Drs Brommage, Liu, and Hanson are employees of Lexicon Genetics. All other authors state that they have no conflicts of interest.
PTH Stimulates Bone Formation in Mice Deficient in Lrp5†
Version of Record online: 4 DEC 2006
Copyright © 2007 ASBMR
Journal of Bone and Mineral Research
Volume 22, Issue 3, pages 394–402, March 2007
How to Cite
Iwaniec, U. T., Wronski, T. J., Liu, J., Rivera, M. F., Arzaga, R. R., Hansen, G. and Brommage, R. (2007), PTH Stimulates Bone Formation in Mice Deficient in Lrp5. J Bone Miner Res, 22: 394–402. doi: 10.1359/jbmr.061118
- Issue online: 4 DEC 2009
- Version of Record online: 4 DEC 2006
- Manuscript Accepted: 30 NOV 2006
- Manuscript Revised: 26 OCT 2006
- Manuscript Received: 8 JUN 2006
- Lrp5 knockout;
Lrp5 deficiency decreases bone formation and results in low bone mass. This study evaluated the bone anabolic response to intermittent PTH treatment in Lrp5-deficient mice. Our results indicate that Lrp5 is not essential for the stimulatory effect of PTH on cancellous and cortical bone formation.
Introduction: Low-density lipoprotein receptor–related protein 5 (Lrp5), a co-receptor in canonical Wnt signaling, increases osteoblast proliferation, differentiation, and function. The purpose of this study was to use Lrp5-deficient mice to evaluate the potential role of this gene in mediating the bone anabolic effects of PTH.
Materials and Methods: Adult wildtype (WT, 23 male and 25 female) and Lrp5 knockout (KO, 27 male and 26 female) mice were treated subcutaneously with either vehicle or 80 μg/kg human PTH(1-34) on alternate days for 6 weeks. Femoral BMC and BMD were determined using DXA. Lumbar vertebrae were processed for quantitative bone histomorphometry. Bone architecture was evaluated by μCT. Data were analyzed using a multiway ANOVA.
Results: Cancellous and cortical bone mass were decreased with Lrp5 deficiency. Compared with WT mice, cancellous bone volume in the distal femur and the lumbar vertebra in Lrp5 KO mice was 54% and 38% lower, respectively (p < 0.0001), whereas femoral cortical thickness was 11% lower in the KO mice (p < 0.0001). The decrease in cancellous bone volume in the lumbar vertebrae was associated with a 45% decrease in osteoblast surface (p < 0.0001) and a comparable decrease in bone formation rate (p < 0.0001). Osteoclast surface, an index of bone resorption, was 24% lower in Lrp5 KO compared with WT mice (p < 0.007). Treatment of mice with PTH for 6 weeks resulted in a 59% increase in osteoblast surface (p < 0.0001) and a 19% increase in osteoclast surface (p = 0.053) in both genotypes, but did not augment cancellous bone volume in either genotype. Femur cortical thickness was 11% higher in PTH-treated mice in comparison with vehicle-treated mice (p < 0.0001), regardless of genotype.
Conclusions: Whereas disruption of Lrp5 results in decreased bone mass because of decreased bone formation, Lrp5 does not seem to be essential for the stimulatory effects of PTH on cancellous and cortical bone formation.
Low-density lipoprotein receptor–related protein 5 (LRP5) is a cell surface Frizzled co-receptor for Wnt proteins involved in activating the canonical Wnt/β-catenin signaling pathway.(1,2) Inactivation of LRP5 through nonsense, frameshift, and missense mutations results in osteoporosis pseudoglioma syndrome (OPPG),(3) a rare disease characterized by low bone mass and congenital or early-onset blindness. Disruption of the mouse Lrp5 gene results in similar skeletal and ocular phenotypes.(4,5) Another protein, LRP6, is believed to fulfill the functions of LRP5 in all nonskeletal, nonocular tissues.
