Conditional deletion of Pkd1 in osteocytes disrupts skeletal mechanosensing in mice

Abstract

We investigated whether polycystin-1 is a bone mechanosensor. We conditionally deleted Pkd1 in mature osteoblasts/osteocytes by crossing Dmp1-Cre with Pkd1^flox/m1Bei mice, in which the m1Bei allele is nonfunctional. We assessed in wild-type and Pkd1-deficient mice the response to mechanical loading in vivo by ulna loading and ex vivo by measuring the response of isolated osteoblasts to fluid shear stress. We found that conditional Pkd1 heterozygotes (Dmp1-Cre;Pkd1^flox/+) and null mice (Pkd1^Dmp1-cKO) exhibited a ∼40 and ∼90% decrease, respectively, in functional Pkd1 transcripts in bone. Femoral bone mineral density (12 vs. 27%), trabecular bone volume (32 vs. 48%), and cortical thickness (6 vs. 17%) were reduced proportionate to the reduction of Pkd1 gene dose, as were mineral apposition rate (MAR) and expression of Runx2-II, Osteocalcin, Dmp1, and Phex. Anabolic load-induced periosteal lamellar MAR (0.58±0.14; Pkd1^Dmp1-cKO vs. 1.68±0.34 μm/d; control) and increases in Cox-2, c-Jun, Wnt10b, Axin2, and Runx2-II gene expression were significantly attenuated in Pkd1^Dmp1-cKO mice compared with controls. Application of fluid shear stress to immortalized osteoblasts from Pkd1^null/null and Pkd1^m1Bei/m1Bei-derived osteoblasts failed to elicit the increments in cytosolic calcium observed in wild-type controls. These data indicate that polycystin-1 is essential for the anabolic response to skeletal loading in osteoblasts/osteocytes.—Xiao, Z., Dallas, M., Qiu, N., Nicolella, D., Cao, L., Johnson, M., Bonewald, L., Quarles, L. D. Conditional deletion of Pkd1 in osteocytes disrupts skeletal mechanosensing in mice.

The skeleton adapts to alterations in mechanical loading by changing bone structure. Increased load results in increased bone mass as a result of increased osteoblast-mediated bone formation (1–4), whereas skeletal unloading, which occurs with immobilization, disuse, and exposure to low gravity, leads to low bone mass from increased bone resorption (5–6). The cellular targets in bone and the molecular mechanisms that sense changes in mechanical load are uncertain. Osteocytes, which are embedded in bone matrix and have a high degree of interconnectivity, are postulated to be an important mechanosensing cell in bone (7–15). The central role of osteocytes in bone metabolism is supported by the development of osteoporosis and defective mechanosensation after the ablation of osteocytes (16), as well as by the fact that osteocytes secrete paracrine and systemic regulators affecting osteoblast-mediated bone formation and resorption, systemic phosphate and vitamin D homeostasis, and bone mineralization, such as sclerostin (Sost; refs. 17–19), dickkopf-1 (Dkk1; refs. 17, 20–22), fibroblast growth factor 23 (FGF23; refs. 23–24), and matrix extracellular phosphoglycoproteins (MEPE; ref. 25), respectively. Mature osteoblasts (26–33) and chondrocytes (34–38) also detect and respond to mechanical loading.

Many cell surface molecules (e.g., Lrp5 and ERα) and intracellular signaling pathways have been postulated to transduce the mechanosensing response in osteoblasts and osteocytes (39–45). Putative mechanosensing molecules include several G-protein-coupled receptors (e.g., prostaglandin and purinergic receptors), integrin receptors, connexins/gap-junctions, hemichannels, and stretch-activated ion channels (1–4). Although the mechanosensing function of many of these pathways has been demonstrated in cell culture models, their mechanosensation in vivo remains to be established (46). In addition, multiple autocrine and paracrine factors and signaling pathways have been implicated, including prostaglandins (e.g., PGE2), nitric oxide (NO), ATP, growth factors (e.g., TGF-β, FGF2, and IGFs), and Wnt/β-catenin signaling (15, 18, 40, 47–55); however, the molecular identity of the environmental stimuli or relevant intracellular signaling pathways involved in mechanosensing remain uncertain.

We recently reported that polycystin-1 (PC1 or Pkd1), which is part of a mechanosensing complex in renal epithelial cells (56–57), is highly expressed in osteoblasts and osteocytes and plays an important role in both skeletal development and postnatal bone homeostasis through intracellular calcium and Runx2-dependent signaling mechanisms (58–60). The function of Pkd1 as a mechanosensor in bone, however, has not been established. In the current study, to determine whether Pkd1 in osteoblasts and osteocytes has a mechanosensing function in bone, we assessed fluid flow-induced intracellular calcium response in Pkd1-null osteoblasts and the mechanical loading response of ulnae from mice in which Pkd1 was conditionally deleted in osteocytes. Using Dmp1-Cre, we demonstrated that conditional deletion of Pkd1 from osteocytes results in osteopenia and a significant decrease in the anabolic response to mechanical loading in vivo. In addition, Pkd1-null osteoblasts markedly lost their response to fluid shear stress in vitro. These results indicate that Pkd1 has a direct mechanosensing role in osteoblasts and osteocytes to regulate osteoblast function and skeletal homeostasis.

MATERIALS AND METHODS

Mice

Dentin matrix protein 1 (Dmp1)-Cre mice (61) were crossed with floxed Pkd1 mice obtained from Dr. Gregory Germino (Johns Hopkins University, Baltimore, MD, USA; ref. 62). The Pkd1^m1Bei heterozygous mice were available in our laboratory, as described previously (60). These mice were bred and maintained on a C57BL/6J background. Because Cre-recombinase-mediated deletion of single flox/m1Bei (which functions as a null allele) allele reduces the risk of mosaicism that may occur because of the <100% efficiency of Cre-recombinase to excise two floxed alleles (flox/flox) (63), we created double-heterozygous Dmp1-Cre;Pkd1^m1Bei/+ mice and homozygous Pkd1^flox/flox mice. Double-heterozygous Dmp1-Cre;Pkd1^m1Bei/+ mice were mated with homozygous Pkd1^flox/flox mice to generate excised floxed Pkd1 heterozygous (Dmp1-Cre;Pkd1^flox/+) and null mice (Dmp1-Cre;Pkd1^flox/m1Bei or Pkd1^Dmp1-cko), as well as Beier Pkd1 heterozygous mice (Pkd1^m1Bei/flox) and Dmp1-Cre-negative control mice (Pkd1^flox/+, equivalent to wild type). These mice were used for skeletal phenotype analysis and mechanical loading experiments. All animal research was conducted according to guidelines provided by the U.S. National Institutes of Health and the Institute of Laboratory Animal Resources, National Research Council. The University of Tennessee Health Science Center (protocol 1889) and University of Kansas Medical Center (protocol 2007-1630) Animal Care and Use Committees approved all animal studies.

Genotyping PCR and real-time PCR to detect mutations and deletions

Genomic DNA was prepared from tail and other tissue specimens by using standard procedures (64). PCR genotyping was performed using the following primers (Fig. 1A; ref. 62): F1, 5′-CTTCTATCGCCTTCTTGACGAGTTC-3′; R1, 5′-AGGGCTTTTCTTGCTGGTCT-3′; and R2, 5′-TCGTGTTCCCTTACCAACCCTC-3′. Pkd1 floxed (Pkd1^flox) alleles were identified in 2% agarose gels as 670-bp bands (Fig. 1B). The Δ floxed Pkd1 (Pkd1^Δflox) allele was detected as a 0.85-kb band in 1% agarose gels (Fig. 1B; ref. 58). The Pkd1^m1Bei allele was genotyped using SYBR Green (Bio-Rad, Hercules, CA, USA) real-time PCR, as described previously (60).

Figure 1.

