For a Commentary on this article, please see Towler (J Bone Miner Res. 2011;26:2579–2582. DOI: 10.1002/jbmr.523).
Intermittent PTH(1–84) is osteoanabolic but not osteoangiogenic and relocates bone marrow blood vessels closer to bone-forming sites†
Article first published online: 20 OCT 2011
Copyright © 2011 American Society for Bone and Mineral Research
Journal of Bone and Mineral Research
Volume 26, Issue 11, pages 2583–2596, November 2011
How to Cite
Prisby, R., Guignandon, A., Vanden-Bossche, A., Mac-Way, F., Linossier, M.-T., Thomas, M., Laroche, N., Malaval, L., Langer, M., Peter, Z.-A., Peyrin, F., Vico, L. and Lafage-Proust, M.-H. (2011), Intermittent PTH(1–84) is osteoanabolic but not osteoangiogenic and relocates bone marrow blood vessels closer to bone-forming sites. J Bone Miner Res, 26: 2583–2596. doi: 10.1002/jbmr.459
- Issue published online: 20 OCT 2011
- Article first published online: 20 OCT 2011
- Accepted manuscript online: 28 JUN 2011 09:12AM EST
- Manuscript Accepted: 20 JUN 2011
- Manuscript Revised: 18 APR 2011
- Manuscript Received: 2 DEC 2010
- bone vessel quantitative imaging;
- Parathyroid hormone;
- Top of page
- Materials and Methods
Intermittent parathyroid hormone (PTH) is anabolic for bone. Our aims were to determine (1) whether PTH stimulates bone angiogenesis and (2) whether vascular endothelial growth factor (VEGF A) mediates PTH-induced bone accrual. Male Wistar rats were given PTH(1–84) daily, and trabecular bone mass increased 150% and 92% after 30 and 15 days, respectively. The vascular system was contrasted to image and quantify bone vessels with synchrotron radiation microtomography and histology. Surprisingly, bone vessel number was reduced by approximately 25% and approximately 40% on days 30 and 15, respectively. PTH redistributed the smaller vessels closer to bone-formation sites. VEGF A mRNA expression in bone was increased 2 and 6 hours after a single dose of PTH and returned to baseline by 24 hours. Moreover, anti-VEGF antibody administration (1) blunted the PTH-induced increase in bone mass and remodeling parameters, (2) prevented the relocation of bone vessels closer to bone-forming sites, and (3) inhibited the PTH-induced increase in mRNA of neuropilin 1 and 2, two VEGF coreceptors associated with vascular development and function. In conclusion, PTH(1–84) is osteoanabolic through VEGF-related mechanism(s). Further, PTH spatially relocates blood vessels closer to sites of new bone formation, which may provide a microenvironment favorable for growth. © 2011 American Society for Bone and Mineral Research
- Top of page
- Materials and Methods
Angiogenesis and osteogenesis are coupled during bone modeling processes such as growth and repair. In mice deficient in vascular endothelial growth factor (VEGF), the almost complete suppression of blood vessel invasion corresponded with impaired trabecular bone formation and limited expansion of the hypertrophic chondrocyte zone.1 In addition, fatigue loading in rat ulnas increased vascularity (ie, vessel number and size) at the periosteal surface, which preceded bone growth at this site during repair.2, 3 The literature, however, is sparse concerning the relation between bone and bone vessels during remodeling despite the fact that vessels are located next to bone remodeling units in the trabecular compartment and run along their central axis in cortices.4 Two weeks of treadmill training in rats increased the number of blood vessels in the tibial metaphysis, which preceded the augmentation of BMD observed at 3 weeks.5 In addition, blockade of VEGF inhibited the gain in metaphyseal BMD, cancellous bone volume, and blood vessel number.5 These data tended to show that when an anabolic mechanical stimulus was applied, bone and vessel growth were coupled in a VEGF-dependent manner.
Vascular factors other than vessel density may contribute to bone remodeling, and evidence suggests that the interplay between the vascular and skeletal systems is quite complex. For example, in addition to capillary rarefaction, osteoporosis in humans has been linked to age-related declines in skeletal perfusion.6 Declines in skeletal perfusion in aged rats, exhibiting a tendency for reduced femoral trabecular bone volume, were associated with diminished vasodilator capacity and nitric oxide bioavailability of the osseous resistance vasculature.7 Subsequent data revealed that stimuli such as exercise training improved not only bone mass but also skeletal perfusion and bone vascular function in aged rats.8 Therefore, mechanisms that alter the ability of the osseous vasculature to provide the skeleton with blood and nutrients7 may be equally modifying to bone mass as agents that inhibit angiogenesis.5 Further clarification of the interplay between the two systems is necessary for the treatment of diseases that alter bone remodeling.
Parathyroid hormone (PTH) is involved in bone cellular communications important for the activation, reversal, and termination phases of bone remodeling.9 When administered intermittently, PTH augments bone mass through stimulation of osteoblast formation and therefore has become a current treatment for osteoporosis.9 Receptors for PTH have been identified on both osteoblasts10 and endothelial cells.11 In addition, PTH stimulates the release of VEGF from human osteosarcoma12 and endothelial cells.13 VEGF administration and inhibition, respectively, augmented and diminished blood flow to distracted bones.14 In this study, we originally hypothesized that intermittent PTH(1–84) administration would promote bone angiogenesis, thus augmenting bone blood flow and providing a microenvironment favorable for bone growth. Yet, despite PTH-induced increases in bone mass and VEGF A expression in bone, bone angiogenesis did not occur with intermittent PTH(1–84) treatment. Since VEGF is a known vasodilator in other tissue beds,15–17 we assumed that it would function similarly in the bone vasculature and provide a mechanism for bone formation related to bone perfusion via enhanced vasodilation. Under this premise, if VEGF A plays a nonangiogenic role in bone formation, its blockade during PTH-induced bone remodelling would modify bone accrual. We then undertook a set of experiments examining the effects of VEGF blockade on bone microarchitectural, static and dynamic properties, and the bone vasculature.
Materials and Methods
- Top of page
- Materials and Methods
Male Wistar rats (3 to 4 months) were obtained from Janvier (Le Genest-St-Isle, France). All procedures were in accord with the European Community standards on the care and use of laboratory animals (Authorization 04827, Ministère de l'Agriculture, France). Rats were housed at 23 °C, were maintained on a 12/12-hour light/dark cycle, and were fed rat chow and water ad libitum.
