Weaning Triggers Decreases in RANKL Expression and Increases in OPG Expression in Bone. L. Ardeshirpour1, P. R. Dann*2, J. N. VanHouten*2, J. J. Wysolmerski2. 1Pediatrics ( Endocrinology), Yale University School of Medicine, New haven, CT, USA. 2Internal Medicine ( Endocrinology), Yale University School of Medicine, New haven, CT, USA

In mammals, bone is lost rapidly during lactation and recovers fully after weaning. In fact the recovery of bone mass post weaning represents the most rapid accrual of bone during adult life. Our previous studies suggested that the elevated rates of bone resorption seen during lactation are promptly suppressed after weaning. Histomorphometry studies have revealed a precipitous decline in osteoclast numbers by 3 days post weaning. Osteoclast differentiation, activity and survival are all regulated by the local production of RANKL (receptor activator of nuclear factor kappa B ligand) and OPG (osteoprotegerin) by cells in the osteoblast lineage and many studies have shown that the RANKL/OPG ratio is an important determinant of bone resorption rates. In order to study the role of these cytokines during lactation and post weaning, twelve-week old CD1 mice were allowed to become pregnant, deliver and lactate. On day 12 of lactation, pups were removed to initiate weaning. Calveria and long bone metaphyses were harvested on day 12 of lactation and 3 days after weaning. Age-matched virgin mice served as controls. RNA was extracted and real time PCR was used to measure RANKL, RANK (receptor activator of nuclear factor kappa B) and OPG expression. Compared to nulliparous controls, there was a 6-fold increase in RANKL mRNA levels during lactation, which returned to baseline following weaning. Likewise, there was a 4-fold increase in RANK expression, which decreased to virgin levels following weaning. Somewhat surprisingly, OPG mRNA levels increased 4.5-fold at day 12 of lactation and increased further to 7-fold over virgins at 3 days after weaning. These changes generated a modest 1.3-fold increase in the RANKL/OPG ratio during lactation, but after weaning there was a 10-fold decline in this ratio compared to virgins. These findings suggest that a sudden decline in RANKL production coupled with a significant increase in OPG levels causes the abrupt halt in bone resorption that marks the transition from skeletal catabolism to anabolism at weaning.

Disclosures: L. Ardeshirpour, None.

This study received funding from: NIH.


Pregnancy Rescues Low Bone Mass and Normalizes Intestinal Calcium Absorption in Vdr Null Mice. N. J. Fudge*, J. P. Woodrow, C. S. Kovacs. Faculty of Medicine, Memorial University of Newfoundland, St. John's, NF, Canada

Intestinal calcium absorption and skeletal metabolism is dependent upon sufficiency of vitamin D and calcitriol. Vitamin D receptor (Vdr) null mice raised on a standard 1% calcium chow have hypocalcemia, hypophosphatemia, and rickets; furthermore, the absorption of 45Ca from an in situ duodenal loop is less than 20% at 10 minutes as compared to 70-80% in WT mice. The skeletal and biochemical abnormalities of Vdr null mice are prevented by a 2% calcium chow given prior to day 16 post birth.

Pregnancy induces a doubling of intestinal calcium absorption and calcitriol levels in humans, with similar increases occurring in rodents. Lactation causes 5-10% losses of skeletal content in humans and 25-30% losses in rodents; these losses are fully restored after weaning.

We studied Vdr null mice to assess whether intestinal calcium absorption during pregnancy and skeletal recovery after weaning require calcitriol or the VDR.

Vdr null mice were raised on 1% calcium chow until 10 weeks, at which age they were switched to the 2% calcium chow to enhance fertility. Baseline measurements of whole body and regional (spine, hindlimb) bone mineral content (BMC) were made over the first ten days prior to mating and were stable. BMC was measured at the end of pregnancy (ED 18.5), end of lactation (21 d) and after 14 days post-weaning. To assess duodenal calcium absorption, the in situ loop technique was used. In brief, ligatures were tied around a 1.5 cm length of duodenum, into which 45Ca was injected. After 10 minutes, the loop was excised, solubilized, and counted. Efficiency of 45Ca absorption was reported as [45Ca remaining / 45Ca injected] × 100.

Vdr null mice entered pregnancy with a lower whole body BMC (0.381 ± 0.003 g vs. 0.519 ± 0.004 g in WT, p <0.02). Vdr nulls increased BMC to 158% relative to baseline during pregnancy vs. 122% in WT (p <0.02), dropped to 122% at the end of lactation vs. 85.8% in WT, and returned to 149% after weaning vs. 104% in WT (p <0.02). Final post-weaning BMC values were no different between Vdr null and WT (0.544 ± 0.003 g versus 0.531 ± 0.004 g). Duodenal 45Ca absorption in our lab is 78.7 ± 3.8% in non-pregnant WT mice; the published value is <20% in non-pregnant Vdr null mice, and in pregnant Vdr null mice we found it was increased to 94.1 ± 8.8%.

In summary, intestinal 45Ca absorption normalized during pregnancy in the absence of the VDR, and enabled accrual to a normal BMC at the peak of pregnancy. Further, the skeleton fully recovered after weaning in the absence of the VDR.

In conclusion, the response of Vdr null mice to pregnancy and lactation demonstrated that intestinal calcium absorption and skeletal mineralization can be potently stimulated independently of calcitriol and its receptor.

Disclosures: C.S. Kovacs, None.


Bone Is the Primary Target of PTH in Regulation of Fetal Blood Calcium. T. Kobayashi1, A. C. Karaplis2, D. Goltzman3, H. M. Kronenberg1. 1Endocrine Unit, Massachusetts General Hospital, Boston, MA, USA. 2Sir Mortimer B. Davis Jewish General Hospital, McGill University, Montreal, PQ, Canada. 3Royal Victoria Hospital of the McGill University Health Centre, Montreal, PQ, Canada

Parathyroid hormone (PTH) regulates blood calcium levels through multiple mechanisms. PTH directly acts on bone to stimulate calcium mobilization into the systemic circulation. In addition, PTH acts on the kidney to increase the blood calcium level by increasing distal tubular calcium reabsorption and proximal tubular 1α hydroxylation of vitamin D metabolites. During fetal stages, calcium is actively transported from the mother to fetuses across the placenta, which plays a significant role in maintaining fetal calcium homeostasis. PTH-related protein (PTHrP) stimulates placental calcium transport through a mechanism distinct from activation of PTH/PTHrP receptor (PPR). Loss of fetal PTHrP results in reduction of fetal calcium levels to the maternal level. On the other hand, loss of PTH or PPR signaling in the fetus was reported to cause severe hypocalcemia. Therefore, a question remains as to the relative importance of placental calcium transport and other fetal PTH-target organs in regulation of fetal blood calcium.

We generated mutant mice in which the PPR was disrupted in cartilage and bone. Transgenic mice in which Cre recombinase was driven by a type II collagen promoter (Col2-Cre) were crossed to mice in which the E1 exon of the PPR gene was flanked by loxP sequences (floxed PPR). Since Cre recombinase in Col2-Cre mice cleaves the floxed PPR gene likely in common precursor cells of chondroctyes and osteoblasts, PPR is lost both in chondrocytes and osteoblasts in endochondral bone. The fetal calcium level of mutant mice at E18.5 was significantly lower than that of control littermates, suggesting that PPR signaling in the bone is a critical regulator of fetal blood calcium. To further confirm the roles of PTH signaling in fetal bone in fetal blood calcium regulation, we analyzed fetal calcium levels of Runx2 null, PTH null and Runx2:PTH double mutant mice. Runx2 null mice had little skeletal mineralization because of lack in osteoblast differentiation. The fetal calcium level of Runx2 null mice was significantly higher than that of control littermates, demonstrating that calcium flux into bone had a considerable influence on the circulating calcium level. PTH null fetuses showed severe hypocalcemia. However, Runx2:PTH double mutant fetuses showed calcium levels similar to those of Runx2 single mutant mice. This observation that loss of PTH had no impact on the fetal blood calcium level in the absence of bone demonstrates that bone is the primary target of PTH in regulation of fetal blood calcium.

Disclosures: T. Kobayashi, None.


Anabolic Responses of Bone to Parathyroid Hormone Are Maintained in Mice with Deficient Osteoclast Formation and Activity. J. A. Lorenzo, J. Kalinowski*, B. Koczon-Jaremko*, S. Jastrzebski*, C. Jacquin*, H. Aguila*, S. Lee. University of Connecticut Health Center, Farmington, CT, USA

Previous studies demonstrated that anabolic responses of bone to parathyroid hormone (PTH) were decreased by prior treatment with bisphosphonates, which inhibit osteoclast (OC) formation and activity. To determine if this occurred when OC were inhibited by another mechanism, we developed a new mouse model with decreased OC responses to PTH. RANK deficient mice, which cannot form OC because they lack a critical receptor, are fragile and die shortly after weaning. To overcome this, we transplanted C57BL/6 RANK deficient hematopoietic cells into wild type (WT) C57BL/6 mice (RANK>WT). Controls were WT mice that received WT hematopoietic cells (WT>WT). Animals were irradiated at 9 weeks of age, transplanted and allowed to recover for 7 weeks. PTH (160 mcg/kg/day) or vehicle (Veh) was then given SC for 10 days. Results in transplanted mice were compared to mice of identical age that were not transplanted. Engraftment for all transplants was 88 to 98%. Responses were monitored by micro-CT, histomorphometry and a serum TRAP5b ELISA.

In non-transplanted WT mice PTH increased femoral osteoblast (OB) surfaces by 46% (OB S/BS, p <0.05) and OC number by 39% (N OC/BPm, p <0.05). PTH also increased TRAP5b by 121% (p <0.05) and vertebral trabecular bone mass by 19% (p <0.05) compared to Veh-treated controls. Bone marrow cells from RANK>WT mice failed to form OC in vitro with M-CSF and RANKL after 4 days; although, a few ( <20/well) OC formed after 8 days. In contrast, abundant OC (>200/well) formed in cultures from WT>WT mice at 4 and 8 days. Surprisingly, bones from RANK>WT mice at 9 weeks post-transplant contained OC. PTH increased OB surfaces by 175% in WT>WT and 122% in RANK>WT femurs (p <0.05 for both). However, it increased OC number (by 54%, p <0.05) only in WT>WT bones without altering this parameter in RANK>WT bones. Serum TRAP5b was decreased by 35% in Veh-treated RANK>WT mice compared to Veh-treated WT>WT mice (p <0.05). PTH increased serum TRAP5b by 125% in WT>WT mice but only by 42% in RANK>WT mice (p <0.05 for both and for the difference between responses of WT>WT and RANK>WT mice). However, PTH had similar anabolic effects on vertebral trabecular bone mass in WT>WT and RANK>WT mice (increases of 13.2% and 13.6% respectively, p <0.05 for both).

These results demonstrate that the anabolic effects of PTH on bone can occur in the absence of increases in OC number and do not correlate with OC activity as measured by serum TRAP5b. While our data do not rule out a role for OC in the anabolic response of bone to PTH, they strongly suggest that some therapies, which inhibit OC, can be given with anabolic regimens of PTH without reducing the PTH response.

Disclosures: J.A. Lorenzo, None.

This study received funding from: National Institutes of Health.


In Vivo over Expression of the Transcriptional Factor Runx2 Abolishes the Anabolic Effect of Parathyroid Hormone. D, Merciris*1, C. Marty*1, J. Launay*2, M. de Vernejoul1, V. Geoffroy1. 1Hǒpital Lariboisière, INSERM U606, Paris Cedex 10, France. 2Biochemistry department, Hǒpital Lariboisière, Paris Cedex 10, France

There is convincing evidence that Runx2 is a potential regulator of anabolic action parathyroid hormone (PTH) in bone. Therefore, we decided to evaluate how Runx2 over expression in osteoblast can affect anabolic PTH response.

One-month-old transgenic females over expressing Runx2 specifically in osteoblast (TG) and their wild type (WT) littermates received single daily subcutaneous doses of vehicle (VEH) or human PTH (1-34) at l00μg/kg/day for 6 weeks. Bone mineral density (BMD) was measured by DEXA at day 0 and day 42 and analysis of trabecular and cortical bone structure and dynamic parameters were evaluated by histomorphometry. The level of expression of osteoblastic genes was evaluated by QRT-PCR. Finally, response of primary osteoblast from WT and TG mice to PTH was evaluated by determination of cAMP production.

At baseline, TG mice exhibited significantly lower BMD and trabecular bone volume compared to WT mice. Furthermore, PTH administration increased significantly BMD only in WT mice, the TG mice being unaffected. The histomorphometric analysis confirmed the anabolic effect of PTH on trabecular and cortical bone in WT mice. Indeed, administration of PTH increased significantly trabecular thickness and bone formation rate (BFR) in WT mice but not in TG mice. However, the trabecular separation was significantly decreased in the both genotypes treated with PTH compared their controls. In vehicle TG mice, we confirmed that the expression of the transgene markedly decreased osteocalcin (OC) expression. This finding suggested that osteoblasts of TG origin are globally less differentiated compared to WT osteoblasts. PTH administration led to an increase of OC expression that was stronger in WT (54%) than in TG mice (27%). Altogether, these results indicated that Runx2 over expression affects the PTH induced bone formation rather than bone resorption.

Our molecular analysis further indicated that the milder response to PTH observed in TG mice might be due to the lower expression of the PTH-R1 in TG compared to WT mice (−43%). In addition, cAMP accumulation induced by PTH (5 and 10nM) exposure was significantly lower in TG osteoblast than in WT osteoblast. But PTH treatment was still able to increase PTH-R1 expression in both TG mice (121%) and WT mice (22%). The increase of the transgene expression by the treatment (83%) could also be part of the lack of PTH anabolic action in TG mice.

In conclusion, our findings demonstrate that high level of Runx2 abolishes the anabolic effect of PTH in vivo probably through the decreased sensitivity of TG osteoblasts to PTH.

Disclosures: D. Merciris, None.


Mice Lacking a Distant Transcriptional Enhancer of the RANKL Gene Display a Blunted Response to PTH and 1,25(OH)2D3 In Vitro and In Vivo. Q. Fu*1, L. A. Zella*2, J. A. Fretz*2, P. E. Cazer*1, C. Galli*1, J. W. Pike2, S. C. Manolagas1, C. A. O'Brien1. 1Center for Osteoporosis & Metabolic Bone Diseases, University of Arkansas for Medical Sciences, and Central Arkansas Veterans Healthcare System, Little Rock, AR, USA. 2Department of Biochemistry, University of Wisconsin-Madison, Madison, WI, USA

Previous in vitro studies utilizing large transcriptional reporter constructs (based on bacterial artificial chromosomes containing the full-length RANKL gene), and ChIP-on-chip assays, indicated that PTH and 1,25(OH)2D3 control murine RANKL gene expression in stromal/osteoblastic cells in part via an evolutionarily-conserved transcriptional enhancer. This enhancer is designated the distal control region (DCR) and is located 76 kb upstream from the transcription start site. To determine the significance of this enhancer for control of the endogenous RANKL gene by calciotropic hormones, we created knockout (KO) mice in which the 2 kb DCR was eliminated by gene targeting. DCR KO mice were obtained at the expected Mendelian frequency and had no obvious growth abnormalities. In addition, basal RANKL mRNA expression in vertebral and long bones was not different between wild-type (WT) and KO mice. Nonetheless, RANKL mRNA in thymus and spleen was significantly reduced (by approximately 40%) in KO compared to WT mice. PTH stimulation of RANKL mRNA in primary bone marrow cultures was reduced by more than 4-fold in KO mice while the stimulation by 1,25(OH)2D3 was reduced by more than 2-fold. Consistent with this, PTH-stimulated osteoclast formation in primary bone marrow cultures from KO mice was dramatically reduced compared to cultures from WT littermates. 1,25(OH)2D3-stimulated osteoclast formation was also significantly reduced, but to a lesser extent than with PTH, in cultures from KO mice. To test the responsiveness of the RANKL gene to PTH in vivo, WT and KO mice were placed on a calcium-deficient diet for 7 days to induce secondary hyperparathyroidism. PTH levels were elevated to the same extent in both WT and KO mice, but RANKL mRNA was significantly elevated only in the bones of WT mice. Finally, to test the responsiveness to 1,25(OH)2D3 in vivo, mice were placed on a calcium-deficient diet for 3 days to remove calcium from the gut and then given a single injection of 1,25(OH)2D3 at 1, 10, or 100 μg/kg to stimulate bone resorption. The 1 μg/kg dose caused a significant rise in serum calcium 24 hr after the injection in WT but not in KO mice. At the two higher 1,25(OH)2D3 doses, the rise in serum calcium was significantly blunted in the KO mice compared to WT mice. We conclude that the DCR is required for responsiveness of the RANKL gene to PTH and that it also plays a role, albeit to a lesser extent, in the response of this gene to 1,25(OH)2D3.

Disclosures: C.A. O'Brien, None.


Very Young and Very Old Bone Collagen Is Associated with Increased Risk of Osteoporotic Fractures in Postmenopausal Women. I. Byrjalsen1, D. J. Leeming*1, P. Qvist1, C. Christiansen2. 1Nordic Bioscience, Herlev, Denmark. 2CCBR, Ballerup, Denmark

It has previously been shown that postmenopausal women with low bone mass and increased bone resorption have an increased risk of sustaining osteoporotic fractures. Today's research on skeletal fragility is focused on bone quality including development of methods for assessment of the micro-architectural structure of bone and identification of biochemical indexes reflecting other aspects of bone quality. The aim of the present study was to assess the relationships between biochemical markers of bone turnover as predictors for osteoporotic fractures in postmenopausal women.

Study subjects included 1,534 healthy postmenopausal women from the prospective PERF cohort study. The study cohort represented a broad spectrum of early and elderly postmenopausal women being healthy at their baseline visit and followed from 7 to 23 years. During a median of 8.5 years of follow-up (7-23 years), 306 women sustained osteoporotic fractures. One or more of the bone turnover markers of serum CTX (type I collagen C-telopeptide), urinary CTX, urinary age specific forms of CTX of newly synthesized collagen (ααCTX), and aged collagen (ββCTX), urinary hydroxyproline, serum osteocalcin, and urinary calcium excretion were measured in samples obtained at baseline.

Bone resorption as assessed by S-CTX and U-CTX at baseline was higher in subjects with incident osteoporotic fractures as compared with those without fractures (p = 0.03 and 0.002, respectively). These markers showed the association of an increased risk of osteoporotic fractures with increased level of concentration. The women with values in the upper quartile had independently of bone mass, a two-fold higher risk of fracture than women with values in lower quartiles. The marker of the bone collagen age, i.e. the ratio between ααCTX and ββCTX, showed a U-shaped risk profile with 1.5 to 2-fold higher risk associated with values in lowest quartile (aged collagen) and highest quartile (young collagen) as compared with the mid-range values (p = 0.04, adjusted for age, length of follow-up, and prevalent fracture). Furthermore, the risk for osteoporotic fractures associated with very low and high bone collagen age remained unchanged after adjustment for rate of bone resorption (S-CTX), suggesting these to be additive.

Combining bone mass measurements with measurements of bone resorption and bone collagen age may prove useful for assessment of bone quality and thereby for assessment of risk of osteoporotic fractures in postmenopausal women.

Disclosures: 1. Byrjalsen. Nordic Bioscience 3.


Plasma Vitamin B6 and Bone Mineral Density in Elderly Men and Women: The Framingham Study. R. R. McLean1, P. F. Jacques*2, J. Selhub*2, L. Fredman*3, K. L. Tucker2, D. P. Kiel4, L. A. Cupples*3, M. T. Hannan4. 1Hebrew SeniorLife, Boston, MA, USA. 2USDA HNRC, Tufts Univ. Boston, MA, USA. 3BU Sch of Public Health, Boston, MA, USA. 4Hebrew SeniorLife & Harv Med Sch, Boston, MA, USA

The relation among homocysteine (Hcy), B vitamins and bone mineral density (BMD) is unclear. Previous studies of B vitamins have focused on folate and B12 which both influence Hcy metabolism via the methylation pathway. Vitamin B6 is a cofactor in the transsulfuration pathway of Hcy metabolism but its relation to BMD has not been evaluated in both men and women. We examined the association of plasma vitamin B6 concentration with BMD in men and women in the Framingham Study Original Cohort. Non-fasting blood samples were drawn from 385 men and 570 women in 1987-90 (baseline). Plasma B6 (nmol/L) was measured by the tyrosine decarboxylase method and classified as normal (>30), low (20-30) and deficient ( <20). Femoral neck (FN) BMD (g/cm2) was measured at baseline with a Lunar DP3 and four years later using a Lunar DPX-L (accounting for technology change). Separately for men and women, we used analysis of covariance to compare least-squares (LS) mean baseline and four-year % BMD change among B6 groups and test for linear trend, adjusting for baseline measures: age (yr), height (in), weight (lbs), smoking (cig/d), physical activity index and in women current estrogen use (y/n). BMD change analyses were further adjusted for four-year weight change. Mean age ( ± SD) among men was 75.1 (4.9) and among women it was 75.6 (5.0). Normal, low and deficient group frequencies (%) were 77, 12, 11 and 76, 12, 12 for men and women, respectively. 

Table  . LS mean ( ± SE) baseline BMD (g/cm2) and four-year % BMD change per B6 group
inline image

At baseline, men in the low B6 group and women in the deficient group had lower mean FN BMD than their respective normal groups (P≤0.01). The decreasing trend across B6 groups in women was significant. In longitudinal analyses among men, FN BMD loss was greater in the low (P = 0.05) and deficient groups (P = 0.01) compared to the normal group and increased significantly with decreasing B6. B6 was not significantly associated with longitudinal bone loss in women. These findings suggest that vitamin B6 may be a potentially modifiable risk factor for low FN BMD among both older men and women. This is the first study to examine the relation of B6 with BMD in elderly men and women. Additional studies are needed to determine whether B6 may play a role in the observed association between Hcy and fracture.

Disclosures: R.R. McLean, None.


The Association between Hyperglycaemia, Diabetes and Fracture Risk in Middle Age: A Prospective, Population-Based Study of 22,444 Men and 10,902 Women. A. H. Holmberg1, P. M. Nilsson*2, O. Johnell1, K. Akesson1. 1Department of Clinical Sciences, Malmö, Clinical and Molecular Osteoporosis Research Unit, Malmö, Sweden. 2Department of Clinical Sciences, Malmö, Department of Internal Medicine, Malmö, Sweden

Diabetes type 1 is associated with increased fracture risk, whereas the risk effect associated with diabetes type 2 is less obvious. High lasting blood glucose (FBG) and a high 2-hour blood glucose level of an oral glucose tolerance test (OGTT) can indicate impaired glucose tolerance or diabetes. We investigated the associations between FBG and 2-hour blood glucose and risk of low-energy fracture.

The Malmö Preventive Project consists of 22 444 men (44 ± 6.6 years) and 10 902 women (50 ± 7.4 years) at inclusion. Baseline assessment included multiple examinations and lifestyle information. Mean follow-up: 19 ( ± 3.9) years for men and 15 ( ± 4.5) years for women, for incident fractures. FBG was measured in the whole cohort, 2-h blood glucose in 57 % and fasting serum insulin in 22% of the cohort The HOMA index was calculated using fasting serum insulin. A logistical regression model was used, with the subgroups divided in quartiles, adjusted for age and BMI. Due to shorter gastric emptying time in smokers and thus decreased 2-h glucose levels, adjustment for smoking was added.

At least one low-energy fracture occurred in 1262 men and 1257 women, 5.6% and 11.5% of the study population. Osteoporotic fractures (forearm, hip, proximal humerus, vertebrae, pelvis, proximal tibia) were identified in 704 (47%) men and 955 (74 %) women of the fracture population. A 2-h blood glucose above 4.2 mmol/L in men and above 5.4 mmol/L in women, the 2nd, 3rd and 4th quartile in both genders, were associated with decreased risk of multiple fractures, in men (Odds Ratio, OR 0.50, Confidence Interval CI 95% 0.36–0.70) and in women (OR 0.34, CI 95 % 0.22–0.52), adjusted for age and BMI. In women, 2-h blood glucose was associated with a 50 % risk decrease of osteoporotic fracture (OR 0.52, CI 95% 0.41–0.68), in men with a 25 % risk reduction (OR 0.75, CI 95% 0.58–0.98). FBG above 6.1 mmol/L, an indication of diabetes, was found in 879 (3.9%) men and 365 (3.4%) women. FBG was associated with small but significantly increased risks of fracture (OR 1.06-1.09), independent of age and BMI, in both genders. No associations between tasting serum insulin levels, HOMA and fracture risk were found.

In middle-aged men and women, impaired glucose tolerance was associated with a decreased risk of multiple low-energy fractures, as well as of osteoporotic fractures, independent of age, BMI and smoking. These findings may indirectly suggest a positive effect on bone from elevated insulin levels.

Disclosures: A.M. Holmberg, None.


Successful Skeletal Aging: A Marker of Low Fracture Risk and Longevity. J. Cauley1, L. Lui2, D. Barnes*3, K. E. Ensrud4, A. B. Newman*1, J. M. Zmuda1, T. Hillier5, M. C. Hochberg6, A. V. Schwartz3, M. C. Nevitt3, K. Yaffe*3, S. R. Cummings2. 1University of Pittsburgh, Pittsburgh, PA, USA. 2CPMC Research Institute, San Francisco, CA, USA. 3UCSF, San Francisco, CA, USA. 4University of Minnesota, Minneapolis, MN, USA. 5KCHR, Portland, OR, USA. 6University of Maryland, Baltimore, MD, USA

The primary objective of this analysis was to determine if there is a distinct subset of women who maintain their BMD into old age and to identify the unique characteristics that distinguish these women from those who lose BMD. Further, we examined the relationship between successful skeletal aging, fracture incidence and mortality rates. We studied 8,514 Caucasian women enrolled in the Study of Osteoporotic Fractures. Total hip BMD was measured at 5 clinic visits over 14 years of follow-up. The median number of BMD measurements per subject was 4 (range 1-5). Random effects regression was used to determine a unique age-adjusted BMD slope and intercept for each subject. Three groups were formed: “optimal” BMD change/year: slope ≥ 0, n = 758 (9%), mean age baseline = 68.9y; “expected” BMD change/year: slope <0 to < 1 standard deviation (SD) below mean, n = 6,700, (79%), mean age = 71.5y; “accelerated” BMD change/year: slope ≥ 1SD below mean, n = 1,056 (12%), mean age = 71.9y.

Multinomial logistic regression was used to compare two groups with the “expected” group (Table). The independent baseline predictors of “optimal” bone loss were younger age, greater weight and weight increase, not being a diabetic, avoidance of smoking, hormone therapy (HT), and better neuromuscular function. “Accelerated” bone loss was associated with greater weight loss, diabetes and no HT. Correlates remained the same when we excluded hormone users. “Optimal” bone loss was associated with a lower rate of total fractures and hip fractures; the age-adjusted incidence of total and hip fracture (per 1000 person years), respectively was: “optimal” 38.9 and 5.1; “expected” 40.6 and 9.0; “accelerated” 49.1 and 13.7. Age-adjusted total mortality rates were also lower: “optimal”, 16.2; “expected”, 32.9; and “accelerated”, 36.0. We conclude that a subset of older women maintain their BMD, experience fewer fractures and have a lower total mortality rate over 14 years. This phenotype could be a marker of successful aging.

Table Table:. Multivariate Analyses: Factors associated with “Optimal” and “Accelerated” Bone Loss: Odds Ratio (95% CI Confidence Interval); “expected” = referent group
inline image

Disclosures: J. Cauley, Merck 2, 8; Eli Lilly 2; Pfizer Pharmaceuticals 2; Novartis 2, 5.