Dmp1-Cre-mediated conditional deletion of Pkd1 from the floxed Pkd1^flox allele in different tissues. A) Schematic illustration of wild-type Pkd1⁺, mutant Pkd1^m1Bei, and floxed Pkd1 allele before (Pkd1^flox) and after (Pkd1^Δflox) deletion of the lox P cassette containing exons 2–4 via Cre-mediated recombination. Slashes indicate all the introns and exons omitted between exons 5 and 25. B) Genotype PCR analysis of different tissues that were harvested from 16-wk-old Dmp1-Cre;Pkd1^flox/m1Bei mice. Pkd1 Beier and floxed alleles existed in all tested tissues, including bone. However, Dmp1-Cre-mediated recombination of excised floxed Pkd1^Δflox allele occurred in bone tissues, such as calvarias and femur, but also had a leakage in the brain, muscle, and intestine. C) Real-time RT-PCR analysis of total Pkd1 transcripts. Expression of total Pkd1 transcripts was performed using Pkd1-allele-specific primers, as described in Materials and Methods. Normal Pkd1⁺ vs. cyclophilin A is normalized to the mean ratio of 5 control mice, set to 1. Percentage of conditional deleted and mutant transcripts was calculated from the relative levels of the normal Pkd1⁺ transcripts in different Pkd1 exons. Data are expressed as the percentage of wild-type Pkd1⁺ and Pkd1^flox, mutant Pkd1^m1Bei, and conditionally deleted Pkd1^Δflox allele expressions in Pkd1^flox/+ control and Dmp1-Cre;Pkd1^flox/− conditionally null mice. D) Histology of adult kidney. Hematoxylin-and-eosin (H&E)-stained sections from 6-wk-old mice failed to identify any cystic tubules in either cortical or medullary regions of kidney from Dmp1-Cre;Pkd1^flox/+ or Pkd1^Dmp1-cKO mice, consistent with the absence of Dmp1-Cre expression in the kidney. In contrast, ablation of Pkd1 in the kidney of 6-wk old Col1a1(3.6)-Cre;Pkd1^flox/flox caused massive cyst formation, which served as a positive control. Cy, cyst. Scale bars = 100 μm.

Bone densitometry and histomorphometric and microcomputed tomography analyses

Bone mineral density (BMD) of femurs was assessed at 6 and 16 wk of age with a Lunar Piximus bone densitometer (Lunar Corp., Madison, WI, USA). Calcein (Sigma, St. Louis, MO, USA) double labeling of bone and histomorphometric analyses of periosteal mineral apposition rate (MAR) in tibiae were performed using the osteomeasure analysis system (OsteoMetrics, Decatur, GA, USA) (64–65). The distal femoral metaphyses were also scanned with a Scanco μCT 40 (Scanco Medical AG, Brüttisellen, Switzerland). Three-dimensional images were analyzed to determine bone volume/trabecular volume (BV/TV) and cortical thickness (Ct.Th), as described previously (64).

Biomechanical testing

On the day of testing, the femurs were thawed and rehydrated with 1× PBS for 3 h. Bending tests were performed using a 3-point fixture on an electromechanical testing system (ELF 3200, EnduraTEC, Inc., Minnetonka, MN, USA). The bones were flexed in the anterior-posterior plane by displacing the loading point fixture at 5 mm/min until failure (66). Bending force-deflection curves were constructed and analyzed for stiffness (S) and maximum force to failure (F_max). The flexural rigidity (EI) was calculated from the classical equations of beam theory for a simply supported beam with a central concentrated load.

Serum biochemistry

Serum osteocalcin levels were measured using a mouse osteocalcin EIA kit (Biomedical Technologies, Stoughton, MA, USA). Serum urea nitrogen (BUN) was determined using a BUN diagnostic kit (Pointe Scientific, Lincoln Park, MI, USA). 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 Rank ligand (RankL) were measured using mouse ELISA kits (Quantikine; R&D Systems, Minneapolis, MN, USA), and serum tartrate-resistant acid phosphatase (TRAP) was assayed with the ELISA-based SBA Sciences mouse TRAP assay (Immunodiagnostic Systems, Fountain Hills, AZ, USA).

Intracellular calcium ([Ca²⁺]_i) measurements in vitro

Calvaria from wild-type (Pkd1^+/+), heterozygous Pkd1^null/+ and Pkd1^m1Bei/+, and homozygous Pkd1^null/null and Pkd1^m1Bei/m1Bei embryos at embryonic day 15.5 (E15.5) were used to isolate osteoblasts by sequential collagenase digestion. To engineer immortal osteoblast cell lines, isolated primary osteoblasts were infected using a retroviral vector carrying SV40 large and small T antigen, as described previously (60). For flow experiments, the immortalized cells were cultured on type I rat tail collagen-coated 40-mm diameter glass slides at 80–90% confluency in α-minimal essential medium (α-MEM) containing 2% FBS and 1% penicillin and streptomycin (P/S) for 3 d. Cells were rinsed 2 times with Hanks' balanced saline solution (HBSS). The cells were then loaded with 3 μM Fura-2-AM (Molecular Probes, Eugene, OR, USA), a fluorescent Ca²⁺ probe, in HBSS for 30 min at 37°C. Loaded cells were incubated for an additional 45 min with HBSS alone to ensure complete deesterification of the fluorescent molecule. A glass slide was then placed in an FCS2 parallel plate flow chamber (Bioptechs, Butler, PA, USA), 0.25 × 14 × 22 mm. A fresh bolus of flow medium was added to the chamber, and the cells were left undisturbed for 30 min. The flow medium consisted of phenol-free α-MEM and 2% FBS equilibrated with 5% CO₂/95% air at 37°C. The chamber was mounted on the stage of an inverted microscope with CCD camera to allow real-time record of fluorescence intensity (F340/F380 ratio) to generate ratiometric video images of individual static cells or cells exposed to flow (Intracellular Imaging, Cincinnati, OH, USA). The F340/F380 ratios were converted to concentration using standard calibration curve. Baseline levels of [Ca²⁺]_i were obtained for 3 min before exposing to flow. The slides of cells were exposed to 6.24 dyn/cm² pulsatile laminar fluid flow for 3 min with a peristaltic pump (Bioptechs). A ratiometric video image analysis at individual cells was used to determine changes in [Ca²⁺]_i.

Ulna loading experiments in vivo

The forearm compression ulna loading was performed starting at 16 wk of age in male control Pkd1^flox/+ mice, single Pkd1^flox/m1Bei heterozygous mice, single Dmp1-Cre;Pkd1^flox/+ heterozygous mice, and Dmp1-Cre;Pkd1^flox/m1Bei conditionally null mice, as described previously (40, 48, 67–71). Briefly, each animal was anesthetized with 4% isoflurane inhalation, and the right forearm was positioned between two brass cups and loaded in compression with a mechanical testing machine (ElectroForce 3200; Bose Corp., Eden Prairie, MN, USA). The right ulnae were loaded at −3.0 N, haversine waveform, 2 Hz, 180 cycles, on Mondays, Wednesdays, and Fridays for 2 wk. To measure strain-induced new bone formation, calcein (20 mg/kg) was injected into the mice 12 and 3 d before sacrifice. The ulnae and radii were removed for further analysis of bone formation rate and bone histomorphometry. The femurs were dissected and cleaned from muscle and soft tissue attachments and then frozen at −20°C for biomechanical property testing.

RNA isolation and real-time RT-PCR

For quantitative real-time RT-PCR, 2.0 μg total RNA isolated from whole tibiae of 6-wk-old mice, no load ulnae of 16-wk-old mice, whole loaded and no load ulnae of 18-wk-old mice, and 10-d cultured primary osteoblasts in differentiation medium was reverse transcribed, as described previously (72). PCR reactions contained 100 ng template (cDNA or RNA), 300 nM each forward and reverse primers, and 1× iQ SYBR Green Supermix (Bio-Rad) in 50 μl. The threshold cycle (C_t) of tested gene product from the indicated genotype was normalized to the C_t for cyclophilin A. Expression of total Pkd1 transcripts was performed using the following Pkd1-allele-specific primers: in exon 26, forward primer of normal Pkd1⁺ transcript: 5′-CTGGTGACCTATGTGGTCAT-3′, forward primer of mutant Pkd1^m1Bei transcript: 5′-CTGGTGACCTATGTGGTCAG-3′, and common reverse primer: 5′-AGCCGGTCTTAACAAGTATTTC-3′; in exon 2–4, forward primer of normal Pkd1⁺ transcript: 5′-ATAGGGCTCCTGGTGAACCT-3′ and reverse primer: 5′-CCACAGTTGCACTCAAATGG-3′. The normal Pkd1⁺ vs. cyclophilin A was normalized to the mean ratio of 5 control mice, which was set to 1. The percentage of conditionally deleted and mutant transcripts was calculated from the relative levels of the normal Pkd1⁺ transcripts in different Pkd1 exons (73). All primer sequences of other genes used in real-time RT-PCR are provided in Table 1.

