IGF-I induced phosphorylation of PTH receptor enhances osteoblast to osteocyte transition

Parathyroid hormone (PTH) regulates bone remodeling by activating PTH type 1 receptor (PTH1R) in osteoblasts/osteocytes. Insulin-like growth factor type 1 (IGF-1) stimulates mesenchymal stem cell differentiation to osteoblasts. However, little is known about the signaling mechanisms that regulates the osteoblast-to-osteocyte transition. Here we report that PTH and IGF-I synergistically enhance osteoblast-to-osteocyte differentiation. We identified that a specific tyrosine residue, Y494, on the cytoplasmic domain of PTH1R can be phosphorylated by insulin-like growth factor type I receptor (IGF1R) in vitro. Phosphorylated PTH1R localized to the barbed ends of actin filaments and increased actin polymerization during morphological change of osteoblasts into osteocytes. Disruption of the phosphorylation site reduced actin polymerization and dendrite length. Mouse models with conditional ablation of PTH1R in osteoblasts demonstrated a reduction in the number of osteoctyes and dendrites per osteocyte, with complete overlap of PTH1R with phosphorylated-PTH1R positioning in osteocyte dendrites in wild-type mice. Thus, our findings reveal a novel signaling mechanism that enhances osteoblast-to-osteocyte transition by direct phosphorylation of PTH1R by IGF1R.


INTRODUCTION
Osteocytes make up over 90% of bone cells and play a major role in control of skeletal tissue homeostasis 1 . Osteocytes regulate bone remodeling, maintain phosphate homeostasis, serve as mechanosensors, and secrete endocrine hormones to communicate with other organs 2 . Osteocytes are terminally differentiated osteoblasts derived from mesenchymal stem cells (MSCs) that become embedded in bone matrix. During the differentiation of osteoblasts to osteocytes, the cells undergo morphological changes, transitioning from a polygonal shape into cells with dendritic extensions 3,4 . While signaling mechanisms that direct differentiation of MSCs to osteoblasts have been extensively studied, the characterization of the transition of osteoblasts to osteocytes is just beginning to be elucidated 2 , but the mechanism regulating changes in cytoskeletal proteins, enzymes, and hormones remains unclear. As osteocytes can survive for up to decades 2 , further studies elucidating factors that influence osteocyte differentiation are essential for understanding disease conditions and therapeutics.
Insulin-like growth factor type 1 (IGF-1) plays a key role in MSC to osteoblast differentiation [5][6][7][8][9] . Osteocytes, similarly to MSCs, osteoprogentior cells, and mature osteoblasts, express IGF-1 and the insulin-like growth factor type 1 receptor (IGF1R) 10 . Disruption of IGF1R in mature osteoblasts and early osteocytes in mice impairs bone formation 11,12 , whereas transgenic overexpression of IGF-1 in mature murine osteoblasts increases osteocyte lacunae occupancy, indicating a potential role in osteoblast-to-osteocyte transition 13 . Clinical and mouse observations suggest an interdependent role of IGF-1 and parathyroid hormone (PTH) for anabolic effects 14 . Both patients who are growth hormone deficient and mice that have been hypophosectomized show a blunted response to PTH, with restoration of the PTH response with co-administration of growth hormone [15][16][17] . More specifically, global IGF-1 knockout mice and osteoblast-specific IGF1R knockout mice fail to show an anabolic response to PTH in trabecular bone [18][19][20] . While it is well known that PTH increases IGF-1 mRNA and protein expression [21][22][23][24][25] , Yamaguchi et al. found that downstream IGF-1 signaling events could be detected in response to PTH long before IGF-1 mRNA transcription occurred 26 , suggesting a more complicated relationship between PTH and IGF-1 signaling.