Five rats were euthanized at baseline, and 15 rats were treated with PTH(1–84) (Preotact, a kind gift from Nycomed, Roskilde, Denmark) at a dose of 100 µg/kg subcutaneously and euthanized 2, 6, and 24 hours after injection. Tibias were dissected and processed for measurement of mRNA expression of VEGF A (see below).
According to baseline trabecular bone volume/total volume ratios (BV/TV) in the tibias assessed by high-resolution microtomography (see below for methods), 24 rats were assigned to two experimental groups: (1) PTH and (2) control (CON). Rats were subcutaneously administered PTH at a dose of 100 µg/kg per day and control rats were given a vehicle (VEH; 0.9% sterile saline) 5 days per week for 4 weeks. Injection volumes were adjusted weekly to correspond with increasing body weights. A dose of 100 µg/kg per day of PTH(1–84) is molecularly equivalent to 43 µg/kg per day of PTH(1–34).18 PTH and vehicle were administered between 9 and 10 am daily.
According to baseline tibial BV/TV, 72 rats were divided into either one of four groups: (1) control, (2) PTH, (3) PTH + anti-VEGF antibody (Ab), and (4) anti-VEGF Ab alone. Rats were subcutaneously administered PTH(1–84) (100 µg/kg per day) or vehicle 5 days per week for 2 weeks. Rats administered intraperitoneal injections of anti-VEGF Ab received 0.70 mg/kg per day of bevacizumab (Avastin, Roche, Neuilly/Seine, France). At the end of the experiment, 10 rats per group were perfused with a barium sulfate solution for contrasting of bone vascularization (see methodology below). For the 8 remaining rats in each group, the left tibias were dissected and stored for VEGF A (Vegfa), VEGF receptor 1 and 2 (Vegfr1 and Vegfr2, respectively), and neuropilin 1 and 2 (Np1 and Np2, respectively) mRNA analyses (see methodology below). For determination of bone static and dynamic properties, rats were double-labeled with tetracycline (50 mg/kg of body) by intraperitoneal injection on days 5 and 1 prior to killing.
Vascular network perfusion
For contrasting of the bone vasculature, rats were injected subcutaneously 1 hour prior to killing with heparin (50 UI/kg) to prevent blood coagulation. Subsequent to acquiring the postexperimental bone microarchitectural properties via micro–computed tomography (µCT), the rats were anesthetized with pentobarbital (40 mg/kg of body weight) and prepared for vascular injection of a barium sulfate solution for opacification of the bone vascular network.19 Briefly, a needle was inserted into the left ventricle, and a small incision was made in the right atrium. Using a perfusion pump (Masterflex L/S, Cole-Parmer, Vernon Hills, IL, USA), blood of the vascular system was flushed with physiologic phosphate-buffered saline (PBS). A cannula then was inserted into the abdominal aorta, and barium sulfate solution was pumped into the blood vessels. The vascular system was considered entirely perfused when the liver and feet of the animal appeared white. The right tibias and femurs then were dissected free and stored in 10% formalin at 4 °C for 3 days.
The left proximal tibias and distal femurs were removed from the formalin, dehydrated in absolute acetone, and embedded in methyl methacrylate at low temperature. The proximal tibial metaphysis was sectioned frontally with a microtome (Reichert-Jung Polycut, Heidelberg, Germany), and histologic slides were created. Adipocyte cell number and area and sinusoidal lobule (Sin.L) number and area were quantified in the secondary spongiosa using an ocular integrator with 100-point grid at ×250, as described previously, for the measurement of bone vasculature.5
Evaluation of bone parameters
Five 9-µm-thick histologic sections stained with Goldner's trichrome were used for measurement of BV/TV (%), trabecular thickness (Tb.Th, µm), trabecular number (Tb.N /mm2), and trabecular separation (Tb.Sp, µm) using an automatic image analyzer (BIOCOM, Lyon, France). Osteoid surface per bone surface (OS/BS, %) and osteoblast surface per bone surface (Ob.S/BS, %) were measured on Goldner-stained sections. Five 9-µm-thick sections were stained for tartrate-resistant acidic phosphatase (TRACP) activity to measure osteoclast number and surface (Oc.N/B.Ar, /mm2, and Oc.S/BS, %, respectively). Histodynamic parameters were determined under ultraviolet light from tetracycline labeling of bone-forming surfaces on five unstained 12-µm-thick sections: mineral apposition rate (MAR, µm/d), single-labeled surface per bone surface (sLS/BS, %), double-labeled surface per bone surface (dLS/BS, %), percent of mineralizing surface per bone surface (MS/BS, %; calculated as dLS/BS + ½sLS/BS), and bone-formation rate (BFR/BS, µm3/µm2/day; calculated as MS/BS × MAR). These parameters of bone resorption and formation were measured with a semiautomatic system using a digitizing table (Summasketch-Summagraphics, London, UK) connected to a personal computer and to a Reichert Polyvar microscope (Reichert Jung, Germany) equipped with a drawing system (Camera Lucida Roper Scientific, Friedland, Germany).
Evaluation of vascular structural parameters
Vessel volume to tissue volume (Ves.V/TV, %), vessel thickness (Ves.Th, µm), vessel number (Ves.N, /mm), and vessel separation (Ves.Sp, µm) were measured on the Goldner-stained sections using an automatic image analyzer with the software generally used for bone parameter assessment on unstained sections (BIOCOM). Sinusoidal lobule (Sin.L) number and volume were quantified in the secondary spongiosa using the ocular integrator with 100-point grid at ×250, as described previously, for the measurement of bone vasculature.19
Analysis of the spatial distribution of bone vasculature in relation to trabecular surface and osteoid seams
Goldner staining of barium sulfate–perfused tibial samples allows simultaneous visualization of blood vessels, trabecular bone, and osteoid seams. Five Goldner's trichrome–stained sections per rat were used for determination of distances between bone marrow blood vessels, trabeculae, and osteoid seams. Five serial photographs of the secondary spongiosa in the tibial metaphysis were taken per slides. Briefly, using an appropriate color deconvolution/separation algorithm with Image J Software (NIH, Bethesda, MD, USA), blood vessels (brown), trabecular bone (green), and osteoid (pink; ie, new bone formation) were extracted separately from the photographs to obtain binarized images of the structures of interest. Vessels were analyzed according to size (≤29 µm, 30 to 100 µm, and 101 to 200 µm), and distance maps were calculated automatically according to size and distances (µm) between each vessel profile compared with either total trabecular surface or surfaces covered by osteoid seams. The particular sizes of vessels were chosen to roughly differentiate between vessels that direct blood flow (ie, arterioles), venules, and larger conduit vessels (ie, arteries and veins) and vessels responsible for nutrient exchange with the surrounding tissue (ie, capillaries). With recognition that arteriolar diameters can be as small as 10 µm, for the sake of simplicity, vessels 29 µm or less will be referred to as “capillaries.” Intraindividual coefficients of variations for distances between blood vessels and trabeculae and blood vessels and osteoid seams were 6% and 7%, respectively.