Similar Subsequent Fracture Risk following Initial Osteoporotic Fracture in Men and Women: A 15 Year Follow-Up Study. J. R. Center, D. Bliuc*, T. V. Nguyen, J. A. Eisman. Bone and Mineral Research Department, Garvan Institute of Medical Research, Sydney, Australia.

Although several studies have reported that prior osteoporotic fracture results in an increased risk of subsequent fracture, there are few long-term, age-specific data on subsequent fracture risk in women and even fewer in men. The aim of this study was to examine the effects of age, gender and time for a variety of clinical fracture types on subsequent fracture risk in a community-dwelling cohort of men and women over a 15-year interval.

All clinical low trauma fractures and deaths from 2245 women and 1760 men aged 60+ participating in the Dubbo Osteoporosis Epidemiology Study were recorded from 1989 to 2005. Subsequent fracture rates were compared with the population initial fracture rates according to sex, age group (60–69, 70–79 and 80+), type of fracture and time since fracture. Relative risks and 2-sided confidence intervals were based on Poisson assumption and the Kaplan Meier survival distribution was calculated for subsequent fracture risk.

There were 880 incident fractures and 253 subsequent in women (27,687 and 4076 person-yrs of follow-up, respectively) and 329 incident and 71 subsequent fractures in men (20,054 and 1248 person-yrs of follow-up, respectively). Following initial fracture, risk of subsequent fracture increased 1.6–2.4 fold for women and 2.8–4.3 fold for men, resulting in an overall subsequent fracture risk that was comparable in women and men and worse than that of a women 10 years older. Thus, 60–69 year old women or men had subsequent fracture rates of 36 and 37/1000 p-yrs; (CI 26–48 and 23–59), respectively, compared with an initial fracture risk of 27/1000 p-yrs (CI 24–30) for a 70–79 year old woman. Fracture rates decreased over time towards the increasing population baseline fracture rates but differed for different age and sex groups. However, fracture rates were increased for 7–12 years by which time 40–60% of women and men had re-fractured. All fracture types, apart from rib fractures in men and ankle fractures in women, resulted in increased subsequent fracture risk with the highest relative risks following hip and clinical vertebral fractures in younger men (RR = 10.0; CI 1.4–72 and RR = 15;CI 6.1–38 respectively).

After an initial fracture, risk of subsequent fracture was increased more so in men than women such that subsequent fracture rates were similar between the sexes and higher than of an initial fracture rate of a woman 10 years older. This increased risk followed almost all types of fractures and persisted for 7–12 years, by which time about half of all subjects had re-fractured. These data indicate the high re-fracture risk after all clinical osteoporotic fractures.

Disclosures: J.R. Center, None.


Family Aggregation of Hip Fractures - In the Swedish Population, an Analysis of the Swedish Multiple-Generation Register. O. Johnell1, L. Lidgren*2, P. Wagner*2, D. Mellstrom3. 1Clinical Sciences, Malmö, Lund University, Malmö, Sweden. 2Lund University, Lund, Sweden. 3University of Göteborg, Göteborg, Sweden

Several studies have suggested that hereditary factors influence the risk of hip fracture. These studies have been designed both as general epidemiological cohort studies and as genetic twin studies, but typically with small and selected study populations.

We have explored parent-children relations in a nation-wide case-cohort study based on a record linkage between four Swedish national registers: 1) the national multiple-generation register, which contains information on 11 million individuals, and makes it possible to link all subjects with their 1st degree relatives (i.e. biological parents, siblings and children, both biological and adopted) from 1932, 2) the national mortality register, which contains causes of death for all diseased subjects since 1961, 3) the register on hospitalisations, which contains information on hip fractures, and finally, 4) the population register with information on occupation, socio-economy, etc. The number of identified cases was 3 365.

We modelled statistically, using multiple logistic regression analysis to adjust for differences in age, the relative risk (RR) of hip fracture for 1st degree relatives to hip fracture cases. 

Table  .  
inline image

All relations but two were statistically significant; the one with lowest relative risk including adopted children.

We conclude that heredity is an important factor in the physiological development leading to a hip fracture, and that this factor might have a more pronounced effect in daughters and female siblings.

Disclosures: D. Mellstrom, None.


Risk Factors for Hip Fracture, WHI. J. Robbins1, A. Agaki*2, C. Kooperberg*2, J. A. Cauley3, Z. Chen4, J. Wactawski-Wende5, M. S. LeBoff6, C. E. Lewis7, R. Jackson8, N. B. Watts9, M. Stefanick*10. 1IM, UC Davis, Sacramento, CA, USA. 2Div. of PHS, Fred Hutchinson CRC, Seattle, WA, USA. 3Epi., U. of Pitt, Pittsburgh, PA, USA. 4Epi. and Biostat, U. of AZ, Tucson, AZ, USA. 5Social and Preventive Med., U. at Buffalo, Buffalo, NY, USA. 6Med., Harvard, Boston, MA, USA. 7Div, of Prev. Med., U. of AL, Birmingham, AL, USA. 8Int. Med., OSU, Columbus, OH, USA. 9U. of Cincinnati, Cinicinnati, OH, USA. 10Stanford U., Stanford, CA, USA

Ability to predict hip fracture (HFx) risk from historical information would have major importance for patients and public health. Prior algorithms have been developed from relatively small populations of mostly Caucasian women. We used data from a multi racial group of 93,676 postmenopausal women age 50-79 in the observational (OS) cohort of the Women's Health Initiative (WHI) to select predictors. Based on these predictors, the probability of suffering a HFx within 5 years was computed from a logistic regression model. We then tested the model in the 64,785 women in the WHI clinical trials (CT) that had a HFx within 5 years. Medical records were reviewed and cases of suspected HFx centrally adjudicated. The average (SD) follow up was 7.2 (1.6) years, 949 HFx were identified in the OS. Potential risk factors from the literature were independently evaluated. Risk factors with a modest statistical significance (p <0.20), after adjusting for age and ethnicity, were entered into a prediction model. Only those factors which improved a cross-validated estimate of prediction risk were used. Significant factors which did not enhance prediction were not included. The factors which remained in the model were; (point scores) Age ((age −50)/2), General health (0-4), Height (max 3), Weight (max 4), Fracture after the age of 50 (2), Race (0-5), Physical activity (0-2), Smoking (0-4), Parental fracture (2) and Steroid use (2). The point score gives a spectrum of risk. BMD was not included. ROC curves were developed to test how well the algorithm preformed in a test population. The area under the curve (AUC) when tested against various groups of WHI CT participants ranged from 79-83%. The AUC was 80% for all WHI CT participants and 83% in all WHI control participants. The ROC curve below shows the “accuracy” at different estimations of risk tested in all trial participants. Thus, thresholds for further screening can be set based on acceptable risk and desire for certainty.

Thumbnail image of

Disclosures: J. Robbins, None.

This study received funding from: NIH.


Vertebral Fractures Are Highly Prevalent in Hip Fracture Patients: Results from a 1 Year Consecutive Cohort. J. Ryg*1, J. Gram*2, S. Qvergaard*3, K. Brixen1. 1Department of Endocrinology, Odense University Hospital, Odense, Denmark. 2Department of Internal Medicine, Ribe County Hospital, Esbjerg, Denmark. 3Department of Orthopedic Surgery, Odense University Hospital, Odense, Denmark.

It is well known that patients with hip fractures and vertebral fractures (VFx) have an increased risk of future fracturesa. Moreover, the prevalence of VFx has been shown to increase with age in the background populationb. The prevalence of VFx in patients with recent hip fracture, however, is unknown. The aim of this study was to elucidate the prevalence of VFx in a cohort of 461 consecutive hip fracture patients. Since January 1 2005 all patients admitted with a low energy fracture of the hip were offered evaluation by Dual energy X-ray Absorptiometry (DXA), Instant Vertebral Assessment (IVA) scanning from T4 through L4, biochemical screening, and clinical assessment followed by specific anti-osteoporotic treatment as appropriate. This was done as part of an ongoing health technology assessment study. During 2005, 461 patients were admitted with a hip fracture. Excluding patients with dementia, severe comorbidities, and pathological fractures, a dedicated nurse approached 336 (73 %) of the patients. During the hospital stay, 28 (6 %) died. A total of 306 patients (66 %) were referred for evaluation and 224 (49 %) patients (145 women and 79 men) showed up for the scanning (median [range] 79 [45;96] years). The number of vertebra assessed using either single or dual energy or both was 3541. A total of 139 (62 %) patients had at least one, 78 (35 %) had at least two, and 44 (20 %) three or more fractures, defined as a reduction of vertebral height of at least 25 %. The prevalence of VFx did not increase significantly with age. Adjusting for age, no significant difference was found in the prevalence of VFx in women and men except in patients aged 85+ years where VFx was significantly more prevalent in men (p <0.05). Our study shows that patients with recent hip fracture have a high prevalence of VFx as assessed by IVA. We propose that radiographic assessment (IVA or X-rays) of the spine should be performed in all hip fracture patients evaluated for osteoporosis irrespective of age and sex to allow optimal assessment of the risk of future fractures. aKlotzbuecher et al. J Bone Miner Res. 2000 Apr;15(4):721–39.bMelton et al. Am J Epidemiol. 1989 May;129(5):1000–11.

Disclosures: J. Ryg, None.

This study received funding from: MSD, Univ. South. DK, Nat. Board of Health.


Ten-Year Probability of Osteoporotic Hip Fracture in 12958 Elderly Women Combining Clinical Factors and Heel Bone Ultrasound: The Combined “SEMOF + EPIDOS” Prospective Cohort. D. B. Hans1, A. M. Schott-Pethelaz*2, C. Durosier*1, J. A. Kanis3, H. Johanson*4, P. Burckhardt5, J. Cornuz*5, P. J. Meunier6, M. A. Krieg5 -, and the EPIDOS and SEMOF study groups*7. 1Nuclear Medicine, Geneva University Hospital, Geneva 14, Switzerland. 2Medical Information, Hospices Civils of Lyon, Lyon, France. 3WHO Collaborating Center for Metabolic Bone Diseases, University of Sheffield, Sheffield, United Kingdom. 4Consulting Statistician, Gothenburg, Sweden. 5Outpatient Clinic, Lausanne University Hospital, Lausanne, Switzerland. 6Inserm U403, HEH, Lyon, France. 7 France, Switzerland

The risk of osteoporotic fracture is multi-factorial and many independent clinical risk factors have been identified that contribute to risk above that reflected by BMD. DXA is not universally available and prospective studies have shown that heel bone ultrasound (HBU) predicts hip fracture as well as hip BMD. Our aim was to develop a model that combined clinical factors and Stiffness HBU to calculate the 10 years probability of hip fracture.

We grouped the EPIDOS (French) and SEMOF (Swiss) prospective cohorts (12958 women aged 70–100 years). At baseline, all the women were assessed by HBU Achilles+ (GE-Lunar) and a structured questionnaire for risk factors. 307 women reported a hip fracture during a mean follow-up of 3.2 ± 0.9 years. The association of risk factors with the risk of hip fracture and possible interactions were examined using a Poisson regression model and only significant ones were included into the model. Stiffness HBU was expressed as cohort-specific Z-score. We estimated the gradient of risk per SD and the ten-year probability of hip fracture of all possible combinations of selected clinical factors and Stiffness HBU.

The independent predictors of hip fracture were: Stiffness HBU (Stiffness), BMI, past fracture, recent fall, chair test, past HRT and current smoking. Gradient of risks at one year for the Stiffness alone was 2.33 and 2.24 for the entire model per standard deviation decrease. An example of 10 years probability of a 70 years old woman is given in the figure below according to the cumulative number of risk factors.

Thumbnail image of

This is the first study showing that combined use of clinical factors and Heel bone ultrasound improves the estimation of 10 years probability of hip fracture. Clinical evaluation is underway to further validate this model.

Disclosures: D.B. Mans, Synarc Inc. 4.


Identification of Osteoporotic Vertebral Fracture and Non-Osteoporotic Short Vertebral Height in Elderly Men. L. Ferrar1, G. Jiang*1, P. Cawthon*2, R. San Valentin*2, R. Fullman*2, L. Lambert*2, L. Palermo*2, S. Cummings2, D. Black2, E. Orwoll3, R. Eastell1. 1Clinical Sciences, University of Sheffield, Sheffield, United Kingdom. 2California Pacific Medical Center, San Francisco, CA, USA. 3University of Oregon, Oregon, OR, USA

Short vertebral height (SVH) is non-osteoporotic deformity that gives the radiological appearance of “reduced” vertebral height. The causes of SVH include normal or developmental variation, degenerative change and traumatic or pathologic fracture. In women, SVH is not associated with low bone mineral density (BMD). The aims of this study were to 1) test the associations between BMD and osteoporotic vertebral fracture (VF) and SVH identified by the algorithm-based qualitative method (ABQ) in elderly men and 2) determine whether discrepancies between methods are explained by SVH. We studied a subset of 732 white Caucasian men ages 65 years or older (mean 74, SD 6 years) participating in the Osteoporotic Fractures in Men Study (MrOS). Spinal radiographs were assessed visually by the ABQ and semiquantitative (SQ) methods and by quantitative morphometry (QM). Criteria for diagnosis of VF were as follows: ABQ, evidence of osteoporotic endplate depression; SQ, vertebral height reduction (estimated) > 20%; QM, vertebral height ratio (measured) > 3 SD below the reference mean. Radiographs were visually triaged prior to QM, as normal, uncertain or certain fracture (by estimated reduction in vertebral height > 20%). We report QM data for those triaged as uncertain or definite VF. The age-adjusted standard deviation units for BMD (z scores) were significantly lower than the population mean (zero) in men with VF (n = 74) but not SVH (n = 402) identified by ABQ. Prevalence of VF was a little higher for SQ and QM (13%) compared to ABQ (10%). Agreement between ABQ and SQ for identification of men with VF was moderate for all fractures (kappa, k = 0.58, 95% CI 0.49, 0.65), but poor for mild fractures (k = 0.36, 95% CI 0.19, 0.52). The SQ method agreed more closely with QM (k = 0.61, 95% CI 0.52, 0.71) than did ABQ (k = 0.49, 95% CI 0.33, 0.61). Bone density was lower than the population mean among men with VF agreed by all three methods (n = 40), but was average or higher in men with VF identified by any one method alone. Most discrepancies between ABQ and other diagnostic methods were related to mild thoracic wedging diagnosed as fracture by SQ or QM, but as SVH by ABQ and SQ or QM osteoporotic fractures diagnosed by ABQ as SVH due to traumatic fracture. We conclude that among elderly men 1) short vertebral height identified by ABQ is not related to low BMD and 2) short vertebral height, particularly in thoracic vertebrae is a major cause of disagreement between diagnostic methods.

Disclosures: L. Ferrar, None.


The Effect of Fracture on Bone Turnover Markers: A Longitudinal Study Comparing Marker Levels before and after Injury in 113 Elderly Women. K. K. Ivaska1, P. Gerdhem1, A. Isaksson*2, K. J. Obrant1. 1Department of Orthopaedics, Malmö University Hospital, Lund University, Malmö, Sweden. 2Department of Clinical Biochemistry, Lund University Hospital, Lund University, Lund, Sweden

Fracture, fracture healing and immobilization may alter serum and urine levels of bone turnover markers (BTM) and thus have implications on the use of markers in clinical practice. The aim of this prospective longitudinal study was to evaluate if BTMs are altered immediately after fracture and to study the changes during fracture healing.

Serum and urine samples were collected at admission to emergency room from 85 women (77.8+/−1.8 years old) who had sustained a major fracture after low-energy trauma. Control samples were collected from 28 women who experienced a similar trauma without a fracture. Follow-up samples were obtained four months after injury. Subjects were members of the Malmö Osteoporosis Prospective Risk Assessment (OPRA) study and all women had thus also been sampled on average 1.05+/−0.85 years before the injury event. This enabled us to compare BTM levels before and after fracture within each individual. Bone formation was assessed by serum osteocalcin (S-OC) and bone resorption by C-terminal crosslinked telopeptides of type I collagen (S-CTX) and tartrate resistant acid phosphatase 5b (S-TRACP5b). Furthermore, urinary osteocalcin (U-OC) was analyzed. All markers from all timepoints were assessed in batches at the same time.

No significant changes were seen in most BTMs in the immediate postfracture period (within one day after injury). Only U-OC was significantly increased when compared to prefracture levels (p <0.001). Four months after fracture both bone formation and resorption markers were significantly elevated. Increase was for S-OC 36–39% (p <0.001), for S-TRACP5b 6% (p = 0.009), and for S-CTX 17% (p = 0.004). Increase was evident especially after hip fractures (n = 30), but the trend was similar after other types of fractures as well. The kinetics of markers during fracture healing were different with an early increase in S-CTX and S-TRACP5b and a later rise in S-OC. The pattern for U-OC resembled the one for resorption markers.

This is the first study where prefracture levels of BTMs have been assessed and compared with postfracture levels. In conclusion, most BTMs sampled immediately after injury, with or without fracture, are not significantly altered compared to preinjury levels. They are, however, significantly elevated during fracture healing and the patterns for formation and resorption markers differ from each other. Ongoing fracture healing and postfracture immobilization clearly seem to have implications on the use of BTMs in clinical practice.

Disclosures: K.K. Ivaska, None.


Validation of DXA Hip Structural Analysis Implementations with a QCT Volumetric Gold Standard. S. Prevrhal1, J. A. Shepherd1, K. G. Faulkner2, K. W. Gaither2, T. F. Lang1. 1Radiology, UCSF, San Francisco, CA, USA. 2Synarc, Inc., San Francisco, CA, USA

Purpose: Hip Structural Analysis (HSA) extends dual x-ray absorptiometry (DXA) by measurement of geometrical and mechanical parameters and is widely used in osteoporosis trials. This study compares results from two implementations of HSA to volumetric Quantitative CT (QCT) measurements in the same population.

Materials and Methods: 121 women (107 white, 14 non-white, mean height 158 cm, weight 69 kg) participated. Each women receive 1 volumetric QCT with 3 mm slice thickness, 0.94 mm pixel size (9800, GE Healthcare) and 1 DXA scans of the left hip with standard hip mode (Prodigy, GE Healthcare). QCT scans were analyzed with in-house software that directly computed geometric and mechanical parameters at the neck and trochanteric regions. DXA HSA was performed with two methods: John Hopkins HSA and GE/Lunar HSA (part of the Prodigy release software). Pairwise linear regression of HSA variables was conducted by method to site matched QCT variables for BMD, cross-sectional area (CSA) and cross-sectional moment of inertia (CSMI) of the femur neck. Results: Table 1 shows correlation results. The Hopkins HSA correlated weakly with QCT, whereas GE/Lunar HSA correlated modestly. In addition, the two DXA HSA methods correlated only weakly with each other. Conclusion: Different implementations of DXA HSA do not produce equivalent and correlate differently to QCT. We conclude that the Hopkins HSA for the femur neck for Prodigy devices should be used with caution. 

Table  . Correlations Coefficients (R2) between methods
inline image

Disclosures: S. Prevrhal, None.

This study received funding from: NIAMS RO1-AR46197.


Connexin43 Is Required for Normal Skeletal Development and Bone Mass Acquisition. M. Watkins1, D. Ornitz*1, K. Willecke*2, R. Civitelli1. 1Bone and Mineral Diseases, Washington University in St. Louis, St. Louis, MO, USA. 2Institüt für Genetik, Universität Bonn, Bonn, Germany

Connexin43 (Cx43) is the major gap junction protein expressed in osteoblasts and osteocytes. Because Cx43−/− mice die soon after birth from cardiovascular defects, we have used a Cx43flox/LacZ gene replacement model and Cre-driving promoters to ablate Cx43 at different stages of osteoblast differentiation. Using the α1(I) collagen promoter, we have previously reported reduced peak bone mass, osteopenia, and dramatically attenuated response to PTH. Instead, deletion of Cx43 using osteocalcin promoters only minimally affects bone mass. In this study, we used the Dermo-1 (DM1) promoter, which is embryonically expressed in cells that give rise to chondrocytes and osteoblasts. DM1-Cre;Cx43-/flox mice are viable and of normal size at birth, but with age they remain substantially smaller than their littermates (body weight: −38% females, −47% males at 1 month), with only a modest tendency to recover at 6 months. Radiography revealed normally shaped but smaller axial skeleton and proximal limbs, with very rarefied trabeculae and thin cortices. Significantly reduced bone mineral density was detected by DEXA in DM1-Cre;Cx43-/flox mice relative to wild type littermates in both genders (−15% at 2 months), differences persisting up to 6 months. On histology, strongly LacZ-positive cells were observed in DMI-Cre;Cx43-/flox mouse tibias up to 6 months, selectively localized all along the endosteal surfaces (cortical and trabecular), and on the primary spongiosa, but excluding cells in the bone marrow space, confirming lineage specific gene deletion by DM1-Cre. Trabecular bone mass and cortical thickness were both reduced >50% in conditionally deleted animals relative to wild type littermates. Patchy alopecia and coarse hair were observed in DM1-Cre;Cx43-/flox mice, but no cardiovascular malformations. These animals also exhibited small, narrow skull and long nasal bones, resulting in a pointed snout. Similar craniofacial abnormalities occur in germline Cx43−/− mice, and are remindful of malformations reported in mice expressing a mutant form of Cx43 (G60S), which partially phenocopies the human disease oculodentodigital dysplasia. Therefore, embryonic Cx43 ablation in osteochondroprogenitor cells results in craniofacial malformations, stunted skeletal development and severe post natal osteopenia. These defects are more severe than those observed when Cx43 is deleted later in osteoblast differentiation, thus revealing a more prominent role of this gap junction protein at earlier rather than later steps of osteogenesis.

Disclosures: M. Watkins, None.

This study received funding from: NIH-NIAMS.


Overexpression of Hey1, a Notch Target Gene, Leads to Osteopenia in Mice Due to Decreased Osteoblast Performance. M. Susa, N. Zamurovic, R. Salie*, D. Rohner*, G. Evans*, M. Vukcevic*, M. Mueller*, B. Kinzel*, M. Kneissel. Novartis Institutes for BioMedical Research, Basel, Switzerland

We have previously shown that the transcription factor Hey1 is strongly up-regulated during BMP-2-induced osteoblastic differentiation. Down-regulation of Hey1 in cell cultures stimulated mineralization, while transient overexpression suppressed the activity of Runx2, the master regulator of osteogenesis. In this study we examined the role of Hey1 in bone metabolism by ubiquitously overexpressing Hey1 in mice within the beta-actin locus.

Heterozygous Hey1 transgenic mice showed no overt phenotype. Using realtime RT-PCR, Hey1 mRNA overexpression up to ∼35-fold higher than wild type levels was measured in several adult tissues, including long bones and calvariae. Hey1 protein overexpression was confirmed in primary osteoblasts by Western blotting. We monitored tibial bone mineral density and architecture in male and female heterozygous Hey1 transgenic mice at 12, 16 and 20 weeks of age by pQCT and micro CT. Hey1 overexpression induced progressive osteopenia in both genders. Ex vivo measurements also indicated decreased bone mass in the femur and the lumbar spine. The osteopenia was mainly due to the reduction of cancellous bone volume, caused by the loss of trabecular number and, in the spine, also of trabecular thickness. To understand the mechanism of bone loss, we have analyzed primary osteoblasts and osteoclasts from Hey1 transgenic mice and performed histomorphometric analyses of the long bones and vertebrae of the 20-week old mice. Osteoclasts derived from adult bone marrow showed a small reduction in activity, while cell number was less affected. Consistent with these in vitro findings, a trend for decreased osteoclast surface was observed in vivo. Osteoblast cultures derived from calvariae of neonatal Hey1 transgenic animals showed no change in their proliferation rate, but alkaline phosphatase was reduced, while mineralization was completely abolished at a low concentration of osteogenic stimulus. These results indicated inhibition of osteoblast differentiation and bone nodule formation. Interestingly, mineral apposition rate was decreased in the vertebrae, in agreement with the reduced trabecular thickness and the reduced osteoblast performance in vitro. Ongoing studies are directed at the identification of genes involved in this process.

In agreement with our previous in vitro findings, our data suggest that Hey1 plays an inhibitory role in the process of osteoblast maturation in vitro and in vivo. As Hey1 is a target gene of the receptor Notch, our data imply the involvement of the Notch pathway in regulation of bone metabolism.

Disclosures: M. Susa, Novartis Pharma AG 3.

This study received funding from: Novartis AG.


Deletion of the G Protein Subunit Gsα in Early Osteoblasts Results in Fractures at Birth. J. Y. Wu1, T. Kobayashi1, S. Rodda2, A. P. McMahon*2, H. M. Kronenberg1. 1Endocrine Unit, Massachusetts General Hospital, Boston, MA, USA. 2Department of Molecular and Cellular Biology, Harvard University, Cambridge, MA, USA

In endochondral bone formation the differentiation of osteoblasts is closely coupled to chondrocyte maturation; however, the pathways regulating osteoblastic differentiation are not well delineated. An important function for parathyroid hormone/parathyroid hormone-related peptide receptor (PPR) signaling has been suggested by the dramatic increase in the amount of trabecular bone in transgenic mice expressing a constitutively active form of PPR in osteoblasts. Gsα is a ubiquitously expressed G-protein alpha subunit that serves as an important downstream effector for several G protein-coupled receptors including PPR. To investigate the impact of signaling pathways downstream of PPR in regulating skeletal development, we generated mice with a conditional deletion of Gsα in osteoblasts expressing the transcription factor osterix (Osx), a gene expressed early in the osteoblastic lineage. OsxCre:Gsα(fl/fl) mice were born at the expected frequency and were morphologically similar to wild-type littermates at birth. However, postnatal growth was significantly impaired, and no mutant mice survived past three weeks of age. Skeletal preparations stained with Alcian blue and Alizarin red did not reveal any defects in skeletal patterning. In contrast, bone mass and strength was dramatically decreased as demonstrated by the presence of fractures of the posterior ribs and femurs by postnatal day 1. Bone mineral density was reduced by greater than 40% in the femurs of mutant mice. Histological analysis of postnatal long bones revealed marked attenuation of both trabecular and cortical bone. To assess osteoblast differentiation in mutant mice, in situ hybridization was performed using markers of progressive stages in the osteoblast lineage. While expression patterns of early markers of osteoblast differentiation were unaltered in mutant mice, there was near-complete loss of osteocalcin expression by two weeks of age. TRAP staining for osteoclasts was not increased in mutant mice, suggesting that the defect lies primarily in bone formation rather than excessive bone resorption. These findings indicate a critical role for Gsα-mediated signaling early in the osteoblast lineage during endochondral ossification.

Disclosures: J. Y. Wu, None.