Table 1.

Primer sequences used in real-time RT-PCR

Gene	Accession no.	Forward primer	Reverse primer
Runx2-II	NM_009820	5′-GCCTCACAAACAACCACAGA-3′	5′-TTAAACGCCAGAGCCTTCTT-3′
Runx2-I	D14636	5′-TCGCTAACTTGTGGCTGTTG-3′	5′-GCTCACGTCGCTCATCTTG-3′
Runx2	NM_009820	5′-TCTGGCCTTCCTCTCTCAG-3′	5′-GGATGAAATGCTTGGGAAC-3′
Osteocalcin	NM_007541	5′-AGCAGGAGGGCAATAAGGTA-3′	5′-CAAGCAGGGTTAAGCTCACA-3′
Osteopontin	AF515708	5′-TCTGATGAGACCGTCACTGC-3′	5′-CCTCAGTCCATAAGCCAAGC-3′
Bsp	NM_008318	5′-GCCTCAGTTGAATAAACATGAAA-3′	5′-TCCTCACCCTTCAATTAAATCCCACAA-3′
Opg	MMU94331	5′-GTTCCTGCACAGCTTCACAA-3′	5′-AAACAGCCCAGTGACCATTC-3′
Rank ligand	NM_011613	5′-GCAGAAGGAACTGCAACACA-3′	5′-TGATGGTGAGGTGTGCAAAT-3′
Mmp13	NM_008607	5′-AGTTGACAGGCTCCGAGAAA-3′	5′-GGCACTCCACATCTTGGTTT-3′
Dmp1	MMU242625	5′-AGTGAGGAGGACAGCCTGAA-3′	5′-GAGGCTCTCGTTGGACTCAC-3′
Phex	NM_011077	5′-CGCCTGACAAACTTTTGAGACC-3′	5′-TGCTCCCTGTTTCTGCTTCC-3′
Sost	NM_024449	5′-AAGCCGGTCACCGAGTTGGT-3′	5′-GTGAGGCGCTTGCACTTGCA-3′
Trap	NM_007388	5′-AACACCACGAGAGTCCTGCT-3′	5′-GTACCAGGGCAGAGAAGCTG-3′
Mmp9	NM_013599	5′-AGTTGCCCCTACTGGAAGGT-3′	5′-GTGGATAGCTCGGTGGTGTT-3′
Collagen II	NM_031163	5′-TGGCTTCCACTTCAGCTATG-3′	5′-AGGTAGGCGATGCTGTTCTT-3′
VegfA	NM_009505	5′-GGGTGCACTGGACCCTGGGTTTAC-3′	5′-CCTGGCTCACCGCCTTGGCTTGTC-3′
PPARγ	NM_009505	5′-CAAGGTGCTCCAGAAGATGA-3′	5′-AGTAGCTGCACGTGCTCTGT-3′
aP2	NM_024406	5′-GCGTGGAATTCGATGAAATCA-3′	5′-CCCGCCATCTAGGGTTATGA-3′
Cyclophilin A	NM_008907	5′-CTGCACTGCCAAGACTGAAT-3′	5′-CCACAATGTTCATGCCTTCT-3′

Strain-gauge testing

The right arms were isolated from male Pkd1^flox/+ control and Dmp1-Cre;Pkd1^flox/m1Bei conditionally null mice at 16 wk of age. Loading strain was determined using a single-element strain gauge (EA-06–015-DJ-120; Vishay Intertechnology, Malvern, PA, USA) attached to the medial surface of the ulnar midshaft, as described previously (67). After applying strain gauges, loading was conducted on the intact right forearm with 1.0-, 1.5-, 2.0-, 2.5-, 3.0-, and 3.5-N compressive forces in a haversine waveform at 2 Hz for 15 cycles. Strains generated from ulnae in the last 5 cycles were averaged.

Primary osteoblast culture for proliferation and differentiation and gene expression profiles

Primary osteoblasts were isolated from the newborn mouse calvarias by sequential collagenase digestion at 37°C, as described previously (60, 72). The cells were cultured in α-MEM containing 10% FBS and 1% P/S. Cell proliferation was detected by BrdU incorporation assays following the manufacturer's directions (QIA58; Calbiochem, Gibbstown, NJ, USA). To induce differentiation, primary osteoblasts were plated at a density of 2 × 10⁴ cells/well in a 12-well plate, and 4 × 10⁴ cells/well in a 6-well plate and grown up to 21 d in α-MEM containing 10% FBS supplemented with 5 mM β-glycerophosphate and 25 μg/ml ascorbic acid. Alkaline phosphatase activity and Alizarin red-S histochemical staining for mineralization were performed as described previously (60, 72). Total DNA content was measured with a PicoGreen dsDNA quantitation reagent and kit (Molecular Probes, Eugene, OR, USA). Protein concentrations of the supernatant were determined with a protein assay kit (Bio-Rad). For gene expression profiles, 2.0 μg of total RNA was isolated from primary osteoblasts cultured 4, 14, and 21 d in differentiation medium. The cDNAs were generated using a Reverse Transcriptase Kit (Perkin-Elmer, Foster City, CA, USA). PCR reactions contained 100 ng template (cRNA or cDNA), 200 nmol each forward and reverse primers, 1× iQ SYBR Green Supermix (Bio-Rad) in 50 μl. The C_t of tested gene product from the indicated genotype was normalized to the C_t for cyclophilin A, as described previously (59–60, 72).

Immunofluorescence

Primary osteoblasts were grown on collagen-coated 4-well chambers at 1 × 10⁵ cells/well and kept at confluence for ≥3 d. At the end of the culture, the cells were washed 3 times with PBS, fixed with cold 4% paraformaldehyde/0.2% Triton for 10 min at room temperature, and washed with PBS 3 times. The cells were incubated for 30 min in 1% BSA before incubation with primary acetylated α-tubulin antibody (1:4000, Sigma Aldrich, T6793) for 1 h at room temperature. After washing 3 times in PBS, they were treated with secondary Texas Red-labeled anti-mouse IgG (715-076-150; Jackson ImmunoResearch, Bar Harbor, ME, USA) in 1% BSA for 1 h at room temperature and washed 3 times in PBS before mounting with ProLong Gold antifade reagent (P36935; Invitrogen, Carlsbad, CA, USA). Nuclei were counterstained with DAPI blue. Photographs were taken under a microscope at × 40 for counting the number of primary cilia in cultured primary osteoblasts, as described previously (60).

Analysis

We evaluated differences between 2 groups by unpaired t test and multiple groups by 1-way ANOVA. All values are expressed as means ± sd. All computations were performed using GraphPad Prism5 (GraphPad Software, La Jolla, CA, USA).

RESULTS

Dmp1-Cre-mediated conditional deletion of Pkd1 in different tissues

We characterized 4 genotypes, including conditional Dmp1-Cre;Pkd1^flox/m1Bei-null mice (hereafter designated Pkd1^Dmp1-cKO), single conditional Dmp1-Cre;Pkd1^flox/+ heterozygous mice, single Pkd1^flox/m1Bei heterozygous mice, and control Pkd1^flox/+ mice. These mice were born at the expected Mendelian frequency, and Pkd1^Dmp1-cKO, Dmp1-Cre;Pkd1^flox/+, and Pkd1^flox/m1Bei exhibited survival indistinguishable from control mice. The normal survival of Pkd1^Dmp1-cKO mice contrasted with the perinatal lethality of homozygous Pkd1^m1Bei/m1Bei mice (74). To establish the conditional loss of Pkd1 from bone, we performed PCR analysis by using a combination of primers that specifically detected floxed Pkd1^flox alleles and the excised floxed Pkd1^Δflox alleles in Dmp1-Cre;Pkd1^flox/m1Bei mice (Fig. 1A). We found that Dmp1-Cre-mediated excision of the Pkd1^Δflox allele occurred in bone tissues, including calvaria and femur, but also occurred in brain, muscle, and intestine, indicating that the Dmp1-Cre promoter is not specific for bone (Fig. 1B). We found no evidence for the Pkd1^Δflox allele in other tissues. Both the floxed Pkd1^flox and mutant Pkd1^m1Bei alleles were detected in all tested tissues (Fig. 1B).