PTH plays a critical role in both osteoblasts and osteocytes by regulating calcium homeostasis and orchestrating bone remodeling 27 . The actions of PTH are mediated by a G-protein-coupled receptor, termed PTH type I receptor (PTH1R) 28,29 , which is expressed in MSCs, osteoblasts, and osteocytes 30,31 . Constitutive activation of PTH1R in murine osteoblasts/osteocytes results in increased trabecular and cortical bone 32,33 , while mice lacking PTH1R in osteoblasts/osteocytes have less trabecular bone formation 34 . PTH1R stimulates downstream signaling events via cyclic AMP production, but can also directly interact with other cell surface proteins to regulate bone remodeling 27 . For example, PTH1R-mediated endocytosis of BMP antagonists and recruitment of low-density lipoprotein-related protein 6 (LRP6) as a co-receptor stabilizes β-catenin and enhances bone morphogenetic protein (BMP) signaling 35,36 . Phosphorylation of PTH1R by transforming growth factor beta 2 receptor leads to endocytosis of both receptors and suppression of TGF-β signaling 37 . In the current study, we investigated if PTH1R and IGF1R directly interact to promote osteoblast-to-osteocyte differentiation. We found in vitro that IGF1R phosphorylated tyrosine 494 (Y494) in the cytoplasmic tail of PTH1R. A novel antibody against this tyrosinephosphorylation was generated, which facilitated subsequent tracking of phosphorylated-PTH1R in osteoblasts and early osteocytes in vitro and in vivo. We found that the Y494phosphorylation mediated the preferential targeting of PTH1R to the actin cytoskeleton. PTH signaling enhanced actin filament polymerization to promote outgrowth of osteocyte dendrites. Our findings suggest that a novel PTH-IGF-1 signaling interaction contributes to dendrite outgrowth during osteoblast-to-osteocyte transition.

RESULTS
Osteoblast and osteocyte differentiation is enhanced by costimulation of IGF-1 and PTH IGF-1 is known to play a key role in osteoblast differentiation; however, little is known about osteoblast-to-osteocyte differentiation. Alkaline phosphatase (Fig. 1a) and alizarin red staining (Fig.  1b, c) showed that IGF-1 promotes osteoblast differentiation, whereas PTH alone does not. However, PTH and IGF-1 together resulted in increased osteoblast differentiation, suggesting a potential interaction between IGF-1 and PTH signaling (Fig.  1a-c), similar to previous reports 14,20 . To explore if the IGF-1 signaling pathway is important for osteocyte differentiation, we deleted IGF1R selectively in preosteoblasts by crossing floxed IGF1R (IGF1R f/f ) mice with osterix-Cre (Osx-Cre) mice. Mice lacking IGF1R (Osx-Cre/IGF1R f/f ) had a decreased length of dendrites on osteocytes (Fig. 1d, e) and decreased number of osteocytes per tissue area (Fig. 1f) relative to wild-type littermates (Osx-Cre/ IGF1R f/f ). While PTH treatment in wild-type littermates increased the number of osteocytes per tissue area, this effect was lost in the Osx-Cre/IGF1R f/f mice (Fig. 1f). These results suggest that both IGF-1 and PTH play a role during osteoblast-to-osteocyte transition, and the role of PTH in osteocytes depends on the presence of IGF1R.
