The cellular and molecular events triggering the anabolic response of the skeleton to exogenous parathyroid hormone (PTH) are not well understood. Despite the numerous bone mass studies in rats, few data are available for mice. Therefore, we treated 10-week-old female intact C57BL/6J mice with human PTH(1-34) delivered subcutaneously at a dose of 40 μg/kg per day 5 days a week for 3 weeks and 7 weeks. Bone mineral density (BMD) of total bone, femur, tibia, and lumbar vertebrae was measured weekly by PIXImus. Bone turnover was examined by histomorphometry, and gene expression of bone formation and resorption markers and osteoclastogenesis regulators in the excised femur and tibia was assessed by reverse-transcription polymerase chain reaction (RT-PCR) at 3 weeks and 7 weeks. The PTH-stimulated increase in BMD was more prominent in the tibia and femur than in the lumbar vertebrae, with an anabolic effect detected within 1-2 weeks and BMD continuing to increase. The appearance of a detectable PTH-stimulated increase in BMD was slower in the lumbar vertebrae where the increase was only significant after 7 weeks of treatment. Histomorphometric analysis of the proximal tibia at both 3 weeks and 7 weeks indicated significant time-dependent increases in trabecular area, trabecular number, trabecular and cortical widths, and osteoblast and osteoid perimeters. In the lumbar vertebrae, these stimulatory effects of PTH on trabecular area, trabecular number, and cortical width were smaller and not detected until 7 weeks. PTH-stimulated increases in bone turnover were evident by increased gene expression of osteocalcin (OC), tartrate-resistant acid phosphatase (TRAP), and receptor of activator nuclear factor κB (NF-κB) ligand (RANKL) in the tibia and femur. No significant difference in gene expression was observed between the two long bone sites. In conclusion, PTH exerts an anabolic action at the tissue and cellular levels in intact mice and the magnitude and temporal pattern of this anabolic action, as assessed by densitometry and histomorphometry, are skeletal site specific.
PARATHYROID HORMONE (PTH) is a major regulator of calcium and phosphate homeostasis. Therapeutically, PTH has been shown to be a potent anabolic agent in humans,(1–3) rats,(4–9) rabbits,(10) and nonhuman primates(11) when it is administered intermittently. In rat models, we(4,5) and others(6–9) have shown that PTH stimulates net bone formation and that PTH restores loss of cancellous bone and trabecular connectivity in the ovariectomized (OVX) rats by itself and when combined with estrogen.(4) The mouse animal model offers several advantages for the study of the mode of action of anabolic agents at the molecular level, including the ability to produce transgenic animals. However, until recently, it has been difficult to measure bone density and conduct histomorphometric studies in mice in vivo because of the smaller skeletal size. Recently, several studies have indicated the feasibility of skeletal manifestations in mice.(12–15) For example, Mohan et al. have shown that PTH(1-34) produces a significantly greater femoral bone mineral density (BMD) response in mice than the truncated N-terminal fragment PTH(1-31) by dual-energy X-ray absorptiometry (DXA).(12) Similarly, Andersson et al. have reported that human PTH(1-84) rescues the ovariectomy-induced bone loss in mice(13) assessed by peripheral quantitative computed tomography (pQCT). However, a number of questions remain to be answered before the utility of this rodent model can be evaluated fully. For example, it needs to be defined clearly how BMD changes with age in mice in vivo, when peak bone mass is achieved, and how the responsiveness to anabolic agents such as PTH differs depending on mouse strain and skeletal site.(15–19)
There seem to be differences in the responsiveness of different skeletal sites to PTH in humans and rats. PTH is more responsive in the lumbar spine than the femoral neck in humans.(1,2) Similarly, the response to PTH was much greater in the moderately osteopenic lumbar vertebra than in the severely osteopenic proximal tibia in aged OVX rats.(8) On the other hand, recently, it has been shown that PTH stimulates cancellous bone formation at all skeletal sites in both young and old OVX rats regardless of marrow composition.(9) It is still unclear how the murine skeleton responds to PTH or whether there is any site-specific effect on the PTH anabolic action in mice.