To quantify the effect of combined use of floxed Pkd1^flox allele with functional Pkd1^m1Bei-null allele to increase the net efficiency of Pkd1 inactivation by Cre-recombinase, we examined the percentage of Pkd1 conditionally deleted and Pkd1^m1Bei mutant allele expressions in bone tissues from 6-wk-old mice by real-time RT-PCR. As expected, both Pkd1^flox/m1Bei and Dmp1-Cre;Pkd1^flox/m1Bei mice expressed 50% Pkd1^m1Bei mutant (functional null) allele, whereas Dmp1-Cre;Pkd1^flox/+ and Dmp1-Cre;Pkd1^flox/m1Bei mice exhibited ∼40% excision of the floxed exon 2–4 from Pkd1, likely reflecting the expression of Dmp1-Cre mature osteoblasts just entering the osteoid, osteoid osteocytes, and mature osteocytes (61). The combined effect of Pkd1^m1Bei and Pkd1^Δflox in Dmp1-Cre;Pkd1^flox/m1Bei resulted in a net reduction of Pkd1 expression by ∼90% in bone (Fig. 1C). However, Dmp1-Cre-mediated conditional deletion of Pkd1 did not alter the appearance of primary cilia in cultured osteoblasts (data not shown). In addition, real-time RT-PCR to assess the expression level of the residual functional Pkd1 transcript confirmed the progressive reduction of Pkd1 message in conditional mutant mice, i.e., Pkd1^flox/+ (100%), Dmp1-Cre;Pkd1^flox/+ (60%), Pkd1^flox/m1Bei (50%), and Pkd1^Dmp1-cKO (10%) mice (data not shown). Consistent with the lack of Cre expression in the kidney, Pkd1^Dmp1-cKO, Dmp1-Cre;Pkd1^flox/+, and mutant Pkd1^m1Bei mice demonstrated no cyst formation in the kidney, whereas use of a bone-restricted promoter Cre with kidney expression (Col1a3.6-Cre) resulted in the development of polycystic kidney disease in 6-wk-old Col1a3.6-Cre;Pkd1^{flox/flox mice} (Fig. 1D).

Additive effects of global mutant (Pkd1^m1Bei) and Dmp1-Cre-mediated conditionally deleted (Pkd1^Δflox) Pkd1 alleles in postnatal bone

We observed no change in body weight, lean body mass, or fat mass in Pkd1^Dmp1-cKO compared to controls (data not shown). In contrast, we observed a reduction of BMD of 9–12% in both female and male Pkd1^m1Bei/flox and Dmp1-Cre;Pkd1^flox/+ heterozygous mice at 6 wk of age and a greater reduction in BMD of 19–27% in Pkd1^Dmp1-cKO mice compared with age-matched Pkd1^flox/+ control mice (Fig. 2A). Micro-CT analysis revealed that the lower bone mass in single Pkd1^m1Bei/flox and Dmp1-Cre;Pkd1^flox/+ heterozygous mice was caused by a reduction in TV (37 and 32%, respectively) and cortical bone thickness (10 and 6%, respectively) (Fig. 2B). Pkd1^Dmp1-cKO had greater loss in both trabecular (48%) and cortical bone (17%) (Fig. 2B), indicating that the conditional deletion of Pkd1 on the Pkd1^m1Bei background resulted in additional bone loss. These reductions in BV were associated with a significant decrease in MAR in single Pkd1^m1Bei/flox and Dmp1-Cre; Pkd1^flox/+ heterozygous mice compared with age-matched control mice and an even greater reduction in Pkd1^Dmp1-cKO mice (Fig. 2C).

Figure 2.

Dmp1-Cre-mediated somatic loss of Pkd1 leads to osteopenia. A) Effects of Dmp1-Cre-mediated Pkd1^Δflox allele on BMD at 6 wk of age. Similar to Beier Pkd1 heterozygous Pkd1^m1Bei/flox mice, there was ∼9–12% reduction of BMD in both female and male single-excised floxed Dmp1-Cre;Pkd1^flox/+ mice compared with age-matched Pkd1^flox/+ control mice, and an even greater reduction (19–27%) in double-heterozygous Dmp1-Cre;Pkd1^flox/m1Bei (Pkd1^Dmp1-cKO) mice, indicating an additive effect of global mutant and conditional deleted Pkd1 alleles on loss of bone mass. B) Effects of Dmp1-Cre-mediated Pkd1^Δflox allele on bone structure of femurs. Micro-CT analysis of the distal femoral metaphyses and midshaft diaphyses revealed that double-heterozygous Pkd1^Dmp1-cKO mice had greater loss in both trabecular and cortical bone than did single Dmp1-Cre;Pkd1^flox/+ and Pkd1^m1Bei/flox heterozygous mice, consistent with additive effects of global mutant and conditionally deleted Pkd1 alleles on bone structure and a direct role of Pkd1 in bone in osteocytes. C) Effects of Dmp1-Cre-mediated Pkd1^Δflox allele on bone MAR. There was a significant decrease in MAR in Pkd1^m1Bei/flox and Dmp1-Cre;Pkd1^flox/+ mice compared with age-matched control mice and an even greater reduction in Dmp1-Cre;Pkd1^flox/m1Bei mice, indicating an additive effect of global mutant and conditionally deleted Pkd1 alleles to impair osteoblast-mediated bone formation. Data represent means ± sd from 5–6 individual mice. *P < 0.05 vs. control Pkd1^flox/+; ^#P < 0.05 vs. Dmp1-Cre;Pkd1^flox/+ and Pkd1^flox/m1Bei.

To investigate whether combined Pkd1^Δflox and Pkd1^m1Bei deficiency resulted in additive effects on gene expression profiles in bone, we examined by real-time RT-PCR the expression levels of a panel of osteoblast lineage-, osteoclast-, and chondrocyte-related mRNAs from the tibiae of 6-wk-old control, Dmp1-Cre;Pkd1^flox/+, Pkd1^flox/m1Bei, and Dmp1-Cre;Pkd1^flox/m1Bei mice (Table 2). Bone derived from heterozygous Dmp1-Cre;Pkd1^flox/+ and Pkd1^flox/m1Bei mice had measurable reductions in the osteoblast-lineage gene transcripts, including Runx2-II, total Runx2, Osteocalcin (Oc), Osteopontin, Bsp, Osteoprotegerin (Opg), Mmp13, Dmp1, and Phex mRNA levels, compared to control mice. Significantly greater reductions of Runx2-II, total Runx2, Oc, Osteopontin, Bsp, RankL, Mmp13, Dmp1, and Phex were observed in Pkd1^Dmp1-cKO mice. In this regard, the Opg/RankL expression ratio was increased in a gene dose-dependent manner (Table 2). Consistent with a ratio of Opg/RankL that favored the reduced osteoclastogenesis, bone expression of Trap and Mmp9, markers of bone resorption, were also reduced in heterozygous Pkd1-deficient mice and to a greater extent in Pkd1^Dmp1-cKO mice (Table 2). In contrast, the transcription of Sost/sclerostin, an osteocyte-derived negative regulator of bone formation, was significantly increased in heterozygous Pkd1-deficient and Pkd1^Dmp1-cKO mice compared to control mice, and transcripts of chondrocyte-related genes did not differ between heterozygous Pkd1-deficient and Pkd1^Dmp1-cKO mice (Table 2). In addition, PPARγ, an adipocyte transcription factor, and adipocyte markers, such as adipocyte fatty acid-binding protein 2 (aP2) were also increased in tibiae of Pkd1-deficient mice in a Pkd1 gene dosage-related manner (Table 2).

Table 2.