To further evaluate the role of PTH and IGF-1 in osteoblast-toosteocyte transition, we developed a cell culture system for early osteocyte-like (EOcyL) cells. We selected primary BMSCs using markers for Sca1, CD45 and CD11b (Fig. 2a). Sca1 + /CD45 − / CD11b − BMSCs were seeded on slices of bovine cortical bones in osteogenic medium (Fig. 2b). By 12 days of culture, the morphological appearance of the cells revealed dendrite-like processes (Fig. 2b) and expressed the early osteocyte marker genes Dmp1, Phex and FGF23, but not the mature osteocyte marker SOST (Fig. 2c). EOcyL cells were exposed to PTH alone, IGF-1 alone, or both PTH and IGF-1. DMP1 expression did not change with PTH alone, increased slightly with IGF-1 alone, and increased more significantly with both PTH and IGF-1 relative to vehicletreated control EOcyL cells (Fig. 2d) Fig. 1 Osteoblast/Osteocyte differentiation is enhanced by co-stimulation of IGF-1 and PTH.a, b Osteogenic differentiation of BMSCs treated with vehicle (-), IGF-1 (50 ng·mL -1 ), PTH (100 nmol·L -1 ), or both as assessed by alkaline phosphatase (a) and alizarin red staining (b). c Quantification of alizarin red staining. d-f Representative F-actin staining (d) and quantification of the number of osteocytes (e) and estimated dendritic length (f) in calvarial osteocytes in two-month-old Osx-Cre/IGF1R f/f mice and wild-type littermates (Osx-Cre/IGF1R +/+ ) treated with either vehicle (veh) or PTH. The data shown as mean ± s.d. n = 5. *P < 0.05; **P < 0.01; NS not significant was noted in regards to dendrite length (2E). We also disrupted the expression of PTH1R in EOcyL cells by isolating and utilizing BMSCs from PTH1R f/f mice with adenovirus-mediated transduction of Adeno-Cre-GFP. The GFP + transductants were PTH1R-deficient cells (ΔPTH1R) and exhibited shortened dendritic processes, relative to negative transductants (GFP − ) (PTH1R +/+ ) (Fig. 2f, g). In contrast, the cells with Adeno-GFP control virus had no such effects (Fig. 2f, inset). While co-treatment with IGF-1 and PTH increased dendrite length in PTH1R +/+ EOcyL cells, this effect was abrogated in ΔPTH1R EOcyL cells (Fig. 2g), again suggesting a  whether IGF1R directly interacts with PTH1R ( Fig. 3b). Reconstitution of YFP fluorescence from IGF1R-YFP1 and PTH1R-YFP2 was induced on the cell membrane only in the presence of exogenous IGF-1, not with either vehicle or PTH (Fig. 3c, d), indicating that IGF-1 induces a direct interaction between PTH1R and IGF1R. The complex formation of PTH1R with IGF1R was confirmed by immunoprecipitation assays in HEK293 cells transfected with IGF1R-Flag and PTH1R-HA plasmids ( Fig. 3e) or primary calvarial osteoblasts (Fig. 3f). The immunoprecipitate was detected only in cells exposed to IGF-1, not vehicle or PTH (Fig. 3e, f). To further determine whether IGF1R phosphorylates PTH1R directly, we prepared a native cytoplasmic domain of PTH1R (cPTH1R) 37 , and tested its potential phosphorylation using the C-terminal of IGF1R (cIGF1R) as a kinase. The cIGF1R exhibited auto-phosphorylation activity and phosphorylated cPTH1R (Fig. 3g). Mass spectrometry analysis of the phosphorylated vs. unphosphorylated peptide identified an 80.11 mass unit shift between the seventeenth and eighteenth tyrosine, corresponding to mono-phosphorylation Y494 of PTH1R (Fig. 3h), indicating that IGF1R kinase directly phosphorylates a specific tyrosine of PTH1R.

PTH1R and IGF1R interaction does not alter downstream signaling pathways
To test the effect of Y494-phosphorylation on downstream PTH and IGF-1 signaling, we generated a phenylalanine mutation (pMSCV-PTH1R-Y494F) in PTH1R to make Y494 non-phosphorylatable, while maintaining a similar amino acid structure. We measured the activation of intracellular cAMP using the biosensor ICUE3 38 in HEK293 and UMR106 cells transfected with wild-type PTH1R or PTH1R-Y494F in response to PTH (Fig. 4a, b). The emission ratio was not altered in PTH1R-Y494F cells relative to control cells (Fig. 4a, b). Phosphorylation of CREB, a cAMP signal transducer, was also similar in cells expressing PTH1R-Y494F compared to those expressing wild-type PTH1R, indicating that PTH/cAMP signaling is independent of Y494-phosphorylation (Fig.  4c). IGF-1 and PTH did not directly cross-activate each other's downstream signaling targets. Specifically, IGF-1 did not induce the phosphorylation of CREB (Fig. 4d), and PTH did not stimulate the phosphorylation of IGF1R or Akt, downstream components of IGF-1 signaling (Fig. 4e, f).