Osteoblast Deficiency in sox4 Heterozygous Mice. N. Rucci1, L. S. H. Nissen-Meyer*2, R. Jemtland*3, V. T. Gautvik*2, M. Pedersen*2, R. Paro*1, D. Fortunati*2, A. Del Fattore*1, A. Rufo*1, S. Reppe*2, D. Pierroz*4, S. Ferrari*4, F. P. Reinholt*5, A. Teti1, K. M. Gautvik6. 1Department of Experimental Medicine, University of L'Aquila, L'Aquila, Italy. 2Department of Biochemistry, University of Oslo, Oslo, Norway. 3Department of Medicine, National University Hospital, Oslo, Norway. 4Department of Rehabilitation and Geriatrics, Geneva University Hospital, Geneva, Switzerland. 5Department of Pathology, National University Hospital, Oslo, Norway. 6Department of Biochemistry, University of Oslo and Ullevaal University Hospital, Oslo, Norway

Sox4 is a BMP2- and PTH-regulated transcription factor expressed by hypertrophic chondrocytes and osteoblasts. Sox4 null mice die in utero due to circulatory failure. Even if sox4+/− mice appear normal, we found a lower BMD and BMC from early age, with a failure to reach the wildtype (WT) peak bone mass. The differences were most prominent in trabecular bone (lumbar spine, trochanter and proximal tibiae) but cortical bone was also affected. MicroCT confirmed impaired bone architecture in sox4+/− mice, and in vivo double fluorochrome labeling demonstrated a low mineral apposition rate. Consistently, histomorphometry showed lower Ob.S/BS and unchanged Oc.S/BS in sox4+/− mice vs. WT. In vitro experiments demonstrated a cell autonomous defect in sox4+/− calvarial osteoblasts, with remarkably less alkaline phosphatase (ALP) mRNA expression and activity, lower proliferation and impaired mineralization relative to WT cells. Sox4+/− osteoblasts also exhibited lower mRNA expression of osterix, osteocalcin and collagen1A2, while transcription of runx-2 and osteopontin was unaffected. Bone marrow cultures from sox4+/− mice confirmed normal osteoclast formation and activity compared to WT cultures, with similar expression of osteoclastogenic cytokines by sox4+/− and WT osteoblasts. Sox4 knockdown by siRNA reproduced sox4 insufficiency in osteoblasts, with reduction of differentiation and function, selective down-regulation of ALP, osterix, osteocalcin and collagen1A2, and again no effect on runx2 and osteopontin. In contrast, runx2 silencing by siRNA, while expectedly reducing osterix and ostecalcin mRNA expression, was without effect on sox4 transcription, thus indicating that runx2 and sox4 affect osterix/osteocalcin expression through independent pathways. In conclusion, sox4 transcriptional activity is upstream of osterix and independent of runx2, and its haploinsufficiency negatively affects the peak bone mass due to an osteoblast failure. Collectively, these results provide a novel insight into the molecular mechanisms of osteoporosis.

Disclosures: N. Rucci, None.


Mice Lacking the Novel Transmembrane Protein Osteopotentia Exhibit Catastrophic Defects in Bone Modeling. M. L. Sohaskey*1, Y. Jiang2, J. Zhao*2, A. Mohr*3, F. Roemer*3, H. K. Genant*3, R. M. Harland*1. 1Molecular and Cell Biology, University of California, Berkeley, CA, USA. 2Division of Musculoskeletal Radiology, University of Michigan, Ann Arbor, MI, USA. 3Osteoporosis and Arthritis Research Group, University of California, San Francisco, CA, USA

Strictly coordinated bone modeling processes generate and ensure the integrity of the developing skeleton. Deregulation of such processes results in pathophysiological changes and metabolic bone disease. In an insertional mutagenesis screen for essential regulators of mammalian development, we identified Osteopotentia (Opt), a gene encoding a novel transmembrane protein widely expressed in the developing embryo. We have investigated the consequences of Opt mutation for postnatal skeletal development using genetic, biochemical, histological and radiological approaches. Mice homozygous for an insertional mutation in the Opt locus die predominantly at birth of undetermined causes. Mice surviving this perinatal period develop striking skeletal abnormalities similar to those seen in humans and genetically engineered mouse models having severe forms of the brittle bone disease osteogenesis imperfecta. Abnormalities include inflammation and hemorrhaging around limbs, deformation of long bones, and aberrantly healed fractures evidenced by the formation of hypertrophic calluses on ribs and long bones. Surviving Opt homozygous mutant mice exhibit a marked decrease in the amount of trabecular and cortical bone relative to normal littermates. Quantitative micro-CT imaging confirms significant reductions in cortical thickness and 3D trabecular bone volume fraction, number and thickness, with increased 3D trabecular spacing. Decreased levels of osteocalcin mRNA and protein in bone and serum, respectively, suggest that low bone mass in Opt homozygous mutants may result from a primary defect in osteoblast differentiation. Consistent with this hypothesis, ex vivo studies using primary calvarial osteoblast cultures reveal a cell-autonomous impairment of differentiation in Opt mutant osteoprogenitors, despite a normal proliferative index. Surprisingly, osteoclast function also appears to be compromised in Opt mutant mice, implying a state of low bone turnover in which increased bone resorption does not contribute to the fragility of the Opt mutant skeleton. Overall, clarification of the physiological role played by Opt will be important in establishing how mutagenesis of this novel gene leads to a dramatic failure of bone modeling in the mouse. Our findings may also provide mechanistic insight into the etiology of recessive forms of osteogenesis imperfecta demonstrating no genetic linkage to either type I procollagen locus.

Disclosures: M.L. Sohaskey, None.

This study received funding from: Osteogenesis Imperfecta Foundation.


The Hypoxia-Inducible Factor Pathway Couples Angiogenesis to Osteogenesis during Skeletal Development. Y. Wang*1, C. Wan1, X. Liu*1, M. L. Bouxsein2, M. Faugere3, R. S. Johnson*4, V. H. Haase*5, L. C. Gerstenfeld6, E. Schipani2, T. L. Clemens1. 1University of Alabama at Birmingham, Birmingham, AL, USA. 2Harvard Medical School, Boston, MA, USA. 3University of Kentucky, Lexington, KY, USA. 4University of California, SD, San Diego, CA, USA. 5University of Pennsylvania, Philadelphia, PA, USA. 6Boston University, Boston, MA, USA

Skeletal development occurs in close spatial and temporal association with angiogenesis. Osteoblasts are ideally positioned in developing bone to sense fluctuations in oxygen and respond to hypoxia by elevating the hypoxia-inducible factor 1 (HIF-1), which activates angiogenic gene programs (e.g. VEGF). To investigate the role of the HIF-1α pathway in developing bone, we used Cre/LoxP recombination methods to constitutively activate HIF-1 by selectively disrupting Von Hippel-Lindau (VHL) protein in osteoblasts, the ubiquitin ligase that degrades HIF-1. Mice lacking VHL (ΔVHL) expressed high levels of HIF-1 and VEGF and developed extremely thick, dense bones (BV/TV: 41.62 ± 8.39 vs. 12.58 ± 2.52 at 6 wks). Double calcein labeling showed that mutant bones had more labeled surfaces compared to controls. Surprisingly, this increased bone modeling was associated with a reduction in osteoblast number. Micro CT imaging of vasculature in Microfil® perfused bones revealed a striking increase in vessel volume, number and connectivity (> 4 fold) with a reduction in vessel separation in ΔVHL mice. To confirm the role of VEGF in this phenotype, we performed angiogenesis assays using explants of E17.5 mouse metatarsals. ΔVHL metatarsals were more vascularized than controls and this response was entirely blocked with VEGF-neutralizing antibody. To examine the effect of HIF-1 activation on osteogenesis independent from angiogenesis, primary osteoblasts from the floxed VHL calvaria were infected with adenoviral vectors expressing Cre or GFP as control. Osteoblasts lacking VHL had increased levels of HIF-1 and VEGF mRNA but proliferated and differentiated similar to controls. These results indicate that the increased osteogenesis seen in the ΔVHL mice results mainly through cell non-autonomous mechanisms which promote increased angiogenesis. Finally, to unequivocally establish the role of HIF-1 we analyzed a second mouse model in which HIF-1 was conditionally disrupted in osteoblasts using the same (OC) promoter to direct Cre. Phenotypes of HIF-1 deficient mice were opposite from the VHL mutants: long bones were significantly thinner and less vascularized compared to controls. Based on these results we propose that activation of the HIF-1 pathway in osteoblasts during bone development initiates increased bone modeling events by coordinating the timing, direction and degree of new blood vessel formation into bone.

Disclosures: Y. Wang, None.

This study received funding from: NIH grant R01 AR049410.


A Novel Function of Mature Osteoblasts: Regulation of Differentiation of Early Precursors through A Glucocorticoid Dependent Pathway. H. Zhou1, W. M. Mak*1, I. Kalajzic2, C. R. Dunstan1, M. J. Seibel1. 1Bone Research Program, ANZAC Research Institute, Sydney University, Sydney, Australia. 2Genetics and Developmental Biology, University of Connecticut health Center, Farmington, CT, USA

Transgenic (tg) expression of 11beta-hydroxysteroid dehydrogenase type 2 (HSD2), a glucocorticoid (GC) inactivating enzyme, under the control of a 2.3Kb collagen type I promoter (Col2.3-HSD2) abrogates intracellular GC signalling in mature osteoblasts. To investigate the effects of disrupted GC signalling on osteoblast differentiation, primary osteoblast cultures were generated from the calvaria of 1-day-old Col2.3-HSD2 tg mice and wild type (WT) littermates. mRNA for HSD2 was only detected in cells cultured from tg mice, with increasing expression levels observed from day 3 to day 7, consistent with a modest increase in numbers of mature osteoblasts. Compared to WT cultures, tg cultures formed 50% fewer nodules but greatly increased adipocyte numbers, suggesting a shift in lineage commitment from osteoblast to adipocyte. Thiram, an HSD2 specific inhibitor, rescued the tg phenotype. At day 1, mRNA for Runx2, ALP, and osteocalcin was expressed at low but similar levels in tg and WT cultures. These levels remained low after 3 days in tg cultures, while they increased with osteoblast differentiation in the WT cultures. Conversely, expression of adipogenic transcription factors, C/EBP alpha and PPAR gamma, increased from day 3 only in tg cultures.

These results suggest cross-talk between mature osteoblasts and their precursors, whereby mature osteoblasts direct precursor lineage commitment. We therefore used transwell techniques to co-culture tg and WT early precursor cells of the first calvarial digestion (P1) with more differentiated cells of the third and fourth digestions (P3–4) as confirmed by Col2.3-GFP primary osteoblast cultures. In this system, P3–4 cells from WT, but not from tg mice induced nodule formation in both WT and tg P1 cells. No nodules were formed in P1 cells cultured alone. Wnt10b and Wnt7b expression increased during osteoblast differentiation in WT cultures but remained low in tg cultures. Conversely, sFRP1, a Wnt inhibitor, expressed at higher levels in tg culture. Beta-catenin protein levels and mRNA expression of its target gene, TCF4, were higher in WT cultures.

We conclude that disrupted GC signalling in mature osteoblasts impairs early osteoblast differentiation and promotes adipogenic differentiation. This suggests the existence of a GC regulated secreted paracrine signal between mature osteoblasts and their mesenchymal precursors that influences lineage commitment. Canonical Wnt could play a central role in this signalling process.

Disclosures: H. Zhou, None.

This study received funding from: NHMRC.


Novel Role of 1,25-dihydroxyvitamin D3 in Inducing Antimicrobial Peptides in Human Bronchial Epithelial Cells. S. Yim*1, S. Christakos2, G. Diamond*1. 1Oral Biology, UMDNJ-New Jersey Dental School, Newark, NJ, USA. 2Biochemistry, UMDNJ-New Jersey Medical School, Newark, NJ, USA

The host innate immune system forms a crucial first line of defense that protects the airway from invading pathogens. A number of effector molecules contribute to this defense system, including antimicrobial peptides (APs) such as β-defensins and cathelicidins. An increase in antibiotic resistant bacteria has become a major public health concern; therefore, the agents that induce the endogenous AP production have become excellent candidates for therapeutics able to eradicate antibiotic resistant pathogens. Recent studies have shown that the hormonal form of vitamin D3, 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) up-regulates AP gene expression in several established cell lines. However, the effect of 1,25(OH)2D3 on host defense gene expression in primary cultures of airway remains unkown. In this study, we investigated the effect of 1,25(OH)2D3 on AP mRNA expression in normal human bronchial epithelial (NHBE) cells by RTQ-PCR. When NHBE cells were treated with 1,25(OH)2D3 (10−8M), cathelicidin antimicrobial peptide (CAMP) mRNA expression level was up-regulated three-fold as early as 3 hours of incubation, and the highest induction (10-fold) was observed at 12–24 hours of incubation. Human β-defensin (HBD) 3 mRNA expression level was also up-regulated six-fold after 12 hours of incubation, and 12-fold increased expression was observed after 48 hours of incubation. In addition, the synergistic role of calcium was studied as calcium is an important regulator of various cellular functions in the epidermis. Pre-treatment of cells with 1.0 mM calcium for 24 hours increased CAMP mRNA levels 1.5 to 2-fold compared to cells treated with 1,25(OH)2D3 alone (Induction with Ca++ + 1,25(OH)2D3 = 15-fold). Moreover, 1,25(OH)2D3 induced antimicrobial activity against airway pathogens including Bordetella pertussis. A 40% reduction in colony forming units (CFU) in culture supernatant was observed when B. pertussis was treated with the medium collected from 1,25(OH)2D3 stimulated cells compared to the unconditioned medium. 1,25(OH)2D3 or the calcium pre-treatment has no affect on expression of other APs, including HBD-1, HBD-2 and HBD-4. Elucidation of the effect of 1,25(OH)2D3 on APs in NHBE cells will aid in the development of novel therapeutic agents for treatment of airway infection.

Disclosures: S. Yim, None.

This study received funding from: NIHROIHL67871.


Transgenic Expression of the Human Vitamin D Receptor (hVDR) in the Proximal Small Intestine Restores Vitamin D-Responsive Active Calcium Absorption but Is Insufficient to Support Adequate Calcium Homeostasis. H. D. Marks*1, Y. Xue*2, J. C. Fleet2, S. Peleg1. 1Endocrine Neoplasia, University of Texas M.D. Anderson Cancer Center, Houston, TX, USA. 2Foods and Nutrition, Purdue University, West Lafayette, IN, USA

The active form of vitamin D, 1,25(OH)2D3, regulates calcium homeostasis by acting on the intestine, bone, and kidney. These actions are mediated through the vitamin D receptor (VDR) as is evident from development of hypocalcemia and severe rickets in hereditary vitamin D resistant rickets in man and in mVDR−/− mice. The purpose of this study was to determine if the vitamin D-regulated active calcium absorption machinery in the proximal small intestine is sufficient to restore calcium homeostasis and prevent rickets. To that end, we generated transgenic mice expressing human VDR (hVDR) exclusively in the proximal small intestine of mVDR−/− mice by using the adenosine deaminase enhancer (hVDR+/mVDR−/−).

Unlike wild-type (WT) mice, which express the VDR throughout the intestine and in other tissues, hVDR+/mVDR−/− mice express hVDR only in the proximal small intestine. Furthermore, while VDR target genes (TRPV6, calbindin D9K, 24-hydroxylase) were up-regulated in the proximal small intestine of hVDR+/mVDR−/− mice injected with 1,25(OH)2D3, up-regulation of these genes by 1,25(OH)2D3 was detected in both the proximal small intestine and in the large intestine of WT mice. Finally, a ligated duodenal loop assay determined that the calcium absorbing machinery in the proximal small intestine of hVDR+/mVDR−/− mice was as responsive to 1,25(OH)2D3 as was that of WT mice. Despite having functional hVDR in the proximal small intestine, hVDR+/mVDR−/− mice were hypocalcaemic and rachitic when fed a normal rodent diet (1% Ca, 0.67% P) at weaning, as were the mVDR−/− mice fed the same diet. The hypocalcaemia and rickets in both mVDR−/− and hVDR+/mVDR−/− mice was prevented if they were given the rescue diet (2% Ca, 0.8% P, 20% lactose) at weaning. However, when 90-day-old rachitic mice were fed a rescue diet, serum calcium was normalized and bone calcification improved in hVDR+/mVDR−/− mice, but not in mVDR−/− mice.

In conclusion, although transgenic hVDR in the proximal small intestine of mVDR−/− mice was transcriptionally active and regulated active calcium absorption, it is likely that VDR actions elsewhere in the intestine are necessary to support adequate calcium homeostasis. In addition, because hVDR+/mVDR−/− mice responded better to the late rescue diet than did the mVDR−/− mice, we hypothesize that expression of VDR in the proximal small intestine protects the calcium absorbing machinery from age-dependent decline.

Disclosures: H.D. Marks, None.

This study received funding from: National Institutes of Health.


Phosphorylation of the Human Retinoid (hRXRα) at a MAP Kinase Consensus Sequence on Serine 260 Inhibits Nuclear Import and Coactivator Recruitment of the Vitamin D Receptor (VDR)/hRXR Complex. L. Nguyen-Yamamoto1, S. Laporte*1, S. Mader*2, M. Marcoritto*1, D. C. Huang1, R. Kremer1. 1Medicine, Royal Voctoria Hospital McGill University, Montreal, PQ, Canada. 2Biochimie, Universite de Montreal, Montreal, PQ, Canada

We previously demonstrated that phosphorylation of the hRXRα by MAP kinase at serine 260 results in partial resistance to the growth inhibitory effect of 1,25 dihydroxyvitamin D3 (1,25(OH)2D3). This resistance can be reversed by treatment with a MAP kinase inhibitor and by overexpression of a non-phosphorylable hRXRα Ala260Ser mutant. In the presence study the mechanism of this resistance was explored using fluorescence resonance energy transfer (FRET) to analyze in detail cellular trafficking and the interaction between the VDR and hRXRα and its coactivators. Immortalized (HPK1 A) and ras-transformed (HPK1Aras) keratinocytes and the human brest cancer cell line (MDA-MB-231) were co-transfected with green fluorescence protein (GFP) labeled hRXRα or a coactivator consensus sequence (LXXLL) together with red fluorescence protein (RFP) labeled VDR. Both HPK1Aras and MDA-MB-231 expressed high levels of p42/p44 as compared to HPK1A cells. First, nucleocytoplasmic distribution was assessed in the presence or absence of 1,25(OH)2D3 (10−7M). Next, cells were treated with the MAP kinase inhibitor UO126 (10−6M) (Promega) or transiently transfected with a vector overexpressing the hRXRα Ala260ser mutant or an empty vector (control). Trafficking and interaction of the VDR/RXR/coactivator complex was then analyzed at timed intervals. In the absence of 1,25(OH)2D3 hRXRα-GFP, LXXLL-GFP and VDR-RFP distribution was 60% nuclear and 40% cytoplasmic. Addition of 1,25(OH)2D3 to HPK1A cells induced a rapid (3 hours) nuclear import of the VDR/hRXRα complex (80% nuclear and 20% cytoplasmic) and recruitment of LXXLL-GFP as indicated by yellow merged images and micro-foci formation. In contrast, addition of 1,25(OH)2D3 to HPK1Aras cells resulted in a diminished nuclear import of the VDR/hRXRα complex and LXXLL-GFP recruitment, with a 55% accumulation in the perinuclear region, as compared to HPK1A cells. Treatment with UO126 or transfection with a vector overexpressing hRXRα Ala260Ser mutant restored nuclear import of VDR/hRXRα complex and LXXLL-GFP recruitment to the levels seen in HPK1A cells (80% nuclear and 20% cytoplasmic). Similar results were observed in MDA-MB-231 cells.

Our study therefore demonstrated that serine 260 phosphorylation of the hRXRα is a central modulator of vitamin D action by regulating nucleocytoplasmic trafficking and interaction with its VDR heterodimeric partner and coactivator complex.

Disclosures: L. Nguyen-Yamamoto, None.


The Stimulatory Effects of Phosphate Depletion on 25-Hydroxyvitamin D-1α-Hydroxylase (Cyp27b1) Expression Are Modulated by the Secreted Form of Klotho Protein. M. Takaiwa, B. Yuan, J. J. Dufek*, M. K. Drezner. University of Wisconsin and GRECC, VAMC, Madison, WI, USA

Although recent studies in mouse strains with mutations in the klotho (Kl) gene suggest that the klotho protein may influence renal 25-hydroxyvitamin D-1α-hydroxylase expression, it remains unknown if the klotho protein plays a role in modulating the effects of known hormonal/metabolic stimuli of Cyp27b1. Nevertheless, independent investigations have revealed that oral P-depletion, a known stimulus of Cyp27b1, does increase klotho protein in the murine kidney. Any conclusions from these studies, however, have been limited by recent reports, which revealed the isolation of two Kl transcripts, secratory (s)-Kl and membrane bound (m)-Kl, arising from alternative RNA splicing, necessitating new approaches to studies of Kl protein in which independent effects of the klotho isoforms are evaluated. In the present study, therefore, we investigated if either of the klotho gene products may modulate the effects of P-depletion and consequent hypophosphatemia on the regulation of renal 1,25(OH)2D synthesis. Initially, we observed that 3 days of P-depletion in C57B16J mice increased renal s-Kl mRNA expression, reaching a 1.9 ± 0.09 fold enhancement (p <0.01) by day 10. In concert with stimulation of s-Kl expression, we noted significant stimulation (p <0.01) of Cyp27b1 expression, reaching a maximum of 2.63 ± 0.46 fold. Subsequently we sought to causally relate the change in s-Kl expression to stimulation of Cyp27b1. In this regard, we documented by immunohistochemistry and micro-dissection of mouse renal tubules the localization of s-Kl (and m-Kl) mRNA and protein in proximal renal tubule cells, the site of Cyp27b1 expression. In addition, we confirmed the P-dependent effects on s-Kl expression in in vitro studies; 6 hours after exposure to low P medium, significant (p <0.05) expression of s-KL (but not m-KL) was induced in HEK293 (renal tubule) cells. Moreover, we found that transfection and subsequent expression of recombinant mouse s-Kl protein in the HEK293 cells caused a 2- to 3-fold increase in CYP27B1 mRNA expression (p <0.01). In contrast, neither the presence of purified s-Kl protein in the cell growth medium nor the transfection and expression of recombinant mouse m-Kl protein altered CYP27b1 expression in the HEK293 cells. These findings indicate that the stimulatory effects of P-depletion on 1,25(OH)2D synthesis in renal proximal tubules are modulated by the positive regulatory actions of s-Kl on Cyp27b1 expression.

Disclosures: M. Takaiwa, None.


1α,25(OH)2D3-Induced DNA Methylation Mediates the Transrepression by VDR. M. Kim*1, A. Murayama2, K. Takeyama1, S. Kato1. 1University of Tokyo, Institute of Molecular and Cellular Biosciences, Tokyo, Japan. 2University of Tsukuba, Graduate School of Life and Environmental Sciences, Tsukuba Science City, Japan

VitaminD receptor(VDR) is a nuclear receptor(NR) regulating bone metabolism and calcium homeostasis. VDR modulates its target gene expression at the transcriptional level in a ligand -dependent manner. To understand the ligand-dependent transrepression mechanism by NR, we have picked up vitaminD3 1alpha-hydroxylase[1α(OH)ase] gene, which is negatively regulated by 1α,25(OH)2D3-bound VDR. Recently, We have demonstrated that a bHLH transcription facor, VDIR directly bonud on the negative vitamin D response element in the promoter[1αnVDRE], and ligand-dependent association of VDR with VDIR recruited an HDAC2 co-repressor complex for ligand-dependent manner(EMBO J. 23, 1598, 2004). Histone deacetylation was a critical step for remodeling chromatin structure of the 1α(OH)ase gene promoter(EMBO J. 24, 3881, 2005). However, TSA(HDAC inhibitor) could not fully abrogate 1α,25(OH)2D3-induced transrepression, suggesting that an additional repression mechanism should support this transcriptional repression. To explore this unkown mechanism, we tried to purify the VDIR-VDR co-repressor complex formed a 1α,25(OH)2D3-dependent manner. By using the biochemical purification approach, we identifed Dnmt1(DNA methyltrasferase) that related to gene repression. The effect of 1α,25(OH)2D3 on transrepression was significantly enhanced by Dnmt1. Dnmt1 interacted with both VDR and HDAC2 in a ligand-dependent manner. Furthermore, a ChIP assay showed that HDAC and Dnmts recruitment occurred sequentially on the 1α(OH)ase gene promoter after an 1α,25(OH)2D3 treatment. In fact, DNA methylation was induced at CpG sites in the promoter and exon regions of 1α(OH)ase gene, by 1α,25(OH)2D3 for 24hours. Using mouse embryonic fibroblasts from VDR knockout mice, we further found that VDR is required for 1α,25(OH)2D3 -dependent DNA methylation of 1α(OH)ase gene. All together, these findings suggest that ligand-induced DNA methylation in the promoter may contribute, at least in part, to continuous transcriptional repression by VDR, and provide a novel mechanism of a ligand-dependent transrepression by NR that links repressive histone modification with epigenetic repression through DNA methylation.

Disclosures: M. Kim, None.

This study received funding from: ERATO, JST.


Effect of Three Month Treatment with the Cathepsin-K Inhibitor, Balicatib, on Biochemical Markers of Bone Turnover in Postmenopausal Women: Evidence for Uncoupling of Bone Resorption and Bone Formation. P. Papanastasiou*1, C. E. Ortmann*1, M. Olson*1, A. Vigneron*2, U. Trechsel1. 1Novartis Pharma AG, Basel, Switzerland. 2Novartis, Rueil-Malmaison, France

Cathepsin K (cathK)is a cysteine protease that is selectively, highly expressed by bone resorbing osteoclasts and is believed to play a key role in bone resorption. Balicatib is an orally bioavailable, novel inhibitor of human cathK. It is predicted that balicatib will prevent bone loss in osteoporosis[dl]. In contrast to the mechanisms of established treatments, cathK inhibitors act late in the resorption process, and should not affect osteoclast/osteoblast interactions. Therefore, bone formation may be less decreased than with other resorption inhibitors, leading to a positive bone balance at the level of the bone remodeling unit. In this 12-week RPC trial, 140 postmenopausal women (28 patients per group) were treated with 5, 10, 25, or 50 mg oral balicatib daily, or placebo. Bone resorption markers were serum CTX, and serum and urinary NTX. Formation markers were serum osteocalcin (OC) and bone-specific alkaline phosphatase (BSAP). They were measured at baseline, and at 1,4,8, and 12 weeks of treatment before the daily dose. sCTX was additionally measured at 10h after the first dose.

Results: sCTX showed the greatest reduction at 10h after the first dose decreasing to −90% of its baseline value, in the 50 mg balicatib dose group. Resorption markers showed a rapid and sustained decrease over the entire treatment period. In contrast, there was little change in formation marker levels. The table below shows percent changes from baseline at 12 weeks (ITT population, LOCF). 

Table  .  
  1. * P <0.05 for treatment comparison balicatib vs. placebo from log-scale ANCOVA Balicatib was generally safe and well tolerated at all doses. However, there was a higher incidence of skin adverse events (mainly pruritus) on balicatib >5mg.

inline image

Conclusion: the decrease on bone resorption markers with little decrease in bone formation markers with balicatib suggests an uncoupling of these processes in favor of formation. Whether this translates into an increase in bone mineral density or antifracture efficacy superior to that seen with other resorption inhibitors is under investigation.

Disclosures: P. Papanastasiou, Novartis Pharma AG 3.

This study received funding from: Novartis.