Gene-expression profiles in 6-wk-old mice

Gene	Accession no.	Dmp1-Cre; Pkd1flox/+	Pkd1flox/m1Bei	Dmp1-Cre; Pkd1flox/m1Bei	P value
Osteoblast lineage
Runx2-II	NM_009820	0.77 ± 0.07*	0.75 ± 0.06*	0.52 ± 0.17*,#	0.0007
Runx2-I	D14636	1.06 ± 0.63	0.96 ± 0.31	0.94 ± 0.17	0.9504
Runx2	NM_009820	0.73 ± 0.08*	0.69 ± 0.13*	0.47 ± 0.15*,#	<0.0001
Osteocalcin	NM_007541	0.69 ± 0.12*	0.66 ± 0.11*	0.41 ± 0.11*,#	0.0002
Osteopontin	AF515708	0.72 ± 0.14*	0.69 ± 0.16*	0.39 ± 0.19*,#	0.0001
Bsp	NM_008318	0.70 ± 0.12*	0.71 ± 0.07*	0.56 ± 0.11*,#	<0.0001
Opg	MMU94331	0.75 ± 0.12*	0.72 ± 0.13*	0.65 ± 0.12*	0.0200
Rank ligand	NM_011613	0.70 ± 0.14*	0.71 ± 0.13*	0.41 ± 0.17*,#	0.0003
Mmp13	NM_008607	0.66 ± 0.14*	0.61 ± 0.12*	0.37 ± 0.16*,#	<0.0001
Dmp1	MMU242625	0.63 ± 0.17*	0.60 ± 0.16*	0.29 ± 0.05*,#	0.0004
Phex	NM_011077	0.67 ± 0.14*	0.68 ± 0.05*	0.54 ± 0.11*,#	<0.0001
Sost	NM_024449	1.62 ± 0.47*	1.67 ± 0.11*	1.64 ± 0.31*	0.0013
Osteoclast
Trap	NM_007388	0.65 ± 0.16*	0.63 ± 0.21*	0.31 ± 0.12*,#	0.0004
Mmp9	NM_013599	0.71 ± 0.11*	0.77 ± 0.09*	0.33 ± 0.09*	<0.0001
Chondrocyte
Collagen II	NM_031163	1.28 ± 0.89	1.30 ± 0.85	0.97 ± 0.38	0.7666
VegfA	NM_009505	1.02 ± 0.55	1.01 ± 0.29	1.19 ± 0.62	0.8889
Adipocyte
PPARγ	NM_009505	1.41 ± 0.35*	1.56 ± 0.38*	1.76 ± 0.41*	0.00168
aP2	NM_024406	1.46 ± 0.16*	1.57 ± 0.23*	1.95 ± 0.26*,#	0.0058

Data are means ± sd from 5–6 tibias of 6-wk-old individual mice and expressed as x-fold change relative to the housekeeping gene cyclophilin A, subsequently normalized to control (Pkd1^flox/+) mice.

^*P < 0.05 vs. Pkd1^flox/+;

^#P < 0.05 vs. Dmp1-Cre;Pkd1^flox/+ and Pkd1^flox/m1Bei.

Changes in gene expression in bone correlated with alterations in serum biomarkers. In this regard, control Pkd1^flox/+ mice had reduced osteoblastic and osteoclastic markers as a function of age, consistent with an age-dependent decrease in bone formation and resorption (Table 3). At 6 wk of age, Pkd1^Dmp1-cKO mice had a further reduction in both osteoblast and osteoclast markers, evidenced by decreased osteocalcin (106±12 vs. 39±7 ng/ml), RankL (179±29 vs. 115±28 pg/ml), and TRAP (3.4±0.73 vs. 2.0±0.29 U/L) compared to age-matched control Pkd1^flox/+ mice. At 16 wk of age, a reduction in osteocalcin was not observed in Pkd1^Dmp1-cKO mice, but decreases in RankL (110±14 vs. 91±7 pg/ml), and TRAP (2.7±0.34 vs. 2.1±0.42 U/L) remains the same compared to age-matched control Pkd1^flox/+ mice. In addition, the RankL/OPG ratio, which is an indicator of osteoclastogenesis, was reduced by ∼43% at 6 wk of age and 6% at 16 wk of age in Pkd1^Dmp1-cKO mice compared to age-matched control Pkd1^flox/+ mice, consistent with the reduction in osteoclastic markers (Table 3). These data suggest that the conditional loss of Pkd1 in osteocytes results in diminished osteoblast and osteoclast function and consequent low-turnover osteopenia. In contrast, Pkd1^Dmp1-cKO mice had no effect on serum calcium and phosphorus levels at either 6 or 16 wk (Table 3).

Table 3.

Biochemistry analysis of serum in 6- and 16-wk-old mice

Serum value	Pkd1flox/+	Pkd1Dmp1−cKO
BUN (mg/dl)
6 wk	31 ± 4.4	30 ± 4.8
16 wk	29 ± 5.1	28 ± 3.5
Ca (mg/dl)
6 wk	8.9 ± 0.26	9.1 ± 0.27
16 wk	8.7 ± 0.39	9.0 ± 0.22
P (mg/dl)
6 wk	8.1 ± 0.33	8.0 ± 0.48
16 wk	7.6 ± 0.61	7.2 ± 0.34
Osteocalcin (ng/ml)
6 wk	106 ± 12	39 ± 7*
16 wk	14.4 ± 5.1	12.3 ± 5.5
OPG (ng/ml)
6 wk	2.4 ± 0.33	2.7 ± 0.53
16 wk	2.5 ± 0.34	2.2 ± 0.42
RankL (pg/ml)
6 wk	179 ± 29	115 ± 28*
16 wk	110 ± 14	91 ± 7*
TRAP (U/L)
6 wk	3.4 ± 0.73	2.0 ± 0.29*
16 wk	2.7 ± 0.34	2.1 ± 0.42*

Data are means ± sd from 5 to 6 individual mice. Osteocalcin is produced by osteoblasts; TRAP is produced by osteoclasts.

^*P < 0.05 vs. Pkd1^flox/+.

Age-dependent effects of global Pkd1^m1Bei mutant and Dmp1-Cre-mediated Pkd1^Δflox alleles on bone mass, structure, geometry, and mechanical properties

Compared with control mice, there was an age-dependent partial recovery of BMD from 27, 16, and 12% reduction at 6 wk of age to 10, 6, and 4% reduction at 16 wk of age in Pkd1^Dmp1-cKO, Pkd1^flox/m1Bei, and Dmp1-Cre;Pkd1^flox/+ mice, respectively, indicating age-dependent effects that attenuate the effects of mutant Pkd1on bone mass (Fig. 3A). Micro-CT analysis revealed that the increase in bone mass was caused by a recovery in cortical bone thickness. Indeed, the differences in cortical bone thickness observed at 6 wk of age were no longer significant in the 4 genotypes at 16 wk of age (Fig. 3B). In contrast, there was less recovery of BV/TV with age. In this regard, BV/TV remained significantly lower in single Pkd1^flox/m1Bei and Dmp1-Cre;Pkd1^flox/+ heterozygous mice, as well as in Pkd1^Dmp1-cKO mice compared to control mice at 16 wk of age (Fig. 3C), suggesting a site-specific interaction between Pkd1 mutations and age-dependent changes in bone structure. Moreover, we found that the mechanism of cortical bone recovery resulted from alterations in bone geometry and led to a compensatory increase in bone mechanical properties in 16-wk-old of Pkd1^Dmp1-cKO mice. In this regard, single Pkd1^flox/m1Bei heterozygous and conditional Pkd1^Dmp1-cKO null mice showed a gene dose-dependent decrease in total bone area but no difference in cortical bone area when compared with age-matched control Pkd1^flox/+ mice (Table 4), consistent with a reduction in the size of the marrow cavity resulting from a smaller midshaft diameter compared with control mice. However, there was a significant reduction in moment of inertia I_x and distance c (distance from the neutral axis to the plane where the load is applied) in Pkd1^Dmp1-cKO mice, but no differences were observed in Pkd1^flox/m1Bei and Dmp1-Cre;Pkd1^flox/+ mice compared to control mice (Table 4). To examine whether changes of femoral bone geometry may affect bone mechanical properties, we used these femurs to perform 3-point bending experiments (Table 5). Compared with the control mice, Pkd1^Dmp1-cKO mice had a higher maximum stress only in 3-point bending, but no significant differences in bending stiffness, maximum force, or energy to failure (Table 5), indicating that the changes of bone geometry and bone structure at 16 wk of age may preserve bone strength in Pkd1^Dmp1-cKO mice. There was no difference in these parameters between single heterozygous mice and control mice (Table 5).

Figure 3.