The importance of LRP5 in determining bone mass is highlighted by the independent discovery by two groups that healthy subjects with extremely high bone mass have an activating G171V mutation in LRP5.(6,7) Subsequent work has identified additional activating mutations,(8) with all mutations described to-date located in the first of four extracellular WV “propeller” regions of LRP5. Transgenic mice expressing Lrp5 with this G171V mutation have similarly high bone mass.(9) In addition, genetic manipulation of proteins that modulate the canonical Wnt/β-catenin signaling pathway also modulate bone mass. Cells expressing the activating LRP5 mutation do not respond to the secreted Wnt antagonist dickkopf (DKK)-1,(6,10) and overexpression of DKK-1 by osteoblasts has been implicated in the bone loss observed in patients with multiple myeloma.(11) In addition to modulation by dickkopfs, Wnt signaling is inhibited by five secreted Frizzled related proteins (sFRPs)(12) and sclerostin.(13,14) Mice with disruption of the sFRP1 gene develop high bone mass after 6 months of age.(15) Humans(16) and mice(17) with disruptions in sclerostin have extremely high bone mass. Furthermore, mice harboring either gain-of function or loss-of-function mutations of β-catenin in osteoblasts exhibit high and low bone mass phenotypes, respectively.(18) Inhibiting glycogen synthase kinase-3β, a component of Wnt/β-catenin signaling, by treating mice with lithium increases bone mass.(19)
Lrp5 knockout (KO) mice have decreased osteoblast proliferation and bone matrix synthesis,(4) whereas increased osteoblast number and life span are observed in the Lrp5 G117V transgenic mice.(9) The osteopenia of Lrp5-deficient mice can be completely rescued when these animals are crossed with mice harboring a constitutively active allele of β-catenin in osteoblasts, supporting the hypothesis that Lrp5 signals through the canonical Wnt/β-catenin pathway.(20) Based on these observations, it would be anticipated that endogenous factors that regulate the expression of Lrp5 would have the potential to influence bone formation through this pathway. PTH has been shown to regulate the expression of components of the Wnt signaling pathway in cultured bone cells and in vivo,(21–23) suggesting that some of its actions on bone may involve Lrp5.
PTH is an important physiological regulator of mineral homeostasis and has bone anabolic and catabolic actions. When administered intermittently, PTH induces substantial increases in osteoblast surface, osteoid surface, and osteoid volume in animals, resulting in increased bone mass and strength.(24–27) In humans, clinical studies have shown that PTH increases bone mass and reduces vertebral and nonvertebral fractures.(28) Based on these findings, PTH was approved by the FDA for treatment of established osteoporosis.(29)
Whereas the skeletal benefits of intermittent PTH are well recognized, the molecular mechanisms underlying its anabolic actions on bone are not completely understood. The purpose of this study was to use Lrp5-deficient mice to evaluate the potential requirement for this gene in mediating PTH-induced bone formation.
MATERIALS AND METHODS
Generation of Lrp5-deficient mice
Mice homozygous for the Lrp5 KO allele were generated at Lexicon Genetics (The Woodlands, TX, USA). Endogenous Lrp5 gene transcription was disrupted using retrovirus-mediated gene trapping, described in detail elsewhere.(30–33) Briefly, from a gene trap library (OmniBank) containing 271,860 sequence-tagged mouse (129SvEv) embryonic stem cell clones, a clone containing an insertion in the Lrp5 gene (OST164883) was injected into C57BL/6J-TyrcBrd host blastocysts. The precise genomic insertion site (located within the first intron) was determined using inverse genomic PCR. Chimeric mice were generated and bred to C57BL/6J-TyrcBrd mice, with the resulting Lrp5+/− offspring interbred to produce F2 homozygous Lrp5 KO and wildtype (WT) littermates for examination. Routine genotyping was performed by quantitative PCR. Disruption of Lrp5 gene expression was confirmed in kidney and spleen (Fig. 1) of a homozygous KO mouse with RT-PCR (30 cycles) using oligonucleotide primers (A, 5′-GCCGCTGCTGCTGCTGGTGCTGTACTGC-3′ and B, 5′-GGTGACACGAGGCCCGAGATGACAATGTTC-3′) complementary to exons flanking the insertion site.
The mice were housed in plastic shoebox cages with four to five animals per cage. Food (LabDiet Autoclavable Mouse Breeder Diet 5021) and acidified water were provided ad libitum to all animals under controlled room temperature with a 12-h/12-h light/dark cycle. The mice were maintained in accordance with the NIH Guide for the Care and Use of Laboratory Animals, and the experimental protocol was approved by the Institutional Animal Care and Use Committee at Lexicon Genetics, where the study was conducted.
Male and female mice at 20–22 weeks of age were chosen to minimize longitudinal bone growth as a confounding variable. The experiment consisted of 101 mice divided among three cohorts to facilitate animal handling and sample collection. Male and female Lrp5 KO and WT littermate control mice were randomized by total body BMD into four groups: WT vehicle, WT PTH, Lrp5 KO vehicle, and Lrp5 KO PTH. Numbers of mice in each treatment group for each cohort are listed in Table 1.
