( Fig. 4A-4B). There was no significant difference in adipocyte formation between single Kif3a +/D and wild-type BMSC cultures. However, superimposed Kif3a deficiency entirely reversed the Table 1. Gene-expression profiles in 6-week-old mice.

Gene
Accession no.  Role of hedgehog/Gli2 signaling in Pkd1 or/and Kif3a deficiency mice Since heterozygous deficiency of Kif3a failed to alter primary cilia formation or Pkd1 expression (Fig. 1E, vide supra), we explored the potential role of hedgehog signaling pathway in mediating the effects of Kif3a deficiency to rescue skeletal abnormalities in heterozygous Pkd1 +/D mice. As a measure of Hedgehog pathway activity in the various mouse models, we assessed expression of total Gli2 transcripts using total RNA from tibias and cultured osteoblasts. We found that Gli2 expression was significantly increased in single heterozygous Kif3a +/D mice compared to wildtype controls (Fig. 5A-5B). Although most studies show that mutations in IFT proteins, including Kif3a, result in a downregulation of Hedgehog signaling [45,46,47], inactivation of Kif3a has been reported to paradoxically increase Hedgehog pathway activity in certain tissues, including cranial neural crest cells and the limb [45,48,49]. In contrast, we found a significant reduction of Gli2 expression in whole tibias and calvarial-derived osteoblasts from single heterozygous Pkd1 +/D mice compared to wild-type mice. The effect of Pkd1 deficiency to suppress Gli2 expression was completely reversed in bone and calvarial cultures derived from compound heterozygous Pkd1 +/D ;Kif3a +/D mice (Fig. 5A and  5B).