Age-dependent effects of Dmp1-Cre-mediated Pkd1^Δflox allele on bone mass. A) Age-dependent effects of Dmp1-Cre-mediated Pkd1^Δflox allele on femoral BMD. B) Age-dependent effects of Dmp1-Cre-mediated Pkd1^Δflox allele on TV of distal femoral metaphyses. C) Age-dependent effects of Dmp1-Cre-mediated Pkd1^Δflox allele on Ct.Th of femoral midshaft diaphyses. Compared with control mice, there was an age-dependent partial recovery of BMD in Dmp1-Cre;Pkd1^flox/+, Pkd1^flox/m1Bei, and Pkd1^Dmp1-cKO mice, respectively. Micro-CT analysis revealed that the increase in bone mass was caused by a recovery in cortical bone thickness, but BV/TV remained significantly lower in these Pkd1-deficient mice at 16 wk of age. Data represent means ± sd from 5 or 6 individual mice. *P < 0.05 vs. control Pkd1^flox/+; ^#P < 0.05 vs. Dmp1-Cre;Pkd1^flox/+ and Pkd1^flox/m1Bei.

Table 4.

Femur bone geometry in 18-wk-old mice

Parameter	Pkd1flox/+	Dmp1-Cre; Pkd1flox/+	Pkd1flox/m1Bei	Dmp1-Cre; Pkd1flox/m1Bei
Total area (mm2)	2.1 ± 0.16	2.2 ± 0.14	1.9 ± 0.11*	1.7 ± 0.13*,#
Cortical area (mm2)	1.03 ± 0.09	1.08 ± 0.08	0.99 ± 0.05	0.98 ± 0.08
Moment of inertia, Ix (mm4)	0.19 ± 0.03	0.21 ± 0.02	0.17 ± 0.02	0.14 ± 0.02*,#
c (mm)	0.71 ± 0.05	0.70 ± 0.04	0.68 ± 0.04	0.63 ± 0.03*,#

c, distance from the neutral axis to the plane where the load is applied. Data are means ± sd from 5 to 6 individual mice.

^*P < 0.05 vs. Pkd1^flox/+ and Dmp1-Cre;Pkd1^flox/+;

^#P < 0.05 vs. Pkd1^flox/m1Bei.

Table 5.

Femur biomechanical properties in 18-wk-old mice

Parameter	Pkd1flox/+	Dmp1-Cre; Pkd1flox/+	Pkd1flox/m1Bei	Dmp1-Cre; Pkd1flox/m1Bei
Stiffness (N/mm)	210 ± 35	174 ± 34	203 ± 41	179 ± 46
Maximum force (N)	22 ± 3.8	23 ± 3.9	23 ± 2.9	22 ± 3.0
Maximum stress (Mpa)	108 ± 16	109 ± 19	122 ± 17	132 ± 14*
Energy to failure (N-mm)	9.7 ± 2.1	9.2 ± 1.9	8.4 ± 1.6	8.1 ± 2.1

Data are means ± sd from 5 to 6 individual mice.

^*P < 0.05 vs. Pkd1^flox/+ and Dmp1-Cre;Pkd1^flox/+.

Effect of conditionally deleted and mutated Pkd1 on osteoblastic function ex vivo

To determine the effect of conditionally deleted Pkd1 on osteoblast function ex vivo, we examined cell proliferation and osteoblastic differentiation and gene expression profiles in primary osteoblast cultures derived from control and Pkd1^Dmp1-cKO mice. Control osteoblasts exhibited a time-dependent increase of total Pkd1 transcripts (wild-type Pkd1⁺ allele) during 21 d of osteogenic culture. In contrast, conditional Pkd1^Dmp1-cKO osteoblasts showed a proportionate time-dependent increase of Pkd1 mutant Pkd1^Δflox and Pkd1^m1Bei alleles consistent with 50, 92, and 150% inactivation of Pkd1 transcripts (Fig. 4A). Consistent with no Dmp1-Cre-mediated deletion of Pkd1 in the early stage of osteoblast culture (61), the Pkd1^Dmp1-cKO osteoblasts showed no alterations on BrdU incorporation and cell proliferation (Fig. 4B). However, Pkd1^Dmp1-cKO osteoblasts displayed impaired osteoblastic differentiation and maturation, as evidenced by time-dependent lower alkaline phosphatase activity (Fig. 4C), diminished calcium deposition in extracellular matrix (Fig. 4D), and reduced osteoblastic differentiation markers, including Runx2, Osteocalcin, and Dmp1, compared to controls (Fig. 4E–G). In agreement with increased adipogenic activity in vivo, the cultured primary calvarial cells under osteogenic condition exhibited a marked increase of adipocyte markers such as aP2 (Fig. 4H), suggesting impairment of osteogenesis and enhancement of adipogenesis in Pkd1^Dmp1-cKO osteoblast cultures.

Figure 4.

Effects of Dmp1-Cre-mediated Pkd1 deletion on osteoblastic proliferation and maturation ex vivo. A) Total Pkd1 transcripts by real-time RT-PCR from control and Pkd1^Dmp1-cKO osteoblasts during 21 d of culture. Control osteoblasts exhibited a time-dependent increase of total wild-type Pkd1⁺ allele, while conditional Pkd1^Dmp1-cKO osteoblasts showed a proportionate time-dependent increase of Pkd1^Δflox and Pkd1^m1Bei mutant alleles. B) BrdU incorporation. There was no significant change in BrdU incorporation for 6 h between control and Pkd1^Dmp1-cKO osteoblasts, indicating Dmp1-Cre-mediated Pkd1 deletion did not affect proliferation of primary cultured osteoblasts. C) Alkaline phosphatase (ALP) activity. Primary cultured Pkd1^Dmp1-cKO osteoblasts displayed time-dependent increments in ALP activity during 21 d of culture, but ALP activity was significantly lower at different time points compared with control Pkd1^flox/+ osteoblasts. D) Quantification of mineralization. Alizarin Red-S was extracted with 10% cetylpyridinium chloride and quantified as described in Materials and Methods. Primary cultured Pkd1^Dmp1-cKO osteoblasts had time-dependent increments in Alizarin Red-S accumulation during 21 d of culture, but accumulation was significantly lower at different time points compared with control Pkd1^flox/+ osteoblasts. E–H) Gene expression profiles by real-time RT-PCR. Primary cultured Pkd1^Dmp1-cKO osteoblasts in osteogenic differentiation media showed time-dependent increments in osteogenesis but significantly lower at different time points compared to control osteoblasts, evidenced by a significant reduction in osteoblastic markers, including Runx2 (E), Osteocalcin (F), and Dmp1 (G). In contrast, a marked increase of adipocyte markers, such as aP2 (H) at different time points was observed from the Pkd1^Dmp1-cKO osteoblasts under the same differentiation medium when compared with control osteoblasts. Data are expressed as means ± sd from 3 independent triplicate experiments. *P < 0.05 vs. control Pkd1^flox/+.

Impairment of Pkd1 deletion/mutation on flow-induced intracellular calcium response in osteoblasts

We found that Pkd1 deletion/mutation had a gene dose effect on basal [Ca²⁺]_i and flow-induced intracellular calcium response in immortalized Pkd1-deficient osteoblasts. In this regard, heterozygous Pkd1^null/+ and Pkd1^m1Bei/+ osteoblasts showed a significantly lower basal [Ca²⁺]_i concentration compared with wild-type Pkd1^+/+ cells, and homozygous Pkd1^null/null and Pkd1^m1Bei/m1Bei osteoblasts had greater reductions of basal [Ca²⁺]_i compared with their respective heterozygous cells (Fig. 5A, B). To study whether polycystin-1-mediated mechanical flow-induced intracellular calcium level changes, these immortalized cells were exposed to 6.24 dyn/cm² pulsatile laminar fluid flow. On fluid stimulation, we detected an immediate rise in intracellular calcium throughout the wild-type Pkd1^+/+ cell population, peaking roughly 10–20 s after stimulation (Fig. 5C, D). The [Ca²⁺]_i levels then rapidly decreased but were maintained at moderate levels for 50–60 s before returning to baseline. In contrast, when we exposed these Pkd1-deficient osteoblasts to an identical flow stimulus, we detected intermediate calcium response curve in the heterozygous cells and little or no calcium influx in either the early or late phase in the homozygous osteoblasts (Fig. 5C, D). However, 10 mM caffeine still resulted in normal calcium influx in Pkd1-null cells after flow stimulus (data not shown), indicating the viability of the Pkd1^null/null and Pkd1^m1Bei/m1Bei cells in the loading chamber. Our data suggest that the flow-induced [Ca²⁺]_i response requires polycystin complex in osteoblasts and loss of Pkd1 to abolish fluid flow sensing in osteoblasts.