Effect of Cathepsin K Inhibition on Bone Resorption Markers in Healthy Postmenopausal Women. S. A. Stoch*, D. L. Miller*, K. Van Dyck*, B. Jin*, D. Panebianco*, O. Liu*, J. Stone*, K. M. Gottesdiener*, J. A. Wagner*. Clinical Research, Merck & Co. Inc., Rahway, NJ, USA

Postmenopausal osteoporosis is a common disorder characterized by an increase in bone resorption relative to bone formation, generally resulting in an increase in bone turnover. The bone loss is brought about by an imbalance between bone resorption and bone formation. Osteoclastic bone resorption requires 2 processes: deminineraliztion of the inorganic bone components and degradation of the organic bone matrix. Cathepsin K is the most abundant cysteine protease expressed in the osteoclast and is necessary for bone matrix degradation required for bone resorption. It was hypothesized that MK-0822, a potent cathepsin K inhibitor, would inhibit bone resorption. A randomized, double-blind, placebo-controlled, sequential group, multiple oral dose study conducted in 30 postmenopausal women to assess safety, tolerability, pharmacokinetics, and the pharmacodynamic profile of MK-0822 during multiple oral dosing. Three panels of 10 postmenopausal female subjects (8 active, 2 placebo) were administered 0.5, 2.5, and 10 mg MK-0822 or placebo once daily for 21 days. Subjects participated in only one panel and received only one dose level. Blood and urine samples were obtained at selected time points pre- and postdose for measurement of serum CTx, urinary NTx/Cre, bone specific alkaline phosphatase (BSAP), and osteocalcin. Robust reductions in urine NTx (∼80%) were seen following daily administration of 2.5 mg and 10 mg for 21 days in postmenopausal women. More modest reductions in urine NTx (∼15%) were seen following 0.5 mg administered once-daily for 21 days in postmenopausal women. The effects of MK-0822 on serum CTx were similar to those seen on urine NTx. Serum CTx was suppressed ∼15%, ∼70%, and ∼80% following 21 days of dosing MK-0822 at 0.5, 2.5, and 10 mg, respectively. No significant effect was seen on bone formation markers (BSAP and osteocalcin). MK-0822 was generally well tolerated, with no subjects discontinuing due to adverse experiences. In conclusion, the cathepsin K inhibitor, MK-0822, results in robust suppression of bone resorption markers (urine NTx/Cr, serum CTx) when administered once-daily to postmenopausal women for 21 days. These data suggest that the cathepsin K inhibitor, MK-0822 may be developed as an effective therapeutic agent for osteoporosis.

Disclosures: S.A. Stoch, None.


Effect of Calcium Supplementation on Fracture Risk: A Double-Blind Randomized Controlled Trial. H. A. Bischoff-Ferrari1, J. R. Rees*2, M. V. Grau*2, J. A. Baron*2. 1Rheumatology, University Hospital Zurich, Zurich, Switzerland. 2Dept. of Community and Family Medicine, Dartmouth Medical School, Hanover, NH, USA.

Context: Calcium supplementation is commonly recommended for the prevention of osteoporotic fractures despite conflicting evidence from randomized controlled trials.

Objective: To determine whether four years of calcium supplementation reduced fracture risk in a randomized, placebo-controlled trial.

Design, setting and patients: This is a secondary analysis of the effects of calcium on fracture risk in a multi-center, double-blind randomized controlled trial of colorectal adenoma chemoprevention between 1988 and 1996. Participants were aged <80 and generally healthy, with a recent diagnosis of colorectal adenoma. After randomization, 930 participants (72% male, mean age 61 years) were followed for a mean of 10.3 years. All reported fractures were confirmed by medical record review.

Intervention: Patients were randomly assigned to receive four years' treatment with 3 g calcium carbonate (1200mg of elemental calcium) daily or placebo.

Main outcome measure: The primary outcomes were all fractures and minimal trauma fractures (caused by a fall from standing height or less while sitting, standing or walking).

Results: Fractures were confirmed among 46/464 participants in the calcium group and 54/466 in the placebo group. The time to first fracture did not differ significantly between groups (hazard ratio [HR], 0.87; 95% confidence interval [CI], 0.58–1.28). Minimal trauma fractures were confirmed in 15/464 participants in the calcium group and 29/466 in the placebo group. The rate of minimal trauma fractures was significantly lower in the calcium group (HR 0.52; 95% CI, 0.28–0.97; P = 0.04), with a similar effect in men (HR, 0.55; 95% CI, 0.25–1.20) and women (HR, 0.50; 95% CI, 0.18–1.42).

Conclusion: Calcium supplementation significantly reduced the risk of minimal trauma fracture among healthy individuals.

Disclosures: H.A. Bischoff-Ferrari, None.


Evidence for Hip Fracture Risk Reduction with Calcium and Vitamin D from a Comparative Meta-Analysis of Randomized Controlled Trials Including RECORD and WHI. S. Boonen*1, R. Bouillon*1, D. Vanderschueren1, P. Haentjens*2, P. Lips3. 1Center for Metabolic Bone Diseases, Katholieke Universiteit Leuven, Leuven, Belgium. 2Department of Orthopaedics and Traumatology, Vrije Universiteit Brussel, Brussels, Belgium. 3Department of Endocrinology, VU University Medical Center, Amsterdam, The Netherlands.

Context: Recent studies - including the Randomised Evaluation of Calcium Or vitamin D (RECORD) and the Women's Health Initiative (WHI) trials have concluded that the effectiveness of calcium and vitamin D alone in hip fracture prevention is unclear.

Objective and Design: To assess the reduction in hip fracture risk in individuals on vitamin D with or without calcium, we first performed 2 separate meta-analyses, 1 of RCTs of vitamin D supplementation (including RECORD), the other of RCTs of vitamin D with calcium supplementation (including RECORD and WHI), both against placebo/no treatment. We subsequently performed an adjusted indirect comparison of the pooled risk estimates obtained in these meta-analyses.

Data sources: MEDLINE and EMBASE (search terms: “vitamin D” AND “hip fracture”), bibliographies of articles retrieved, and the authors' reference files.

Study selection: Randomized controlled trials (RCTs) of oral vitamin D supplementation with or without calcium supplementation vs. placebo/no treatment in postmenopausal women and/or older men (≥50 years) specifically reporting hip fracture risk.

Data extraction: Independent extraction by 2 authors using predefined criteria, including study quality indicators.

Data synthesis: All pooled analyses are based on random-effects models. Based on 4 RCTs (9083 subjects), the pooled relative risk (RR) of hip fracture for vitamin D supplementation alone was 1.10 (95% confidence intervals [CI], 0.89 to 1.36). No between-trial heterogeneity was observed. For the 6 RCTs (45,509 subjects) of vitamin D supplementation with calcium supplementation, the pooled RR for hip fracture was 0.82 (95% CI, 0.71 to 0.94). There was no heterogeneity between trials. In an adjusted indirect comparison of the summary RRs from the 2 meta-analyses, the RR for hip fracture for vitamin D with calcium vs. vitamin D alone was 0.75 (95% CI, 0.58 to 0.96).

Conclusions: Oral vitamin D supplementation reduces the risk of hip fractures when calcium supplementation is added, even when taking into account recent negative trials like RECORD and WHI. Our findings suggest that to optimize clinical efficacy, vitamin D supplementation should be complemented with calcium supplements.

Disclosures: S. Boonen, None.


An RCT of Vitamin D or Placebo on Falls in Elderly Women with Low Vitamin D Status and a Falling History. K. Zhu1, I. Dick2, A. Devine3, D. Bruce*2, R. L. Prince2. 1Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Nedlands, Australia. 2School of Medicine and Pharmacology, University of Western Australia, Perth, Australia. 3School of Exercise, Biomedical and Health Science, Edith Cowan University, Perth, Australia.

The effect size of vitamin D on falls and mobility remains uncertain. Three hundred and two patients with a serum 25(OH)D concentration < 60 nmol/L and at least one fall in the previous 12 months were recruited and randomised to 1000 IU ergocalciferol (Vit. D) or identical placebo (Plac.), all received 500 mg per day of calcium as CaCO3. Falls data were collected every 6 weeks via telephone interview or clinic visit. Mobility was assessed by Timed Up and Go (TUAG) test. The study was approved by the Human Research Ethics Committee of Sir Charles Gairdner Hospital.

At baseline, the groups were matched in age (77.2 ± 4.6 years), weight (72.2 ± 13.6 kg), TUAG (11.3 ± 5.2 secs), and serum 25(OH)D concentrations (44.6 ± 13.0 nmol/L). The Plac. group was significantly taller than the Vit. D group (159.2 ± 6.1 vs 157.8 ± 6.3 cm, p = 0.05). 

Table  .  
  1. *Adjusted for baseline height

inline image

Vitamin D treatment significantly reduced the proportion of patients who sustained one fall over the 12 months but not the proportion of patients with multiple falls or any falls. Adjustment for baseline height resulted significant reduction in the odds of falling in the whole Vit. D treated group. TUAG at 12 months decreased in both groups compared to baseline, and the reduction was greater in the Vit. D group.

These data show that in patients who have Vitamin D deficiency or insufficiency and are at risk of a single fall in 12 months, vitamin D therapy can reduce the odds of falling by 50%. In all patients, despite the 5.4 ± 3.1% improvement in mobility in the Vit. D treated group compared to the Plac. Group, the 63% 12 month absolute risk of falling was significantly reduced in the Vit. D group only after adjustment for baseline differences in body stature.

Disclosures: K. Zhu, None.

This study received funding from: NHMRC, Australia.


Milk Basic Protein (MBP) Increases Bone Mineral Density and Improves Bone Metabolism in Healthy Young Women. A. Itabashi1, S. Aoe*2, Y. Toba*3, Y. Takada*4, H. Ishida5, K. Uenishi6. 1Central Clinical Laboratory, Saitama Medical School, Saitama, Japan. 2Department of Home Economics, Otsuma Women's University, Tokyo, Japan. 3Intellectual Property Department, Snow Brand Milk Products Co., Ltd, Tokyo, Japan. 4Product Planning Department, Snow Brand Milk Products Co., Ltd, Saitama, Japan. 5Laboratory of Administrative Dietetics, Kagawa Nutrition University, Saitama, Japan. 6Laboratory of Physiological Nutrition, Kagawa Nutrition University, Saitama, Japan

Milk has more beneficial effects on bone health than other food sources. Recent in vitro and in vivo studies have shown that milk whey protein, especially its basic protein fraction (milk basic protein, MBP), contains several components capable of promoting bone formation and inhibiting bone resorption. We also found that MBP supplementation was effective in preventing bone loss in healthy adult women and menopausal women, and that this improvement in BMD may be primarily mediated through inhibition of osteoclast-mediated bone resorption, while maintaining the balance of bone remodeling, by MBP supplementation. The objective of this study was to examine the effect of MBP on bone mineral density and bone metabolism in healthy young women. Thirty-five healthy young women were randomly assigned to treatment with either placebo or MBP (40 mg per day) for six months. The bone mineral density (BMD) of lumbar vertebrae L2–L4 of each subject was measured by dual-energy x-ray absorptiometry (DXA) after 0 and 6 months of treatment. Serum and urine indices of bone metabolism were measured at 0, 3, and 6 months. All subjects completed the study in accordance with the protocol. The mean rate of gain of lumbar BMD in the MBP group (1.57%) was significantly higher than in the placebo group (0.13%, P = 0.042). When compared with the placebo group, urinary cross-linked N-telopeptides of type-I collagen (NTx) were significantly decreased and serum osteocalcin was significantly increased in the MBP group at six months. These results suggest that MBP supplementation is effective in increasing BMD in young women, and that this increase in BMD may be primarily mediated through the promotion of bone formation and the inhibition of bone resorption. Our previous studies and the present study suggested that MBP supplementation was effective in increasing BMD in young, adult, and menopausal women.

Disclosures: Y. Toba, None.


Leptin Biphasic and Dose-Dependent Effects on Bone in Tail-Suspended Rats Are Related to Bone Growth, Body Weight and IGF-I Serum Levels. A. Martin*1, V. David*1, L. Malaval1, M. Lafage-Proust2, L. Vico1, T. Thomas2. 1Inserm 0366, University of St-Etienne, St-Etienne, France. 2Inserm 0366, University Hospital of St-Etienne, St-Etienne, France

Published data describing leptin effects on bone remain conflictual, with both anabolic and catabolic effects. Balance between these effects may be related to serum leptin levels. Therefore, we tested the hypothesis that exogenous leptin may modulate bone metabolism depending on the administered dose.

Seventy female Wistar rats were randomly tail-suspended (S) or not (NS) and treated either with intraperitoneal murine leptin (L) at 50 μg/kg*d−1 [L50] or 500 [L500] μg/kg*d−1 or with vehicle (V) during 14 days. A control group (B) was sacrificed at baseline. The L50 dose compensated suspension-induced decrease in serum leptin levels, while L500 increased them in a supraphysiological range.

After 14 days, L50 compensated suspension-induced bone loss at both trabecular and cortical level, as assessed by 3D microtomography. On the opposite, L500 completely inhibited femoral bone growth, and reduced bone mass by decreasing bone formation rate (BFR/BS) measured by histomorphometry (3.9 ± 2.2 and 1.1 ± 0.3 vs. 7.5 ± 1.6 μm3/μm2xd−1, for NS- and S-L500 vs. NS-V respectively, P <0.05) and by increasing bone resorption, assessed by serum CTX levels (16.4 ± 1.3 and 15.7 ± 2.2 vs. 10.4 ± 0.2 ng/mL, respectively, P <0.05). Medullar adipocytic volume was reduced in all leptin treated groups, suggesting that direct leptin effects on reciprocal differentiation between adipocytes and osteoblasts were not the mechanism leading to dose-dependent differences. Body weight and abdominal fat mass significantly decreased with L500, regardless of suspension. Concurrently, serum IGF-I levels significantly decreased in both L500 treated groups (0.76 ± 0.10 and 0.68 ± 0.05 μg/mL in NS and S groups respectively) compared to L50 (1.40 ± 0.20 and 1.08 ± 0.18 μg/mL respectively; P <0.05) and vehicle treated groups (1.48 ± 0.15 and 1.20 ± 0.13 μg/mL respectively; P <0.05).

In summary, our data first confirm that low dose of leptin was able to prevent disuse-induced bone loss. Second, they support the hypothesis that dual effects of leptin on bone may be related to serum leptin levels as bone loss associated with decreased bone formation and increased bone resorption was observed with high leptin level. We suggest that the known central and negative leptin effects on bone mediated by the sympathetic nervous system may be involved in the latter. However, our study provides evidence that other mechanisms may be involved, including decreased in bone growth associated with decrease in body weight and IGF serum levels.

Disclosures: A. Martin, None.


Central Regulation of Bone by Leptin in Rats May Be Related to Alteration of Energetic Status and IGF-I Pathway. A. Martin*1, V. David*1, M. Lafage-Proust2, C. Alexandre2, L. Vico1, T. Thomas2. 1Inserm 0366, University of St-Etienne, St-Etienne, France. 2Inserm 0366, University Hospital of St-Etienne, St-Etienne, France

Previous studies showed that peripheral administration of leptin exerted positive effects on bone metabolism, including prevention of disuse-induced bone loss in tail-suspended rats, whereas opposite effects were reported after intracerebroventricular (icv) leptin administration in mice. Therefore, we tested the hypothesis that dual effects of leptin may be related to route of administration and the subsequent cascade of biochemical or neuronal events.

Fifty female Wistar rats were randomly tail-suspended (S) or not (NS) and treated either with icv perfusion of murine leptin (L) [10 μg/kg*d−1] or vehicle (V) during 14 days. A control group (B) was sacrificed at baseline.

After 14 days, icv leptin treatment completely inhibited bone growth, and induced negative effects on bone mass and structure in addition to those related to suspension. Indeed, leptin treatment induced cortical thinning in femoral diaphysis measured by 3D-μQCT (0.52 ± 0.02 and 0.52 ± 0.03 vs. 0.55 ± 0.03 mm and 0.57 ± 0.03 mm for NS-L and S-L vs. NS-V and S-V respectively, P <0.05) and decrease in tibia metaphysis BMD compared to baseline, as measured by DXA (−7.9% and −11.8% for NS-L and S-L, respectively, P <0.05). Surprisingly, lumbar spine BMD increased in both leptin-treated groups (15.2% and 9.5% for NS-L and S-L vs baseline, respectively, P <0.05). The latter groups had reduced bone remodelling associated to a decrease in RANK-L serum level (0.03 ± 0.05 and 0.51 ± 0.48 vs. 16.38 ± 0.85 and I5.77 ± 0.71 pmol/L, in NS-L and S-L groups vs. NS-V and S-V groups respectively, P <0.05). These effects may be related to leptin-induced alteration of sympathetic tone, as norepinephrine levels were greatly higher in treated groups (154.0 ± 38.9 and 124.7 ± 60.1 vs. 58.4 ± 45.9 and 39.5 ± 11.5 pg/mL, for NS-L and S-L groups vs. NS-V and S-V groups respectively, P <0.05). However, other mechanisms may be strongly involved since, body weight and abdominal fat mass decreased by 25% and 40% respectively in treated S and NS animals. Furthermore, serum IGF-I levels at D14 were 5-fold lower in L groups compared to V and baseline groups (P <0.05).

In summary, our data demonstrate that elevated cerebro-spinal fluid leptin levels in rat induced deleterious effects on bone. In addition to stimulation of sympathetic pathway, these negative effects could be related to a dramatic impairment of energetic status and its hormonal consequences such as a profound decrease in IGF-I serum levels. Further studies are required to determine the respective part of each pathway downstream of the central leptin-mediated regulation of bone metabolism.

Disclosures: A. Martin, None.


Deletion of Steroid Receptor Coactivator (SRC)-2 Is Associated with a Marked Increase in Trabecular Bone and Decrease in Bone Marrow Adipocytes. U. I. L. Moedder1, A. Sanyal1, D. G. Fraser*1, M. Gehin*2, P. Chambon*2, B. W. O'Malley*3, T. C. Spelsberg1, S. Khosla1. 1Mayo Clinic, Rochester, MN, USA. 2Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France. 3Baylor College of Medicine, Houston, TX, USA

SRC-1 and −2 are closely related coactivators known to modulate the activity of a number of nuclear receptors. Previous studies have found that deletion of SRC-1 in mice is associated with impaired estrogen action and trabecular osteopenia. Thus, given the potential redundant functions of SRC-1 and −2, we hypothesized that deletion of SRC-2 may result in similar skeletal deficits. To test this, we analyzed the bone phenotype of 3 month old SRC-2 knock out (KO, n = 22) vs. wild type (WT, n = 24) mice. Surprisingly, as shown in the Table, SRC-2 KO mice displayed marked increases in trabecular volumetric BMD (vBMD) at the tibial metaphysis (+43%) and a 2-fold increase in bone volume/tissue volume (BV/TV) by histomorphometry (performed in a subset; n = 7 WT and 6 SRC-2 KO). By contrast, cortical vBMD was only modestly increased (+5%). 

Table  .  
  1. *P <0.05, ***P < 0.001 vs. WT

inline image

Further histological examination suggested a potential cellular mechanism for this increase in bone mass. Thus, we observed a 43% reduction in adipocyte numbers and a 31% reduction in adipocyte volume/tissue volume as well as a 2-fold increase in the fluorescent labeled perimeter in the SRC-2 KO as compared to WT mice. Since in addition to modulating the activity of sex steroid receptors, SRC-2 has also been shown to regulate PPARγ-mediated adipogenesis, these findings led to the alternate hypothesis that deletion of SRC-2 results in increased bone mass due to impaired PPARγ action and a “switching” of mesenchymal precursor cells away from the adipocytic and towards the osteoblastic lineage. Consistent with this, in vitro culture of bone marrow stromal cells in the presence of the PPARγ agonist, rosiglitazone, resulted in markedly reduced adipocyte formation by cells from SRC-2 as compared to WT mice. Collectively, these findings indicate a potentially novel role for SRC-2 in the regulation of adipocyte vs. osteoblast development. Moreover, given the known increase in bone marrow adipocytes and relative deficit in osteoblasts with aging, variations in SRC-2 expression or activity may contribute to differences in individual susceptibility to age-related bone loss and osteoporosis.

Disclosures: U.I.L. Moedder, None.


Evidence for Follicle-Stimulating Hormone(FSH)-Dependent and FSH-Independent Ovarian Contribitions to Bone Loss in FSH Receptor-Deficient(FORKO) Female Mice. J. Gao1, Y. Yang*2, R. Samadfam1, R. Tiwari-Pandey*2, D. Miao3, A. C. Karaplis1, M. R. Sairam*2, D. Goltzman1. 1Medicine, McGill University, Montreal, PQ, Canada. 2Molecular Reproduction Research Laboratory, Clinical Research Institute of Montreal, Montreal, PQ, Canada. 3Nanjing Medical University, Nanjing, China

We examined the FSH receptor knockout (FORKO) mouse to assess the determinants of bone loss in this genetically modified animal model. Female FORKO mice, which lack functional FSH activity develop impaired ovarian development and infertility. By 3 months of age, they exhibit an osteoporotic phenotype characterized by reduced osteoblastic bone formation and slightly increased osteoclastic bone resorption. This bone loss can be reversed with ovarian transplantation. We performed bilateral ovariectomy(OVX) on FORKO and WT female mice at 3 weeks of age and compared them to each other and to sham-operated FORKO and WT controls. We examined the phenotypes at 3 months of age, by densitometry, contact radiography and histomorphometric analysis. Bone mineral density (BMD) and trabecular bone volume(BV/TV) were reduced by 2.6% and 40% respectively in sham-operated FORKO compared to sham-operated WT mice, representing the contribution of FSH-dependent activity to the bone loss. FORKO+OVX displayed no significant difference in BMD compared to WT+OVX, suggesting that with the ovaries removed, the presence or absence of bioactive FSH produced little difference even though low levels of the FSH receptor have been found in mouse bone. Interestingly FORKO+OVX displayed a reduction of 5.7% in BMD and 43% in BV/TV compared to sham-operated FORKO mice indicating the presence of FSH-independent ovarian activity on bone. Osteoblast numbers and osteoblast surface were decreased (about 39%, p <0.01 and 16%, p <0.05 respectively) in FORKO+OVX compared to sham-operated FORKO mice, along with a slight down regulation of Cbfa 1 and IGF-I mRNA levels in bone. Tartrate resistant acid phosphatase (TRAP)-positive osteoclastic surface relative to bone surface was increased about 5% (p >0.05) in the femoral metaphysis of, FORKO +OVX relative to WT+OVX while the ratio of RANKL to OPG mRNA was slightly increased. Consequently loss of FSH-independent ovarian activity appeared to be due to reduced osteoblastic bone formation and increased osteoclastic bone resorption. These results indicate that two major components of bone loss may be identified in the FSH receptor deficient mouse: an FSH-dependent effect which is likely loss of ovarian estrogen, coupled with an FSH-independent ovarian effect which now remains to be characterized.

Disclosures: J. Gao, None.

This study received funding from: CIHR.


Immune Costimulation by CD80 Receptor Mediates T Cell Activation and Bone Loss in Ovariectomized Mice. F. Grassi, M. Robbie-Ryan, K. Page, X. Yang*, M. N. Weitzmann, R. Pacifici. Endocrinology, metabolism and lipids, Emory University, Atlanta, GA, USA

Ovx causes bone loss by increasing the production of TNF by activated bone marrow (BM) T cells. T cell activation is induced by a BM specific increased production of Reactive Oxygen Species (ROS) and the resulting stimulation of antigen (Ag) presentation by dendritic cells (DC), the most powerful Ag presenting cells. However, the mechanism linking ROS, antigen presentation, T cell activation and bone loss is still unclear. The magnitude of the T cell response to DCs is determined by the expression of costimulatory molecules. Among these, CD80 is essential in the initiation of T cell response. We found that ovx increases by 2.5 fold the number of activated DC expressing CD80 and that the resulting stimulation of DC activity promotes T cell activation, T cell TNF production and bone loss. Attesting to specificity, ovx did not alter the expression of other costimulatory molecules, including CD40 or CD86. In vivo treatment with the antioxidant N-Acetyl Cystein (100mg/Kg/day) prevented the increase in the number of CD80+ DCs, BM T cell activation, and the bone loss induced by ovx. Conversely, in vitro oxidation of the BM by H2O2 mimicked the effects of ovx as it increased the number of CD80+ DC, and that of activated T cells. Thus, ROS-mediated activation of CD80 appears to be critical for the T cell activation and the bone loss induced by ovx. To test this hypothesis sham and ovx mice were treated with CTLA4-Ig (Abatacept), a novel immunosuppressant which blocks costimulation by selectively binding CD80 thereby inducing T cells anergy. Mice were treated with CTLA4-Ig (100 μg 3 times a week) for 4 weeks. Analysis by μCT revealed that CTLA4-Ig completely prevented the loss of trabecular bone induced by ovx. In vivo DXA measurements confirmed that CTLA-4 Ig prevents ovx induced bone loss. Consistent with its bone sparing effect, CTLA4-Ig prevented ovx-induced increase in serum CTX, T cell TNF production, T cell activation and the formation of osteoclasts in BM cultures. Time course experiments revealed that 1 week CTLA-4 Ig treatment blocks for 2 weeks both T cell costimulation and the decrease in μCT indices of trabecular volume and structure induced by ovx. Attesting to specificity, discontinuation of CTLA-4 Ig treatment resulted in the resumption of trabecular bone loss which reached significance at week 4. In summary our findings, demonstrate that 1) oxidative stress causes bone loss in ovx mice by upregulating T cell activation through the CD80 costimulatory pathway, 2) CTLA-4 Ig prevents ovx-induced bone loss in WT mice by silencing T cell activation. Thus, inhibition of costimulatory molecules by CTLA4-Ig may represent a novel therapeutic strategy for osteoporosis.

Disclosures: F. Grassi, None.


Aging C57BL/6 Mice Exhibit Increased Glucorticoid Production in Association with Decreased Bone Formation, Wall Width and Canalicular Circulation: Novel Mechanistic Insights into the Involutional Loss of Bone Mass and Strength. R. S. Weinstein, D. Jia, T. M. Chambers*, E. A. Hogan*, S. B. Berryhill*, R. Shelton*, S. A. Stewart, R. L. Jilka, S. C. Manolagas. Center for Osteoporosis, Central Arkansas Veterans Healthcare System, Univ of Ark for Med Sci, Little Rock, AR, USA

Advancing age is associated with a decline in bone strength that is greater than the decline in bone mass. Because glucocorticoid excess also causes a mismatch between strength and BMD, we determined glucocorticoid production in C57BL/6 mice at 4, 8, 16, 25, and 31 months of age (n = 11–12 per group) as well as spinal BMD by DEXA, strength by vertebral compression, and histomorphometry of L1–4. We report that, as compared to 4-month-old mice, 25- and 31-month-old mice exhibited a 2-fold increase in serum corticosterone, adrenal weight and bone mRNA for 11β-HSD1, the enzyme that amplifies glucocorticoid action. Furthermore, with advancing age, C57BL/6 mice exhibited a progressive decrease in spinal BMD and strength. These changes were temporally associated with decreased cancellous area, trabecular width and number and wall width, as well as increased trabecular separation, decreased osteoblast and osteoclast numbers, increased osteoblast and osteocyte apoptosis, and decreased bone formation rate (BFR). The prevalence of osteoblast apoptosis was inversely related to BFR (r = −0.49, p <0.006). Moreover, consistent with the contention that low bone turnover leads to increased fragility, BFR was an independent predictor of vertebral strength (r = 0.75, p <0.0001). In addition, 11β-HSD1 mRNA levels were inversely related to strength (r = −0.67, p <0.05). Remarkably, advancing age was associated with a decrease in lacunocanalicular circulation, as determined by iv injection of procion red (an inert aqueous tracer). Specifically, procion red epifluorescence was abundantly detected in the osteocytic lacunae and canaliculi of young mice but decreased by more than 40% in the older animals (p <0.001). Similar histomorphometric and lacunocanalicular changes were noted in 8-month-old C57BL/6 mice given excess prednisolone when compared with placebo controls. The age-dependent decrease in BFR and wall width, along with the increase in osteoblast and osteocyte apoptosis - cardinal features of the adverse effects of glucocorticoid excess on bone - along with the increase in 11β-HSD1 mRNA, strongly suggest that local amplification of endogenous glucocorticoids negatively affects bone strength and contributes to the disparity between bone quantity and quality with aging. The defect in bone strength may be due in part to disruption of the lacunocanalicular system leading to osteocyte apoptosis and changes in bone material properties.

Disclosures: R.S. Weinstein, P & G 2; Merck 8; Novartis 2; Radius 1.