Mice were treated with either vehicle (20 mM NaH2PO4 in saline) or human PTH(1-34) (purchased from either Bachem, Torrance, CA, USA or Eli Lilly, Indianapolis, IN, USA) at a dose of 80 μg/kg injected subcutaneously on alternate days for 6 weeks. This injection protocol was based on results from an 8-week pilot study that showed that administering PTH on alternate days increased BMC in the treated mice. Double fluorochrome labeling was used to determine active bone mineralization sites and rates of bone formation in cohort 3 mice only. Each mouse was injected subcutaneously two times per day with calcein (15 mg/kg; Sigma Chemical Co., St Louis, MO, USA) on the fifth and second days before necropsy. For specimen collection, all mice were killed by CO2 asphyxiation. The femora and lumbar vertebrae (LV) 3–5 were excised and cleaned free of soft tissue. Left femora and fifth LV were stored in 70% ethanol for DXA and μCT analyses. LV 3–4 were placed in 10% phosphate-buffered formalin (pH 7.4) for 24 h and subsequently transferred to 70% ethanol for histological processing.
Total body and femoral BMC and BMD were assessed in mice from cohort 1 at 6, 10, 14, and 18 weeks of age using DXA (GE-Lunar PIXImus, Madison, WI, USA) at 0.18-mm resolution while the mice were under tribromoethanol anesthesia. BMC and BMD were also determined ex vivo at termination of treatment in dissected femurs from mice in cohorts 1 and 3.
Cortical thickness of the femoral midshaft (20-μm voxel size) and cancellous bone volume (16-μm voxel size) in the distal femoral metaphysis and fifth LV were determined using a Scanco μCT40 scanner (Scanco Medical AG, Basserdorf, Switzerland). A threshold of 240 was used for evaluation of all scans. Twenty slices (0.4 mm) were analyzed in the midshaft femur. For the femoral metaphysis, 100 slices (1.66 mm) were analyzed, starting with the first slice in which both condyles were no longer visible. The small amount of primary spongiosa present in the first few slices was not analyzed. Analyses of LV 5 included the entire region of secondary spongiosa between proximal and distal slices in which the secondary spongiosa occupied at least 50% of the cancellous bone area. For LV 5, typically 120 slices (∼2 mm) were analyzed.
Quantitative bone histomorphometry
After fixation in phosphate-buffered formalin, LV 3–4 were dehydrated in graded increases of ethanol and xylene and embedded undecalcified in methyl methacrylate.(34) Frontal sections (4 and 8 μm thick) were cut with vertical bed microtomes (Leica/Jung 2065 and 2165) and affixed to slides precoated with a 1% gelatin solution. Two nonconsecutive 4-μm-thick vertebral sections per animal were stained according to the von Kossa method with a tetrachrome counterstain (Polysciences, Warrington, PA, USA) and used for determining cancellous bone volume and cellular endpoints. In cohort 3 mice, one 8-μm-thick vertebral section per animal was left unstained and used for assessing fluorochrome labeling and dynamic measurements of bone formation. Histomorphometric data were collected using OsteoMeasure/Trabecular Analysis System (OsteoMetrics, Atlanta, GA, USA) and the Bioquant Bone Morphometry System (Nashville, TN, USA) and reported in accordance with standard bone histomorphometry nomenclature.(35)
For histomorphometric data collection, the sample area within a vertebral body began 0.3 mm away from the cranial and caudal growth plates and included secondary spongiosa only. Cancellous bone volume was measured in one stained vertebral section at a magnification of ×20 and expressed as a percentage of bone tissue area. Osteoclast and osteoblast surfaces were measured in two stained vertebral sections at ×200 and expressed as percentages of total cancellous bone surface. Fluorochrome-based indices of bone formation, including mineralizing surface (percentage of cancellous bone surface with double label plus one-half single label) and mineral apposition rate, were measured in one unstained vertebral section per animal. Bone formation rate (total surface referent) was calculated by multiplying mineralizing surface by mineral apposition rate.(36)
The effects of genotype (with two levels: WT and Lrp5 KO), treatment (with two levels: vehicle and PTH), sex (with two levels: male and female), and their interactions were analyzed using a multiway ANOVA (SPSS 11.5; SPSS, Chicago, IL, USA). When interactions were significant, a separate analysis was conducted for each fixed level of one factor while varying the other factors. Differences were considered significant at p < 0.05. All data are expressed as mean ± SE.