In addition, it has not been determined how gene expression of bone markers such as osteocalcin (OC) and tartrate-resistant acid phosphatase (TRAP) might be altered by PTH treatment in mice in vivo. Accordingly, we examined the temporal changes in BMD in vivo in intact C57BL/6J female mice between 10 and 17 weeks of age, when animals are growing and reaching maturity. In parallel, we also examined the effect of PTH treatment on bone densitometry, histomorphometry, and gene expression of bone formation and resorption markers.
MATERIALS AND METHODS
Human PTH(1-34) fragment was purchased from Sigma Chemical Co. (St. Louis, MO, USA). Chemicals for cell biology and molecular biology assays were purchased from Sigma Chemical Co. and Fisher Scientific (Pittsburgh, PA, USA). The RNase H-free reverse transcriptase SuperScript II and deoxynucleoside triphosphates (dNTPs) were purchased from Life Science Technologies (Stillwater, NJ, USA). Restriction enzymes were purchased from New England BioLab (Waltham, MA, USA) and Promega (Madison, WI, USA). DNA Mediprep purification kit and DNA gel extraction kit were purchased from Qiagen (Valencia, CA, USA). Taq polymerase or polymerase chain reaction (PCR) kit, oligonucleotide primer pairs for β-actin were purchased from Clontech (La Jolla, CA, USA), and customized oligonucleotides were synthesized by Gene Link, Inc. (Hawthorne, NY, USA).
Seven-week-old, virgin female C57BL/6J mice were purchased from the The Jackson Laboratory (Bar Harbor, ME, USA) and stabilized at the Animal Research Facility of Helen Hayes Hospital for at least 1 week before the start of the experiment. Three to four mice were housed per cage, given free access to water, and fed a standard diet (5001; Purina Mills, St. Louis, MO, USA) in a room maintained at 22°C with 60-75% humidity on a 12-h light/dark cycle.
The experimental protocol is shown in Fig. 1 and was approved by the Institutional Animal Care and Use Committee (IACUC) of Helen Hayes Hospital. Thiry-four 10-week-old animals were randomly divided into two groups; one group (n = 18) was treated with PTH, 40 μg/kg per day 5 days a week by subcutaneous (sc) injections, and the other group (n = 16) was similarly treated with vehicle as described in the following section. Each group was divided further into two subgroups, which were killed at 3 weeks and 7 weeks after the initiation of PTH treatment, respectively. Control and PTH-treated animals were anesthetized alternately for the final BMD measurement and death. Immediately after the death of each animal, right femur and tibia were cleaned quickly in ice-cold phosphate-buffered saline (PBS; pH 7.4) by removing all connective tissues and weighed, frozen, and stored in liquid nitrogen until use for RNA extraction. Left femur, tibia, and lumbar vertebrae were cleaned and fixed in 4% paraformaldehyde (PFA) in PBS overnight for histomorphometric analysis.
Preparation of PTH solution
PTH was dissolved in sterile 10 mM acetic acid and aliquoted as a 400-μg/ml stock solution and stored at −20°C until use. The aliquoted PTH was freshly diluted to 8 μg/ml in sterile PBS, and animals were injected sc with 0.8 μg of PTH per 20 g/0.1 ml (40 μg/kg). For control animals, PBS containing an equivalent volume of 10 mM of acetic acid was injected as a vehicle.
BMD measurement by PIXImus
Mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (0.1 mg/kg) in PBS and placed prone on the platform of the PIXImus (GE Lunar Corp., Madison, WI, USA) for BMD measurements once a week according to the manufacturer's instruction using mouse-specific software (version 1.43). In some experiments, the variability in measurements was examined by repeating scans after repositioning the animals. Intra- and interobserver variations were examined also and were found to be very small, suggesting high precision as reported elsewhere.(20) Percent CV of BMD for the repeated scans was 1-3% at all four skeletal sites examined.(20) These CVs are compatible with those in humans.(21,22) Moreover, for the analysis of more localized regions such as the distal femur and proximal tibia, CV% of BMD for the repeated scans after repositioning was as small as 2-5%.(20)
To determine the anabolic rate (mg/cm2; increase in BMD per week), the BMD values from each individual animal at different time points were plotted for three skeletal sites (tibia, femur, and lumbar vertebrae). Because some animals were killed at 3 weeks, values from the four time points (0, 1, 2, and 3 weeks) were used to determine the mean slopes of the initial anabolic rates up to 3 weeks. The slope value of each animal was estimated by plotting computer-analyzed linear regression curve, and the mean of slopes in each group up to 3 weeks was determined.