CD47−/− Mice Display Osteoclast Defects and Decreased Tumor Growth in Bone. Ö. Uluçkan*1, A. Hirbe*1, J. Prior*2, D. Piwnica-Worms*2, K. Weilbaecher1. 1Molecular Oncology, Washington University School of Medicine, St Louis, MO, USA. 2Mallinckrodt Institute of Technology, Department of Molecular Biology and Pharmacology, Washington University School of Medicine, St Louis, MO, USA

CD47 (Integrin-Associated Protein) is a five-transmembrane segment protein that is known to interact with the β3 integrins. Our finding that the β3 integrin deficient mice are protected from bone metastasis led us to evaluate tumor growth and metastasis in bone in CD47 deficient mice. When B16 luciferase-labeled murine osteolytic melanoma cells were directly delivered into the mouse tibia, there was decreased tumor growth in the CD47−/−mice compared to WT controls. As in the β3−/− mice, there was decreased tumor associated bone destruction in the CD47−/− mice compared with WT controls suggestive of a defect in osteoclast (OC) function. CD47 has been implicated in macrophage fusion, thus we evaluated OC formation in the CD47−/− mice. The number of multinucleated TRAP-positive OCs formed by differentiating bone marrow macrophages with M-CSF and RANKL was decreased in CD47−/− mice. Additionally, CD47−/− macrophages and pre-osteoclasts exhibited decreased RANKL induced proliferation as measured by BrdU incorporation. In contrast, CD47−/− mature OCs demonstrated decreased apoptosis at later stages of development. Phospho-p38 signaling was decreased in RANKL stimulated CD47−/− macrophages, whereas NFκB signaling was intact. Our current model is that disrupted CD47 mediated signaling in macrophages leads to reduced osteoclast differentiation possibly through a RANKL dependent pathway. The decreased tumor growth seen in this model suggests that host CD47 expression can modulate tumor growth in bone. We hypothesize that CD47 could be a novel molecular target for treating bone metastasis.

Disclosures: Ö. Uluçkan, None.


Inhibition of TGFβ Signaling Enhances Bone Mass, by Increasing Osteoblast and Reducing Osteoclast Number and Activity. E. G. Stebbins*1, K. S. Mohammad2, M. Niewolna*2, C. R. McKenna*2, H. Walton*2, X. H. Peng*2, G. Li*1, A. Murphy*1, S. Chakravarty*1, L. S. Higgins*1, D. H. Wong*1, T. A. Guise2. 1Scios, Inc., Fremont, CA, USA. 2Division of Endocrinology, Univ. of Virginia, Charlottesville, VA, USA

TGFβ is one of the most abundant growth factors in bone, and it mediates pathological bone loss associated with breast cancer metastases to bone. TGFβ blockade reduces the development and progression of osteolytic breast cancer bone metastases in mouse models by inhibiting tumor cell production of osteolytic factors such as PTHrP and IL-11. However, the role of TGFβ in normal bone remodeling is complex and ill-defined. Transgenic mice with increased TGFβ signaling have decreased bone mass, mineral content and mechanical properties, whereas transgenic mice with decreased TGFβ signaling have bones with increased mass and strength. We determined the effects of pharmacological blockade of TGFβ signaling with SD-208, a small molecule inhibitor of the TGFβ receptor I kinase, on bone remodeling in nude mice bearing MDA-MB-231 breast cancer cells implanted in the mammary fat pad. In this model, the mice do not develop bone metastases. The model thus provides an opportunity to study the effects of TGFβ blockade on bone unaffected by local tumor. Mice treated with SD-208 (20 or 60 mpk po qd) for 4 weeks had increased bone mineral density as measured by DXA, at the tibia, femur and total body sites. Bone histomorphometry revealed an increase in trabecular bone volume, increased osteoblast and reduced osteoclast numbers. To understand the mechanisms by which TGFβ blockade with SD-208 altered osteoblast and osteoclast activity, cell culture systems were utilized. In human mesenchymal stem cell cultures, TGFβ inhibited osteoblast differentiation and mineralized matrix deposition. The inhibition of osteoblast differentiation and matrix deposition by TGFβ was relieved in the presence of SD-208. Blockade of TGFβ signaling with SD-208 in a mouse bone marrow cell - osteoblast co-culture system inhibited osteoclast differentiation. In addition, SD-208 inhibited in vitro bone resorption, independent of the effect on osteoclast number. Together, these data show that pharmacological inhibition of TGFβ signaling in mice bearing mammary fat pad tumors, increases bone mass with a corresponding increase in the number of osteoblasts and decrease in osteoclasts. Thus, TGFβ blockade may have dual benefit for the bones of breast cancer patients; 1) to reduce the development and progression of bone metastases by reducing tumor production of osteolytic factors, including PTHrP and IL-11; and 2) to increase bone mass by stimulating osteoblast activity and reducing osteoclast activity in bone unaffected by local tumor metastases.

Disclosures: E.G. Stebbins, Scios. Inc. 3.

This study received funding from: Scios. Inc.


Inhibition of Gli2 Activity Decreases Tumor-Induced Osteolysis. J. A. Sterling, S. S. Padalecki, B. G. Grubbs, B. O. Oyajobi, G. R. Mundy. Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

Breast cancers metastasize to bone with catastrophic effects on the skeleton inducing a vicious cycle in which tumor produced factors stimulate bone resorption. One such factor is parathyroid hormone-related peptide (PTHrP) produced by the breast tumors. Inhibition of PTHrP transcription by small molecules has previously been demonstrated to reduce bone lesions and tumor burden (Gallwitz et al, 2002). While the regulation of PTHrP expression is clearly an important regulatory step in tumor-induced osteolysis, the mechanisms driving transcription of PTHrP in breast cancer cells are unknown. We have found that Hedgehog signaling is an important regulator of PTHrP expression. Specifically, we have demonstrated that Gli2 mRNA is expressed in human cancer cells that cause osteolytic lesions and express PTHrP, but not in non-osteolytic cell lines which do not secrete PTHrP. Transient tansfections of breast cancer cell lines with Gli2 increased PTHrP promoter activity, and stable expression of Gli2 increased PTHrP production approximately 10-fold compared to empty vector controls. In addition, these stable cell lines showed an increase in bone destruction when inoculated into nude mice via intracardiac injection. We hypothesized therefore that Gli2 is a major regulatory factor in breast cancer cells causing PTHrP expression, and that by reducing Gli2 expression in breast cancer cells we would also decrease PTHrP expression and osteolysis in vivo. In order to address this hypothesis, we stably transfected human breast cancer MDA-MB-231 cells, which express PTHrP and cause osteolysis in nude mice, with a construct that inhibits Gli2 transcriptional activity, Gli2-EnR. Two high expressing clones were selected for the expression of the vector, and real-time PCR for PTHrP was performed to determine the fold repression of PTHrP in these cells. Both cell lines showed a significant decrease in PTHrP mRNA expression of 72–93%. These cells were then inoculated into athymic nude mice via intracardiac inoculation. Results of Faxitron x-ray analysis indicated a greater than 50% reduction in lesion and lesion area in the cell lines stably expressing Gli2-EnR. In other experiments, we have shown that increasing Gli2 expression does not alter MDA-MB-231 proliferation or tumor burden in non-bone sites. These data indicate that inhibiting Gli2 activity can decrease PTHrP expression and subsequent tumor induced osteolysis. Furthermore, these data strongly suggest that inhibiting Hh signaling and Gli2 expression will be a promising therapeutic target for the inhibition of PTHrP and osteolytic bone destruction.

Disclosures: J.A. Sterling, None.


Histone deacetylase 1 (HDAC1) Regulates Dickkopf 1 (DKK1) Expression: Implications for Myeloma Bone Disease. C. Wideman*, A. Gupta*, J. Sterling, C. M. Edwards, M. Zhao, S. E. Harris, G. R. Mundy, I. R. Garrett, B. O. Oyajobi. Cell & Structural Biology, University of Texas Health Science Center, San Antonio, TX, USA

Dkk1, a secreted extracellular inhibitor of canonical Wnt signaling, is preferentially overexpressed in biopsies from myeloma patients. Dkk1 transgenic mice are osteopenic and previous work has implicated Dkk1 in the osteolysis and osteoblast dysfunction in myeloma bone disease. We found that proteasome inhibitors including bortezomib (currently used in the clinical management of myeloma patients), which stimulates bone formation, inhibits Dkk1 expression in bone cells. However, the mechanism by which inhibition of the proteasome modulates Dkk1 expression remains unclear. Because slective inhibitors of Class I HDACs, which are also in clinical trials for myeloma, have anti-tumor and anti-osteoclastogenic effects similar to proteasome inhibitors, we asked (i) whether HDAC1, a Class I HDAC, modulates Dkk1 levels; (ii) whether proteasomal inhibition regulates Dkk1, in part, by changes in HDAC1. Firstly, we examined the effect of HDAC1 on Dkk1 expression in a series of osteoblast-like and mesenchymal cell lines. Co-transfection of pre-osteoblast-like 2T3 cells, bone marrow stromal 14M1 cells, C3H10T1/2 and C2C12 cells with an HDAC1 expression construct increased activity of a full-length 2.4kb Dkk1-luciferase reporter construct. A catalytically-inactive dominant-negative HDAC1 mutant (H141A) had a similar effect to the wild type HDAC1 suggesting that this effect of HDAC1 on Dkk1 expression does not require its deacteylase enzymatic activity. HDAC1 increased Dkk1 mRNA levels in 14M1 and 2T3 cells as assessed by RT-PCR and Northern blot analysis. Next, to determine whether bortezomib modulates Dkk1 by altering HDAC1 levels, 14M1 cells and 2T3 cells pre-treated with bortezomib (1 and 10nM), were co-transfected with HDAC1 and Dkk1-luciferase reporter constructs or empty vector. Bortezomib inhibited Dkk1 promoter activity in a dose- and time-dependent manner in the presence and absence of HDAC1. Furthermore, immunoblotting revealed that endogenous HDAC1 protein level is significantly reduced in nuclear extracts from 14M1 cells treated with bortezomib. Taken together, our data strongly indicates that proteasome inhibition regulates Dkk1 expression, in part, through HDAC1. Thus, HDAC1 inhibition represents another therapeutic strategy in the clinical management of myeloma aimed at reducing Dkk-1 expression in the tumoral bone microenvironment with the potential to promote new bone formation while concomitantly reducing tumor burden and blocking osteolysis.

Disclosures: B.O. Oyajobi, None.

This study received funding from: NIH/NCI.


Proteasome Activity Correlates with Gli2 Stability and PTHrP Expression in Breast Cancer Cell Lines. J. A. Sterling1, M. Gaczynska*2, P. Osmulski*2, B. O. Oyajobi1, G. R. Mundy1. 1Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA. 2Molecular Medicine, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

The mechanisms responsible for increased PTHrP expression in metastatic breast cancer remain unknown. We have previously demonstrated that Hedgehog (Hh) signaling, and specifically the transcriptional activator Gli2, regulates PTHrP expression in human cancer cells, and therefore subsequent tumor-induced osteolysis in cancers that metastasize to bone. However, the reasons that Gli2 is overexpressed in these tumors is unknown. Since proteasomal processing of Gli2 is an important regulatory step in controlling intracellular Gli2 concentrations, we hypothesized that differences in proteasomal activity and/or subunit composition may influence PTHrP production by tumors, and be responsible for differences between lytic and non-lytic tumor cell lines. In order to test this hypothesis, we analyzed the proteasomal activity in the human breast cancer cells, which are associated with Gli2/PTHrP/osteolysis (MDA-MB-231) or not (MCF-7). We found that MCF-7 cells, which do not express Gli2 or PTHrP, displayed twice the proteasomal chymotrypsin-like activity in the microsomal fraction when compared MDA-MB-231, which do express Gli2 and PTHrP. (Independently, we have found that PTHrP expression in these cells is dependent on Gli2). To determine if this difference in proteasomal activities between the two cell types correlated with a change in Gli2 stability, we stably transfected MDA-MB-231 and MCF-7 cells with a Gli2-GFP fusion construct. We found that there was no expression of the Gli2 fluorescent protein in the MCF-7 cells under basal conditions, compared with MDA-MB-231 cells, which in comparison produced obvious Gli2 fluorescent protein. Both types of breast cancer cells were then treated with the proteasome inhibitor MG-262. MG-262 caused markedly enhanced expression of the fluorescent Gli2-GFP protein over time in MCF7 cells. However, the MDA-MB-231 cells were far more responsive to the proteasome inhibitors, as demonstrated by an increase in Gli2-GFP at lower dosages and earlier timepoints, and more cell death at the same concentrations. These data suggest that proteasomal activity varies quite markedly in different human breast cancers, and that the proteasome is an important regulator of both Gli2 and subsequent PTHrP expression in cancer cells that metastasize to bone. Endogenous differences in proteasome activity thus control the expression of Gli2 in these human breast cancer cell lines, which subsequently leads to an increase in PTHrP expression and subsequent osteolysis.

Disclosures: J.A. Sterling, None.


Bone-Derived IGF-1 Enhances Bone Pain through Up-Regulation of Expression and Activation of Acid-Sensing Nociceptors. T. Sakurai*1, H. Yonou*1, H. Wakabayashi*2, A. Hiraki*1, P. Williams*1, A. Farias*1, T. Yoneda1. 1Div Endocrinology, University of Texas, San Antonio, San Antonio, TX, USA. 2Biochem, Osaka University Graduate School of Dentistry, Osaka, Japan

Bone pain is a major complication of bone metastasis. The mechanism underlying bone pain is unknown and control of bone pain is unsatisfactory. Recent studies have shown that the acid-sensing nociceptors (ASNs) such as the transient receptor potential vanilloid 1 (TRPV1) and acid-sensing ion channels (ASICs) expressed in calcitonin gene-related protein (CGRP)-positive nociceptive fibers play a role in transducing peripheral pain signals to CNS. Of note, IGF-1 is recently found to regulate TRPV1 activation. Since IGF is released from bone due to increased bone resorption during the progression of bone metastases, we reasoned bone-derived IGF modulates bone pain. To approach this, we established a mouse model of cancer-induced bone pain by inoculating the 4T1 mouse breast cancer cells into tibiae. Tumor-inoculated tibiae showed significantly increased pain behaviors including hyperalgesia, mechanical allodynia and spontaneous flinching compared with the contra-lateral non-tumor-bearing tibiae, indicating 4T1 breast cancer cells induced bone pain. Immunohistochemical examination revealed the expression of TRPV1/CGRP-positive fibers and increased numbers of these fibers in the tumor-bearing tibiae. RT-PCR showed elevated mRNA expression of TRPVI, ASIC3 and CGRP in the ipsi-lateral dorsal root ganglions (DRGs). To study the role of bone-derived IGF-1 in vitro, we established organ cultures of mouse DRGs. The culture supernatants harvested from the resorbing neonatal mouse calvariae up-regulated TRPV1, ASIC3 and CGRP mRNA expression in organ cultures of mouse DRGs. AG1024, a specific inhibitor of IGF receptor tyrosine kinase, decreased these. IGF-1 (100ng/ml) up-regulated TRPVI, ASIC3 and CGRP mRNA expression in DRGs, while TGFβ, another abundant growth factor in bone, showed marginal effects. IGF-1 also elevated mRNA expression of c-fos, a marker of neural activation, in the F-11 rat DRG-like cells. These results collectively suggest bone-derived IGF-1 influences bone pain. To verify this, we tested AG1024 in vivo. AG1024 (2mg/kg, ip) significantly reduced hyperalgesia, mechanical allodynia and spontaneous flinching in tumor-bearing tibiae. In addition, mRNA expression of TRPVI, ASIC3 and CGRP and c-fos was decreased by AG1024 treatment in the ipsi-lateral DRGs and spinal dorsal horn, respectively. In conclusion, our results suggest that bone-derived IGF-1 enhances cancer-induced bone pain through stimulating ASN expression and activation and that inhibition of bone resorption may reduce bone pain by limiting IGF release from bone.

Disclosures: T. Sakurai, None.


Inhibition of the Osteoblastic Metalloproteinases Prevents the Bone Loss Induced by Oestrogen Deficiency. C. Schiltz*, C. Marty*, M. de Vernejoul, V. Geoffroy. Hǒpital Lariboisière, INSERM U606, Paris Cedex 10, France

In vitro and in vivo studies showed the implication of the osteoblastic matrix metalloproteinases (MMPs) in the process of bone remodelling and in particular in osteoclastic resorption. Our laboratory recently showed that the inhibition of osteoblastic MMPs in vivo led to a reduction of the trabecular bone formation rate in long bones which results in a positive balance and in a gain of bone volume and density in female mice. Oestrogens (E2) play a major role in bone remodelling. We investigated here if inhibition of osteoblastic MMPs would prevent the bone loss induced by E2 deficiency.

We ovariectomized (OVX) or sham-operated (SHAM) 4-month-old mice over expressing the TIMP-1 specifically in the osteoblasts (TG mice) and their wild type littermates (WT mice). One group of OVX mice of each genotype was treated with oestrogen (15μg/kg/day) and used as controls (OVX+E2). Bone density and histomorphometry analysis were carried out one month after ovariectomy at the vertebrae. Urinary desoxypyridinoline (DPD) was measured before and after ovariectomy. Primary osteoclasts (multinucleated (MNC) TRAP+ cells) were differentiated ex vivo from bone marrow in presence of RANK-L and MCSF.

In WT mice, OVX led to a significant reduction in the bone mineral density (BMD). The histomorphometric parameters showed in WT OVX mice a significant reduction of the trabecular bone volume (BV/TV) with a reduction in bone formation (decreased trabecular thickness (Tb.Th) and mineralizing surfaces (MS/BS)) and a significant increase of bone resorption (increased trabecular spacing (Tb.Sp) and osteoclastic surfaces (Oc.S/BS)). By contrast, in TG OVX mice, the bone loss measured by densitometry was abolished and BV/TV, Tb.Th and Tb.Sp were unchanged compared to TG SHAM mice. These results were confirmed by the significant increase in the urinary DPD after OVX in the WT mice only. Treatment with E2 corrected the bone loss in WT OVX mice. Our ex vivo study showed that the number of MNC TRAP+ cells was significantly elevated in bone marrow cultures from WT OVX mice compared to WT SHAM or TG OVX mice.

In conclusion, our results show that inhibition of osteoblastic MMPs blunt the bone loss induced by E2 deficiency through inhibition of bone resorption. Therefore, our study validates osteoblastic MMPs as potential targets to prevent post menopausal osteoporosis.

Disclosures: C. Schiltz, None.


Wdr5, a BMP-2 Induced Gene, Regulates Bone Volume Postnatally. F. Gori, E. D. Zhu*, M. B. Demay. Endocrine Unit, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

Wdr5 is a novel member of a family of structurally conserved proteins, the WD-40 repeat proteins, involved in numerous cellular functions including signal transduction, mRNA processing, gene regulation and chromatin remodeling. Wdr5 accelerates osteoblast and chondrocyte differentiation in vitro, and is developmentally expressed in osteoblasts as well as in proliferating and hypertrophic chondrocytes. We have recently demonstrated that transgenic mice overexpressing Wdr5 under the control of the 2.3-kb fragment of the mouse α(I) I collagen promoter (Col I-Wdr5 mice) display accelerated osteoblast differentiation during embryonic development and that this acceleration is associated with earlier activation of the canonical Wnt signaling pathway. Because osteoblasts play an important role in bone formation and regulate osteoclast differentiation, the effect of overexpression of Wdr5 on osteoblasts was examined postnatally. Histological analyses of tibiae isolated from 7 and 9 day-old mice demonstrated accelerated formation of the secondary ossification center confirming persistent acceleration of skeletal development in transgene positive mice postnatally. Histological analyses of sections of tibiae isolated from 70 day-old mice showed an increase in trabecular bone and thicker cortical bone in transgene positive mice compared to their wild type (wt) littermates. A significant increase (p <0.05) in bone mineral density (BMD), measured by dual energy x-ray absorptiometry (DEXA), was also observed in Col I-Wdr5 mice compared to their wt littermates. Because osteoprotegerin (OPG) levels are increased in primary calvarial osteoblasts isolated from transgene positive mice, the effects of Wdr5 on OPG expression were investigated in adult mice. OPG mRNA expression was increased by 2-fold in the long bones of Col 1-Wdr5 mice compared to their wt littermates. In addition, serum levels of OPG were also significantly increased (p <0.005) in transgene positive mice compared to their wt littermates, suggesting that the observed increase in BMD is a consequence, at least in part, of OPG-mediated inhibition of bone resorption.

These findings indicate that Wdr5 plays a regulatory role in skeletal homeostasis. Targeted expression of this protein to osteoblasts results in an increase in trabecular and cortical bone postnatally. The increase in OPG suggests that the phenotype observed in Col I-Wdr5 mice may be caused by modification of bone resorption leading to an increase in bone volume. Further investigations focused on identifying the role of Wdr5 in bone formation and/or bone resorption are underway.

Disclosures: F. Gori, None.


Ifitm5, a Novel Osteoblast Membrane Protein Stimulating Mineralisation. G. P. Thomas*1, P. Salois*2, K. Sellin*2, M. Bessette*2, E. Godin*2, P. Tambasco*3, A. Nanci*4, C. Lanctǒt*2, P. Moffatt*4. 1Centre for Immunology and Cancer Research, University of Queensland, Woolloongabba, Australia. 2Phenogene Therapeutics, Montreal, PQ, Canada. 3University of Sau Paulo, Ribeirao Preto, Brazil. 4Department de Stomatologic, Université de Montréal, Montreal, PQ, Canada

To identify novel secreted and membrane bound proteins in bone we attempted to define the “secretome” of bone utilising a signal trap screening system. Screening of 8 different bone cell and tissue libraries identified over 400 secreted and membrane proteins. One of the genes identified had been previously described in a screen of haemopoeitic stem cells and was putatively classified through in silico homology as a member of the fragilis family, and named fragilis 4 or interferon-induced transmembrane protein 5 (Ifitm5). No putative function had been assigned to this protein and no extensive localisation undertaken.

We identified Ifitm5 from a screen of the UMR106 osteoblastic cell line, suggesting a role in osteoblasts and so we further investigated the function of this protein in bone. The Ifitm5 gene encodes a 134 amino acid protein with a molecular mass of 14.8 kDa with two transmembrane domains. The protein is very well conserved across species, exceptionally so in the transmembrane domains. Immunostaining confirmed the membrane localisation of Iftm5 and indicated a topology with both N- and C- terminal extracellular domains.

Northern blot analysis showed bone-specific expression with no expression in any other embryonic or adult organs or tissues. In-situ hybridisation in e14.5 and e16.5 mouse embryos showed expression localised on the periosteum and in the developing trabecular bone. Immunohistochemistry in both newborn and adult bone showed high Ifitm5 expression in areas of high bone activity such as the developing primary spongiosa, adult vertebrae and mandibular bone.

Expression of Ifitm5 in MC3T3 and UMR106 osteoblast cell lines as well as in primary rat osteoblasts further confirmed the osteoblastic nature of Iftm5. Importantly expression was associated with the onset of mineralisation in these cells suggesting a role in bone formation which was further confirmed by Ifitm5 protein co-localisation with the mineralised nodules.

Functional evidence of a role for Ifitm5 in mineralisation was demonstrated by adenovirus Ifitm5 overexpression in both UMR106 and primary osteoblasts resulting in 40% and 70% increases in mineralisation respectively.

Thus, we have identified Ifitm5 as a novel osteoblast protein and demonstrated an effect on mineralisation providing further insights into the regulation of bone formation and novel potential therapeutic approaches.

Disclosures: G.P. Thomas, None.


Functional Study of a Unique Polyglutamine Stretch in RUNX2. P. M. Fonseca, G. Zhou*, Y. Chen*, B. Lee. Molecular Human Genetics, Baylor College of Medicine, Houston, TX, USA

RUNX2 is a non-redundant transcriptional activator involved in osteoblast differentiation and chondrocyte maturation. It contains a unique N-terminal polyglutamine-alanine (QA) domain consisting of a stretch of 23 residues of glutamine and 17 residues of alanine. Mutations in RUNX2 cause an inherited skeletal malformation syndrome, cleidocranial dysplasia (CCD). Until now, only one expansion in the alanine tract has been reported in an individual with CCD, while shortenings of the tract are benign polymorphic variants. Polyglutamine expansions are known to cause several neurodegenerative diseases, but the function of the QA tract in RUNX2 is still unclear. In this study, we generated a series of mutant forms of RUNX2 with a deleted (7QE2A-RUNX2) or expanded (72QE23A-RUNX2) QA domain. In vitro studies show that 7QE2A-RUNX2 results in reduced transactivation capacity of a reporter gene, while expansions only reduce transactivation when they exceed 70 repeats. Mice over expressing 7QE2A-RUNX2 under the osteoblast-specific action of promoter Collal present with short stature at birth. Skeletal preparations revealed delayed closure of fontanel while clavicles appeared normal. Radiographic analysis suggests lower bone mass. Transgenic mice also show absence of upper teeth and overgrown lower teeth with malocclusion suggesting this mutant was able to interfere with RUNX2 function during intramembranous ossification thereby phenocopying CCD in the skull. Interestingly, a fragment containing only the QA domain exhibited a stronger dominant-negative effect on full-length RUNX2 in a transactivation assay in ROS17 cells. To test this in vivo, we also generated transgenic mice over expressing only the QA domain in osteoblasts. Two independent transgenic lines exhibit dwarfism and lower bone density. Histomorphometry analysis reveals severe osteopenia at 6 weeks of age (average BV/TV (%) and Tb Wi (mcm) values for transgenic females are 6.5725 ± 0.7483 and 18.8730 ± 0.65586, respectively; 22.6034 ± 0.64386 and 25.5661 ± 0.97651 for control littermates, respectively; p values: 0.1 × 10−7 and 3 × 10−6). These mice show similar dental anomalies as the 7QE2A-RUNX2 transgenic mice as well as abnormal craniofacial features. These data suggest that over expression of normal length QA domain may have dominant negative effect on osteoblast function independent of the RUNT domain. The analysis of these transgenic mice lines will lead to a better understanding of the function of the QA tract in RUNX2.

Disclosures: P.M. Fonseca, None.


Bcl-2 Prevents Bone Loss Induced by Glucocorticoids. W. Zhang*, A. G. Pantschenko*, M. Nahounou*, M. McCarthy*, G. A. Gronowicz. Orthopaedic Surgery, Uconn Health Center, Farmington, CT, USA

One of the causes for glucocorticoid-induced decrease in bone formation is increased apoptosis of osteoblasts via decreased Bcl-2 expression, a cell survival factor. Whether the inhibition of osteoblast apoptosis would prevent osteopenia is not known. In this study, we evaluated the glucocorticoid-induced decrease in bone mass in mice in which apoptosis was blocked by overexpression of Bcl-2. A transgenic mouse, Col2.3Bcl-2, was developed with a 2.3 kb fragment of the type I collagen gene promoter driving the 1.8 kb region of human Bcl-2 (hBcl-2) cDNA. Four-month old wild-type (+/+) and Col2.3Bcl-2 (tg/+) mice were injected subcutaneously with slow release pellets of placebo and 2.1 mg/kg/day prednisone, and sacrificed 35 days later. Calcein and xylenol orange were injected 7 and 2 days, respectively, before sacrifice. Osteoblast apoptosis was determined by terminal deoxynuleotidyl transferase-mediated dUTP nick end labeling (TUNEL). Osteoblast/osteoclast parameters, bone formation, and bone mass were evaluated by static and dynamic histomorphometry. In +/+ animals, prednisone produced a 3.9-fold increase in TUNEL-positive apoptotic osteoblasts per bone surface (p <0.05). There was a significant decrease in % trabecular bone volume, % BV/TV (30.4%, p = 0.02), with a significant increase in trabecular spacing. TbSp (32.8%, p <0.05), and a significant decrease in trabecular number, TbN (21.3%, p <0.05), in the +/+. Prednisone seemed to increase mostly bone resorption in the +/+ animals, as indicated by 1.9-fold increase (p = 0.01) in osteoclast number/bone surface, but no significant increase in the bone formation rate (BFR/BS) or mineral apposition rate (MAR). A small but significant increase was found in the mineralizing surface per bone surface, MS/BS, of the +/+ mice. However, in the tg/+ mice, Bcl-2 overexpression prevented the prednisone-induced changes in all parameters, including the increase in osteoblast apoptosis, decrease in bone mass, and increase in osteoclast parameters. Thus, glucocorticoids induced bone loss by elevating apoptosis of osteoblasts as well as bone resorption, whereas bone-targeted Bcl-2 overexpression prevented osteoblast apoptosis and the increase in bone resorption. Bone mass was maintained in the glucocorticoid treated tg/+ mice. Our study is the first to demonstrate that the prevention of osteoblast apoptosis by Bcl-2 can inhibit glucocorticoid-induced bone loss.