IGF-1-induced PTH1R phosphorylation (P Y -PTH1R) enhances actin polymerization As the signaling pathways of PTH and IGF-1 were unaffected, we then sought to explore if the phosphorylation of PTH1R by IGF1R played a role in localization of the receptor. We developed and validated a polyclonal antibody against P Y -PTH1R to localize and track PTH1R after Y494-phosphorylation. The antibody recognized the antigen only when the cPTH1R peptide was incubated with cIGF1R in a kinase assay (Fig. 5a). To determine the specificity of the PTH1R Y494 phosphorylation, we utilized a small peptide of PTH1R that was either not phosphorylated (spPTH1R) or phospho-saturated (P Y -spPTH1R) at the Y494 position. cIGF1R was only able to induce phosphorylation of spPTH1R, but not P Y -spPTH1R (Fig. 5b). We further tested the P Y -PTH1R antibody in HEK293 cells co-expressing IGF1R and PTH1R fused with CFP (PTH1R-CFP). PTH1R was detected on the cell membrane in vehicle and IGF-1 exposed cells, whereas P Y -PTH1R was observed on the cell membrane only after exposure to IGF-1 (Fig. 5c). BMSCs expressing GFP and plated in osteogenic medium for EOcyL cell culture were supplemented with vehicle or IGF-1. P Y -PTH1R signals were rarely detected in BMSCs treated with vehicle (Fig. 5d, left panels). In the IGF-1 supplemented BMSCs in EOcyL cell culture system, by day 2 of culture, P Y -PTH1R was observed on the cell surface (Fig. 5d, upper panels) and by day 10 of culture, P Y -PTH1R was observed on the tips of the dendritic-like processes (Fig. 5d, lower panels). Barbed ends of polymerizing actin filaments push the membrane and promote the formation of plasma membrane protrusions or invaginations [22][23][24] . Co-staining of P Y -PTH1R and Factin revealed that P Y -PTH1R was trans-localized to the barbed ends of F-actin filaments of BMSCs and treated with IGF-1 (Fig. 6a, left panels). Treatment with PTH increased actin polymerization, whereas PTH plus cyclolignan picropodophyllin (PPP), an IGF1R kinase inhibitor, blunted actin polymerization (Fig. 6a, b). P Y -PTH1R staining and actin polymerization were also both decreased in pMSCV-PTH1R-Y494F transfected HEK293 cells relative to wild-type control (pMSCV-PTH1R) (Fig. 6c, d), indicating that the Y494-phosphorylation site specifically contributed to polymerization of actin filaments. Finally, we transfected Sca1 + / CD45 − /CD11b − BMSCs with either pMSCV-PTH1R-Y494F or pMSCV-PTH1R and subjected the cells to EOcyL conditions. Mineralization as assessed by alizarin red was reduced, whereas cell viability was not affected (Supplementary Fig. S1A, B). As expected, disruption of the Y494-phosphorylation site resulted in a decrease in P Y -PTH1R signal relative to wild-type control cells (pMSCV-PTH1R BMSCs) (Fig. 6e). The length of dendrites was decreased in Y494F cells relative to wild-type controls (Fig. 6e, f). Inhibition of the actin bundling using cytochalasin D, an actin polymerization inhibitor in wild-type controls also had a decreased dendrite length relative to vehicle-treated wild-type control cells (Fig.6f, right two). These results suggest that P Y -PTH1R contributes to the outgrowth of dendrites during the transition of osteoblasts into early osteocytes. P Y -PTH1R localizes to dendrites during osteoblast-to-osteocyte transition in vivo To understand the in vivo role of PTH1R on osteoblast-toosteocyte transition, we analyzed calvarial and cortical bone in mice with Cre-mediated loss of PTH1R driven by the osteocalcin promoter (OC-Cre/PTH1R f/f ) relative to wild-type littermates (OC-Cre/PTH1R +/+ ). In calvarial