Discussion
We previously reported the presence and co-localization of polycystins and primary cilia in osteoblasts [8] and proposed that polycystins and primary cilia have an interdependent role in regulating osteoblast differentiation and skeletal development [9]. In the current study, we attempted to assess the interaction between Pkd1 and primary cilia by disrupting Kif3a in Pkd1deficient mice. We created new mouse models of Pkd1 and Kif3a deficiency by crossing CMV-Cre with either Pkd1 flox/+ mice to achieve a germline deletion of exon 2-4 or Kif3a flox/+ mice to attain deletion of exon 2 [51]. Quantitative RT-PCR analysis revealed that the expression of total Pkd1 transcripts in single Pkd1 +/D heterozygous mice was reduced by ,50% compared to wild-type control. We found that this reduction in Pkd1 expression in heterozygous Pkd1 +/D mice was sufficient to cause osteopenia due to impairment of osteoblast-mediated bone formation and enhanced adipocyte differentiation. These results are in agreement with our previous reports in heterozygous Pkd1 +/m1Bei and conditional Pkd1 Oc-cKO null mice [33,34], indicating that Pkd1 deficiency favors adipogenesis but inhibits osteogenesis in bone tissues. In contrast, we found that single global Kif3a +/D heterozygous mice had no demonstrable bone or kidney phenotype, even though the level of total Kif3a transcripts from tibias and cultured osteoblasts was decreased by ,50%. Hence, haploinsufficiency of Pkd1 had greater effects on bone than partial deficiency of Kif3a.
Our apriori assumption was that an interaction between primary cilia and Pkd1 would be revealed by additive effects of Kif3a deficiency to worsen the effects of Pkd1-deficiency to impair bone development in compound Pkd1 +/D ;Kif3a +/D mice. We found, however, the opposite result. Superimposed Kif3a deficiency in compound Pkd1 +/D ;Kif3a +/D mice rescued the skeletal abnormalities and defects in osteoblast differentiation observed in Pkd1 +/D mice. Indeed, reduction of Kif3a in compound Pkd1 +/D ;Kif3a +/D mice completely reversed the effects of deficiency of Pkd1 on bone mass and osteoblast and adipocyte differentiation, indicating deficiency of Kif3a rescues the bone effects of Pkd1 haploinsufficiency. In addition, neither single Pkd1 and Kif3a nor double Pkd1 and Kif3a mutant mice developed polycystic kidney disease, unlike the reported additive effects of Pkd1 and Pkd2 deficiency to enhance renal cyst formation [52].
This salutary effect of Kif3a deficiency was associated with alterations in the osteogenic and adipogenic pathways. Whereas calvarial and bone marrow derived cells from heterozygous Pkd1 +/D mice had impaired osteoblastic and increased adipocytic differentiation in culture associated with decreased Runx2 and increased PPARc expression, superimposed Kif3a +/D deficiency increased Runx2 and decreased PPARc expression restoring the normal balance between osteogenic and adipogenic differentiation pathways. Thus, rather than primary cilia and Pkd1 interactions per se, we have shown that two components housed in primary cilia have counter regulatory effects on skeletogenesis.
We have previously shown that loss of Pkd1 inhibits osteoblast development through reductions in intracellular calcium-mediated Runx2 expression [33] and that primary osteoblasts derived from conditional Pkd1 Oc-Cre null mice displayed impaired osteoblastic differentiation and enhanced adipogenesis via suppression of PI3K-Akt-Gsk3b-b-catenin signaling pathway [34]. In the current study, we have found evidence that Hedgehog (Hh)/Gli signaling is a possible mechanism mediating the interactions between Pkd1 and Kif3a on skeletogenesis. Even though partial deficiency of Kif3a did not affect primary cilia number or formation, haploinsufficiency of Kif3a might disrupt IFT necessary for normal hedgehog signaling [18,45,51,53,54,55], which in turn could counter the effects of deficiency of Pkd1 signaling [56,57]. In this regard, Gli2 expression was reduced in both bone and cultured primary osteoblasts from single Pkd1 +/D heterozygous mice, but was normalized in compound Pkd1 +/D ;Kif3a +/D heterozygous mice. Since inhibition of microtubule assembly in osteoblasts is reported to stimulate the hedgehog signaling  molecule Gli2 expression [58], the increase of Gli2 in single Kif3a +/D heterozygous mice may have offset the effects of Pkd1 haploinsufficiency. Indeed, we found that overexpression of gainof-function PC1 C-tail construct promotes Gli2 and Runx2-II expression, and we also observed that administration of Shh or overexpression of Gli2 up-regulates Runx2-P1 promoter activity and expression, consistent with a stimulatory role of Gli2 hedgehog signaling in osteoblast differentiation and bone development [22,59,60,61]. In contrast, either administration of Shh or overexpression of Gli2 or gain-of-function PC1 C-tail construct down-regulates PPARc2 promoter activity and expression, consistent with an inhibitory function of Gli2 hedgehog signaling on adipogenesis and fat formation [22,23,62,63,64]. In agreement with these findings, knock-down of Kif3a by siRNA suppresses adipocyte differentiation in 3T3-L1 preadipocyte cultures [7]. Kif3a deficiency reduces expression of PPARc and CEBPá proteins and has anti-adipogenic role in 3T3-L1 adipogenic cells [7]. Since adipogenesis and osteogenesis are inversely related [65,66,67], the reduced adipogenesis in Kif3a haploinsufficient mice may have promoted osteoblasts differentiation to offset the effects of deficiency of Pkd1. In support of this hypothesis, we found that Pkd1 and Kif3a haploinsufficiency had opposite effects on adipogenic and osteogenic transcription factors and related gene expressions. Kif3a +/D heterozygous mice exhibited significant reductions in adipocyte-related markers such as PPARc and aP2, whereas Pkd1 deficient mice had reductions in osteoblastic markers such as Runx2 and Osteocalcin expression, indicating a counterbalance mechanism involved in Pkd1 and Kif3a haploinsufficient mice. These findings suggest that Pkd1 or/and Kif3a deficiency might affect cell fate decision between osteoblastogenesis and adipogenesis through regulation of hedgehog signaling pathway.
Kif3a mediates intraflagellar transport (IFT), which is necessary for normal formation of the appendicular skeleton through hedgehog signaling pathways [2,4,19]. Intraflagellar transport has been shown to be essential for both endochondral bone formation as well as perichondral bone formation, which is a form of intramembranous bone formation. This suggests that IFT may be important in both chondrocytic and osteoblastic lineages, a finding supported by abnormalities of intramembranous bone formation observed in mice after the conditional deletion of Kif3a or Tg737 in mesenchymal skeletal progenitor cells [3,4,5,19]. While most studies suggest that inactivation of Kif3a reduces Hh signal transduction [45,46,47,53], Kif3a may also activate Hh signaling pathways in some settings [68], as we observed in heterozygous Kif3a +/D mice. For example, conditional deletion of Kif3a in chondrocytes stimulates Indian hedgehog (Ihh) expression in the perichondrium [4]. Also, conditional deletion of Kif3a in neural crest cells, which give rise to the facial skeleton, leads to gain of Hedgehog function that results in hypertelorism and iprosopus [48]. Other observations indicate that Kif3a and other IFT mutations can have paradoxical effects on hedgehog signaling [45,49]. This inconsistency has been attributed to the differential effects of Gli proteins (i.e., Gli1, Gli2, or Gli3) that may predominate in different tissues [18,45,53,69].
Another possible explanation for the offsetting effects of Kif3a deficiency to alter the bone phenotype in Pkd1 haploinsufficient mice may be due to Kif3a altering the amount of Pkd1 in primary cilia. This possibility is supported by the fact that Kif3a binds to Pkd2 and regulates the amount of Pkd2 localized in primary cilia [70,71] and the related observation that the ratio of Pkd1 and Pkd2 determines the net function of the polycystins [72]. Thus, deficiency of Kif3a may lead to a disproportionate reduction in Pkd2 compared to Pkd1 in combined Kif3a/Pkd1 mutant mice, leading to ''normalization'' of the polycystin signaling by restoring the ratio of Pkd1 and Pkd2 in primary cilia. Consistent with this possibility, transferring Kif3a-deficiency onto Pkd1-deficiency restores PI3K-Akt signaling in bone and osteoblasts (Fig. 3E).
In summary, Kif3a haploinsufficiency counteracts the negative effects of Pkd1 deficiency on osteoblast function, indicating an interdependent relationship between Pkd1 and Kif3a in postnatal bone formation. While the precise mechanism remains to be defined, Kif3a haploinsufficiency up-regulates hedgehog/Gli2 signaling pathways, leading to both increased Runx2 and osteoblastogenesis and significant inhibition of PPARc and adiopogenesis. In contrast, Pkd1 haploinsufficiency inhibits hedgehog/Gli2 signaling, Runx2 expression and osteogenesis but increases PPARc and adiopogenesis. Thus, cross-talk between Pkd1 and Kif3a may play a counterbalancing role on bone formation through the differential regulation of ostoegenesis and adipogenesis in bone. These findings suggest that activation of hedgehog/Gli2 signaling may provide a mechanism to counteract the effects of lost Pkd1 signaling in bone and possibly other tissues, such as the kidney, where defective hedgehog signaling is associated with cystic kidney disease in Oral-facial-digital syndrome type 1 [73]. With regards to our original question of the interdependence of polycystin and primary cilia function in osteoblast development, additional experimental approaches to more completely ablate Kif3a and/or primary cilia in the osteoblasts lineage will be required.