Disclosures: W. Zhang, None.

This study received funding from: AR38933 from NIAMS.


Hypothalamic Neuropeptide Y Exerts a Negative Effect on Cortical Bone Formation. P. A. Baldock1, S. Allison*1, A. Sainsbury*2, R. F. Enriquez*1, E. M. Gardiner3, H. Herzog*2, J. A. Eisman1. 1Bone and Mineral Research Program, Garvan Institute of Medical Research, Sydney, NSW, Australia. 2Neuroscience Program, Garvan Institute of Medical Research, Sydney, NSW, Australia. 3Department of Medicine, University of Queensland, Brisbane, Qld, Australia.

Neuropeptide Y (NPY) is a neural signalling molecule implicated in two separate models of centrally-mediated regulation of bone mass. NPY is a key ligand for the Y2 receptor and loss of hypothalamic Y2 function results in an increase in cortical bone mass. Moreover, NPY is negatively regulated by leptin in the hypothalamus, being markedly elevated in brains of ob/ob mice, which display a reduction in cortical bone mass. The effect of hypothalamic NPY on cortical bone mass and formation was therefore investigated.

Bone mass and cortical osteoblast activity was investigated in NPY KO mice and following viral-mediated NPY over expression in the hypothalamus of wild type [vNPY wt] and Y2 KO [vNPY Y2KO] mice. These models were compared to non treated Y2 KO and ob/ob mice. Bone mineral content (BMC) was determined using dual X-ray absorptiometry of excised long bones and osteoblast activity (MAR) by histomorphometry on the distal femoral endocortical surface. 

Table  .  
  1. Values % Δ from control, * p <0.05 vs control. ∧ p = 0.07.

inline image

Loss of NPY signaling resulted in an increase in bone mass and cortical osteoblast function, with changes similar to Y2 KO mice. Consistent with its role in energy homeostasis, elevated central NPY increased fat mass in viral treated mice, as evident in ob/ob. Despite the gain in weight, elevated hypothalamic NPY reduced tibial BMC in wt and Y2KO, supporting the central nature of NPY signaling to bone and the negative action being independent of the Y2 receptor. These obese models had reduced bone mass similar to ob/ob despite the intrinsic difference in serum leptin. The acute increase in central NPY following viral injection produced marked inhibition of cortical bone formation.

Hypothalamic NPY expression has an important role in cortical bone formation. Reduced central NPY signaling, as evident in obesity, is associated with greater osteoblast activity and cortical bone mass.

Disclosures: P.A. Baldock, None.


Arrestin Complexes with Connexin (Cx) 43 and Anchors ERKs Outside the Nucleus: A Requirement for the Cx43/ERK-Mediated Anti-Apoptotic Effect of Bisphosphonates in Osteocytes. L. I. Plotkin, K. Vyas*, A. R. Gortazar, S. C. Manolagas, T. Bellido. Center for Osteoporosis and Metabolic Bone Diseases, University of Arkansas for Medical Sciences, and Central Arkansas Veterans Healthcare System, Little Rock, AR, USA

The anti-fracture efficacy of bisphosphonates (BPs) may be due in part to preservation of the osteocyte network resulting from inhibition of osteocyte apoptosis. This effect requires opening of connexin (Cx) 43 hemichannels and phosphorylation of the extracellular signal regulated kinases (ERKs). However, unlike ERK activation by other stimuli such as 17β-estradiol (E2), the Cx43/ERK pathway activated by BPs does not result in nuclear ERK accumulation. Thus, anti-apoptosis by E2 requires nuclear functions of ERKs whereas anti-apoptosis by BPs does not. Instead, the anti-apoptotic effect of BPs depends on the phosphorylation of cytoplasmic targets of ERKs and it is abolished by forced nuclear retention of the kinases. Therefore, ERK localization outside the nucleus is indispensable for osteocyte survival by BPs. We now report that ERKs and the scaffolding protein β arrestin co-immunoprecipitated with Cx43 in MLO-Y4 osteocytic cells and that the BP alendronate increased this association. Moreover, ERK2 fused to red fluorescent protein (ERK2-RFP) co-localized with Cx43 fused to green fluorescent protein (GFP) outside the nucleus in both untreated and alendronate-treated cells; whereas ERK2-RFP localized in the nucleus in E2-treated cells. In addition, as expected, alendronate did not induce ERK nuclear accumulation in cells transfected with wild type β arrestin (wtARR) or vector control. In contrast, alendronate did induce ERK nuclear accumulation in cells expressing a dominant negative β arrestin mutant (dnARR) that consists of the clathrin binding domain of β arrestin (amino acids 319–418) and competes with endogenous β arrestin for binding to clathrin. Importantly, alendronate activated ERKs in dnARR-transfected cells as effectively as in cells transfected with wtARR, demonstrating that the dnARR mutant only interferes with subcellular localization but not with activation of ERKs by BPs. Remarkably, whereas alendronate inhibited apoptosis induced by etoposide in cells expressing wtARR or vector control, it was ineffective in cells expressing the dnARR mutant. On the other hand, ERK nuclear localization or anti-apoptosis induced by E2 were not affected by either dnARR or wtARR indicating the specific cross-talk between β arrestin and the Cx43/ERK pathway. We conclude that BPs induce the formation of a complex comprising Cx43, β arrestin and clathrin, which directs ERKs outside the nucleus and is indispensable for osteocyte survival induced by these agents.

Disclosures: L.I. Plotkin, None.


Notch Signaling Represses Osteoblast Activity but Promotes Osteoclast Function In Vivo. M. J. Hilton1, S. Bai*2, R. Kopan*3, F. P. Ross2, S. L. Teitelbaum2, F. Long1. 1Internal Medicine - Division of Bone and Mineral Diseases, Washington University School of Medicine - St. Louis, St. Louis, MO, USA. 2Department of Pathology, Washington University School of Medicine - St. Louis, St. Louis, MO, USA. 3Internal Medicine - Division of Molecular Biology and Pharmacology, Washington University School of Medicine - St. Louis, St. Louis, MO, USA

The role of Notch signaling in bone biology in vivo is not known. To directly address this issue, we have manipulated Notch signaling selectively in the skeletogenic tissues in the mouse, by using the Cre-loxP technique. Specifically, we utilized the PrxICre transgenic mice that express Cre predominantly in the calvarial and limb primordia, both containing precursors of osteoblasts. In the first set of experiments, we eliminated all Notch signaling by removing Presenilin 1 and 2, the catalytic subunits of the γ-secretase complex indispensable for signaling from all Notch receptors. The mutant animals (PrxICre: PSIc/c; PS2n/n, where “c” stands for the floxed allele and “n” stands for the conventional null allele) exhibited a profound phenotype in their long bones. As early as E18.5, the mutant animals showed a marked increase in bone volume. The increase in bone mass was neither due to a shortage of osteoclasts, nor due to abnormal osteoblast differentiation. Rather, our preliminary results suggest that an increase in osteoblast activity, and a compromise in osteoclast activity may be the underlying causes. In keeping with a defect in osteoclast function, the hypertrophic cartilage was markedly expanded in the mutant. Importantly, other aspects of the growth plate cartilage were remarkably normal, indicating that deficiency of Notch signaling in chondrocytes was unlikely responsible for the bone phenotype. The phenotypes in the mutant animals persisted throughout their postnatal life. At 8-weeks of age, the excessive bone and expanded hypertrophic cartilage were readily detectable by contact autoradiography, histomorphometry, and micro CT analysis. In the second set of experiments, we specifically removed Notch1 and Notch2 activities from the skeletogenic tissues. Remarkably, the Notch1 and Notch2 mutant (PrxICre; Notchln/c; Notch2c/c) mice also showed an excess in bone mass as well as an expansion in hypertrophic cartilage, at both E18.5 and 8-weeks of age, strikingly similar to the PS1 and PS2 mutant animals. These results indicate that disruption of Notch signaling was primarily responsible for the skeletal phenotypes in both animal models. Taken together, our data support a model that Notch signaling in osteoblasts maintains bone homeostasis by suppressing osteoblast activity while promoting osteoclast function.

Disclosures: M.J. Hilton, None.


Cloning, Expression and Function of NHE-oc: A Novel Sodium/Proton Exchanger Expressed in Osteoclast Mitochondria. R. A. Battaglino1, L. Pham*1, L. R. Morse*2, M. Joe*1, A. Sharma*1, P. Stashenko*1. 1Cytokine biology, Forsyth Institute, Boston, MA, USA. 2Rehabilitation Medicine, Harvard Medical School, Boston, MA, USA

We identified a gene, nhe-oc, that codes for a novel sodium/proton exchanger expressed in osteoclasts. The goals of this study are to characterize nhe-oc expression in vitro and in vivo, to elucidate the activity of the NHE-oc protein and to evaluate the effect of nhe-oc blockade on osteoclast differentiation and function.

NHE-oc has strong homology with other putative NHEs (human, dog and rat), which have been identified as orthologues by reciprocal BLAST analysis, nhe-oc contains 13 exons that span ∼40 Kb on chromosome 3, giving rise to a 2.0 Kb transcript that codes for a 547 aa protein. Despite of the fact that NHE-oc has no significant sequence similarity to other NHEs, their hydropathic profiles are very comparable. Northern Blot and RT-PCR analysis showed no nhe-oc mRNA expression in tissues other than bone. In situ analysis showed nhe-oc mRNA expression in 17.5 dpc femur sections, adjacent to the growth plate, a localization consistent with osteoclast expression. An NHE-oc expression vector was transfected into both PS120 fibroblasts lacking all endogenous plasma membrane NHEs, and RAW264.7 monocytes. Confocal microscopy and cell fractionation analysis revealed mitochondrial localization of NHE-oc. Mitochondria isolated from transfected and untransfected PS120 cells were loaded with a pH-sensitive dye (BCECF). pH was monitored after abrupt removal of Na+. A sudden increase in mitochondrial H+ followed Na+ removal in transfected cells, indicating Na+/H+ exchange. This activity was sensitive to NHE inhibitors. Passive mitochondrial swelling, induced by the addition of Na+ acetate, can be monitored by a decrease in absorbance at 546 nm. Mitochondria from transfected cells underwent Na+ acetate-induced swelling that also was inhibitor-sensitive. Finally, RAW 264.7 cells were transfected with a cocktail of NHE-oc specific or control small inhibiting (si) RNAs and stimulated with RANKL. Si treatment reduced osteoclast differentiation and resorption on osteologic disks by ∼60%. Accordingly, nhe-oc, mmp-9 and c-src mRNA expression was reduced by ∼90% suggesting a role for NHE-oc in terminal osteoclast differentiation.

We conclude that NHE-oc is a novel osteoclast-specific mitochondrial Sodium Proton Exchanger. NHE-oc mediates Na+-dependent changes in mitochondrial pH and matrix swelling in isolated mitochondria. NHE-oc activity is required for normal osteoclast differentiation and function. As a bone specific novel protein, NHE-oc has the potential to become a powerful anti-resorptive therapeutic target that may help prevent pathological bone loss.

Disclosures: R.A. Battaglino, None.

This study received funding from: NIH/NICDR grant DE007378-18.


TGF-β Activates a Novel Smad-Independent Pathway to Promote Osteoclast Survival. A. Gingery1, L. Pederson*2, M. Ruan*2, N. J. Horwood3, M. J. Oursler2. 1Biochemistry and Molecular Biology, University of Minnesota, Duluth, MN, USA. 2Endocrinology, Mayo Clinic, Rochester, MN, USA. 3Kennedy Institute of Rheumatology Division, Faculty of Medicine, Imperial College, London, United Kingdom

Defining the intracellular pathways through which growth factors promote osteoclast survival could reveal strategies to block pathological bone disease. We have therefore investigated TGF-β influences on osteoclast survival using in vitro generated murine osteoclasts. After reaching maturity, osteoclasts apoptose and addition of TGF-β repressed this apoptosis, suggesting that TGF-β promotes osteoclast survival. TGF-β increases neuronal cell survival through stimulation of the Activin receptor, ALK1, not the canonical ALK5 receptor. We have therefore investigated the role of ALK5 in TGF-β-mediated osteoclast survival using SB-431542, which blocks ALK5 but not ALK1 activation. SB-431542 blocked TGF-β-mediated osteoclast survival. Real-time PCR was used to explore TGF-β effects on ALK1 and ALK5 gene expression in osteoclasts and documented a 4-fold increase of ALK5 message in TGF-β treated osteoclasts, but no change in the ALK1 message, strengthening that TGF-β-mediated osteoclast survival is through ALK5 activation, not ALK1. Since osteoclast survival involves MEK, AKT, and NFκB activation, we examined TGF-β effects on their activation and observed rapid phosphorylation of MEK, AKT, and NFκB. The timing of activation coincided with SMAD activation, indicating that activation is independent of SMAD signaling. An adenoviral NFκB luciferase reporter expression system was used to confirm TGF-β-mediated activation of NFκB and two-fold increase in reporter activity was observed after a 30 minute TGF-β treatment. We have used chemical and adenoviral-mediated inhibition of MEK and AKT to explore the pathways by which TGF-β activates NFκB and found that MEK inhibition abrogated TGF-β-mediated AKT activation. Moreover, adenoviral inhibition of MEK or AKT abrogated TGF-β-mediated activation of NFκB. These data support that TGF-β-mediated NFκB activation is by sequential activation of MEK and AKT. To explore the roles of MEK, AKT, and NFκB in TGF-β-mediated osteoclast survival, we used chemical and adenoviral inhibition approaches and found that inhibiting these signaling components repressed TGF-β-mediated osteoclast survival.

Our major findings are that (1) TGF-β promotes osteoclast survival, (2) the unique signaling pathway by which TGF-β promotes osteoclast survival is by ALK5-mediated activation of MEK, which activates AKT to increase NFκB activation, and (3) NFκB activation occurs independent of Smad signaling.

Disclosures: A. Gingery, None.


Classical NF-κB Subunit RelA Regulates Osteoclast Precursor Survival but not Differentiation. S. Vaira, M. Alhawagri*, D. V. Novack. Division of Bone and Mineral Disease, Washington University in St. Louis, St. louis, MO, USA

RANKL activates both the classical and the alternative NF-κB pathways. The classical pathway activates primarly, RelA/p50 dimers while alternative pathway is responsible for activation of RelB/p52 dimers. Blockade of both pathways abolishes osteoclast (OC) differentiation, but the specific contribution of each of these two pathways is not clear. Mice lacking RelA exhibit TNF-induced liver apoptosis and embryonic lethality, rescued by deletion of TNFR1. To address the role of RelA in the osteoclast, we transplanted either RelA+/+ or RelA−/− marrow (both on the tnfr1−/− background) into wild-type recipients. Four months after engraftment, sections of tibiae from RelA−/− recipients showed higher bone mass than RelA+/+ recipients (BVATV%: 17.1 + 6.5 vs. 5.3 ± 3.0, n = 4–6). Additionally, after 5 days of RANKL injection, RelA−/− recipients had lower serum TRAP5b levels and fewer OCs on calvariae sections, compared to RelA+/+ controls. Thus RelA is important for OC recruitment in vivo.

To better analyze the mechanism by which RelA impacts OC differentiation and function, we turned to in vitro culture. In the presence of MCSF and RANKL, precursors from RelA-deficient mice formed few OCs. Additionally, we noted fewer cells present in RelA−/− cultures treated with RANKL, but not M-CSF, as confirmed by MTT assay. Since RelA is an important cell survival factor, we examined markers of apoptosis. We found increased DNA fragmentation and caspase3 activation in both RelA+/+ and RelA−/− cultures treated with RANKL for 36–48 h, but the levels were 2.5–3-fold higher in the absence of RelA. In order to determine if the failure to generate OCs is secondary to apoptosis, we used the caspase inhibitor ZVAD, and found that differentiation was completely rescued. Furthermore, this anti-apoptotic function is specific for RelA because retroviral expression of RelA, but not RelB, in RelA−/− OCs precursors rescued the apoptosis and thus differentiation.

In embryonic fibroblasts lacking RelA, TNF induces apoptosis due to downregulation of Gadd45β, an inhibitor of MKK7, with subsequent prolonged JNK activation and initiation of the caspase cascade. In our RelA−/− OC precursors, RANKL induced enhanced JNK activation but failed to upregulate Gadd45β. RANKL-induced apoptosis was blocked by the JNK inhibitor SB600125, and retroviral expression of Gadd45β rescued RelA−/− OC formation. We conclude that 1) RANKL initiates an apoptotic signal counteracted by RelA dependent gene expression, 2) RelA is not required for OC differentiation and 3) RelB cannot compensate for the absence of RelA.

Disclosures: S. Vaira, None.


Microphthalmia-Associated Transcription Factor (MITF): Integration of Signaling to Transcriptome during Osteoclast Differentiation. S. M. Sharma*1, R. Hu*1, A. Bronisz*1, K. Mansky*2, M. C. Ostrowski1. 1MVIMG, The Ohio State University, Columbus, OH, USA. 2University of Minnesota, Minneapolis, MN, USA

The regulation of osteoclast (OCL) differentiation in the bone microenvironment is critical for normal bone remodeling, as well as for various human bone diseases. Two major signaling cascades initiated by CSF1/CSF1R and RANK-L/RANK modulate the fine balance of various transcription factors during OCL differentiation. Previously it has been shown that both these signaling events control the function of MITF, a basic-helix-loop-helix zipper factor, that regulates distinct target gene expression in several cell types including osteoclasts.

In an earlier report from our laboratory we reported that PU.1, an Ets family member interacts with MITF and this interaction induced activation of MITF target genes in OCLs. Here we demonstrate genetic interaction between MITF and PU.1 using hypomorphic (Mitf-vga9) and a semi dominant (MITF-mi) allele of MITF in a Pu.1 null allele background. Mitf-vga mice show very low incidences of osteopetrosis. Reduction of PU.1 levels results in higher penetrance of osteopetrosis. This observation was confirmed by histomorphometric analysis and subsequent analysis of MITF target gene expression.

We utilized a combination of techniques to study the effect of MITF and PU.1 interactions with respect to CSF1 and RANK-L signaling in primary bone derived OCL/macrophage precursors. Surprisingly, removal of CSF-1 from primary cells exported MITF from nucleus to cytoplasm, while Pu.1 remained localized in the nucleus. Using Biochemical techniques to study the subcellular localization of MITF in response to cytokine treatment, we were able to demonstrate that CSF1 alone could partially drive MITF to the nucleus; however the combination of CSF-1 and RANKL was much more efficient.

Chromatin immunoprecipitation assays demonstrated that while CSF1 alone could recruit both MITF and PU.1 to Cathepsin K (Ctsk) promoter, this was insufficient to elicit expression of the gene. However, CSF-1/RANKL treatment resulted in robust recruitment of both MITF and PU.1 to the promoter and transcription of Ctsk gene. Examination for the presence of chromatin remodeling proteins and other transcriptional activators at cathepsin K promoter indicated a sequential recruitment of MITF and PU.1 followed by the mammalian Swi/Snf complex, and finally recruitment of NFATc1 and CBP/p300 before high-level transcription of the gene. These results provide a mechanism to explain why osteoclast precursors express both MITF and PU.1 without activating the osteoclast-specific gene program, and also redefine the key role of MITF as a regulator of terminal differentiation in osteoclasts.

Disclosures: S.M. Sharma, None.

This study received funding from: NIAMS Grant # AR-44719.


The Small G-Protein Rap1 Inhibits the Differentiation and Promotes the Survival of Osteoclasts. A. Ueda, T. Hiraga, T. Yoneda. Biochem, Osaka Univ Grad Sch of Dent, Osaka, Japan.

Evidence is accumulating that small G-proteins play crucial roles in osteoclast differentiation, function and survival. The small G-protein Ras and Rho are shown to promote osteoclast survival and actin ring formation, respectively. Rap1 is a family member of Ras with high homology to Ras. Recent studies have reported the amino-containing bisphosphonates (BPs) inhibit Rap1 activation, suggesting an involvement of Rap1 in bone resorption. However, the role of Rap1 in osteoclasts still remains unclear. To determine this, we studied bones of mice deficient in SPA-1, a GTPase-activating protein that specifically inactivates Rap1. Histomorphometrical examination demonstrated significant increase in trabecular bone volume and decrease in TRAP+ osteoclasts in SPA-1-deficient (SPA-1−/−) mice compared with wild-type (WT) mice. Ral-GDS pull-down assay showed that Rap1 was markedly activated in spleen cells and osteoclasts isolated from SPA-1−/− mice compared with those from WT mice. TRAP-positive multinucleated osteoclast-like cell (TRAP+ MNOC) formation in the presence of soluble RANKL and M-CSF was significantly less in the spleen cell cultures of SPA-1−/− mice than WT mice. To further examine the role of Rap1 on osteoclastogenesis, we introduced a constitutive-active (Rap1V12) or a dominant-negative (Rap1N17) mutant of Rap1 into the osteoclast precursor cell line RAW264.7 using adenovirus vector expression systems. Overexpression of the constitutive-active Rap1V12 suppressed RANKL-induced TRAP+ MNOC formation, while the dominant-negative Rap1N17 enhanced it. To determine the role of Rap1 in osteoclast function, we next examined the effects of Rap1 on actin ring formation, a critical prerequisite to osteoclastic bone resorption. Immunocytochemical examination revealed no difference in actin ring formation in mature osteoclasts harvested from WT and SPA-1−/− mice. Overexpression of Rap1V12 and Rap1N17 had little effects on actin ring formation in TRAP+ MNOC. Finally, we examined the effects of Rap1 on osteoclast survival. The constitutive-active Rap1V12 significantly increased the survival of TRAP+MNOC with decreased caspase 3 expression, while the dominant-negative Rap1N17 decreased survival with increased caspase 3 expression. Of note, alendronate inhibited Rap1 activation. In conclusion, our results suggest that Rap1 activation suppresses the differentiation of osteoclasts, whereas Rap1 promotes the survival of osteoclasts. They also suggest that one of the mechanisms by which BP inhibits bone resorption may be suppression of osteoclast survival through an inhibition of Rap1 activation. Regulation of Rap1 activity may be critical to the control of osteoclastic bone resorption.

Disclosures: A. Ueda, None.


The Inositol Polyphosphate 4-Phosphatase Type II as a Negative Regulator of Osteoclastogenesis. J. Vacher, M. Ferron*, M. Pata*. Clinical Research Institute of Montreal, Montreal, PQ, Canada

Osteoclasts are unique hematopoietic cells that derive from the monocyte/macrophage lineage and are responsible for bone resorption. Loss of osteoclast activity leads to malignant osteopetrosis, a genetic disease characterized by increase in bone mass and severe reduction in bone marrow compartment. We are studying the grey-lethal (gl) mouse mutant that displays a severe recessive osteopetrotic phenotype closely related to human infantile malignant osteopetrosis. In this model, osteopetrosis results from osteoclast maturation/function defect with reduced bone resorption. To identify putative target genes of gl and osteopetrosis-associated genes a differential-display study was undertaken. One down-regulated transcript in gl/gl mice corresponds to the murine homologuc of Inositol Polyphosphate 4-phosphatase type II (Inpp4b). We cloned the mouse Inpp4b cDNA that encodes a 105kDa protein including a C2 domain and a consensus C(X)5R catalytic site. Inpp4b expression is normally detected during late osteoclast differentiation and in mature cells. Interestingly, Inpp4b colocalizes with F-actin in osteoclast podosomes. To understand Inpp4b role in osteoclast, ex vivo overexpression of Inpp4b in RAW 264.7 was generated and upon RANKL stimulation, caused a decrease in osteoclast-like cells (OCL) differentiation. This effect is dependent of Inpp4b phosphatase activity as overexpression of a phosphatase dead Inpp4b mutant potentiated differentiation. Importantly, we showed that this differentiation was mainly regulated by NFATcl activation and nuclear translocation. Moreover, ectopic phosphatase dead Inpp4b expression in RAW cells, as opposed to native Inpp4b, interfered with normal cytoskeletal organization and resulted in formation of multiple actin rings and poor bone resorption. To investigate the role of Inpp4b in vivo, TRAP-Inpp4b transgenic mice were produced with no adverse phenotype. In contrast, TRAP-Inpp4b-gl/gl transgenic mice displayed a significant increase in bone density with a reduced osteoclast population, leading to a more severe osteopetrotic phenotype than gl/gl mice. Altogether our ex vivo and in vivo results demonstrated that the phosphatase activity of Inpp4b is critical for osteoclast cytoskeleton organization and that Inpp4b protein is a major regulator of osteoclast differentiation, most likely through the NFATcl pathway.

Disclosures: J. Vacher, None.

This study received funding from: IRSC.


Pyk2−/− Mice Are Osteopetrotic Due to Defective Regulation of Microtubules and Actin in Osteoclasts. H. Gil-Henn*1, O. Destaing*2, A. Bruzzaniti3, L. Neff*3, N. Sims*4, K. Aoki*5, N. Alles*5, A. Sanjay.*6, R. Baron3, J. Schlessinger*1. 1Pharmacology, Yale University, New Haven, CT, USA. 2Cell Biology, Yale University, New Haven, CT, USA. 3Orthopaedics, Yale University, New Haven, CT, USA. 4Medicine, University of Melbourne, Victoria, Australia. 5Pharmacology, Tokyo University, Tokyo, Japan. 6Anatomy and Cell Biology, Temple University, Philadelphia, PA, USA

The non-receptor protein tyrosine kinase Pyk2 is highly expressed in osteoclasts, where it localizes to podosomes and sealing zones, and contributes to the recruitment and activation of Src kinase downstream to integrins. In order to explore the physiological role of Pyk2 in bone in vivo, we generated Pyk2-deficient mice. Skeletal examination of these mice revealed an osteopetrotic phenotype which is caused by a cell autonomous defect in mature osteoclasts. In vitro, pyk2−/− osteoclasts formed only small actin rings, failed to form a proper ruffled border and to efficiently resorb dentin, generating only shallow pits. Mutational analysis of Pyk2 functional domains by adenovirus-mediated infection into pyk2−/− osteoclasts revealed that the phenotype could be fully rescued by expression of wild-type Pyk2, whereas podosome belt formation and bone resorption were only partially rescued by a kinase-negative mutant (Pyk2-K457A), by a mutant that is unable to recruit Src (Pyk2-Y402F), or by a deletion mutant lacking the FERM domain. In contrast, a FAT-domain deletion mutant, in which the region that allows localization to adhesion sites is absent, failed to restore podosome belt and sealing zone formation, and showed substantial reduction of bone resorption.

Given that the defect in pyk2−/− osteoclasts appeared to affect primarily the transition between podosome clusters/rings and belts, we then determined whether stabilization and acetylation of microtubules, a pathway which involves Rho-mDIA2-HDAC6, was affected. Examination of microtubules properties in pyk2−/− osteoclasts revealed a markedly decreased microtubules stability compared to wild-type cells, with pyk2−/− osteoclasts showing almost complete absence of nocodazole-resistant microtubules and a reduction of acetylated microtubules. Furthermore, videomicroscopy and FRAP analysis of GFP-actin in osteoclasts revealed alterations in actin dynamics in pyk2−/− cells.