Histomorphometric measurement of the proximal tibias and lumbar vertebrae
Excised left proximal tibias and third lumbar vertebrae were cut longitudinally to expose the bone marrow with a low-speed metallurgical saw and dehydrated in graded ethanol, defatted in toluene, and embedded in methylmethacrylate. Four-micrometer-thick sagittal sections of the central regions of proximal tibias and lumbar vertebral bodies were cut with a Polycut microtome (Reichert-Jung, Heidelberg, Germany) and stained with Goldner's trichrome. Histomorphometry of the proximal tibia was performed on the cancellous and cortical bone in an area between the lowest point of the growth plate and 1.2 mm distal to the growth plate. Histomorphometry of lumbar vertebra was carried out in a region 0.1 mm from the cranial and caudal growth plates. The variables of cortical width, cancellous bone area, and perimeter and cancellous osteoid, osteoblast, and osteoclast perimeters were measured using a digitizing image analysis system and a morphometric program OsteoMeasure (OsteoMetrics, Atlanta, GA, USA) at a magnification of ×200. All parameters were calculated and expressed according to standard formulas and nomenclature.(23)
Total RNA was extracted using PURESCRIPT (Gentra Systems, Minneapolis, MN, USA) according to the manufacturer's instructions. The right femora or tibias were ground with a sterile mortar and pestle in liquid nitrogen and homogenized on ice with five strokes in a Lysis buffer, and the extracted RNA was precipitated with 2-propanol, washed with 70% ethanol, air-dried, and reconstituted in DNase- and RNase-free water. After determining the concentrations of RNA spectrophotometrically at 260 nm, the reconstituted RNAs were stored in 2-propanol at −80°C until use for RT-PCR.
Two micrograms of RNA was reverse-transcribed at 42°C for 50 minutes using SuperScript II in the presence of oligo dT primers (0.5 μg), dNTPs (10 μM each), and dithiothreitol (DTT; 1 mM). The reaction was terminated by heating the mixture at 70°C for 15 minutes. The RT products were first treated with RNase at 37°C for 30 minutes. However, we used the RNase-untreated RT products as a template, because the preliminary data showed that there was no significant difference between RNase-treated or -untreated samples (data not shown). One hundred nanograms of the RT products were then amplified by PCR, using Taq polymerase. The primer pairs for OC (forward primer, 5′-CCATCTTTCTGCTCACTCTGCTG-3′, and backward primer, 5′-CTTCAAGCCATACTGGTCTGATAGC-3′, predicted size 245 base pairs [bp]), TRAP (forward primer, 5′-ACTTGCGACCATTGTTAGCCAC-3′, and backward primer, 5′-GTTGGGGACCTTTCGTTGATG-3′, predicted size 161 bp), calcitonin receptor (CTR; forward primer, 5′-AGTTCTTCAGGCTCCTACCAATCTC-3′, and backward primer, 5′-CTGTCAGGGTGTCTAAACCACTCTC-3′, predicted size 309 bp), osteoprotegerin (OPG; forward primer, 5′-CCTTCTTCAGGTTTGCTGTTCCTAC-3′, and backward primer, 5′-GCAGGTCTTTCTCGTTCTCTCAATC-3′, predicted size 344 bp), and receptor of activator nuclear factor κB (NF-κB) ligand (RANKL; forward primer, 5′-AGCACGAAAAACTGGTCGGG-3′, and backward primer, 5′-AAGGGTTGGACACCTGAATGC-3′, predicted size112 bp) were specifically designed by using MacVector software (6.5 version; Oxford Molecular Ltd., Madison, WI, USA) and National Center of Biotechnology Information (NCBI) blast search to obtain the specific sequences with the most conserved region of the target proteins among species. Primers for β-actin (forward primer, 5′-GTGGGCCCGC TCTAGGCACCAA-3′, and backward primer, 5′-CTCTTTGATGTCACGCACGATTTC-3′, predicted size 540 bp; Clontech) were used for standardization of the band intensity. The PCR was conducted generally at 95°C for 5 minutes for the first denaturing and then at 95°C for 45 s, at 55-65°C for 45 s for annealing, and at 72°C for 1-2 minutes for extension for 30-36 cycles. The last extension was generally 7 minutes.