Genotyping PCR
Genomic DNA was prepared from tail specimens using standard procedures [33]. Genotyping PCR was performed using the following primers as previously described [39,51,74] wild-type and null alleles were identified in 2% agarose gel as 250bp and 850-bp bands (Fig. 1B, upper two panels), while Kif3a wildtype and null alleles were detected as 360-bp and 200-bp bands (Fig. 1B, lower two panels), respectively.
Bone Densitometry, histomorphometric and Micro-CT analysis BMD of femur from 6-week-old mice were scanned using a PIXImus bone densitometer (Lunar Corporation, Madison, WI) with dual-energy X-ray absorptiometry technology. BMD within a defined area of the whole femur was analyzed using proprietary Lunar PIXImus software. Calcein (Sigma, St. Louise, MO) double labeling of bone and histomorphometric analyses of periosteal MAR in tibias were performed using the osteomeasure analysis system (Osteometrics) [75,76]. Femurs were isolated from 6-weekold mice and fixed in 70% ethanol for mCT analysis. The long axis of the femur was oriented orthogonally to the rotation axis of the scanner. Entire femur scans were performed at an isotropic voxel size of 12 mM using a mCT 40 scanner (Scanco Medical AG, Brüttisellen, Switzerland). A 3D image analysis was done to determine BV/TV in the distal metaphyses and Ct.Th in the midshaft of diaphyses area as previously described [8,33,75,77].