Figure 5.

Effects of Pkd1 deletion and mutation on baseline and flow-induced [Ca²⁺]_i response in osteoblasts. A, B) Gene dose-dependent reduction of basal [Ca²⁺]_i level was observed in cultured heterozygous Pkd1^null/+ and homozygous Pkd1^null/null cells (n=32; A), as well as heterozygous Pkd1^m1Bei/+ and homozygous Pkd1^m1Bei/m1Bei cells (n=32; B) compared with wild-type Pkd1^+/+ cells (n=32). C, D) Flow-induced [Ca²⁺]_i responses are also impaired in a gene dose-dependent fashion in cultured heterozygous Pkd1^null/+ and homozygous Pkd1^null/null cells (n=10; C) as well as heterozygous Pkd1^m1Bei/+ and homozygous Pkd1^m1Bei/m1Bei cells (n=10; D) compared with wild-type Pkd1^+/+ (n=10) cells. Immortalized osteoblasts from newborn wild-type Pkd1^+/+, heterozygous Pkd1^null/+ and Pkd1^m1Bei/+, and homozygous Pkd1^null/null and Pkd1^m1Bei/m1Bei mice were cultured on type I rat tail collagen-coated 40-mm diameter glass slides at 80–90% confluency. Cells were loaded with 3 μM Fura-2-AM, and the slide was then placed in an FCS2 parallel plate flow chamber. Real-time record of fluorescence intensity (F340/F380 ratio) was performed in the cells when exposed to 6.24 dyn/cm² pulsatile laminar fluid flow. The 340/380 ratios were converted to concentration using standard calibration curve. Data represent means ± sd from individual cells.

Impaired mechanical loading response in single Pkd1 heterozygous and Pkd1^Dmp1-cKO mice

To investigate whether Pkd1 has a mechanosensing function in bone, we performed in vivo ulnae loading experiments using 16-wk-old control, single Pkd1^flox/m1Bei heterozygous mice, single Dmp1-Cre;Pkd1^flox/+ heterozygous mice, and conditional Pkd1^Dmp1-cKO null mice. We applied the same strain to the control, Pkd1^flox/m1Bei, Dmp1-Cre;Pkd1^flox/+, and Pkd1^Dmp1-cKO mice. In contrast to difference in MARs among control, single Pkd1 heterozygous mice, and Pkd1^Dmp1-cKO mice at 6 wk of age, we did not observe a difference in baseline MAR of cortical bone as assessed by calcein double labeling in the no-load ulnae of control among 4 groups (Fig. 6A). Consistent with previous researchers reporting this method of loading (40, 48, 67, 71), we found a robust bone formation response in the loaded ulna of control mice (Fig. 6A). In contrast, the bone formation response was markedly attenuated in the loaded ulna from these Pkd1-deficient mice (Fig. 6A). In this regard, load induced a 3-fold increase in MAR in the loaded ulnae from control mice; a reduction in ulna-loading response was proportionate to the gene dose in single Pkd1 heterozygous and conditional Pkd1-null mice (Fig. 6B). Similar to the observation in 6-wk-old mice, no load ulnae of 16-wk-old Pkd1^flox/m1Bei and Dmp1-Cre;Pkd1^flox/m1Bei mice expressed 50% Pkd1^m1Bei mutant (functional null) allele, whereas Dmp1-Cre;Pkd1^flox/+ and Dmp1-Cre;Pkd1^flox/m1Bei mice exhibited ∼42% excision of the floxed exon 2–4 from Pkd1. Again, a net reduction of Pkd1 expression in Pkd1^Dmp1-cKO mice was >90% in no-load ulnae bone samples (Fig. 6C). To examine the expression of mechanical load responsive genes in vivo, we performed real-time RT-PCR using RNAs from no-load and loaded ulnae of control and Pkd1^Dmp1-cKO mice. Consistent with the known anabolic response using this loading method, we found that Runx2-II, Cox-2, c-Jun, and Wnt-related genes (such as Wnt10b, FzD2, and Axin2) were significantly increased in the loaded ulnae from control mice (Fig. 6D), whereas these transcripts were no longer responsive to loading in Pkd1^Dmp1-cKO mice (Fig. 6D). To determine whether the applied stress was similar between the two groups, we performed ulna strain-gauge testing from control and Pkd1^Dmp1-cKO mice at 16 wk of age. A linear relationship between peak compressive force (N) and peak tension microstrain (με) at the lateral ulna midshaft during cyclic axial compression loading ex vivo was observed from control Pkd1^flox/+ and Pkd1^Dmp1-cKO mice. We also found that ulnae from Pkd1^Dmp1-cKO mice had almost 1.5-fold higher microstrains at −3.0 N load compared with control mice (Fig. 6E), indicating that the attenuated anabolic response occurred despite increased bone stress to mechanical loading in Pkd1^Dmp1-cKO.

Figure 6.

Conditional deletion and mutation of Pkd1 in osteocytes impairs anabolic response to mechanical loading. A) Representative images of midshaft ulnar cross sections from no-load and loaded ulnae of male control Pkd1^flox/+ mice, single Pkd1^flox/m1Bei heterozygous mice, single Dmp1-Cre;Pkd1^flox/+ heterozygous mice, and conditional Pkd1^Dmp1-cKO null mice after loading. There was a robust bone formation response on the medial (inset) and lateral surfaces of loaded control Pkd1^flox/+ ulna. Intermediate bone formation response was observed in the loaded ulnae of Pkd1^flox/m1Bei and Dmp1-Cre; Pkd1^flox/+ mice, but almost no response can be observed in the loaded Pkd1^Dmp1-cKO ulnae. B) MAR on periosteal surface of the midshaft ulna in response to applied mechanical strain in 4 genotypes. Pkd1 gene dose-dependent impairment of MAR was observed in response to −3.0-N loading strains. C) Real-time RT-PCR analysis of total Pkd1 transcripts in no-load ulnae of control and Pkd1-deficient mice at 16 wk of age. Expression of total Pkd1 transcripts was performed using Pkd1-allele-specific primers, as described in Materials and Methods. Similar to the observation in 6-wk-old mice, reduction in total functional Pkd1 transcripts was proportionate to the gene dose in single Pkd1 heterozygous and conditional Pkd1-null mice. D) Expression of mechanical load responsive genes in control and Pkd1^Dmp1-cKO mice 4 h after loading. Real-time RT-PCR analysis from both male genotypes shows that mRNA levels of Runx2-II, Cox-2, c-Jun, and Wnt-related genes (such as Wnt10b, FzD2, and Axin2) were significantly increased in the loaded ulnae from control Pkd1^flox/+ mice, but there was much less change in loaded ulnae from Pkd1^Dmp1-cKO mice. E) Strain gauge measurement. Linear relationship between peak compressive force (N) and peak tension microstrain (με) at the lateral ulna midshaft during cyclic axial compression loading ex vivo was observed from Pkd1^flox/+ and Pkd1^Dmp1-cKO mice. This relationship was significantly different between Pkd1^flox/+ and Pkd1^Dmp1-cKO ulnae using 2-way ANOVA analysis (P<0.0001). Ulnae were loaded in vivo using a peak compressive force of −3.0 N; corresponding microstrain was 1.5-fold greater in Pkd1^Dmp1-cKO ulnae than in control Pkd1^flox/+ ulnae. Data represent means ± sd from 5 or 6 individual mice. *P < 0.05 vs. control Pkd1^flox/+.