bone, OC-Cre/PTH1R f/f mice had less intense fluorescence staining, decreased dendrite length, and decreased number of dendrites per osteocyte relative to OC-Cre/ PTH1R +/+ mice (Fig. 7a-d). In the femur, the number of osteocytes in femoral bone was significantly decreased in OC-Cre/PTH1R f/f mice relative to Oc-Cre/PTH1R +/+ mice (Fig. 7e). P Y -PTH1R was localized to the dendrites in wild-type femoral endocortical bone (Fig. 7f, left panel). Furthermore, PTH1R and P Y -PTH1R co-localized in osteocytes (Fig. 7f, right 3 panels) and occurred at similar frequencies (Fig. 7g). Immunofluorescence of PTH1R in femoral metaphyseal trabecular bones of osteocytes at various stages of maturity, as defined by phenotypic appearance previously described 3 , revealed positive staining on the side of osteoblasts (OB) along the bone surface (Fig. 7h, panel 1). The dendritic cytoskeletons of embedding osteoblasts (EOB) and osteoid osteocytes (OOC) also revealed positive staining for PTH1R towards the bone surface (Fig. 7h, panels 2  cell body projecting towards the bone surface (Fig. 7h, panels 4 and 5).

DISCUSSION
Osteocytes are dispersed throughout the mineralized matrix, communicating with other osteocytes and cells in the bone marrow via dendritic processes. Extension of dendritic process is thought to be key to the osteocytes role in bone remodeling and mechanosensory signaling. Factors associated with osteocyte dendrite growth, the major morphological change noted in the transition of osteoblasts to osteocytes are under investigation. Podoplanin/gp38 expression has been noted in embedded osteoid osteocytes 39,40 . While fluid-flow shear stress increases the number and length of dendrites, blockade of podoplanin/ gp38 blocks this response. Other important factors involved in eliciting and maintaining osteocyte dendrites include cytoskeleton proteins (i.e., tubulin, vimentin, actin, CapG, and destrin) [41][42][43] , enzymes, and hormones. Specifically, during the transition of osteoblasts to osteocytes, alkaline phosphatase expression is reduced while casein kinase II and osteocalcin are increased 44 . In the current study, we found that PTH1R can be phosphorylated by IGF1R. The phosphorylation mediated preferential targeting of PTH1R to the cytoskeleton, specifically, the barbed ends of actin filaments. PTH-stimulated polymerization of these actin filaments at the barbed ends and promoted the dendrite outgrowth of osteoblasts. Disruption of the phosphorylation inhibited the cytoskeletal targeting of PTH1R and led to decreased dendrite length. PTH regulates bone remodeling by orchestrating the signaling of local factors, such as TGF-β, Wnts, bone morphogenetic protein (BMP), and IGF-1 27 . PTH elicits these effects via cell membrane receptor interactions, stabilization of downstream signaling events, and gene transcription [21][22][23][24][25][35][36][37] . Our group and others have previously reported that serine residues on the PTH1R cytoplasmic tail can be phosphorylated 37,[45][46][47][48] . The serinephosphorylation promotes the internalization and desensitization responses of PTH1R 49,50 and regulates cAMP, calcium and phosphate homeostasis 37,51,52 . In the current study, we noted two tyrosines in close proximity to the serine residues. We found that a specific tyrosine residue, Y494, on the cytoplasmic domain of PTH1R can be directly phosphorylated by the IGF1R tyrosine kinase. The Y494-phosphorylation was induced by IGF1R at the cell surface. Disruption of the Y494-phosphorylation resulted in decreased actin polymerization and dendrite length.