The results are presented in two parts. The effects of targeted disruption of Lrp5 on bone mass and turnover will be described first, followed by a description of the effects of Lrp5 disruption on the skeletal response to PTH.
Effects of targeted disruption of Lrp5 on body mass and bone
At the termination of the study when the animals were 26–28 weeks old, Lrp5 KO mice weighed, on average, 9% less than WT mice (Fig. 2). In both genotypes, females were ∼27% lighter than males.
Femoral BMC was 30% lower in Lrp5 KO mice than in WT mice (Fig. 3A). Interestingly, BMC differed between male and female mice in a genotype-specific manner. Whereas femoral BMC was 16% lower (p < 0.003) in female compared with male WT mice, significant differences in BMC were not detected between the two sexes in Lrp5 KO mice.
Femoral BMD was 19% lower in Lrp5 KO mice than in WT control mice (Fig. 3B). There was a tendency (p < 0.1) for femoral BMD to also differ between male and female mice in a genotype-specific manner. Femoral BMD was ∼7% lower (p < 0.028) in WT females compared with WT males. Significant differences in BMD were not detected between male and female Lrp5 KO mice.
A time-course assessment of femoral and total body BMC and BMD in mice between 6 and 18 weeks of age in a subset of animals (cohort 1) showed that the low bone mass phenotype associated with Lrp5 deficiency was evident by 6 weeks of age. Total body BMC and BMD were 21% (p < 0.0001) and 14% (p < 0.0001) lower in 6-week old Lrp5 KO than in WT mice of the same age, respectively (Fig. 4). At this age, femoral BMC and BMD were 23% (p < 0.0001) and 16% (p < 0.0001) lower in Lrp5 KO than in WT mice, respectively (data not shown). These patterns persisted throughout the 12-week course of assessment (Fig. 4).
Cortical thickness in the femoral diaphysis, as measured by μCT, was ∼11% lower in Lrp5 KO mice compared with WT mice. There was a tendency (p < 0.08) for cortical thickness to be lower in female compared with male mice, irrespective of genotype (Fig. 3C).
Cancellous bone volume in the distal femur, as measured by μCT, was 54% lower in Lrp5 KO mice than in WT mice (Fig. 3D). Furthermore, sex differences in cancellous bone volume in the distal femur were observed in WT mice only. Femoral cancellous bone volume was ∼68% lower (p < 0.0001) in WT females compared with WT males. Cancellous bone volume was low (<5%) in both male and female Lrp5 KO mice.
Cancellous bone volume in the lumbar vertebra, as measured by μCT, was 38% lower in Lrp5 KO mice than in control WT mice, irrespective of gender (Fig. 5A). This decrease was associated with a decrease in trabecular number and thickness in the Lrp5 KO animals (data not shown). The lower cancellous bone volume determined by μCT was consistent with a similar decline (47%) measured by histomorphometry (Fig. 5B).
Osteoclast surface, an index of bone resorption, was, on average, 24% lower in lumbar vertebrae of Lrp5 KO mice compared with WT mice (Fig. 6A). In both genotypes, osteoclast surface was greater in female than male mice. Osteoblast surface, an index of bone formation, was, on average, 45% lower in Lrp5 KO mice in comparison with WT mice, irrespective of gender (Fig. 6B). Bone formation rate, a dynamic index of bone formation, was likewise lower in Lrp5 KO mice compared with WT mice (Fig. 6C). The decreased bone formation rate was caused by decreases in both mineralizing surface and mineral apposition rate (data not shown).
Effects of targeted disruption of Lrp5 on the skeletal response to PTH
Treatment of WT or Lrp5 KO mice with PTH for 6 weeks had no effect on body weight in either genotype (Fig. 2).
Administration of PTH resulted in higher femoral BMC and BMD in WT and Lrp5 KO mice (Figs. 3A and 3B). In both genotypes, BMC and BMD were, on average, 7% higher in PTH-treated than in vehicle-treated animals.
PTH treatment increased femoral cortical thickness in both genotypes (Fig. 3C). Cortical thickness was 11% higher in mice treated with PTH as opposed to mice treated with vehicle alone. In contrast to its effect on cortical bone, PTH treatment did not augment cancellous bone volume in the distal femur in either WT or Lrp5 KO mice (Fig. 3D).