High-resolution µCT and synchrotron radiation microtomography (SR-µCT)
In vivo follow-up of structural and microarchitectural bone parameters by µCT
At the beginning and end of the experimental protocols, the left tibias were scanned in vivo using high-resolution µCT Viva CT40 (Scanco Medical AG, Brüttisellen, Switzerland). At baseline, the rats were assigned to the control and treated groups so that the mean BV/TV did not differ between or among groups. Rats were maintained under anesthesia (isoflurane 2%, O2 balance) during acquisition. Four hundred slices were acquired at 55 kVp starting just beneath the end of the growth plate, thus including both the primary and secondary spongiosa. Quantification of BV/TV, Tb.Th, Tb.N, and Tb.Sp was performed on 130 slices in the secondary spongiosa as described previuosly.20
Synchrotron radiation microtomography
Sample processing: The femurs were removed from the formalin, dehydrated in absolute acetone, and embedded in methyl methacrylate at low temperature. 3D synchrotron radiation microtomography (SR-µCT) imaging was performed on beamline ID19 at the European Synchrotron Radiation Facility (ESRF, Grenoble, France), where a parallel-beam 3D µCT setup has been developed.21 Two thousand radiographs of each sample were recorded under different angles (over 360 degrees) by use of a 2048 × 2048 pixel CCD-based detector.22 The voxel size was 2.8 µm3, yielding a cylindrical field of view (FOV) of diameter 5.6 mm. The energy was set to 25 keV with a beam height of approximately 2 mm. Exposure time was 0.25 second per image with a large dynamic range (the detector provides 14 bits). Acquisition time per scan was approximately 18 minutes. Two images per bone were acquired that corresponded to a region of interest in the secondary spongiosa of the metaphysis (ROI1) and diaphysis (ROI2). Because sample dimensions did not fit the FOV, a nonconventional acquisition procedure was used: recording images over 360 degrees with the axis of rotation displaced to the edge of the FOV and enabling the reconstruction of a larger FOV encompassing the sample. This procedure yielded reconstructed 3D images of 2500 × 2500 × 700 voxels.
Image segmentation: Segmentation procedures required the identification of vessels, bone, and background.23 Since a separate analysis of trabecular and cortical bone was planned, trabecular and cortical bone envelopes were segmented as separate compartments. Segmentation of bone and vessels was performed using 3D region growing. Volumes were used to separate bone into trabecular and cortical bone and vessels into cortical and bone marrow vessels by arithmetical operations detailed in Langer and colleagues.23 After segmentation, each volume was partitioned into subvolumes so that the percent of vasculature in the metaphysis and diaphysis could be determined. Therefore, the following parameters were assessed: metaphysis: vessel volume to marrow volume (Ves.V/Mar.V, %), vessel thickness (Ves.Th, µm); and diaphysis: vessel volume to cortical bone volume (Ves.V/Ct.V) and vessel volume to marrow volume (Ves.V/Mar.V, %). Vessel thickness (Ves.Th, µm) was assessed in both marrow and cortical compartments.
Total RNA was extracted from the left proximal tibia by a modified technique using a single-step total RNA isolation reagent developed by Chomczynski and Mackey (1995). Briefly,24 frozen bone fragments were ground to a fine powder in 10 mL of TRI reagent, followed by purification according to the manufacturer's instructions (Sigma Chemical Company, St Louis, MO, USA). The nucleic acid pellet then was retreated using the Rneasy Mini-Kit (Qiagen, Courtaboeuf, France). Monitoring of mRNA purity showed a 260:280 nm ratio of absorbance comprised between 1.7 and 2. The integrity of RNA was checked by electrophoresis on 1% agarose-1× Tris-acetate EDTA gel after ethidium bromide staining. Complementary DNA (cDNA) was synthesized from 4 µg of total RNA using a first-strand cDNA synthesis kit for RT-PCR (AMV; Roche, Indianapolis, IN, USA) in a reaction mix containing 2 µL of 10× reaction buffer, 3.1 mM MgCl2, 20 nmol each of dNTP, 1,6 µg oligo-p(dT)15 primer, 50 U RNase inhibitor, and 20 U AMV reverse transcriptase. The reaction was incubated for 60 minutes at 42 °C. Then 8 µL of cDNA mixture diluted in water was subjected to real-time PCR using SYBR Green I dye (Light Cycler-FastStart DNA Master SYBR Green I, Roche, Penzberg, Germany). Reactions were performed in 20 µL of PCR mixture containing 4 µL of 5× Master Mix (dNTP mixture with dUTP instead of dTTP, MgCl2, SYBR Green dye, Taq DNA polymerase, and reaction buffer), 0.5 to 1.0 µL of 10 µM primers, and 6 or 7 µL of PCR-grade water. Amplifications were performed in a Light Cycler 3.0 Instrument (Roche, Meylan, France). Primer sequences were as following: Vegfa: forward (F), 5'-GGG AGC AGA AAG CCC ATG AA-3', reverse (R), 5'-CAG GAA GCT CAT CTC TCC TA-3', product size (S) 272 bp, Genebank Identification Number (ID) NM_031836; Vegfr1: F, 5'-TCA CCA CGG ACC TCA ATA CA-3', R, 5'-CGA TGC TTC ACG CTG ATA AA-3', S 255 bp, ID NM_019306; Vegfr2: F, 5'-TGT CAC TTT GTG CTA GGT AT-3', R, 5'-AGT CCC AGG AAG GGA TTT CA-3', S 374 bp, ID NM_013062; Np1: F, 5'-GGA GCT ACT GGG CTG TGA AG-3', R, 5'-ATG TCG GGA ACT CTG ATT GG-3', S 203 bp, ID NM_145098; Np2: F, 5'-ACA CAA GGA GCC ATT TCC AG-3', R, 5'-CGG ATC CTG ATG AAA CGA GT-3', S 200 bp, ID NM_030869; and Cyclophilin: F, 5'-AGC ACT GGG GAG AAA GGA TT-3', R, 5'-TTC ACC TTC CCA AAG ACC AC-3', S 281 bp, ID NM_017101. A typical protocol included a denaturation step at 95 °C for 480 seconds followed by 40 to 50 cycles at 95 °C for 15 seconds, temperature annealing for 10 seconds, and a 72 °C extension for 14 seconds. Gene expression levels were normalized using cyclophilin; reference and control were set at 100%.