DISCUSSION

An essential role of mechanosensing in maintaining bone homeostasis is well established, but the cellular and molecular mechanisms responsible for sensing strain and signal transduction pathways leading to the anabolic bone responses appear to be complex. Our studies are the first to show that the selective deletion of Pkd1 from mature osteoblasts and osteocytes results in osteopenia and impaired mechanosensing both in vitro and in vivo. These results establish an important role for polycystin 1, which is part of a known mechanosensing complex involving polycystin 2 and primary cilia in renal epithelial cells, in mechanosensory responses in the skeleton. Indeed, we found that Dmp1-Cre effectively and selectively resulted in the conditional deletion of Pkd1 in mature osteoblasts and osteocytes in bone and that loss of Pkd1 in osteocytes was associated with reduced BMD, TV, Ct.Th, and impaired bone formation rates in vivo under normal physiological loading conditions. More important, the reduction in Pkd1 in osteocytes resulted in reduced bone formation and impaired up-regulation of mechanoresponsive gene expression in response to ulna loading in the conditional Pkd1^Dmp1-cKO null mice in vivo. Finally, a direct effect of Pkd1 in osteoblast/osteocytes is supported by the ex vivo experiments showing that Pkd1-deficient primary osteoblasts have impaired response to flow-induced intracellular calcium response. Together, these data suggest that polycystin-1 (Pkd1) in mature osteoblasts/osteocytes functions as a mechanosensor essential for bone cell responses to flow and mechanical loading.

We and others have previously shown that Pkd1 and Pkd2 form a polycystin complex, which is colocalized to plasma membrane and primary cilia (56, 59). In addition, Pkd1 is coupled to multiple intracellular signal transduction pathways and cellular responses that have been linked to mechanosensing responses in bone, including increments in intracellular calcium, calcineurin/NFAT, Wnt/β-catenin, G-protein, and AP-1 activation (56–57, 75–79). Consistent with Pkd1 stimulation of Runx2 expression in osteoblasts through intracellular calcium and Akt-Gsk3β-β-catenin-dependent pathways (58–59), we found that load-induced expression of Runx2-II was attenuated in Pkd1^Dmp1-cKO mice. In addition, up-regulation of other mechanoresponsive genes, including Cox-2, c-Jun, and Wnt-related genes (such as Wnt10b, FzD2, and Axin2) were markedly attenuated in the loaded ulnae of Pkd1^Dmp1-cKO mice when compared with control mice. Thus, Pkd1 may act as a hub to channel environmental cues into multiple signaling pathways necessary for the bone anabolic response to load. The mechanosensing function of Pkd1 in bone is also consistent with its mechanosensing function in kidney and other tissues (56, 80–81). In this regard, polycystins are proposed to form a flow-sensitive ion channel complex in the primary cilium of both renal epithelial and vascular endothelial cells (56, 80). Polycystins are also reported to have an essential role in pressure sensing in cardiovascular smooth muscle cells, where it is hypothesized that the PC1/PC2 ratio controls stretch-activated ion channel (SAC) mechanosensitivity through filamin A coupled to the actin cytoskeleton, which converts intraluminal pressure to local tension (81). The proximate external forces that activate Pkd1 in osteocytes are not known. The stimulus for Pkd1 activation might involve primary cilia (59, 82), cell to cell (83–84), and/or cell to matrix interactions (85–86). Recent studies have shown that Pkd1 is a cell adhesion protein with mechanosensing properties and has abilities to interact with components of the extracellular matrix (heterophilic), neighboring PC1 (homophilic interactions), and the cytoskeleton via intermediate filaments as proposed in other tissues (83–88).

Our findings do not distinguish between mature osteoblast and osteocytes as mechanosensing cells in bone, because Dmp1-Cre would be expected to reduce Pkd1 expression in both. Pkd1 also functions in early osteoblasts, as we have previously shown that conditional deletion of Pkd1 in mature osteoblasts using Oc-Cre results in osteopenia (58). However, osteocytes make up >90–95% of the cells in the adult skeleton (89), are embedded in bone matrix, and form an intercellular network via their dendritic processes (8) that make them suitable for mechanosensing (7–8, 17, 23, 90–96). In addition, osteocytes have both intercellular communications with other osteocytes, as well as secrete autocrine and paracrine factors capable of regulating osteoblast and osteoclast function (17–24). In addition, the reduction in bone mass observed with conditional deletion of Pkd1 using Dmp1-Cre (Pkd1^Dmp1-cKO) is more severe than the conditional deletion of Pkd1 earlier in the osteoblast lineage using Oc-Cre (Pkd1^Oc-cKO) (58). Differences in the strength of Cre-recombinase expression or tissue specificity might also account for the differential responses to Oc- and Dmp1-Cre-mediated Pkd1 ablation. Consistent with our findings, other studies that ablated osteocytes in mice report resistance to unloading-induced bone loss (16), suggesting a principal role of osteocytes in bone mechanosensing responses.

How closely these studies in mice reflect human physiology remains uncertain, since patients with autosomal dominant polycystic kidney disease (ADPKD) do not have clinically apparent skeletal abnormalities. Unlike our studies of homozygous loss of Pkd1 function in bone, ADPKD is caused by a heterozygous PKD1 or PKD2 mutation and a somatic “second hit” affecting the other allele, leading to loss of both alleles in only a subset of renal cells, which is sufficient to produce cystic renal disease (62, 97). A second hit may not occur in bone, and if it did, an insufficient fraction of osteoblasts may be affected to cause clinically detectable skeletal abnormalities. Indeed, broadly disrupting both alleles in bone cells causes clinically apparent bone disease in humans, as evidenced by the skeletal malformation observed in rare families with early onset of polycystic kidney disease in newborns (99–100). The greater genetic diversity of humans may mask the small effects on bone observed in Pkd1 heterozygous mice on a genetically homogenous background (60). Alternatively, skeletal abnormalities in humans with ADPKD might be detected using more sensitive techniques. For example, FGF23, an osteocyte-derived hormone, is 4-fold higher in ADPKD compared to chronic kidney disease controls (98). Finally, the concomitant presence of renal disease and hyperparathyroidism seen in patients with ADPKD might mask a skeletal phenotype associated with heterozygous inactivation of PKD1 in bone (101). Further studies will be necessary to demonstrate that polycystins are key components of human skeletal mechanosensing pathways.

Our studies have several limitations that will require further research. First, we observed impaired osteoblast differentiation and increased adipocyte formation ex vivo, consistent with a role of Pkd1 in regulating osteoblasts and adipocyte development (102–104). Also, the conditional deletion of Pkd1 in neural crest cells using Wnt1-Cre reduced the bone formation response to tensile stress across the suture that was associated with impaired differentiation of osteochondroprogenitor cells (102). Although we did not observe any gross abnormalities of skeletogenesis in our Pkd1^Dmp1-cKO mice, subtle differences in osteoblast development might have altered the mechanosensing response in vivo. Loss of Pkd1 may also result in additional age-dependent changes in bone structure, consistent with the changes of bone geometry in rats during the spaceflight (105). These age-dependent changes in Ct.Th more likely reflect the alterations in signal in osteocytes that lead to lower osteoclastogenesis and bone resorption and Runx2 isoform expression in Pkd1^Dmp1-cKO mice, which differentially regulate cortical and trabecular bone (59, 72). The adaptive alterations in bone geometry observed in Pkd1^Dmp1-cKO mice resulted in a significantly higher microstrain under the same load force. Thus, the impaired mechanical loading responses after conditional deletion of Pkd1 in osteocytes occurred despite greater applied loads resulting from alterations in bone geometry. Finally, our studies do not preclude the presence of other mechanosensing mechanisms in bone that may respond to different types of mechanical loads (1–4) or additional functions of Pkd1 in osteocytes, osteoblasts, and chondrocytes in the skeleton (81, 102). Indeed, the presence of other mechanosensors is suggested by studies using Osx-Cre to delete Pkd1 gene in osteoblasts and a midpalatal suture expansion loading model, which failed to show a mechanosensing defect in the bone surrounding the suture.

In summary, ex vivo studies of Pkd1-deficient osteoblasts and in vivo mouse ulnar loading in single Pkd1 heterozygous and conditional Pkd1^Dmp1-cKO null mice demonstrate for the first time that polycystin-1 in mature osteoblasts/osteocytes is required for anabolic bone responses to mechanical stimulation in vitro and in vivo. Because primary cilia and polycystin-2 are also expressed in the osteoblast lineage, additional studies will be needed to determine whether Pkd1 participates in a mechanosensing complex with primary cilia and polycystin-2. In addition, the upstream environmental cues that activate Pkd1, as well as the downstream signaling pathways mediating the mechanotransduction signals from Pkd1 in mature osteoblasts/osteocytes remain to be defined.