These results suggest that Pyk2 plays a critical role in the stabilization of microtubules and the regulation of actin dynamics in podosomes. We conclude that pyk2−/− mice are osteopetrotic as a result of impaired microtubules-dependent stabilization of actin structures in osteoclasts, leading to defective bone resorption.

Disclosures: H. Gil-Henn, None.


Loss of Anion Exchanger 2 (Ae2) in Mice Results in Osteopetrosis. I. Jansen*1, T. De Vries1, J. Ravesloot*2, V. Everts1, R. Oude Elferink*3. 1Periodontology and Oral Cell Biology, ACTA, Amsterdam, The Netherlands. 2Physiology, AMC, Amsterdam, The Netherlands. 3Liver Center, AMC, Amsterdam, The Netherlands

Osteoclasts are multinucleated cells specialized in bone resorption. Central to this is extracellular acidification by proton secretion. When protons are secreted the internal pH of the cell changes. An anion exchanger is thought to maintain the intracellular pH (pHi) by removing bicarbonate and taking up chloride ions.

To evaluate the role of anion exchanger 2 (Ae2) in osteoclastic bone degradation, Ae2−/−mice were generated by deleting three out of the five Ae2 transcripts. The a, b1 and b2 transcripts were deleted, whereas the c1 and c2 isoforms, until now exclusively found in the stomach remained intact. The calvariae and long bones were examined by X-ray, μCT, and light and electron microscopy. In vitro generated osteoclasts from wildtype and Ae2−/− mice were tested for bone resorption, TRACP activity and intracellular pH regulation.

X-ray and μCT showed a very high bone density in the long bones of the Ae2−/− mice, resulting in an osteopetrotic phenotype. The calvariae, however, were normal. Multinucleated osteoclasts were present in Ae2−/− mice but those in the long bones lacked a ruffled border and were inactive. In contrast, normal functional osteoclasts were present in calvariae. Osteoclasts generated in vitro from bone marrow contained lower numbers of multinucleated TRACP+ cells; the majority of the TRACP+ cells were mononuclear and bone resorption did not occur. pHi measurements demonstrated that the Ae2−/− cells were impaired in chloride/bicarbonate exchange.

Thus, in the absence of Ae2 a, b1 and b2 isoforms, osteoclasts in long bone lack the capacity to resorb, which results in a severe osteopetrotic phenotype. The pHi is not properly regulated, and in vitro the osteoclasts are not readily formed. However, osteoclasts in calvariae resorb normally, an extensive ruffled border is present. We hypothesize that calvarial osteoclasts either make use of another anion exchanger or of the c1 and/or c2 Ae2 isoforms. Since in vitro generated calvarial osteoclasts show a similar phenotype as long bone osteoclasts, the data suggest that calvarial osteoclasts primarily use the a, b1 or b2 isoform but in vivo they appear to have the possibility to compensate for this lack.

Our findings indicate that Ae2 is essential to the function of osteoclasts. Calvarial osteoclasts differ from those of long bones in their capacity to compensate for the absence of Ae2.

Disclosures: I. Jansen, None.


Different BMD Values in the US, Sweden and Hong Kong - A Study in 11,000 Men - Mr OS International. O. Johnell1, E. Barrett-Connor2, P. Cawthon3, P. Leung4, O. Ljunggren5, D. Mellstrom6, E. S. Orwoll*7, S. Cummings3. 1Clinical Sciences, Malmö, Lund University, Malmö, Sweden. 2University of California, San Diego, San Diego, CA, USA. 3California Pacific Medical Center Research Institute, San Francisco, CA, USA. 4Chinese University of Hong Kong, Hong Kong, Shatin, China. 5University of Uppsala, Uppsala, Sweden. 6University of Gothenburg, Gothenburg, Sweden. 7Oregon Health & Science University, Portland, OR, USA

Background: Rates of hip fracture in men are highest in Scandinavia and lowest in Asia, and intermediate in the U.S. This suggests that, after considering known differences in BMI, BMD might differ between these regions and follow the same pattern.

Methods To test this hypothesis we developed large prospecive studies of osteoporosis in 11,000 men age 65 and older: MrOS U.S. (5,995), MrOS Hong Kong (2000), and MrOS Sweden (3008). The studies used identical core assessments of risk factors, measurements of height and weight, and cross-calibrated measurements of hip and spine BMD.

Results: BMI was highest in the U.S. and lowest in Hong Kong (Table). Before adjustement for differences in weight and age, U.S. men had the highest levels of total hip and L-spine BMD. After these levels were adjusted for age and BMI, there were still significant differences in total hip BMD between men in the three countries, however the differences were less. The differences in L-spine BMD also narrowed. Even though there are statistical significances, we regard clinical significance of the differences as minor, the power for comparison with the large sample size between HK and Sweden is almost 80% to find a difference of 0.01mg/cm2 and even smaller with US. The resulting differences were too small to account for any differences in fracture rates between these regions. 

Table  .  
  1. *P <.001 vs. US; † P<.01 vs US

inline image

Conclusions: Differences in BMD between three regions of the world are largely due to differences in BMI. After adjustment for BMI, levels of hip BMD did not differ and levels of spine BMD were very similar between nations. Thus, the lower rate of hip fractures in men in Hong Kong, and the smaller differences between the U.S. and Sweden cannot be explained by differences in BMI or BMD and, therefore, must be due to other genetic factors controlling bone structure or environmental factors that determine risk of falls and hip fractures.

Disclosures: E.S. Orwoll, Procter & Gamble, GSK, Aventis, TAP Pharma 5: Merck 8.


Hip Fracture Is Associated with Increased Long-Term Mortality as well as Short-Term Mortality in Older Women: The Study of Osteoporotic Fractures. T. A. Hillier1, K. L. Pedula*1, J. H. Rizzo*1, H. Fink2, P. Cawthon3, D. Bauer3, J. A. Cauley4, S. R. Cummings3, W. S. Browner3. 1Center for Health Research, Kaiser Permanente, Portland, OR, USA. 2VA Medical Center, Minneapolis, MN, USA. 3San Francisco Coordinating Center, CPMC and UCSF, San Francisco, CA, USA. 4U Pittsburgh, Pitt, PA, USA

Most studies evaluating mortality after hip fracture are limited to hospitalized samples, and by lack of a representative comparison group. No study has evaluated long-term mortality risk with hip fracture, and potential mediation with BMD.

Among a prospective cohort of 9,704 community dwelling women age 65 and older, follow-up for incident hip fractures and mortality was ascertained every 4 months (>95% complete) and physician adjudicated from medical records. We compared mortality rates in 3,303 women (1101 incident hip fractures after measured total hip DXA BMD (HFx), 2202 without hip fracture (NoHFx), after randomly matching 2 women without hip fracture to fracture cases by age at hip fracture (+/− 5 yrs). We used logistic regression models to compare mortality rates at 1, 5, and 10 year intervals among matched HFx:NoHFX groups where all survived to the start of the interval.

Overall 643 women (58.4%) died over a mean of 3.5 yrs after the hip fracture (range 0–15.3 years), 203 (18.4%) by year 1 after hip fracture. Mortality rates were similar after both intertrochanteric and femoral neck hip fractures. 

Table  .  
  1. * Multivariable model includes age at fracture, bmi, smoking, fair/poor health status, history of diabetes or stroke at baseline exam.

inline image

Although overall mortality was significantly increased over a range up to 15 years after hip fracture, this increased risk was not uniform across the years (Table). Relative to similarly aged women, hip fracture was associated with a 2-fold increased 1 year mortality risk independent of BMD. Increased associated risk of mortality was again present 10 years later, but in contrast was completely attenuated by BMD. It is possible that other diseases or frailty may account for some of the increased mortality observed.

Our results suggest that, among community-dwelling women age 65 and older, decreased BMD is an important mediator of long-term, but not short-term, increased mortality risk among women who suffer a hip fracture.

Disclosures: T.A. Hillier, None.

This study received funding from: NIAMS and NIA.


Increased Mortality Following All Major Osteoporotic Fractures: A 15 Year Follow-Up Study. D. Bliuc*, T. V. Nguyen, J. A. Eisman, J. R. Center. Bone and Mineral, Garvan Institute of Medical Reasearch, Sydney, Australia.

Increased mortality over the first 5 years following osteoporotic fracture has been documented for hip, vertebral and major fractures in both women and men but long term patterns of mortality are unclear. The aim of this study was to examine mortality over a 15 year period following all osteoporotic fractures in elderly women and men.

All subjects sustaining an osteoporotic fracture from the Dubbo population (Dubbo Osteoporosis Epidemiology Study) were followed from 1989 to 2005. Fractures were classified into hip, vertebral, major (proximal humerus, distal femur, proximal tibia, pelvis, multiple rib) and minor (all other peripheral fractures excluding fingers and toes) fractures. Mortality data was obtained from searches of funeral directors and local Dubbo media records. Age and sex-specific standardized mortality ratios following fracture were calculated based on yearly mortality rates for the Dubbo population using data from the Australian Bureau of Statistics. Life tables were constructed for the Dubbo population and fracture subjects in 5-year age categories and by fracture type. Changes in mortality over time were calculated over 5-year age intervals with adjustments made for increasing age brackets.

There were 880 incident fractures in women and 329 fractures in men over 27,687 and 20,054 person-yrs of follow-up, respectively. There were 399 deaths in women and 181 deaths in men following fracture. Standardised mortality ratios (SMR) were increased for both women and men in all fracture groups except those with minor fractures. In women the ratios were 3.8 (95% CI 3.1, 4.6) for proximal femur, 2.2 (1.9, 2.7) for vertebral and 1.8 (1.4–2.3) for major fractures. For men the SMRs were 4.8 (3.6, 6.5) for proximal femur, 3.1 (2.4, 4.0) for vertebral and 1.7 (1.2, 2.4) for major fractures. The greatest increases in mortality were observed for the youngest age groups in both women and men with SMRs of 9.2 (4.5, 18.8) and 3.9 (2.2, 7.1) for hip and vertebral fractures in women and 7.5 (1.9, 30.2) and 6.4 (3.2, 13.1) for hip and vertebral fractures in men. This equated to 6.7 and 12.2 years of life lost in the younger women and men with hip fractures. The majority of deaths occurred in the first 5 years but increased mortality rates were observed for up to 10 years following a fracture.

Thus this study has demonstrated that increased mortality following hip, vertebral and major fractures persists for up to 10 years after the initial fracture with the greatest increases in mortality and years of life lost seen in the younger age groups. Thus, although the fracture burden lies predominantly in the older age groups, the associated mortality has its greatest impact in younger old people.

Disclosures: D. Bliuc, None.


Clinical Risk Factors for Fracture in Minority Women: The Women's Health Initiative (WHI). J. A. Cauley1, G. Talavera*2, L. Wu*3, N. S. Wampler*4, M. Allison*5, J. Barnmhart*6, Z. Chen7, J. Robbins8, R. D. Jackson9. 1University of Pittsburgh, Pittsburgh, PA, USA. 2San Diego State University, San Diego, CA, USA. 3FHCRC, Seattle, WA, USA. 4UCHSC, Denver, CO, USA. 5UCSD, La Jolla, CA, USA. 6Albert Einstein College of Medicine, Bronx, NY, USA. 7University of Arizona, Tucson, AZ, USA. 8UC Davis, Sacramento, CA, USA. 9OSU, Columbus, OH, USA

Fracture rates are lower in minority women, but recent gains in life expectancy have been greater for non-whites which may lead to an increase in the number of fractures in these women. To identify clinical risk factors for fractures in minority women, we used the combined dataset of the WHI Clinical Trials and Observational Study, n = 161,808. Over an average 7.1 years of follow-up, total incident fractures occurred in 21,083 (15.8%) White women; 1,112 (7.6%) Black women; 580 (8.9%) Hispanic women; 391 (9.3%) Asian women; 104 (14.6%) American Indian women: the annualized rates (%) of fracture in these groups were 2.0, 0.9, 1.3, 1.2, and 2.0, respectively. We used Cox Regression Models to calculate the Multivariate (MV) Hazard Ratio (95% Confidence Interval) for fracture. Models were run for each ethnic group separately and then combined to identify race by risk factor interactions and evaluate the degree to which risk factors explain ethnic differences in fracture rates. We highlight the results for each minority group. Significant (p < 0.05 in MV models) predictors of fracture in each group were: Blacks: ≥ high school education, 1.22 (1.0, 1.5); (+) fracture history (hx), 1.75 (1.4, 2.2); and > 2 falls, 1.67 (1.9, 2.0); Hispanics: height (> 162 cm), 1.55 (1.1, 2.2); (+) fracture hx, 1.85 (1.4, 2.5); > 2 falls, 1.81 (1.4, 2.3); rheumatoid arthritis, 1.29 (1.1, 1.6); corticosteroid use, 3.88 (1.9, 8.0); and parental hx of fracture, 1.28 (1.0, 1.6); Asians: age (per 5 y), 1.15 (1.0, 1.3); (+) fracture hx, 1.5 (1.1, 2.0); current hormone therapy (HT), 0.66 (0.5, 0.8); parity (> 1), 1.78 (1.1, 3.0); > 2 falls, 1.42 (1.1, 1.9); American Indian: (+) fracture hx, 2.89 (1.5, 5.7); current HT, 0.5 (0.3, 0.9). Two race × risk factor interactions were identified in the MV model (Table). The risk of fracture differed by age, (no increase observed in Blacks) and fall hx (enhanced risk in Blacks and Hispanics). Despite these associations, ethnic differences in fracture rates are independent of the clinical risk factors examined in this study. 

Table  . Race by Ethnicity Interactions: Hazard Ratio for Fracture by Age and > 2 Falls
inline image

Disclosures: J.A. Cauley, Merck 2. 8: Eli Lilly 2; Pfizer Pharmaceuticals 2: Novartis 2, 5.


Anemia and Fractures. Z. Chen1, L. A. Arendell*1, C. A. Thomson*2, T. I. Bassford*3. 1Epidemiology, University of Arizona, Tucson, AZ, USA. 2Nutritional Science, University of Arizona, Tucson, AZ, USA. 3Family and Community Medicine, University of Arizona, Tucson, AZ, USA

Anemia is highly prevalent in older people and it presents a significant public health problem in the growing aging population. The health consequences of anemia are still under investigation. Here, the relationship between anemia and fractures was examined in a prospective multiethnic cohort of postmenopausal women in the Women's Health Initiative (WHI) Observational Study (OS). Participants (n = 93,676) were from 40 WHI clinical centers across the United States. They were age 50–79 years old at baseline and have been followed for an average of 6.5 years as of February 2004. Only women with baseline hemoglobin (Hg) measurements were included in this analysis (n = 91,999). Anemia was defined as Hg concentrations less than 12 gm/dl. During the follow-up, WHI-OS participants reported incident fractures in their annual medical updates. Self-reported hip fractures were further adjudicated by trained physicians using clinical records. A good agreement between self-reported fractures and medical records was found in the WHI. The numbers of fractures used in this study were: 773 hip fractures, 2,776 forearm or wrist fractures, 12,358 all fractures combined. Over 6% of women were anemic at baseline. The age adjusted fractures rates were 20.00 at hip, 46.02 at forearm or wrist, and 218.02 for total per 10,000 person-years in women with anemia, while the rates observed in non-anemic women were 12.59 for hip, 46.89 for forearm or wrist, and 207.90 for total. Cox proportional hazards models were used to test the association between anemia and fracture incidence. Table 1 presents the crude as well as the adjusted hazards ratios (HR) and 95% confidence intervals (95%CI). In conclusion, women with anemia had significant increased risk for hip factures and all fractures combined; suggesting anemia is a new risk factor for osteoporotic fractures. This is the first study linking anemia with increased incidence of fractures in U.S. postmenopausal women from diverse race/ethnic backgrounds. Whether treating anemia will reduce fracture risk remains to be tested.

Table Table 1.. Relative risk of fractures in postmenopausal women with anemia
  1. * Adjusted tor age, race/ethnicity, and BMI.

inline image

Disclosures: Z. Chen, None.

This study received finding from: NIH.


Vascular Calcification and Risk of Hip Fracture: The Framingham Study. E. J. Samelson1, K. E. Broe1, M. Lu*1, M. T. Hannan1, L. A. Cupples*2, C. J. O'Donnell*3, D. P. Kiel1. 1Institute for Aging Research, HSL, Boston, MA, USA. 2Biostatistics, BUSPH, Boston, MA, USA. 3NHLBI, Boston, MA, USA

Evidence for shared pathogenesis between osteoporosis and atherosclerosis has been emerging. We examined the association between aortic calcification (AC) and risk of hip fracture (HF) in women and men in the population-based Framingham Study.

Participants included 1453 women and 1046 men with lumbar x-rays taken in 1967–70. AC was scored (0–24) using a validated method (ICC = 0.96). HF was ascertained by interview and record review and confirmed by x-ray report. Incidence was calculated as cases divided by person-years of follow-up (date of x-ray to first of: HF, last contact, death, or 12/31/01). Cox regression was used to calculate hazard ratios (HR) and 95% confidence intervals (CI) for relation between AC and HF adjusted for age, BMI, smoking, alcohol, BP, cholesterol, diabetes, CHD, and estrogen. Using same predictors, we calculated HRs for mortality to evaluate the relation between AC and death. Due to skewed distribution, AC was transformed to log (AC+1).

At baseline, mean age was 61 years (range 47–80), and median AC score was 1 in women and 2 in men. 41% of women and 33% of men were free of AC (score 0), and 12% of women and 9% of men had AC score 11+. Individuals with prevalent AC (score 1+) were older, had higher levels of cholesterol and BP, and higher prevalence of diabetes and CHD than those free of AC. Mean duration of follow-up was 25 years during which time 215 (15%) women and 48 (5%) men sustained HF. Incidence of HF was 16% (132/851) in women and 5% (36/706) in men with prevalent AC (score 1+), and 14% (83/602) in women and 4% (12/340) in men without AC (score 0). Risk of HF was similar in persons with AC (score 1+) and without AC (score 0). HR was 1.2 (CI:0.9–1.7) in women and 1.7 (CI:0.9–3.4) in men. No increasing trend in HF risk was observed with increasing AC score (Table). However, AC score was associated with increased risk of death (Table). Mortality was double in those with AC score 11+ compared to those with score 0.

Lack of association observed between AC and hip fracture is possibly due to competing risk of death, as individuals with AC were more likely to have died than those free of AC. The common radiographic finding of aortic calcification in middle aged and older adults cannot be recommended in the clinical setting for identifying persons at increased risk of HF. 

Table  . Adjusted HR (95% CI) for Hip Fracture and Death by AC Score
inline image

Disclosures: E.J. Samelson, None.


High Trauma Fractures in Men Are Associated with Low Bone Mineral Density. D. C. Mackey*1, S. R. Cummings1, P. C. Bauer1, P. M. Cawthon1, C. Lewis2, E. Barrett-Connor3 for the Osteoporotic Fractures in Men (MrOS) Study*1. 1San Francisco Coordinating Center, California Pacific Medical Center Research Institute and University of California, San Francisco, San Francisco, CA, USA. 2University of Alabama at Birmingham, Birmingham, AL, USA. 3University of California, San Diego, San Diego, CA, USA

It is widely believed that fractures that result from significant trauma in men are not “osteoporotic”. This assumption has not been studied. To test this belief, we measured bone mineral density (BMD) of the hip and lumbar spine with dual-energy x-ray absorptiometry (DXA) (Hologic QDR 4500) in 5995 older men (mean ± SD age: 73.7 ± 5.9 yrs) who were enrolled in the prospective Osteoporotic Fractures in Men (MrOS) Study.

We confirmed incident non-spine fractures by central review of radiology reports. Fractures were defined as high trauma if they resulted from a car crash, falls from more than standing height, or similarly severe trauma. Low trauma included falls from standing height or less severe trauma. Degree of trauma was classified by experienced osteoporosis clinical investigators based on patient histories using a standardized rating system. We used proportional hazards regression models to estimate the hazard ratio (HR) and 95% confidence interval (CI) of non-spine fracture per standard deviation (SD) decrease in BMD. During an average of 4.3 years of 99% complete follow-up, 82 men suffered incident high trauma non-spine fractures and 277 men suffered low trauma fractures. After adjusting for age, the risk of high trauma fracture increased 1.5-fold with each SD decrease in total hip BMD and 1.6-fold with each SD decrease in total lumbar spine BMD (Table). This association between BMD and high trauma fracture was similar in magnitude to the association between BMD and low trauma fracture. 

Table Table.. High and low trauma fracture: Age-adjusted relative hazard (95% CI) per SD decrease in BMD.
inline image

Our results suggest that high trauma fractures in men are related to low BMD and should be considered osteoporotic. High trauma fractures should not be excluded in clinical trials, and should be treated similarly to low trauma fractures in clinical practice.

Disclosures: D.C. Mackey, None.


Loop Diuretic Use Is Associated with Increased Rates of Hip Bone Loss in Older Men. L. S. Lim*1, H. A. Fink2, M. A. Kuskowski*2, J. T. Schousboe*3, K. E. Ensrud2. 1Erlanger Bledsoe Clinic, Pikeville, TN, USA. 2GRECC & CCDOR, VA Medical Center, Minneapolis, MN, USA. 3Park Nicollet Clinic, Minneapolis, MN, USA

Loop diuretics exert a calciuric effect. However, whether use of loop diuretics is associated with increased rates of bone loss is unknown.

We examined the association between loop diuretic use and loss of bone mineral density (BMD) using data from the Osteoporotic Fractures in Men (MrOS) Study, a multi-center, prospective cohort study of osteoporotic fractures in men aged ≥65 years. Based on standardized evaluations of all prescription bottles at the baseline and follow-up clinic visits (mean interval 3.4 years +/− 0.5 SD), subjects were categorized as continuous users of loop diuretics (use at both visits), intermittent users of loop diuretics (use at 1 but not both visits), or loop diuretic nonusers. Subjects taking a non-loop diuretic at either visit were excluded from all analyses.

A total of 2328 men (2168 nonusers, 86 intermittent users and 74 continuous users) with mean age 73 years at baseline were included in the analyses. BMD was measured at total hip and hip subregions using DXA at both the baseline and follow-up visit. For each skeletal site, we computed the adjusted mean annualized % change in BMD in multivariate adjusted models. All covariates were measured at baseline. 

Table  . Mean Annualized Rate Of Change (%/year) In Hip BMD (95% CI) By Loop Diuretic Use
  1. Adjusted for age, baseline site-specific BMD, body mass index, Physical Activity Scale for the Elderly score, dietary potassium intake, diabetes mellitus, chronic obstructive pulmonary disease, congestive heart failure, statin use, and clinic site.

  2. *p <0.05 for comparison with nonuser loop diuretic group.

inline image

We conclude that older men taking loop diuretics have increased rates of hip bone loss. If these findings are confirmed in other studies, older people taking loop diuretics should be targeted for bone density testing.

Disclosures: L.S. Lim, None.


Comparison of 3 Sequential Treatment Regimens of Teriparatide: Final Results from the EUROFORS Study. R. Eastell1, P. Hadji2, J. Farrerons*3, M. Audran4, S. Boonen5, K. Brixen6, J. Melo Gomes*7, B. Obermayer-Pietsch*8, A. Avramidis*9, G. Sigurdsson10, C. Glueer11, S. Cleall*12, F. Marin12, T. Nickelsen12. 1University of Sheffield, Sheffield, United Kingdom. 2Klinikum der Philipps-Universitaet, Marburg, Germany. 3Hospital Santa Creu i Sant Pau, Barcelona, Spain. 4Centre Hospitalier Universitaire, Angers, France. 5University of Leuven, Leuven, Belgium. 6Odense Universitetshospital, Odense, Denmark. 7Servimed, Lisboa, Portugal. 8Universitaetsklinik fuer Innere Medizin, Graz, Austria. 9Hippokration Hospital, Thessaloniki, Greece. 10Landspitalinn University Hospital, Reykjavik, Iceland. 11UKSH, Kiel, Germany. 12Eli Lilly & Company Europe, Windlesham, United Kingdom

The prospective, controlled, randomized EUROFORS trial was designed to compare 3 sequential treatment regimens of teriparatide (TPTD) in postmenopausal women with established osteoporosis. In year 1, all patients received open-label TPTD 20μg/d and supplements of 500 mg/d calcium and 400-800 IU/d vitamin D (Ca/D). After 12 months, women were randomized at a ratio of 3:1:1 to (1) continue TPTD, (2) switch to raloxifene (RLX) 60 mg/d, or (3) receive no pharmacologic treatment (substudy 1). A fourth cohort of women with documented inadequate response to prior antiresorptive therapy - as defined previously1 - continued to receive TPTD throughout the second year (substudy 2). Ca/D supplements were continued in all groups. 706 of the 865 originally enrolled patients completed year 1; of these, 507 were in substudy 1 and were randomized. The Table lists the mean BMD changes in the second year by substudy and treatment arm [absolute and % change, (95%C.I.)] and the p-values for within-group differences from the randomization visit at 12 months (*p<.05, **p<.001) and between-group differences at 24 months (§ P<.001 vs Ca/D, $ P<.001 vs RLX, ° P = .001 vs Ca/D, # P<.05 vs Ca/D). 

Table  .  
inline image

In contrast to Ca/D, TPTD further increased lumbar spine BMD in year 2, while RLX maintained a stable level. At the total hip, both regimens were associated with a further statistically significant BMD increase. We conclude that both treatment sequences involving active pharmacologic treatment for 2 years (TPTD-TPTD and TPTD-RLX) produce superior overall BMD results compared with TPTD-Ca/D.

1Farrerons J et al. J Bone Min Res 19 Suppl 1:F441.

Disclosures: R. Eastell, Lilly 2, 5.

This study received funding from: Eli Lilly and Company.


The DIVA Study: Substantial Hip Bone Mineral Density Improvements with Intermittent Intravenous Ibandronate Injections. E. Lewiecki1, S. Adami*2, G. Bianchi*3, P. Sambrook*4, C. Hughes*5, N. Mairon*5, I. Jonkanski*5, M. Bolognese6. 1New Mexico Clinical Research and Osteoporosis Center, Albuquerque, NM, USA. 2University of Verona, Verona, Italy. 3Ospedale La Colletta, Azienda Sanitaria Genovese, Genova, Italy. 4University of Sydney, Sydney, Australia. 5F. Hoffmann-La Roche Ltd, Basel, Switzerland. 6Bethesda Health Research, Bethesda, MD, USA

A potential treatment option for women with postmenopausal osteoporosis who are unable to take or tolerate oral bisphosphonates is an intravenous (i.v.) bisphosphonate injection. Daily oral ibandronate has proven antifracture efficacy (3-year vertebral fracture risk reduction: 52%1). DIVA, a randomized, double-blind study, enrolled 1,395 postmenopausal women (aged 55–80 years and ≥5 years postmenopause) with osteoporosis (lumbar spine bone mineral density [BMD] T-score <−2.5 and ≥−5.0). Participants received ibandronate for 2 years administered as a 2mg every 2 months (q2mo) or 3mg every 3 months (q3mo) i.v. injection over 15–30 seconds or a 2.5mg daily oral regimen (active comparator). Supplements of calcium (500mg) and vitamin D (400IU) were also provided. At 1 and 2 years, the i.v. ibandronate regimens were prospectively proven non-inferior and also superior (p <0.001) to daily oral ibandronate for improving lumbar spine BMD.2.3 Robust gains in hip BMD were also obtained. In all treatment arms, sizeable gains in proximal femur (total hip, femoral neck and hip trochanter) BMD were observed at 2 years (Table). BMD increases were greater at all hip sites in the i.v. groups than the oral group, and post-hoc analyses revealed that the BMD gains at the total hip and hip trochanter were significantly greater (p <0.001) (Table). For all spine and hip BMD endpoints, post-hoc analyses revealed no statistically significant differences between the two i.v. regimens. At the lumbar spine and hip, intermittent i.v. ibandronate injections provide consistently greater BMD increases compared to the efficacious daily oral regimen. Ibandronate i.v. injections are therefore an effective treatment option for women with postmenopausal osteoporosis.