One-way ANOVAs (SPSS Version 17, SPSS, Inc., Chicago, IL, USA) were used to compare differences between (experiment 1: control versus. PTH) and among (experiment 2: control, PTH, PTH + anti-VEGF Ab, and anti-VEGF Ab) the different groups. Post-hoc analyses were conducting using the Student-Newman-Keuls test. To examine the relationships between bone remodeling and vascular-osteoid separation, Pearson product-moment correlations were measured between vascular-osteoid separation of capillaries and bone-remodeling parameters for both controls (15- and 30-day data pooled) and PTH-treated (15- and 30-day data pooled) groups. Alpha levels of p ≤ 0.05 were considered statistically significant, and group differences of p < 0.10 were acknowledged as tendencies. Data are represented as mean ± SEM.
- Top of page
- Materials and Methods
Body masses were not statistically different between and among groups at the start of each protocol (data not shown). While body weight increased significantly in all animals across time for each experiment, there were no differences between PTH and controls or among controls, PTH, PTH + anti-VEGF Ab, and anti-VEGF Ab groups at end of the experiments (data not shown).
Effects of PTH(1–84) and/or anti-VEGF Ab administration on gene expression in bone
After one single dose of PTH(1–84) (100 µg/kg), Vegfa mRNA expression in the tibia increased significantly by 1.6-fold after 2 hours, with a further 2.5-fold increase after 6 hours. Values returned to baseline by 24 hours (Fig. 1A). mRNA expressions for Vegfa, Vegfr1, and Vegfr2 did not differ among groups following 15 days of intermittent PTH administration and/or anti-VEGF Ab treatment (Table 1). In contrast, there were significant increases in mRNA expressions of both neuropilin 1 (Np1) and neuropilin 2 (Np2) in the 15-day PTH-treated animals. Interestingly, this PTH effect was completely blocked by anti-VEGF Ab administration (Fig. 1B).
|Control||PTH||PTH + anti-VEGF Ab||Anti-VEGF Ab|
|Vegfa/cyclophilin ratio||0.45 ± 0.10||0.37 ± 0.07||0.30 ± 0.06||0.45 ± 0.07|
|Vegfr1/cyclophilin ratio||0.20 ± 0.03||0.26 ± 0.05||0.22 ± 0.07||0.17 ± 0.04|
|Vegfr2/cyclophilin ratio||1.36 ± 0.17||1.33 ±0.25||1.19 ± 0.51||1.17 ± 0.29|
The PTH-induced increases in trabecular bone mass and formation are prevented by anti-VEGF Ab administration
Effects on bone mass and structural parameters
As expected, intermittent PTH administration enhanced tibial BV/TV versus the control groups by 92% and 150% in animals treated for 15 and 30 days, respectively (Fig. 2A). This resulted from higher Tb.N and Tb.Th and lower Tb.Sp in treated animals (Table 2). Administration of the anti-VEGF Ab completely blunted the rise in BV/TV induced by PTH treatment alone, whereas BV/TV of the PTH + anti-VEGF Ab group did not differ from those the anti-VEGF Ab and control groups (Fig. 2A). Importantly, anti-VEGF Ab administration alone and at this dose did not delay bone growth, as evidenced by similar BV/TV and microarchitecture of the primary spongiosa in the anti-VEGF and control groups (Fig. 2A and Table 2).
|Control||PTH||PTH + Anti-VEGF Ab||Anti-VEGF Ab|
|Tb.N/mm||2.3 ± 0.1||3.9 ± 0.2a|
|Tb.Th (µm)||71.7 ± 1.7||90.6 ± 2.8a|
|Tb.S (µm)||462.5 ± 24.8||236.9 ± 12.5a|
|Conn.D (/mm)||29.2 ± 1.8||60.7 ± 6.0a|
|Tb.N (/mm)||2.4 ± 0.3||3.2 ± 0.1a||2.1 ± 0.1||2.0 ± 0.1|
|Tb.Th (µm)||90.1 ± 3.5||103.9 ± 2.1a||95.2 ± 2.4||86.3 ± 2.1|
|Tb.S (βm)||474.3 ± 59.0||329.8 ± 16.1b||528.2 ± 24.6||534.0 ± 40.8|
|Conn.D (/mm)||35.7 ± 9.2||57.9 ± 7.3a||28.6 ± 1.9||26.6 ± 3.8|
|Tb.N (/mm)||8.5 ± 0.9||10.5 ± 0.7||9.9 ± 0.5||11.1 ± 0.6|
|Tb.Th (µm)||23.5 ± 2.1||24.8 ± 1.5||25.9 ± 1.3||27.2 ± 2.0|
|Tb.S (µm)||120.1 ± 30.6||81.5 ± 9.2||78.5 ± 6.5||67.1 ± 6.4|
Effects on bone remodeling and bone marrow adipocytes
As expected, at the end of the 30-day experiment in PTH-treated rats, we observed a 122% increase in BFR versus controls owing to elevated MAR and MS/BS (Table 3). In addition, PTH treatment augmented OS/BS and Ob.S/BS while reducing osteoclasts parameters (Oc.N/B.Ar and Oc.S/BS) versus the control group. Adipocyte number was significantly diminished by 48% with PTH treatment (Table 3). Following 15 days of PTH, rats exhibited a 106% increase in BFR owing to 2.3-fold augmentation of the dLS/BS, whereas MAR did not differ from the control group (Table 3). Both enhancements in BFR and dLS/BS were fully prevented by anti-VEGF Ab coadministration, whereas anti-VEGF Ab alone had no effect on these parameters. Similar to observations in the 30-day study, OS/BS and Ob.S/BS were augmented significantly with PTH treatment versus the other groups (Table 3). Bone-resorption parameters (ie, Oc.N /B.Ar and Oc.S/BS) following 15 days of treatment did not differ among groups (Table 3). Adipocyte number did not differ from controls and was not affected by anti-VEGF Ab administration.