Quantitative Real-time RT-PCR
A 2.0 mg of total RNA were isolated from tibia of 6-week-old mice or 10 days cultured primary osteoblasts in differentiation media. The cDNAs were generated using a Reverse Transcriptase Kit (Perkin-Elmer, Foster City, CA). PCR reactions contained 100 gg template (cRNA or cDNA), 200 gmol each forward and reverse primers, 1X iQ TM SYBR H Green Supermix (Bio-Rad, Hercules, CA) in 50 ml. The threshold cycle (Ct) of tested-gene product from the indicated genotype was normalized to the Ct for cyclophilin A as previously described [8,33,77]. Expression of total Pkd1 transcripts was performed using the following Pkd1-specific primers: forward primer of Pkd1 transcripts in exon 2-4, 59-TAG GGC TCC TGG TGA ACC TT -39, and reverse primer of Pkd1 transcripts in exon 2-4, 59-CCA GAC CAC AGT TGC ACT CA-39. Expression of total Kif3a transcripts was performed using the following Kif3a-specific primers: forward primer of Kif3a transcripts in exon 2, 59-GCT ATA GAC AGG CCG TCA GC-39, and reverse primer of Kif3a transcripts in exon 2, 59-GTC TTT GGA GGT TCG TTG GA-39. The value of Pkd1 or Kif3a vs. cyclophilin A from the indicated genotype was normalized to the mean ratio of 5 wild-type mice, which has been set to 1.

Primary Osteoblast Culture for Proliferation and Differentiation, and western blot analysis
Primary osteoblasts were isolated from the newborn mouse calvarias by sequential collagenase digestion at 37uC as previously described [8,77]. The cells were cultured in alpha-minimal essential medium (a-MEM) containing 10% fetal bovine serum (FBS) and 1% penicillin and streptomycin (P/S). Cell proliferation was detected by BrdU incorporation assays as the manufacturer describes (QIA58, Calbiochem, Gibbstown, NJ). To induce differentiation, primary osteoblasts were plated at a density of 2.5610 4 cells/cm 2 in a 6-well plate, and grown for period of up to 21 days in a-MEM containing 10% FBS supplemented with 5 mM b-glyceropho-phate and 25 mg/ml of ascorbic acid. Alkaline phosphatase activity and Alizarin red-S histochemical staining for mineralization were performed as previously described [8,77]. Total DNA content was measured with a PicoGreenH dsDNA quantitation reagent and kit (Molecular Probes, Eugene, OR). Protein concentrations of the supernatant were determined with a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA).
To examine the amounts of cytoplasmic phosphorylated Akt, we isolated cytoplasmic protein using NE-PER Nuclear and Cytoplasmic Extraction Kit (Pierce Biotechnology, Rockford, IL) according to the manufactory's instruction. Protein concentrations of the supernatant 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, Carlsbad, CA) and were analyzed with standard Western blot protocols (HRP-conjugated secondary antibodies from Santa Cruz Biotechnology and ECL from Amersham Biosciences, Buckinghamshire, UK). Antibodies against phospho-Akt (ser-473) and Akt were from Cell Signaling Technology (Beverly, MA). Anti-bactin (sc-47778) antibodies were from Santa Cruz Biotechnology.