We observed decreased PTH1R on dendritic protrusions with disruption of Y494-phosphorylation site relative to wild-type PTH1R, suggesting that the phosphorylation may serve as a sorting or packaging signal for the dispatch of PTH1R to its destiny on the actin tips. The mechanism by which P Y -PTH1R is transported from the cell membrane to the cytoskeleton is not yet fully understood. It is possible that alternative downstream signaling pathways other than PTH/cAMP or IGF-1/Akt may be involved. We also observed a preferential expression of PTH1R in osteocyte dendrites oriented towards the bone surface, implying a potential functional role of PTH1R on tips of dendrites, which warrants further investigation. Current knowledge indicates that the dynamic actin cytoskeleton plays an important role in clathrinmediated endocytosis of member receptors 53,54 . The actin network can drive the departure of endocytosed vesicles away from the cell membrane and the transport of the vesicles into intracellular compartments. At the same time the membrane receptors in the vesicles can regulate structural dynamics of the actin network. The role of axon extension in neurons, including the polymerization of the barbed ends of actin filaments generating a pushing force for the extension of the cytoplasmic processes 55,56 , provide insight into future directions to pursue this study.
It is well established that local IGF-1/IGF1R signaling is necessary for the full anabolic effect of PTH on bone 14, 18-21, 25, 26,57 . Mice with global deletion of IGF-1 are largely unresponsive to PTH 18,19 . Liver specific IGF-1 knockout mice maintain skeletal PTH responsiveness 58 , whereas osteoblast-specific deletion of IGF1R is not responsive to PTH 20 in regards to osteoblast differentiation and markers of bone formation. The effects have largely been contributed to increased IGF-1 mRNA and protein expression after PTH treatment [21][22][23][24][25] , which is supported in that IGF-1 alone is sufficient to stimulate the differentiation of MSCs/preosteoblasts to osteoblasts 59,60 and regulate the attachment of preosteoblasts onto the bone surface 9 . The effect of PTH and IGF-1 on osteoblastto-osteocyte transition has not previously been reported. We found that deletion of either IGF1R in osteoblast precursors or PTH1R in mature osteoblasts in vivo resulted in a decreased number of osteocytes and shortened dendritic processes. in vivo studies found that early osteocytes with a point mutation resulting in tyrosine to phenylalanine transcription at amino acid position 494 (Y494F) reduced the length of dendrites. Generation of a site specific Y494F mouse model would be needed to confirm these findings in vivo. Additionally, considering that the dendritic trafficking of PTH1R was only partially suppressed in Osx-Cre/ IGF1R flf mice, other tyrosine receptor kinases may also be able to phosphorylate PTH1R and mediate these effects. Other tyrosine receptors reported in osteoblasts include FGF-2 and EGF receptors. Ablation of either FGF-2 receptor or EGF receptor is also known to attenuate actions of PTH on bone [61][62][63] , suggesting that tyrosine receptor kinases targeting Y494 may not be limited to IGF1R. Finally, whether IGF-1 phosphorylation of PTH1R impacts dendrite length of mature osteocytes and function also remains to be explored.
Our data suggest that the Y494 phosphorylation of PTH1R by IGF1R plays a role in the morphological transition of osteoblasts into osteocytes during bone remodeling. Our results provide new insight into the molecular basis and pathways allowing specific functions and regulation of PTH signaling during osteoblast differentiation. We found that PTH1R is a unique substrate for IGF1R, which phosphorylates it on a site related to its preferential targeting to barbed ends of actin filaments. The phosphorylation integrates PTH and IGF-1 signaling during osteoblast-to-osteocyte transition and directs the outgrowth of dendrites. Further studies elucidating the roles of the P Y -PTH1R and understanding how it mediates the transition of osteoblasts into osteocytes are needed to better understand the process of bone remodeling and the mechanism for PTH anabolic effects on bone.