|Control||PTH||PTH + Anti-VEGF Ab||Anti-VEGF Ab|
|BFR (µm3/µm2/d)||0.09 ± 0.02||0.20 ± 0.03a|
|MAR (µm/d)||1.7 ± 0.1||2.4 ± 0.1a|
|sLS/BS (%)||3.3 ± 0.7||3.3 ± 0.8|
|dLS/BS (%)||3.5 ± 0.4||6.3 ± 0.8a|
|MS/BS (%)||5.2 ± 0.7||8.0 ± 0.9a|
|OS/BS (%)||1.3 ± 1.2||4.2 ± 1.1a|
|Ob.S/BS (%)||5.0 ± 1.7||8.2 ± 1.5a|
|Oc.N/B.Ar (/mm2)||122.6 ± 30.4||23.6 ± 5.3a|
|Oc.S/BS (%)||3.6 ± 0.9||1.0 ± 0.2a|
|Adipocyte Nb/1.5 mm2||186.8 ± 29.6||96.6 ± 15.9a|
|BFR (µm3/µm2/d)||0.16 ± 0.03||0.33 ± 0.06a||0.13 ± 0.02||0.16 ± 0.03|
|MAR (µm/d)||1.3 ± 0.1||1.3 ± 0.0||1.4 ± 0.1||1.2 ± 0.1|
|sLS/BS (%)||12.0 ± 1.5||14.4 ± 1.5||7.1 ± 1.1a||12.9 ± 1.7|
|dLS/BS (%)||6.7 ± 1.4||15.8 ± 2.2a||5.8 ± 1.4||5.8 ± 1.2|
|MS/BS (%)||12.7 ± 1.9||23 ± 2.9a||9.3 ± 1.3||12.3 ± 1.6|
|OS/BS (%)||1.8 ± 0.4||4.6 ± 0.8a||2.5 ± 0.5||1.6 ± 0.4|
|Ob.S/BS (%)||9.5 ± 3.8||18.2 ± 3.2a||7.9 ± 3.2||9.6 ± 2.1|
|Oc.N/B.Ar (/mm2)||34.3 ± 5.4||25.1 ± 3.2||24.8 ± 3.6||34.7 ± 4.8|
|Oc.S/BS (%)||0.8 ± 0.1||0.6 ± 0.1||0.9 ± 0.1||0.7 ± 0.1|
|Adipocyte Nb/1.5 mm2||218.7 ± 24.1||223.6 ± 26.9||220.4 ± 38.1||265.0 ± 15.2|
Intermittent PTH(1–84), while osteoanabolic through VEGF signaling, has no proangiogenic effects on bone vascularization
As shown in Fig. 3 and described previously in detail,19 the homogeneous infusion of barium throughout the bone vasculature resulted in an excellent contrast between bone and vessel networks with both SR-µCT (A–C, E, F) and histologic (D) techniques in trabecular and cortical bone. SR-µCT allowed detailed 3D imaging of the vasculature in the femoral diaphysis (Fig. 3C), revealing various vascular structures, including the larger and smaller bone marrow blood vessels and sinusoidal lobules. The presence of sinusoidal lobules in the bone marrow has been described previously,25, 26 and we showed previously that barium contrasting allows their easy detection, imaging, and quantification.19 Figure 3E shows 3D imaging of the bone and vascular networks after segmentation in trabecular and cortical bone. The cortical bone was removed in Fig. 3F, revealing the dense vascular network in this bone compartment.
Unexpectedly, tibial metaphyseal vessel number (Ves.N) per tissue volume was lower with PTH treatment versus the control group (Fig. 4A). This difference was not present (p > 0.05) when vessel number was normalized to marrow volume in PTH and control rats (6.2 ± 0.2/mm versus 5.1 ± 0.9/mm, respectively). Vessel volume per tissue volume (Ves.V/TV) and mean vessel thickness (Ves.Th) remained similar between the two groups; however, there was a greater separation between the vessels (Ves.Sp) (Table 4). Interestingly, the number of Sin.L increased with PTH treatment (Fig. 4A), whereas their volume did not differ between the two groups (Table 4). Vascular analysis of the distal femoral metaphysis revealed, as found in the tibias, no differences in vessel volume/marrow volume (Ves.V/Mar.V) between PTH and control groups (Table 4). In contrast, in the femoral diaphyseal marrow, the vessel volume (Ves.V/Mar.V) was significantly lower (39%), and the vessels were thicker (26%) in the PTH versus control groups (Fig. 4B and Table 4), whereas in the diaphyseal cortex, no differences were observed between the two groups.