Transient transfection
Pluripotent C3H10T1/2 mesenchymal cells (American Type Culture Collection, Manassas, VA) were maintained in Eagle's Minimum Essential Medium (EMEM) containing 10% FBS and 1% P/S. A 2,300 bp promoter region of PPARc2 gene was amplified by polymerase chain reaction (PCR) using a set of primers 59-TCC CCG GGG GTA TGT GG A GCC CA A CCC A -39 and 59-ACG TCG ACA GAT TTG CTG TAA TTC ACA CTG G -39 containing SmaI and SalI sites and the resultant product was subcloned into pluc4 luciferase reporter construct and confirmed by subsequent analysis [78,79]. To examine if hedgehog signaling regulates Runx2-P1 and PPARc2 promoter activity, a number of 1610 6 of C3H10T1/2 cells were transfected with either Runx2-P1 promoter luciferase reporter (p1.4Runx2-P1-Luc) or PPARc2 promoter luciferase reporter (p2.3PPARc2-Luc) conducted by electroporation using Cell Line Nucleofector Kit R according to the manufacturer's protocol (Amaxa Inc, Gaithersburg, MD). A total of 6.6 mg of plasmid DNA was used for each electroporation, with 3.0 mg of pcDNA3.1 empty expression vector, 3.0 mg of either p1.4Runx2-P1-Luc or p2.3PPARc2-Luc reporter, and 0.6 mg of Renilla luciferase-null (RL-null) as internal control plasmid. The cells were cultured in EMEM supplemented with 1%FBS and promoter activity will be assessed by measuring luciferase activity 72 hours after transfection in the presence or absence of 1 mg/ml of recombinant mouse sonic hedgehog Nterminus (Shh-N) treatment. The total RNA was also isolated for real-time RT-PCR analysis. To examine if Gli2 or PC1 has a direct effect on Runx2-P1 and PPARc2 promoter activity, a Gli2 or PC1 C-tail (PC1-AT) expression construct along with either p1.4Runx2-P1-Luc or p2.3PPARc2-Luc were co-transfected into C3H10T1/2 cells. The cells were cultured for 48 hours in EMEM supplemented with 1%FBS and the relative luciferase activities of cell lysates were measured by a luciferase assay kit (Promega, Madison, WI) and normalized with Renilla luciferase activity and empty expression vector (pcDNA3.1 or sIgØ) as previously described [8,33]. We also isolated the total RNAs for real-time RT-PCR analysis.

Immunofluorescence
Primary osteoblasts were grown on collagen-coated 4-well chamber at 1610 5 cells per well 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 cells were incubated for 30 minutes in 1% BSA before incubation with the primary acetylated alpha-tubulin antibody (1:4000, 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 ProLongH Gold antifade reagent (Invitrogen, P36935). Nuclei were counter-stained with DAPI blue. Photographs were taken under a microscope with magnifications of 406 for counting the number of primary cilia in cultured primary osteoblasts as previously described [8].

Adipocyte Differentiation of Bone Marrow-derived Mesenchymal Stem Cells
Primary BMSCs were harvested from the long bone of 8-weekold mice as previously described [75]. Briefly, the epiphysis of long bone was cut off, the bone marrow was flushed out with a-MEM containing 10% FBS and 1% P/S and went through a 70-mm filter mesh. The cells were plated in 100 mm culture dishes at a density of 2.5610 7 cells per cm 2 and cultured in a humidified incubator with 5% CO 2 and 95% air at 37uC. On day 3, a half of cultured medium was removed and replaced with fresh a-MEM growth medium. On day 5, the adherent cells (representing BMSCs) were detached with 0.25% trypsin/1 mM EDTA and seeded onto 6-well plates at 3610 4 cells per well for up to 8 days.
To induce adipocyte differentiation, the BMSCs were treated with adipogenic differentiation medium supplemented with 10% FBS, 5 mM beta-glycerophophate and 50 mg/ml of ascorbic acid for 4 days and then added 1 mmol/L rosiglitazone for another 8 days. At the endpoint, the cells were processed for Oil Red O lipid staining as previously described [44]. Breifly, the cells were rinsed with 1xPBS and fixed with cold 4% paraformaldehyde solution for 1 hour at 4uC. The cells were rinsed with distilled water followed by 60% isopropanol for 5 minutes at room temperature. Then the cells were stained in 0.5% Oil Red O-isopropanol working solution for 15 minutes at room temperature, differentiated in 60% isopropanol, and rinsed in tap water. For quantitative analysis, Oil Red O was extracted with 1 ml 99% isopropanol for 2 min, and optical density of each sample was determined at 510 nm.

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 TRAP was assayed with the ELISA-based SBA Sciences mouseTRAP TM assay (Immunodiagnostic Systems, Fountain Hills, AZ).

Statistical Analysis
We evaluated differences between two groups by unpaired t-test and multiple groups by one-way analysis of variance. All values are expressed as means 6 S.D. All computations were performed using the GraphPad Prism5 (GraphPad Software Inc. La Jolla, CA).