Plasmids and antibodies
The PTH1R-HA/pcDNA3.0 and IGF1R-Flag/pcDNA3.0 plasmids were cloned from human PTH1R and IGF1R. Zipper-YFP1/ pcDNA3.1 and zipper-YFP2/pcDNA3.1 have been described previously 64 . PTH1R-HA was introduced into pMSCVneo retroviral expression vector (BD Biosciences, San Jose, CA) to generate pMSCV-PTH1R-HA. PTH1R-CFP was generated by inserting PTH1R into pECFP vector (Clontech Laboratories, Mountain View, CA). IGF1R-YFP1 and PTH1R-YFP2 were generated by replacing the zippers in zipper-YFP1 or zipper-YFP2. GST-tagged cytoplasmic domain of PTH1R (amino acids 461-593, cPTH1R) was generated by fusing GST at N-terminals using pGEX-KG prokaryotic gene fusion vector (Pharmacia Corp., Pfizer Inc., New York, NY). GSTtagged cytoplasmic domain of IGF1R (cIGF1R) was purchased (Cell Signaling Technology, Danvers, MA). The mutants PTH1R-Y494F-HA and pMSCV-PTH1R-Y494F-HA were generated by primermediated PCR mutagenesis and verified via DNA sequencing. Adeno-Cre and Adeno-Cre-GFP were also purchased (Vector Biolabs, Malvern, PA). Primary antibodies used for immunoblotting, immunoprecipitation, or immunostaining included rabbit P Y -  31 , IGF1R flox/flox and Osx-Cre (Jackson Laboratory) mouse lines were used in the current study. Oc-Cre/PTH1R f/f mice were generated by crossing PTH1R flox/flox mice with Oc-Cre (referred as Oc-Cre/PTH1R +/+ ) mice. To generate Osx-Cre/IGF1R f/f mice, hemizygous Osx-Cre/IGF1R +/+ mice were crossed with IGF1R flox/flox mice to generate heterozygous IGF1R + / f offspring with or without a Cre allele, these mice were then crossed with IGF1R flox/flox mice to bring out Osx-Cre/IGF1R f/f genotype. Female Osx-Cre/IGF1R +/+ and Osx-Cre/IGF1R f/f whereas both male and female Oc-Cre/PTH1R +/+ and Oc-Cre/PTH1R f/f were used in our studies. The genotype of transgenic mice was determined by PCR analyses of genomic DNA using the primers as previously described 9,65 . The bone phenotypes of Oc-Cre/PTH1R f/f and Osx-Cre/IGF1R f/f mice have been reported previously 9,34 . For PTH effects, 4-week-old mice were randomly divided to receive either, PTH (40 μg·kg -1 ) (Bachem Bioscience Inc., Torrance, CA) or vehicle (equivalent volume of 1 mmol·L -1 acetic acid in sterile PBS) in a final volume of 100 μL given daily by subcutaneous injection for 4 weeks. Both Oc-Cre/PTH1R f/f (12.4 g ± 0.7 g) and Osx-Cre/ IGF1R f/f (16.7 g ± 1.2 g) 4-week-old mice were smaller than their wild-type littermates (19.8 g ± 0.8 g). Mice were maintained in the Animal Facility of Johns Hopkins University School of Medicine, housed in gangs in sterile, ventilated cages (Allentown Caging Equipment, Allentown, NJ), allowed free access to water and standard mouse chow, on a 14/10 h light/dark cycle. No adverse events occurred in the experimental animals. The Animal Facility complies with the Animal Welfare Act and maintains appropriate policies and procedures to ensure humane care and use of animals. The animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University, Baltimore, MD, USA.