|Control||PTH||PTH + Anti-VEGF Ab||Anti-VEGF Ab|
|Ves.V/TV (%)||7.2 ± 0.7||5.4 ± 0.9|
|Ves.Th (µm)||17.9 ± 0.6||17.5 ± 1.2|
|Ves.Sp (µm)||292.7 ± 27.2||464.4 ± 72.2a|
|Sin.LV/TV (%)||2.2 ± 0.3||2.4 ± 0.3|
|Ves.V/MarV (%)||4.3 ± 0.4||5.4 ± 0.7|
|Ves.Th (µm)||55.0 ± 4.0||67.7 ± 9.0|
|Ves.V/Mar.V (%)||4.5 ± 0.8||2.7 ± 0.2a|
|Ves.Th (µm)||72.3 ± 3.1||91.2 ± 5.6a|
|Ves.V/CtV (%)||0.25 ± 0.07||0.18 ± 0.04|
|Ves.Th (µm)||18.1 ± 0.6||19.2 ± 0.7|
|Ves.V/TV (%)||8.6 ± 1.5||6.7 ± 1.1||5.4 ± 1.1||4.8 ± 0.9|
|Ves.Th (µm)||16.8 ± 0.6||17.8 ± 0.6||16.7 ± 0.6||17.1 ± 0.7|
|Ves.Sp (µm)||286.4 ± 89.8||309.3 ± 40.3||555.4 ± 162.9||517.0 ± 111.8|
|Sin.LV/TV (%)||1.7 ± 0.7||1.8 ± 0.7||0.6 ± 0.2||0.6 ± 0.5|
Effects of 15 days of PTH(1–84) and/or anti-VEGF Ab administration on bone vascularization of the secondary spongiosa of the tibial metaphysis
Fifteen days of PTH and/or anti-VEGF Ab administration did not change Ves.V/TV, the mean thickness of the vessels (Ves.Th), or the distance between individual vessels (Ves.Sp) (Table 4). As found after 30 days, the number of blood vessels present in the secondary spongiosa was decreased significantly by PTH treatment in comparison with controls (Fig. 4A). However, the anti-VEGF Ab did not prevent the PTH-induced decrease in blood vessel number. In addition, while the volume of sinusoidal lobules (Sin.LV/TV) did not differ among the groups (Table 4), their number was reduced in all treated groups versus controls (Fig. 4A).
Spatial relation between bone and vessels under PTH(1–84) and/or anti-VEGF Ab administration
Regardless of vessel size or treatment, distances between blood vessels and trabecular surfaces as a whole did not differ between or among groups after 30 or 15 days of PTH treatment (Table 5). Interestingly, when distances to osteoid seams (ie, sites of new bone formation) were specifically assessed, capillaries were closer to osteoid seams in the PTH-treated versus all other groups at both time points (Fig. 5A, B). When coadministered with PTH, anti-VEGF Ab blocked the spatial redistribution of these small vessels (Fig. 5B). In general, when data from the control 15- and 30-day experiments were pooled and vascular-trabecular and vascular-osteoid separations were compared, capillaries were spatially closer to sites of new bone formation than to resting trabecular surfaces (Table 6). In contrast, larger vessels (≥30 µm) remained at equal mean distances to active and trabecular surfaces. When the data from the PTH-treated rats were pooled (ie, 15- and 30-day experiments) and the same comparison made, capillaries and 30- to 100-µm vessels were spatially closer to sites of new bone formation than to resting trabecular surfaces (Table 6). In this instance, only vessels greater than 100 µm maintained a fixed mean distance regardless of whether the bone surface was active or nonactive. Further analyses revealed that capillaries and 30- to 100-µm vessels were closer to their respective osteoid seams than control vessels of similar size from their respective osteoid seams.
|Mean distance from total trabecular surfaces (µm)|
|≤29 µm||30 to 100 µm||101 to 200 µm|
|Control||97 ± 7||136 ± 14||183 ± 13|
|PTH||98 ± 10||141 ± 13||188 ± 13|
|Control||76 ± 8||138 ± 18||186 ± 15|
|PTH||66 ± 4||125 ± 13||166 ± 17|
|PTH + anti-VEGF Ab||89 ± 6||131 ± 13||162 ± 12|
|Anti-VEGF Ab||71 ± 10||124 ± 20||176 ± 14|
|Distance from total trabeculae||Distance from osteoid|
|≤29 µm||88 ± 6||67 ± 6a|
|30 to 100 µm||137 ± 10||126 ± 12|
|101 to200 µm||185 ± 9||176 ± 11|
|≤29 µm||84 ± 7||43 ± 6a,b|
|30 to 100 µm||134 ± 6||98 ± 9a,b|
|101 to 200 µm||179 ± 11||165 ± 13|
Pearson product-moment correlations between the vascular-osteoid separation of capillaries and bone-remodeling parameter
A negative correlation (r = –0.78, p = 0.005) between the number of osteoclasts and the osteoid-vascular distance of capillaries was observed in PTH-treated rats (Fig. 6A). In addition, positive correlations were observed between and OS/BS (PTH: r = 0.63, p = 0.04) and MAR (PTH: r = 0.64, p = 0.04 and control: r = 0.65, p = 0.02) and the osteoid-capillary distance (Fig. 6B, C).
- Top of page
- Materials and Methods
There are five major findings to this investigation: (1) A single dose of PTH(1–84) generated an early expression of VEGF A in bone, (2) 15 days of intermittent PTH(1–84) enhanced mRNA expression of Np1 and Np2 that was blunted by anti-VEGF Ab, (3) the bone gain and increase in bone formation induced by intermittent PTH(1–84) were prevented by anti-VEGF Ab therapy, (4) PTH diminished the number of bone marrow blood vessels, and (5) most intriguingly, PTH spatially redistributed the smallest vessels (≤29 µm) of the bone marrow closer to sites of new bone formation in a VEGF-dependent manner. These data provide evidence that the distance between bone marrow blood vessels and sites of new bone formation is presumably important for the magnitude of bone accrual. Thus, under PTH, vascular-osteogenic interaction and the functional capacity of the bone vasculature, that is, its ability to perfuse the skeleton adequately without angiogenesis, presumably contributed to bone remodeling through VEGF-related mechanisms.
The coupling of osteoclasts and osteoblasts during bone remodeling may be regulated by the vascular system.27 Accordingly, capillaries, critical components of bone-remodeling units (BMUs), remain within 100 µm of osteoclasts and osteoblasts of the BMU.27 The data presented in this investigation support and extend this contention. To our knowledge, this is the first investigation to demonstrate that bone marrow blood vessels have the capacity to alter their spatial location in response to stimuli and relocalize to the vicinity of bone-forming sites (Fig. 5A, B). Most interestingly, intermittent PTH administration reduced these distances even further, whereby capillaries were 17 to 75 µm away from active bone formation (Table 6). PTH-treated rats with 30- to 100-µm vessels were spatially relocated closer to active bone tissue as well. Further, these distances for capillaries and 30- to 100-µm vessels also were reduced in comparison with vessel distance from osteoid seams in control animals, illustrating once again the ability of bone vasculature to alter spatial location in response to stimuli (Table 6). Blood vessels greater than 100 µm maintained distances from bone surfaces regardless of bone activity in both control and PTH-treated rats. Vascular-osteoid distances appear decisive for optimal bone accrual; that is, the smallest distances between the bone surface and blood vessels were observed at sites where bone accrual occurred with PTH treatment.