All values for BMD measurement represent means ± SEM of at least eight animals. For the weekly changes in BMD, bone mineral content (BMC), bone area, body weights (n = 8-18), and histomorphometric analysis (n = 4-6) two-way analysis of variance (ANOVA) were applied to determine the significance of differences in treatment using the Number Cruncher Statistical System 2000 version (NCSS, Salt Lake City, UT, USA).(24) To determine the statistical significance of difference in the anabolic rates between control and the PTH-treated animals, means of the slopes in control and PTH-treated groups were calculated and statistically analyzed using a two-sample t-test.
Changes in body weight with age
Body weight increased between ages 10 and 14 weeks and then reached a plateau after 14 weeks (21-24 g). There were no significant differences in body weight between control and PTH-treated groups at any time points.
Effects of hPTH(1-34) on BMD and bone mineral content
Figure 2 shows the temporal changes in BMD (Fig. 2A, left panel) and bone mineral content (BMC; Fig 2B, right panel) in control (open circles) and PTH-treated animals (closed circles) at the tibia, femur, lumbar vertebrae, and total body skeletal sites, respectively. There were no significant differences in BMD or BMC between control and PTH-treated groups at baseline (10 weeks) at any skeletal site.
An anabolic effect of PTH (increases in BMD) was evident within 1 week at all skeletal sites except the lumbar vertebrae, but at 1 week the difference was significant only in the tibia. By 2 weeks, a significant increase in BMD over control was observed in the tibia, femur, and total body bone. However, there was no significant increase in BMD in the lumbar vertebrae until 7 weeks (Fig. 2A). There was a trend of slight decline of BMD ∼15-17 weeks of age in both control and/or PTH-treated animals in the tibia, femur, and lumbar vertebrae.
PTH also increased BMC at all the skeletal sites examined (Fig. 2B). When the BMC values were corrected by body weight, the trend was similar to that of noncorrected values (data not shown). There were no significant changes in bone area or fat percent at any skeletal sites or at any time points (data not shown).
Figure 3 shows the mean slopes for the initial anabolic rate for 3 weeks in control (solid line) and PTH-treated (broken line) animals at three sites. PTH increased the anabolic rate by 70.7% (2.0499 ± 0.1133 mg/cm2/week vs. control 1.2006 ± 0.1677 mg/cm2/week; p < 0.001) and 85.4% (2.4816 ± 0.2642 mg/cm2/week vs. control 1.3388 ± 0.1312 mg/cm2/week; p < 0.001) in the tibia and femur, respectively. There was no significant difference in the BMD increasing rate in the lumbar vertebrae between control (1.3246 ± 0.2886 mg/cm2/week and PTH-treated animals (1.6935 ± 0.3268 mg/cm2/week). Moreover, there was no significant difference in the anabolic rate in control animals among the three different skeletal sites.
Histomorphometry of proximal tibias and lumbar vertebrae
Table 1 summarizes the histomorphometric data in the proximal tibia and lumbar vertebrae, respectively.
Table Table 1.. Histomorphometric Analysis of the Proximal Tibia and the Lumbar Vertebrae
In control mice, there were no statistically significant, age-related differences in cancellous bone volume, trabecular or cortical width, osteoclast, or osteoblast perimeters in the proximal tibia between 13 and 17 weeks of age. However, osteoid perimeter was significantly higher in the older mice than in the younger ones. PTH significantly increased trabecular area, trabecular number, and osteoblast perimeter after 3 weeks of treatment. These anabolic effects of PTH were more prominent after 7 weeks. At 7 weeks, there also were significant increases in trabecular and cortical widths. In the lumbar vertebrae, where bones were examined only at 7 weeks; the anabolic effects of PTH were less marked than those in the proximal tibia. Similar trends were seen in trabecular area, trabecular number, and osteoblast perimeter, but only the increase in cortical width was significant.