Fluorescence staining, confocal imaging, and histomorphometric analysis Cells plated on glass coverslips were washed with cold PBS, fixed with 4% paraformaldehyde for 10 min, and then carefully permeabilized with 0.2% Triton X-100 for 6 min. The cells were then blocked with 10% BSA and incubated with primary antibodies in 2% BSA at 4°C overnight, followed by incubation with a secondary antibody for 1 h. The stained cells were mounted with Vectashield medium (Vector Laboratories, Burlingame, CA) for confocal imaging. For bone specimens, femurs were dissected from animals and fixed in 10% neutral buffered formalin for 2 days and decalcified in 10% EDTA (pH 7.4) for 2 weeks. The decalcified bones were submerged in 30% sucrose to adjust their osmotic pressure and frozen in OCT compound, then sectioned sagittally into 8-μm-thick slices. The bone slices were permeabilized and blocked by incubation in 3% BSA, 0.1% Triton-100, 0.1% NaN3 for 2 h and stained overnight at 4°C with primary antibodies, Texas red-, or FTIC-conjugated Phalloidin to stain for filamentous actin (F-actin) (1:200, Molecular Probes, Eugene, OR). For P Y -PTH1R staining, bone specimens were fixed in formalin overnight and treated with 10% EDTA for 30 min, then directly submitted for frozen sectioning. Confocal imaging was performed with an LSM510 confocal laser scanning microscopy system (Carl Zeiss, Oberkochen, Germany) at Ross Confocal Facility, Johns Hopkins University. The images were taken with 0.3-μm step size and processed four times with Kalman averaging. Three-dimensional reconstruction and the analysis of osteocyte dendrite length were conducted using software IMARIS (Bitplane, Switzerland). Additionally, quantitative histomorphometric analysis was conducted at the trabecular bone in a blinded fashion with OsteoMeasure XP software (OsteoMetrics, Inc., Decatur, GA). Four randomly selected visual fields per specimen, in three specimens per mouse in each group were measured.
Dendritic trafficking of PTH1R Confocal laser scanning images from 25-μm-thick calvarial bone slices were reconstructed by software IMARIS for estimating the number and length of dendrites. The number and length of dendrites was measured by first rotating the osteocyte into a position such that the dendrite of interest extended only in the X-Y plane using Surpass View mode of IMARIS. The number of dendrites was counted manually. The length of dendrite was measured using Polygon Type in IMARIS. The five longest dendrites (from tip to cell body) in each osteocyte in 25-μm-thick slices and 20-30 osteocytes for each bone samples were measured. The number of dendrites reaching to the bottom layer of bone surface was used to estimate the connectivity between osteocytes and bone surface. To analyze the punctuate signals of PTH1R/ P Y -PTH1R, confocal laser scanning images from cultured cells or 8-μm-thick trabecular bone slices were reconstructed. The number of punctuations was analyzed using Spot function in IMARIS. Background subtraction was turned on. Quality threshold was 15. For cultured cells, estimated diameter for spots was setup to 3 μm. For bone specimens, the diameter for spots was setup to 2 μm. Osteocytes in the depth of 20 μm from bone surface were counted to estimate the number of PTH1R punctuations. About 5-10 osteocytes in four randomly selected visual fields in three specimens per mouse were measured.

Statistical analysis
Cell culture experiments were run in triplicate. For mouse studies, five mice were analyzed per group. Statistical differences between the two groups of data were analyzed with Student's t test. Multiple comparisons were analyzed by one-way ANOVA for wildtype cells exposed to vehicle, IGF-1 and/or PTH and two-way ANOVA for wild type or knockout cells/mice exposed to either vehicle, IGF-1, or PTH with post-hoc Tukey procedure. The data are presented as mean ± s.d. All the data demonstrated a normal distribution and similar variation between groups. For all experiments, P < 0.05 were considered to be significant and indicated by *P < 0.01 were indicated by '**'. All inclusion/ IGF-I induced phosphorylation of PTH receptor enhances osteoblast to osteocyte transition Qiu et al. exclusion criteria were preestablished and no samples or animals were excluded from the analysis. The experiments were randomized. The investigators were not blinded to allocation during experiments and outcome assessment unless otherwise specified.