Reduced vessel number with PTH treatment supports our hypothesis that the capillaries relocated toward the osteoid seams, particularly because these distances were not reduced for capillaries located next to quiescent trabecular surfaces. Had angiogenesis occurred site specifically (ie, at the osteoid seam), we would anticipate a fixed distance between the capillary and the bone surface following completion of the remodeling cycle, that is, a return to a quiescent state. However, capillary distance from quiescent trabeculae in control and PTH-treated rats was similar (Table 5) but significantly increased in comparison with osteoid/vessel distances (Table 6). Thus, presumably, the vascular system optimized efficiency of blood flow and nutrient exchange by relocating capillaries to sites of augmented bone metabolism. This reorganization of blood vessels closer to sites of bone formation probably is mediated by VEGF because blockade of VEGF binding to its receptors inhibited the redistribution of the vasculature (Fig. 5B) because it blunted the gain in BV/TV experienced with PTH administration alone (Fig. 2). To date, the exact mechanism(s) by which this occurs have not been identified. However, this may represent a new and exciting avenue of bone-vascular interaction.
VEGF is a known vasodilator in other tissue beds15–17 and presumably modulates the bone vasculature via similar mechanisms. In addition to being a major hormone for serum calcium homeostasis and a treatment to alleviate declines in bone mass in osteoporotic patients,28 PTH is also a known stimulator of VEGF in vitro,13 a mitogen for endothelial cells and promoter of angiogenesis. In this investigation, intermittent PTH treatment resulted in reduced vascularity coupled with augmented bone mass, which suggests that PTH modulated the functional capacity of the bone vasculature. Presumably, PTH augmented bone perfusion via enhanced vasodilation and/or diminished vasoconstriction, which may or may not be related to VEGF. Interestingly, vessel number remained depressed in rats injected only with the anti-VEGF Ab, suggesting that one or several other mediators contribute to bone blood vessel density and function. These mediators are currently unknown. Vascular function and skeletal perfusion were not measured in this investigation, and data in the literature are obscure because of the effect of PTH on skeletal perfusion. For example, PTH administration was shown to reduced both systemic blood pressure and femoral and tibial blood flows.29 Boelkins and colleagues30 demonstrated, in hens, a biphasic response to bone blood flow following PTH(1–34) administration, which corresponded to hypo- and hypercalemia. A decline in the percentage of cardiac output delivered to the combined femur, tibia, and metatarsus was observed 3 minutes (ie, the hypocalcemic state) subsequent to PTH injection but significantly increased following 30 minutes (ie, the hypercalcemic state).30 The influence of intermittent PTH(1–84) administration delivered over a 15- or 30-day period of skeletal blood flow remains unknown. Whether PTH(1–84) exerts similar bone vascular effects to those of PTH(1–34) is in question, although both isoforms of PTH decreased perfusion pressure in coronary arteries.31
VEGF A is one of the major osteogenic-angiogenic coupling factors for bone-vascular interactions during bone modeling processes.32 In this investigation, a bolus dose of PTH(1–84) induced a transient rise in mRNA expression of VEGF A, which returned to baseline levels 24 hours later. Also, following 15 days of intermittent treatment, mRNA levels of Vegfa were at baseline 24 hours after the last PTH injection. This suggests that the PTH-induced elevation of VEGF A in bone occurs within a transient window that is sustained at least for 6 hours and that may be critical for PTH anabolic effects on bone. Angiogenesis was not observed following intermittent PTH administration in this investigation. Further, administration of the anti-VEGF Ab blocked the PTH-induced bone anabolic effects but not the PTH-induced decrease in blood vessel number. Therefore, perhaps osteoblasts or preosteoblasts respond directly to VEGF by increasing their proliferation and recruitment, respectively, to activated bone surfaces.33 Indeed, we observed an inhibition of the rise in MS/BS and Ob.S/BS (ie, osteoblast recruitment) under the coadministration of anti-VEGF Ab and PTH(1–84), although this had no effect on MAR (ie, osteoblasts activity). Most interestingly, there were no differences in mRNA expression of Vegfr1 and Vegfr2 in bone. Yet a significant augmentation of the coreceptors Np1 and Np2 was observed subsequent to 15 days of intermittent PTH administration (Fig. 1B). Data in the literature indicate that NP1 and NP2, which are expressed by endothelial and other cell types, play a significant role in vasodilator capacity, for example, blockade of VEGF binding to VEGFR1 and NP1 attenuated vasodilator capacity of coronary arterioles.15 Intriguingly, blockade of VEGF signaling in the current investigation not only prevented redistribution of the bone marrow vasculature but also diminished the PTH-induced augmentation of bone volume. In addition, anti-VEGF Ab blunted the augmentation of mRNA expression of Np1 and Np2 induced by PTH (Fig. 1B) suggesting that VEGF is responsible for upregulation of neuropilin expression, as shown previously in other cell types or models.34, 35 Neuropilin 1 is also expressed in osteoblasts at the developing edge of trabeculae within the marrow cavity of young mice.36 NP1 modulates migration of cell types other than neurons (eg, endothelial cells) through mechanisms involving VEGF or class III semaphorins (ie, the other ligands of NPs), which are expressed in bone.37, 38 Therefore, PTH may augment the recruitment of osteoprogenitors toward bone-formation sites, whose migration presumably is controlled by a VEGF-NP1 signaling pathway.39
The interaction of the vascular and skeletal systems results in, among other things, nutrient and O2 delivery and waste removal during increased bone metabolism. These interactions are complex and involve factors such blood flow, angiogenesis, and osteoid-vascular separation. The mechanism(s) responsible for homing of the blood vessels to trabecular (or cortical) surfaces during bone remodeling are unknown but likely depend on the oxygen-sensing machinery in bone. This investigation demonstrated that during intermittent PTH administration, angiogenesis is not one of the mechanisms by which bone mass is augmented. In fact, reduced blood vessel number with PTH administration corresponded with enhanced bone mass and microarchitectural properties and bone static and dynamic parameters, Sin.L number, and reduced Oc.S/BS. The specific organization of a subclass of bone marrow capillaries called sinusoid lobules was reported repeatedly by anatomists,40 who used contrast infusion techniques similar to our approach. However, to date, no biologic roles for these structures have been identified. That PTH exhibits a biphasic effect on Sin.L number and anti-VEGF Ab treatment alone and is able to decrease Sin.L number suggests their potential involvement in bone vascular metabolism. Even though OcS/BS was reduced with 30 days of intermittent PTH administration, osteoclasts potentially release factors responsible for the homing of blood vessel to bone-remodeling sites. Previous investigations reported similar observations in bone-resorption parameters with PTH treatment, for example, erosion surface in 20-week-old c57BL/6 mice femurs were reduced by 50% following 5 weeks of intermittent PTH(1–34) administration.41 In a human paired biopsy study, endocortical wall thickness increased as a result of reduced eroded perimeter at this site in both men and women following 18 and 36months, respectively, of PTH(1–34) administration.42 We found a significant negative correlation between the number of osteoclasts and the osteoid-capillary distance only in PTH-treated animals, whereas a similar positive correlation was observed between MAR and the osteoid-capillary distance in both control and PTH-treated groups. These data suggest that the change in bone remodeling under PTH stimulus may be linked functionally to the relocation of vessels with regard to the bone surfaces. To establish the exact mechanism(s) of blood vessel homing to remodeling sites, more in-depth investigations are required.