No change in cancellous bone volume with PTH treatment was detected in lumbar vertebrae of either WT or Lrp5 KO mice using μCT or histomorphometry (Figs. 5A and 5B).
Osteoclast surface tended (p = 0.053) to be higher in PTH-treated mice compared with vehicle-treated mice, irrespective of genotype (Fig. 6A). Osteoblast surface was 59% higher in mice injected with PTH than in mice injected with vehicle, again irrespective of genotype (Fig. 6B). The osteoblast response to PTH treatment was greater in males than females in both genotypes. Bone formation rate was also increased with PTH treatment in WT and Lrp5 KO mice (Fig. 6C). On average, bone formation rate was 30% higher in PTH-treated mice in comparison with vehicle-treated mice, irrespective of genotype. This increase in bone formation rate was caused by an increase in mineralizing surface (data not shown) Significant differences in mineral apposition rate, an index of osteoblast activity, were not detected with PTH treatment (data not shown).
Effects of targeted disruption of Lrp5 on bone
Our data confirm and extend findings by Kato et al.,(4) Holmen et al.,(5) Clément-Lacroix et al.,(37) and Sawakami et al.(38) that disruption of Lrp5 in mice results in decreased bone mass that mimics defects observed in patients with OPPG. Histologic, DXA, and μCT measurements have all shown that both cortical and cancellous bone mass are reduced in male and female mice from 2 weeks through at least 6 months of age. Use of gene trap technology(30–33) to generate Lrp5 KO mice in this study produced the equivalent skeletal phenotype as homologous recombination procedures in previous reports.(4,5,37)
We observed reduced osteoblast surface and bone formation rate in mice lacking Lrp5. The decrease in bone formation was caused by decreases in mineralizing surface and mineral apposition rate, suggesting that Lrp5 deficiency affects osteoblast number and activity. Kato et al.(4) concluded that bone resorption was not overtly affected in Lrp5-null mice, but recent observations(39) indicate that Wnt signaling inhibits osteoclast differentiation in vitro. In this study, osteoclast surface, an index of bone resorption, was lower in Lrp5 KO than WT animals, suggesting that bone resorption may be decreased by Lrp5 deficiency. Because of these discrepancies, further studies are required to elucidate the precise role of Lrp5 on bone resorption.
As expected, femoral BMC and BMD and cancellous bone volume in the distal femur were greater in male than female WT mice. However, these sex differences were not detected in Lrp5 KO animals. This suggests possible involvement of Lrp5 and Wnt signaling in the sexual dimorphism of the skeleton. Sawakami et al.(38) also reported sex differences in severity of the osteopenic bone phenotype in Lrp5 KO mice, with males exhibiting a more severe osteopenia than females in comparison with WT animals. Interestingly, sex-specific skeletal effects of Lrp5 polymorphisms have been documented in humans. In a population-based study, Ferrari et al.(19) showed a significant association between Lrp5 polymorphisms and bone mass in males but not females. Taken together, these studies suggest that gonadal hormones may influence the effect of Lrp5 on bone mass. In this regard, interactions between canonical Wnt signaling and androgens have been reported in tumor cell lines.(40–42) It remains to be determined whether estrogens and androgens influence the skeletal response to Lrp5 in vivo.
Effects of targeted disruption of Lrp5 on the bone response to PTH
PTH regulates the expression of components of the Wnt signaling pathway,(21,22) including sclerostin,(43,44) suggesting that the hormone's skeletal actions may in part involve Lrp5. However, this study clearly indicates that Lrp5 is not essential for the skeletal response to intermittent PTH in mice. Treatment with the hormone increased cancellous osteoblast surface and bone formation rate to a comparable magnitude in Lrp5 KO and WT mice. Furthermore, PTH increased femoral BMC and BMD, irrespective of genotype. In both genotypes, this increase resulted from greater femoral cortical thickness. Because total tissue area was not affected by PTH treatment, we infer that the increase in cortical thickness was caused by increased endocortical bone formation. This is supported by previous studies reporting increased endocortical bone formation and decreased medullary area in PTH-treated mice.(45)
In agreement with our results, Sawakami et al.(38) report no significant differences in the bone response to PTH between 12-week-old Lrp5 KO and WT mice treated with the hormone (40 μg/kg, 5 days/week) for 4 weeks. PTH increased whole body and hindlimb BMC effectively and equally in both genotypes. As in our study, male and female mice did not exhibit differences in the bone response to PTH treatment.