PTH significantly increased OC gene expression at 3 weeks both in the tibia and in the femur (Fig. 4). However, by 7 weeks, basal OC expression had declined and there was no significant increase with PTH treatment in the tibia. In the femur, however, PTH significantly increased OC expression at both 3 weeks and 7 weeks. PTH also stimulated expression of the bone resorption marker gene TRAP at 3 weeks both in the tibia and in the femur, whereas PTH had no effect on CTR. However, at 7 weeks, there was no stimulation of TRAP expression by PTH in the tibia, whereas there was a slight increase in TRAP expression in the femur. Gene expression of RANKL, a stimulator of osteoclastogenesis, was significantly increased by PTH at both time points in the tibia and the femur, whereas there was no significant difference in OPG (a decoy receptor for RANKL) expression at either site. Table 2 summarizes the PTH action on gene expression of bone markers.
Table Table 2.. Summary of the PTH-induced Changes in Gene Expression
In this study, we have shown the feasibility of using mice as a tool for longitudinal studies of the skeleton. We also have shown that a daily injection of human PTH(1-34) fragment (PTH) increased gene expression of the bone formation and resorption markers in postadolescent, adult virgin female C57BL/6J mice. The bone anabolic action occurred quickly (within 7 days) and generally increased linearly with duration of PTH treatment up to at least 3 weeks. However, the magnitude and temporal pattern of the anabolic effects of PTH were skeletal site specific at the tissue and cellular levels.
To characterize the longitudinal changes in BMD with age and under the influence of PTH, we chose C57BL/6J mice because this strain is widely used and is a base for many transgenic mice. The C57BL/6J strain appeared to be suitable for rapid temporal detection of the anabolic actions of PTH, although this strain has the lowest BMD among several different strains.(18) An anabolic response to PTH was detected within 1-2 weeks at most skeletal sites. However, after 15 weeks of age, BMD plateaued or slightly declined both in control and in PTH-treated animals, suggesting that age-related effects occur about this age in this strain. Moreover, because we only examined short-term (7 weeks maximum) effects of PTH on BMD, it is not clear if the anabolic effects will be seen long term and whether a continuous infusion of PTH also can produce a similar effect.
PTH significantly increased BMD at all sites, including the tibia, femur, and total body, similarly to the effects observed in humans and rats.(1–5) PTH also significantly increased BMC in the tibia but there was no significant change in BMC in the femur or total body. However, there was a noticeable trend toward increased BMC values in the PTH-treated animals at all time points in both femur and total body bone. There was no significant difference in the bone area between the control and PTH-treated animals (data not shown), suggesting that the PTH-stimulated increase in BMD may be caused by, solely, the increased BMC in the tibia. In addition, magnification error, projection error, and other factors affecting the imaging fields and edge detection of PIXImus have a greater impact on BMC and bone area than BMD.(25–27) Consistent with this is the smaller CV% for BMD than that for BMC in mice in vivo.(20)
The BMD response to PTH was detected more rapidly in the long bones, tibia, and femur than the lumbar vertebrae. Detailed analysis showed that the initial anabolic rates were significantly higher in the tibia and femur than those at the lumbar vertebrae (Fig. 3), confirming skeletal site specificity in the responsiveness to PTH. Overall, there appeared to be no significant difference in the anabolic action of PTH between tibia and femur (Fig. 2). Thus, it can be concluded that PTH exerts its anabolic action most rapidly and profoundly in the long bones, perhaps related to weight bearing.