Osteoblasts and adipocytes are derived from common mesenchymal progenitor cells. PTH/PTHrP signaling is a major ontogenic regulator of the bone marrow and stromal tissue. In vitro, intermittent PTH treatment inhibits adipocyte differentiation of human bone marrow stromal cells,43 and in vivo, Kuznetsov and colleagues44 reported that 3-month-old mice expressing constitutively active PTH/PTHrP receptors lacked bone marrow adipocytes. Finally, VEGF blockade was associated with inhibited peripheral adipocyte differentiation in vivo.45 In this study, adipocyte number was reduced with long-term (ie, 30 days) PTH treatment and was not affected by short-term (ie, 15 days) VEGF blockade. These kinetic responses suggest that the early increase in osteoblast recruitment at the bone surface presumably is independent of reduced commitment of mesenchymal progenitor cells toward the adipocyte lineage, as suggested by Dobnig and Turner.46 Further experiments are needed to explore the relation between PTH-induced VEGF production and diminished number of marrow adipocytes observed after long-term intermittent PTH administration.
A limitation to this study is the use of a growing rat model (ie, age of 3 to 4 months), whereby the age-related growth of the animal erroneously may augment the magnitude of the PTH-induced bone anabolic response. For example, Friedl and colleagues47 demonstrated age-associated declines in BFR in Sprague-Dawley rats with BFR 3.5 times and 13 times lower at 3 and 13 months, respectively, than at 1 month. However, this was reversed with intermittent PTH administration, whereby bone formation was augmented as a function of age with 1.5-, 3-, and 4.7 fold increases in BFR, respectively.47 In addition, the mechanisms by which PTH induced bone growth were altered. Osteoblast cell activity, as measured by MAR, was highest in the 1-month-old group and declined as a function of age and PTH treatment. However, there were significant PTH-associated enhancements in Ob.S/BS (%) as a function of age.47 In this investigation, both MAR and Ob.S/BS (%) were augmented versus controls following 30 days of PTH treatment, which represents both increased osteoblast activity and increased recruitment, respectively. However, following 15 days of PTH treatment, osteoblast recruitment (Ob.S/BS) was augmented versus all other groups, whereas osteoblast activity (ie, MAR) did not differ among groups. While the age of the rats between this study and that of Friedl and colleagues47 were similar, the divergent results may stem from the length of the experimental protocols (15 and 30 days versus 7 days), use of different PTH analogues (1–84 versus 1–34), and rat species (Wistar versus Sprague-Dawley).
In conclusion, we observed a decrease in bone marrow blood vessel density following intermittent PTH administration and a VEGF-dependent redistribution of the smallest bone marrow blood vessels closer to sites of new bone formation. These results suggest that distances between vessels and active bone sites are critical for overall bone accrual.
- Top of page
- Materials and Methods
All the authors state that they have no conflicts of interest.
- Top of page
- Materials and Methods
We thank C Olivier and E Boller for their help in SR-µCT image acquisition at the ESRF.
This study was supported by Agence Nationale de la Recherche, France.
Authors' roles: Study design: RP and M-HL-P. Study conduct: RP. Data collection: RP, AG, AV-B, FM-W, M-TL, MT, NL, ML, Z-AP, and FP. Data analysis: RP, FM-W, M-TL, ML, Z-AP, and FP. Data interpretation: RP, M-HL-P, ML, Z-AP, and FP. Drafting manuscript: RP. Revising manuscript content: RP, M-HL-P, LM, AG, FM-W, ML, Z-AP, and FP. Approving final version of manuscript: RP, AG, AV-B, FM-W, M-TL, MT, NL, LM, ML, Z-AP, FP, LV, and M-HL-P. RP takes responsibility for the integrity of the data analysis.
- Top of page
- Materials and Methods
- 16Effects of coronary artery disease on expression and microvascular response to VEGF. Am J Physiol. 1998; 275: H1411–H1418., , , , .
- 28Parathyroid hormone treatment for osteoporosis. In: Miller P, Papapoulos S, (eds.) Primer on the Metabolic Bone Disease and Disorders of Mineral Metabolism Washington DC, USA: American Society for Bone and Mineral Research; 2008. pp. 244–249., .
- 32Vascular endothelial growth factor and osteogenic-angiogenic coupling. In: Bilezikian J, Raisz LG, Martin TJ, (eds.) Principles of Bone Biology San Diego, CA, USA: Elsevier; 2008. pp. 1133–1144..
- 40Blood Supply of Bone: Scientific Aspects London: Springer-Verlag; 1998., .
- 44The interplay of osteogenesis and hematopoiesis: expression of a constitutively active PTH/PTHrP receptor in osteogenic cells perturbs the establishment of hematopoiesis in bone and of skeletal stem cells in the bone marrow. J Cell Biol. 2004; 167: 1113–1122., , , , , , , , , .