In contrast to cortical bone, cancellous bone volume in both the distal femur and the lumbar vertebra was not augmented by intermittent PTH treatment in either WT or KO mice. Whereas fluorochrome-based cancellous bone formation was increased with PTH treatment in WT and Lrp5 KO mice, there was evidence in both genotypes for bone resorption to also increase with PTH treatment.
Increased cancellous bone mass with PTH treatment has been reported in mice in many,(46,47) but not all,(48) studies. Although we observed a large increase in bone formation, we conclude that our failure to observe an increase in bone mass was caused by a comparable increase in bone resorption. This conclusion is supported by the observed increase in osteoclast surface. Our failure to detect an increase in cancellous bone volume may have been caused by the differences in strain of mouse used or the dosing schedule for hormone administration. Our mice were C57BL6/129SvEv hybrids, and strain-dependent variations in the skeletal response to PTH have been reported.(49) We dosed PTH every second day at 80 μg/kg, and the influence of this treatment regimen has not been thoroughly examined. Other studies used a daily or 5 days/week PTH injection protocol at dose rates ranging from 4 to 160 μg/kg/day.(24,48,50–52) A cyclic injection regimen (40 μg/kg/day, daily injection 1-week-on and 1-week off for 7 weeks) increased femoral, tibial, and vertebral BMD, albeit to a lesser extent than daily injections for the same duration.(53) All of these studies, including ours, observed a large increase in bone formation in PTH-treated mice. Most of the studies observed an increase in osteoclasts. Thus, the precise balance between the bone anabolic and catabolic effects of PTH is likely sensitive to the precise dosing schedule of the hormone.
We used 20- to 22-week-old mice to minimize longitudinal bone growth as a confounding variable. Animals of different ages may exhibit differences in cancellous bone response to PTH treatment. However, it should be noted that, in rats, older animals are as or more responsive to the bone anabolic effects of PTH than younger ones. Less data are available in mice. However, Knopp et al.(54) have recently shown that older (18 months old) C57BL/6 mice respond as well as or better to the anabolic effects of PTH than young (3 months old) mice. Thus, PTH induces a similar skeletal response over a wide age range.
In summary, disruption of the Lrp5 gene in mice results in skeletal pathology similar to patients with OPPG. The skeletal anabolic actions of intermittent PTH treatment did not differ between WT and Lrp5 KO mice, showing that the Frizzled co-receptor LRP5 in the Wnt canonical signaling pathway may not be a component of the effects of PTH on bone. In addition to antiresorptive treatment with bisphosphonates,(55) intermittent PTH may be a useful therapy for patients with OPPG.
The Target Validation group at Lexicon Genetics (The Woodlands, TX) identified the bone phenotype during initial high-throughput screening of knockout mouse lines. Fieby Abdelmessih from Lexicon Genetics helped inject cohort 3 mice with PTH. This research was supported by the National Osteoporosis Foundation and Lexicon Genetics.
- 172003 High bone mineral density in SOST knock-out mice demonstrates functional conservation of osteocyte mediated bone homeostasis in mouse and human. J Bone Miner Res 18: S1: S7., , , , , , , , ,
- 202003 Regulation of bone formation by Wnt signaling. J Bone Miner Res 18: S2: S14., , , , ,
- 271998 PTH: Skeletal effects in the ovariectomized rat model for postmenopausal bone loss. Whitfield JF, Morley P (eds.) Anabolic Treatments for Osteoporosis. CRC Press, Boca Raton, FL, USA, 59–81.,
- 292003 Forteo approved for osteoporosis treatment. FDA Consum 37: 4.
- 341983 Processing of undecalcified bone specimens for bone histomorphometry. Recker RR (ed.) Bone Histomorphometry: Techniques and Interpretation. CRC Press, Boca Raton, FL, USA, 13–35., , , ,
- 361983 Bone histomorphometry: Analysis of trabecular bone dynamics. Recker RR (ed.) Bone Histomorphometry: Techniques and Interpretation. CRC Press, Boca Raton, FL, USA, 109–131.
- 491998 Bone effects in mice of ovariectomy and recombinant human PTH (1-34) are highly strain dependent. Bone 23(Suppl 5): S448., , , ,
- 501999 Intermittant administration of human parathyroid hormone (1-34) prevents immobilization-related bone loss by regulating bone marrow capacity for bone cells in ddY mice. J Bone Miner Res 14: 1691–1699., , , , , , ,