Skeletal site specificity in response to PTH is well known in humans, where the anabolic effect of PTH is first detected in the spine, followed by the hip and then total body.(1–3) In aged OVX rats, Qi et al. have shown that PTH markedly stimulates bone formation and restores the lost cancellous bone in the lumbar vertebrae of aged mice, but it fails to restore completely the lost bone in the severely osteopenic proximal tibia.(8) More recently, Li et al. have examined the site-specific effects of PTH according to the differences in marrow composition using young and old OVX rats and have shown that PTH (80 μg/kg per day for 6-10 weeks) stimulates cancellous bone formation at all skeletal sites regardless of marrow composition.(9) However, they have suggested that the magnitude of the responsiveness to PTH may somewhat vary among skeletal sites in these.(9) The difference of the initial target site of PTH anabolic action between mice (long bones) and humans (spine) may be caused by the difference in posture. Skeletal site specificity in mice was confirmed also by the histomorphometric data. At 7 weeks, a robust anabolic action was observed in the proximal tibia, whereas the effect was much less in the lumbar vertebrae (Table 1). Possibly, the skeletal site where the most mechanical stress or weight bearing occurs may be the primary target of anabolic action of PTH. In support of this observation, recently, two studies have shown synergy between the effects of PTH and mechanical stress.(28,29)
Our RT-PCR results showed that PTH significantly increased gene expression of OC, a bone formation marker, and TRAP, a bone resorption marker, both in the tibia and in the femur, indicating PTH-stimulated bone turnover at the molecular level. Increased OC messenger RNA (mRNA) expression is consistent with the finding in rats by Schmidt et al.(30) Although there were certain slight differences in the temporal patterns in OC and TRAP gene expression between the tibia and femur, the trends appeared to be similar. It needs to be determined whether the differences in PTH-stimulated gene expression can be correlated to the differences in PTH-stimulated increases in BMD at different skeletal sites. Unfortunately, in this study, most specimens of the lumbar vertebrae were processed for histomorphometric analyses, so it is unclear how PTH regulates gene expression of bone markers in the mouse lumbar vertebrae. A study to determine whether any skeletal site specificity exists for the PTH anabolic action at the molecular level is currently in process.
PTH also significantly stimulated gene expression of RANKL(31) in mice in vivo. This is consistent with the findings by Suda et al., who have reported that osteotropic factors, including PTH, up-regulate mRNA expression of RANKL in osteoblasts/stromal cells.(32) Moreover, Itoh et al. have shown that PTH up-regulates mRNA expression of RANKL but down-regulates mRNA expression of its decoy receptor OPG in human osteosarcoma SaOS-4/3 cells highly expressing recombinant PTH/PTH-related peptide (PTHrP) receptors.(33) In this study, we killed animals at 22-28 h after the last injection of PTH or vehicle. To determine whether we may have missed any early effects of PTH on gene expression of these markers, we examined the effects of shorter treatments with PTH for 1, 3, 6, and 24 h and found that PTH did not significantly change gene expression before 24 h (data not shown). Thus, the present data show that PTH turns on gene expression of both bone formation and resorption markers and an osteoclast differentiation factor at 3 weeks and 7 weeks in mice.
The results from our histomorphometric analysis confirmed the observation of increased gene expression of bone formation markers at the molecular level. Although PTH also increased TRAP gene expression, histomorphometric analysis revealed that PTH slightly but not significantly reduced osteoclast perimeter percent at 7 weeks in the proximal tibia. This mainly is because the value is expressed as percent of osteoclast perimeter per unit of bone surface. The raw values of osteoclast perimeter were increased in the PTH-treated group as well as bone surfaces and osteoblast perimeter, whose increases were more marked. As a result, percent of osteoclast perimeter was observed as a slight decrease.
Accumulating evidence using gene ablation experiments suggests that the anabolic action of PTH is predominantly induced by its cognate receptor activation of cyclic adenosine monophosphate (cAMP)/adenylyl cyclase (AC)/protein kinase A (PKA) pathway,(34–37) but also may be regulated permissively by Ca/phospholipase C (PLC)/PKC pathway.(38) Interestingly, unlike PTHrP or PTH/PTHrP receptor null mice, PTH null mice live considerably normal lives with increased trabecular number and cortical width, although bone turnover is significantly reduced.(39,40) However, these genetic ablated mice have abnormal physiology. Longer and/or continuous administration of intact biologically active PTH fragments and/or specific fragments of PTH in intact mice with different ages using a similar experimental protocol may provide additional information regarding the molecular mechanism of PTH anabolic action in normal physiology.
In conclusion, this study has shown that mice are a good model for the longitudinal study on skeleton, that PTH exerts an anabolic action at the tissue and cellular levels in intact mice, and that the magnitude and temporal pattern of its anabolic action are skeletal site specific.
We thank Mr. Ken Yokoyama for his assistance in analysis of densitometric scans and statistical analyses and Mr. S. Daisuke Iida-Klein for his assistance in biochemical assays and densitometric measurement. This work was supported by the National Institutes of Health (NIH) grant AR39191 (to R.L.) and by an unrestricted research grant from the Bristol Myers Squibb Co. Foundation.
Presented in part at the 22nd annual meeting of the American Society for Bone and Mineral Research, Toronto, Canada, 2000.