  • Chesnut C, et al. Curr Med Res Opin 2005;21:39-401.

  • Recker R, et al. Arthritis Rheum 2004;50(Suppl.):4095.

  • Emkey R, et al. Arthritis Rheum 2005;52:4060 (Abstract L8).


Table  .  
  1. *n = 334, 320 and 334 for daily, q2mo and q3mo, respectively

  2. p <0.001 vs daily oral ibandronate (IBN)

inline image

Disclosures: E. Lewiecki, F. Hoffmann-La Roche Ltd 2; GlaxoSmithKline 2.

This study received funding from: F. Hoffmann-La Roche Ltd/GlaxoSmithKline.


Microdamage Is Self-Limiting in Beagles Treated for Three Years with Clinical Doses of Alendronate. M. R. Allen, D. B. Burr. Anatomy & Cell Biology, Indiana University School of Medicine, Indianapolis, IN, USA

Clinical doses of alendronate reduce vertebral bone turnover by 71% and increase microdamage by 4-fold in skeletally mature beagles after 1 year of treatment. The goal of this study was to determine if microdamage would continue to accumulate or be self-limiting with up to three years of alendronate treatment. One-year-old female beagles were treated with daily oral doses of vehicle (VEH) or alendronate (ALN; 0.2, or 1.0 mg/kg/day) for either 1 year (n = 12/group) or 3 years (n = 5/group). The alendronate doses correspond to the clinical (0.2) and 5x clinical (1.0) dose, on a mg/kg/basis, used for treatment of postmenopausal osteoporosis. Lumbar vertebrae density (BMD) and trabecular bone volume (BV/TV), activation frequency (Ac.f), and crack surface density (Cr.S.Dn) were quantified. After 1 year of treatment, BMD, BV/TV, and Cr.S.Dn were all significantly higher, and Ac.f was significantly lower, in both ALN 0.2 and ALN 1.0 groups compared to VEH. There was no significant difference between the two ALN doses. Similarly, after 3 years of treatment with either ALN dose, BMD, BV/TV, and Cr.S.Dn were significantly higher, and Ac.f was significantly lower, compared to VEH. Compared to 1 year of treatment, 3 years significantly increased BMD in both ALN 0.2 (+9%) and ALN 1.0 (+8%) groups while VEH-treated animals were unchanged. There was no difference in trabecular BV/TV between 1 and 3 years treatment for any group. Between years 1 and 3 both VEH and ALN 0.2 groups experienced a significant reduction in Ac.f (−55 and −35%, respectively). VEH-treated animals tended to have higher Cr.S.Dn after 3 years compared to 1 year (+2-fold; p = 0.10), perhaps a reflection of the age-associated decline in Ac.f. There was no significant difference in Cr.S.Dn. between 1 and 3 years treatment for either dose of ALN. Based on these results, we conclude that despite continued suppression of turnover and increases in bone density, microdamage does not continue to accumulate in alendronate-treated animals. This suggests microdamage accumulation is self-limiting with maximum accrual during the first year of treatment.

Disclosures: M.R. Allen, None.


Prophylactic Bisphosphonate Treatment Prevents Bone Fractures after Liver Transplantation. M. Bodingbauer*1, T. Wekerle*1, B. Pakrah*2, G. Silberhumer*1, P. Roschger3, M. Peck-Radosavljevic*4, S. Grampp*5, S. Rockenschaub*1, R. Steininger*1, G. Berlakovich*1, R. Oberbauer*6, K. Klaushofer3, F. Muehlbacher*1. 1Surgery, Transplantation, Vienna, Austria. 2Surgery, Neurosurgery, Vienna, Austria. 3Ludwig Boltzmann Institute of Osteology of WGKK and AUVA Trauma Centre Meidling, Vienna, Austria. 4Department of Internal Medicine IV, Division of Gastroenterology and Hepatology, Medical University, Vienna, Austria. 5Department of Radiology, Medical University, Vienna, Austria. 6Department of Nephrology, Medical University, Vienna, Austria

Vertebral fractures in patients who have undergone liver transplantation are associated with high morbidity and increased mortality. No effective prophylaxis for fractures has been described so far.

A randomized controlled prospective open-label single center trial was performed. At the time of transplantation patients were randomly assigned to one of two treatment arms (treatment from time of transplant through 12 months): group 1 (47 patients) received zoledronic acid, (8 infusions at 4mg) calcium (1000mg/d) and vitamin D (800 IE/d): group 2 (49 patients) received calcium and vitamin D at same doses. The frequency of bone fractures or death was pre-defined as the primary endpoint. Secondary endpoints included mean trabecular calcium concentration and trabecular morphometry measured in bone biopsies (d0 and 6 months after LT), bone mineral density (BMD) of the femoral neck and the lumbar spine evaluated by dual energy x-ray absorptiometry, and serum biochemical markers of bone metabolism (determined before, 3, 6, 12 months after LT). Patients were followed for 12 months. Analysis was performed on an intention to treat basis.

The primary endpoint fracture or death was reached in 46% in the control group vs. only 26% in the zoledronic acid group (p = 0.047, log rank test). Densitometry results were different between the groups at the femoral neck at six months after engraftment (mean ± SD BMD control: 0.73 ± 0.14g/cm2 vs. 0.80 ± 0.19g/cm2, p = 0.036). Mixed linear models of biochemical bone markers showed significantly lower bone turnover in the zoledronic acid group at the four different time points investigated (0, 3, 6, and 12 months). Results from this prospective trial show that prophylactic treatment with the bisphosphonate zoledronic acid during the first 12 month after liver transplantation significantly improves parameters of bone metabolism and substantially reduces morbidity.

Disclosures: M. Bodingbauer, None.


The Letrozole (L), Exemestane (E), and Anastrozole (A) Pharmacodynamics (LEAP) Trial: A Direct Comparison of Bone Biochemical Measurements between Aromatase Inhibitors (AIs) in Healthy Postmenopausal Women. E. McCloskey1, R. Hannon*1, G. Lakner*2, G. Clack*3, R. D. Finkelman*4, R. Eastell*1. 1Academic Unit of Bone Metabolism, University of Sheffield, Sheffield, United Kingdom. 2MAV Hospital, Budapest, Hungary. 3AstraZeneca, Macclesfield, United Kingdom. 4AstraZeneca, Wilmington, DE, USA

It has been suggested there are differences between the effect of the steroidal (E) and nonsteroidal AIs (A and L) on bone turnover, with steroidals having a less-negative effect. The LEAP trial is an open, randomized, pharmacodynamic study comparing the effects of the AIs L, E and A on safety parameters, such as serum markers of bone formation and resorption, lipid profiles, and adrenal function in healthy postmenopausal women with normal spine and hip bone mineral density. This analysis reports the effects on bone turnover markers.

Healthy volunteers from the UK and Hungary were randomized to receive A (1 mg/day), L (2.5 mg/day), or E (25 mg/day) orally, once daily for 24 weeks, and then followed up for a further 12 weeks after cessation of therapy. Changes from baseline in log-transformed bone alkaline phosphatase (ALP), serum C-telopeptide crosslinks (CTX), parathyroid hormone (PTH) and propeptide of type I procollagen (PINP) at 24 weeks on A, were compared with those on L and E by ANCOVA, adjusting for treatment, baseline value, BMI, smoking status and baseline estradiol. No adjustments were made for multiple comparisons.

102 healthy volunteers were recruited, 90 of whom were evaluable (29 A, 29 L, 32 E). Participant demographics were similar between groups in terms of age, years since menopause, and history of hysterectomy and oophorectomy. Bone biochemical measurement changes at 24 weeks are presented (Table). There were no statistically significant differences between AIs, with the exception of PTH, where there was a greater decrease with E than with A (p = 0.04).

12 weeks after cessation of therapy, the changes seen in the markers of bone turnover appeared to persist in all treatment groups, though the degree of inter-subject variability remained high, while PTH levels had returned to baseline.

The steroidal and non-steroidal AIs appear to have similar effects on bone biochemical measurements, and thus bone turnover. All three licensed AIs result in increases in bone turnover. 

Table Table.. Bone biomarker changes at 24 weeks
  1. *From baseline (computed from back transformed LS mean from ANCOVA); **p-value versus anastrozole = 0.04

inline image

Disclosures: E. McCloskey, AstraZeneca 2.

This study received funding from: AstraZeneca.


Offset of Effect of Bisphosphonates on Fracture Incidence - Evidence from a Prospective Controlled Trial of Clodronate. E. McCloskey1, M. Beneton*1, D. Charlesworth*1, J. Orgee*1, K. Pande*1, L. Kersh*1, J. Cliffe*1, N. Peel1, S. Aropuu*2, T. Jalava*2, J. Kanis1. 1University of Sheffield, Sheffield, United Kingdom. 2Schering, Helsinki, Finland

Few prospective studies have provided data on fracture incidence in the immediate post-treatment years but the offset on fracture risk remains a critical topic for treatment strategy and modelling cost-effectiveness. We examined the effects of stopping treatment on fracture incidence for up to 2 years in a double-blind placebo-controlled study of the bisphosphonate, clodronate.

Women aged 75 years and older, randomly selected from the local population were randomised, regardless of the presence or absence of osteoporosis, to receive 3 years of therapy with clodronate (BONEFOS®) 800mg daily p.o. or an identical placebo. Thereafter, all women were followed until the last woman enrolled had completed the 3-year medication period so that follow-up was available for up to 2 years off treatment under blinded conditions/The efficacy of clodronate was determined using Cox regression analysis and the treatment effect was evaluated for the 3 years of therapy, the entire study follow-up and for the post-treatment phase separately.

At baseline, the two treatment groups were well-matched for the prevalence of risk factors and the median follow-up in both groups was 4 years. Of 5579 women evaluable for efficacy, 601 (10.8%) had an incident clinical fracture during the 3-year treatment phase; 264 (9.5%) women in the clodronate group and 337 (12.1%) women in the placebo group. The hazard ratio was 0.80 (95%CI 0.68–0.94), indicating that clodronate treatment reduced the risk of any clinical fracture by 20%. During the whole study period, 830 (14.9%) of the women had an incident clinical fracture; 382 (13.7%) women in the clodronate group and 448 (16.1%) patients in the placebo group giving a hazard ratio of 0.87 (95%CI 0.76–1.002), but the difference between the study groups was not significant (p = 0.054). The incidence of fractures in the 2 years following treatment cessation was similar in women who had previously received clodronate (8.7%) or placebo (7.8%) (HR 0.99, 95%CI 0.80–1.22). The results were similar if the analysis was confined to osteoporosis-associated fractures.

We conclude that clodronate reduced the risk of any clinical fracture during the three years of treatment After stopping treatment, the incidence of fracture was similar to that in women who had previously received placebo. The parallel rates of fracture off-therapy suggest the achieved reduction in fractures during therapy persists for several years thereafter.

Disclosures: E. McCloskey, Schering Ag 2, 8.

This study received funding from: Schering Ag.


Denosumab Improves the Structural Geometry of the Proximal Femur in Postmenopausal Women with Low Bone Mass. T. J. Beck1, P. D. Miller2, E. M. Lewiecki3, D. Felsenberg4, L. Sun*5, C. Libanati5, Y. Liu5, L. A. Fitzpatrick5. 1Johns Hopkins University, Baltimore, MD, USA. 2Colorado Center for Bone Research, Lakewood, CO, USA. 3New Mexico Clinical Research and Osteoporosis Center, Albuquerque, NM, USA. 4Centre for Muscle and Bone Research Charite–Univ Medicine, Berlin, Germany. 5Amgen Inc., Thousand Oaks, CA, USA

Denosumab is a fully human monoclonal antibody that inhibits receptor activator of NF-κB ligand (RANKL). In a phase 2 study in postmenopausal women with low bone mass, denosumab treatment decreased markers of bone turnover and increased BMD as measured by dual-energy x-ray absorptiometry (DXA) compared with placebo. The increases in DXA-BMD in response to denosumab were similar to or greater than for open-label alendronate, with differences particularly at predominantly cortical sites (McClung et al. NEJM 2006;354:821–831). In this post-hoc analysis, structural geometry of the hip was determined for patients treated with denosumab 60 mg 6-monthly (N = 47), placebo (N = 46), or open-label alendronate 70 mg once weekly (N = 47) for up to 24 months.

Hip DXA scan images from baseline, 12 months, and 24 months were analyzed with hip structural analysis (HSA) software. DXA-BMD and cross-sectional geometry were evaluated for 3 regions of the proximal femur: the narrowest point of the femoral neck, the intertrochanter, and the proximal shaft. Geometric parameters included bone cross-sectional area (CSA), section modulus, and buckling ratio. Least square mean (LSM) percent changes between groups were compared by using an Analysis of Covariance (ANCOVA) model.

At 12 and 24 months denosumab and alendronate improved DXA-BMD and geometry compared with placebo (shown in Figure for 24 months). Denosumab effects, compared with alendronate, reached statistical significance at the proximal shaft, a purely cortical site. Greater effects of denosumab than alendronate at the purely cortical shaft are consistent with previous results that highlight the cortical effects of denosumab. The changes in structural geometry observed with this analysis suggest that denosumab improved bone mechanical properties in this study. Figure. Effects of Denosumab and Alendronate on the Proximal Femur at 24 Months

Thumbnail image of

Disclosures: T.J. Beck, Hologic Inc. 5.

This study received funding from: Amgen Inc.


Denosumab (AMG 162, a Fully Human RANKL Antibody) Increases Cortical and Cancellous Bone Mass and Density in Aged Ovariectomized Cynomologus Monkeys. M. S. Ominsky1, J. Schroeder*1, S. Y. Smith2, D. J. Farrell*2, P. J. Kostenuik1, J. E. Atkinson1. 1Amgen Inc., Thousand Oaks, CA, USA. 2Charles River Laboratories Preclinical Services Montreal, Inc., Montreal, PQ, Canada

Denosumab, a fully human monoclonal antibody against RANKL, decreases bone resorption and increases bone mineral density (BMD) in postmenopausal women and in young gonad-intact male and female cynomolgus monkeys (cynos). The effects of denosumab have not been previously described in aged primates after estrogen depletion. We now describe for the first time the effects of long-term (16 month) denosumab therapy on cancellous and cortical bone mass, density, geometry, and turnover in aged (9–16 year old) ovariectomized (OVX) cynos. One month after OVX or sham surgery, OVX animals began treatment with either vehicle (OVX-V) or denosumab at 25 or 50 mg/kg (SC, monthly) for 16 months (n = 14–20/group). Sham controls were treated with vehicle. Serum was collected for measurement of bone turnover markers, and DXA and pQCT scans were taken at baseline and after 3, 6, 12, and 16 months of treatment. Percent changes from baseline were compared between treatment groups. Serum CTx, serum BSAP, and urine NTx were significantly reduced by denosumab treatment throughout the 16-month study (p <0.05 vs OVX-V and Sham). Denosumab treatment resulted in significantly greater areal BMD of the lumbar spine, total hip, femur neck, and central tibia compared with OVX-V or Sham (p <0.05, months 6–16). Total volumetric BMC (vBMC) and BMD (vBMD) were significantly increased in the metaphyses of the distal radius and proximal tibia with denosumab (p <0.05 vs OVX-V and Sham, month 16). Trabecular vBMD was significantly reduced at these sites in OVX-V animals (p <0.05 vs Sham, month 16), and denosumab fully prevented this cancellous bone loss. Cortical vBMC and vBMD at the radius and tibial diaphyses were significantly greater in denosumab-treated animals vs OVX-V (p <0.05, months 6–16). Trends toward increased periosteal and decreased endocortical circumference with denosumab treatment resulted in a significantly greater cross sectional moment of inertia (CSMI) at the tibia diaphysis (p <0.05 vs OVX-V, months 6–16). Bone strength index (a surrogate of strength equal to the product of CSMI and cortical vBMD) was significantly increased by denosumab at the radial and tibial diaphyses (p <0.05 vs OVX-V, months 6–16). These results indicate that long-term denosumab therapy is associated with significant gains in the mass and density of cancellous and cortical bone in aged OVX monkeys. The continuous accrual of bone mass throughout the study led to increases that were consistently greater than OVX controls and frequently greater than sham controls.

Disclosures: J.E. Atkinson. Amgen Inc. 1, 3.

This study received funding from: Amgen Inc.


The Role of α5 Integrin as a Mechanosensor in the Regulation of Connexin 43 Hemichannel Release of Prostaglandin in Response to Mechanical Stress. A. J. Siller-Jackson1, S. Burra*1, S. Gu*1, S. E. Harris2, L. F. Bonewald3, E. Sprague*4, J. X. Jiang1. 1Biochemistry, University of Texas Health Science Center, San Antonio, TX, USA. 2Periodontics, University of Texas Health Science Center, San Antonio, TX, USA. 3Oral Biology, School of Dentistry, University of Missouri, Kansas, MO, USA. 4Radiology, University of Texas Health Science Center, San Antonio, TX, USA

Mechanosensing osteocytes express large amounts of Cx43, yet gap junctions are only active at the tips of their dendritic processes. Fluid flow shear stress (FFSS) induces the opening of Cx43-forming hemichannels on the cell surface of MLO-Y4 osteocyte cells through which prostaglandin is released. Opening of hemichannels and release of prostaglandin are magnitude dependent with increasing response to increased shear, but no differences were observed with either steady (16 dynes/cm2) or pulsatile (8 ± 8 dynes/cm2) flow after 2 hrs. A potent Cx43-hemichannel blocking antibody was developed which targets the second extracellular domain of Cx43. The opening of hemichannels induced by FFSS was significantly inhibited by this blocking antibody, as was the release of PGE2. However, gap junction activity was unaffected. To identify the potential mediators responsible for the opening of hemichannels, gene microarray analysis was performed showing FFSS up-regulation of α5 integrin that was partially inhibited by the gap junction and hemichannel inhibitor, 18 b-glycyrrhetinic acid. Following confirmation by Northern analysis, immunostaining showed co-localization of α5 integrin with Cx43 not only on the cell surface but also throughout the cell body and cell processes, becoming more prominent when exposed to FFSS. α5 integrin co-immunoprecipitated with Cx-43 and vice-versa Cx43 co-immunoprecipitated with α5 integrin using the respective antibodies. α5 integrin co-sedimented with hexamers, not monomers, of Cx43 using sucrose gradient sedimentation, implying the association of α5 integrin with assembled hemichannels. Cx43 and α5 integrins were not co-localized with vinculin and paxillin, components of focal adhesions, suggesting that the association between Cx43 and α5 integrin is unrelated to the function of integrins in focal adhesions. More importantly, α5 integrin antibody significantly decreased FFSS induced dye uptake. Together, these results suggest that α5 integrin is a likely mechanical sensor or tether that conveys the effects of mechanical stress on Cx43, leading to the opening of hemichannels providing a novel portal for passage of prostaglandin and potentially other signaling factors into the bone fluid in response to mechanical stress.

Disclosures: A.J. Siller-Jackson, None.

This study received funding from: National Institue of Health Grant AR46798.


Mechanical Signals Regulate VEGF-A Alternative Splicing in Osteoblasts through Actin Polymerisation. C. Faure*, M. Linossier*, M. Lafage-Proust, L. Vico, A. Guignandon*. LBTO, INSERM E366, Fac Med, St Etienne, France

Mechanical strain and the VEGF signalling pathways control common elements such as focal contact adhesion and actin stress fiber organization, which both depend of the Rho GTPases activity. Some VEGF isoforms bind to the matrix and activate integrin alphaV-beta3. In endothelial cells matrix-bound isoforms are known for their anti-apoptotic effects and angiogenic action after release by matrix proteases. In this context, we hypothesized that osteoblasts under mechanical strain will produce matrix-bound VEGF-A isoforms in a cytoskeleton dependent manner.

Human osteosarcoma (MG-63) and primary human osteoblasts were plated onto flexible type I collagen-coated membranes (Flexcell Strain Unit Fx-3000, Flexcell, NC, US) and subjected to 1% deformation at various frequencies (0.05-5Hz). The different VEGF-A isoforms were analysed at the mRNA level by quantitative RT-PCR and at the protein level by ELISA of supernatants (soluble VEGF121 and VEGF165) and matrix/cell lysates (matrix-bound forms VEGF145, 189, 206, and 165). We also quantified the high VEGF affinity protein, fibronectin (mRNA). We then challenged cytoskeletal structures by actin surpolymerisation (jasplakinolide, jasp) or depolymerisation (latrunculinA, latA) and by gene silencing of RhoA and Rac1 by small interfering RNA.

A low frequency regimen (LF, 0.05Hz-150min) stimulates soluble VEGF 121 and 165 production (>2x, p <0.0001) while a high frequency regimen (HF: 5Hz-1.5min) stimulates matrix-bound VEGF 145, 165, 189 and 206 production (>2x, p <0.001) and fibronectin synthesis (60%, p <0.001) in both cell types. Only HF maintains actin stress cable polymerisation 12 hrs after strain. LatA treatment suppresses HF strain-related matrix-bound VEGF production and has no effect on the LF-induced soluble forms. Jasp treatment increases matrix-bound VEGF by itself thus mimicking HF effects. RhoA inactivation prevents HF induced matrix-bound VEGF isoforms production, fibronectin synthesis and cytoskeletal alteration but does not change LF effects. Rac1 inactivation has no influence on HF effects. In conclusion, the cyclic strain regimen is a potent agent for VEGF production and the frequency range modulates VEGF-A alternative splicing in osteoblasts. HF-related increase in matrix-bound VEGF production is mediated by RhoA signalling and actin polymerisation.

Disclosures: A. Guignandon, None.

This study received funding from: ESA and INSERM.


Mechanical Loading Reduces Osteocyte Expression of Sclerostin Protein. A. G. Robling1, T. M. Bellido*2, C. H. Turner3. 1Department of Anatomy and Cell Biology, Indiana University, Indianapolis, IN, USA. 2University of Arkansas School of Medicine, Little Rock, AR, USA. 3Department of Biomedical Engineering, Indiana University, Indianapolis, IN, USA

The population density, distribution, and extensive communication networks among osteocytes make them ideal mechanosensors in bone's adaptive process. Despite these attributes, very little data have been generated that implicate the osteocyte network as the primary mechanosensory cell type, to the exclusion of the other cell types (e.g., osteoblasts, bone lining cells). Sclerostin, the protein product of the Sost gene, is an osteocyte-specific cysteine-knot glycoprotein that is a potent inhibitor of bone formation. Sclerostin can bind and antagonize Lrp5, a Wnt co-receptor that we have shown to be required for load induced bone formation. We investigated the expression of sclerostin in mechanically loaded and control (nonloaded) mouse ulnar diaphyses to determine whether this osteocyte-specific factor was under mechanoregulation. To this end, we subjected the right forelimbs of 4 male C57BL/6 mice at 18-wk of age to a single 30 sec bout of in vivo mechanical loading (60 cycles, 2 Hz, ∼1800 μω) using a noninvasive rodent ulna loading model. Sclerostin expression was measured using immunohistochemistry 24 hrs after loading. We measured the number of sclerostin-positive osteocyte cell bodies, and the cortical area occupied by prominently-stained (sclerostin-positive) canaliculi in each ulnar section. Consistent with other reports, the sections revealed strong reactivity in osteocytes but not in bone lining cells, osteoblasts, or marrow cells. Sections from loaded ulnae revealed a striking decrease in sclerostin immunoreactivity compared to non-loaded controls. Loading reduced the number of sclerostin-positive osteocyte cell bodies by 27% (p = 0.03), with a significant reduction in the intensity of staining in the osteocytes that remained positive. Furthermore, loading induced a dramatic reduction in the percent of bone area exhibiting canaliculi labeled with anti-sclerostin antibody in all 4 mice studied, resulting in an average decrease of 75% (p = 0.01).

In conclusion, mechanical loading resulted in a clear and significant reduction in sclerostin protein levels in the ulnar diaphysis after 24 hrs. Therefore, sclerostin appears to be a mechanically modulated osteocyte-specific factor. Sclerostin is secreted by osteocytes and therefore can reach nearby osteoblasts and suppress Wnt signaling through Lrp5. The suppression of sclerostin expression by mechanical loading may provide a mechanism by which Wnt signaling is enhanced and bone formation is increased.

Disclosures: A.G. Robling, None.


Physical Exercise Reduces Remodeling Imbalance and Preserves Trabecular Microarchitecture in Male Mice with Androgen Receptor Disruption. J. Ophoff*1, K. De Gendt*1, C. Ohlsson2, S. Boonen1, R. Bouillon1, G. Verhoeven*1, D. Vanderschueren1. 1KULeuven, Leuven, Belgium. 2Sahlgrenska University Hospital, Göteborg, Sweden

Disruption of the androgen receptor (AR) in mate mice reduces muscle mass and radial (cortical) bone growth during puberty. Simultaneously, AR inactivation increases cancellous bone resorption. We hypothesized that sustained exercise may restore cortical bone acquisition independently of androgen action.

Aim of this study was to find out whether (voluntary) running during growth rescues effects of AR disruption on bone and body composition. To this end, 3 week old androgen receptor knockout (ARKO) and wild type (WT) mice were put in running cages (cage with treadmill and monitoring system) for a learning period of 2 weeks and then randomized to runners (in running cage) and controls (normal cage without treadmill) at 5 weeks of age. At 5, 8, 12 and 16 weeks of age whole body composition (by DEXA) and tibial bone mass (by in vivo pQCT) were measured prospectively. At the endpoint (16 weeks of age), mice were sacrificed and measurements were performed of urinary deoxypyridinoline (DPD), serum osteocalcin (oc), femur microCT, gonadal fat and quadriceps weight.

As expected, ARKO mice showed a failure of periosteal bone expansion with less thickening of cortex (−9% and −15% respectively compared to WT, P0.05), and lowering of trabecular density already from week 5 on (−54% compared to WT at week 16, P <0.05). During growth, ARKO mice also increased their weight and lean mass less than WT (−15% and −18% respectively, P <0.05) and had lower quadriceps weight (−43% compared to WT, P <0.05) at 16 weeks.

Running decreased weight in both WT and ARKO by reducing total body fat mass (−29% in WT, −12% in ARKO, P <0.05) and gonadal fat (−54% in WT, −46% in ARKO, P <0.05), but did not alter the effects of AR disruption on muscle mass and cortical bone. In addition, and unexpectedly, running decreased bone turnover in ARKO mice (DPD and oc of ARKO runners respectively −21% and −24% lower than controls, P <0.05) and hereby prevented cancellous bone loss (14% more in ARKO runners then in ARKO controls at 16 weeks). MicroCT of the trabecular bone of the femur revealed that running preserved trabecular number (P <0.05) without affecting thickness. In contrast, bone and muscle characteristics in WT were unaffected by physical exercise suggesting that running had anti-resorptive properties only during high bone remodeling. Contrary to our hypothesis, running did not increase muscle and bone formation in WT and ARKO mice. However, running reduced cancellous bone resorption in ARKO mice only. Sustained exercise may therefore compensate the effects of androgen disruption on cancellous but not cortical bone.

Disclosures: J. Ophqff, None.


Low-Level Mechanical Signals Anabolic to Bone Suppress Adiposity in Growing Mice. C. T. Rubin1, B. Busa*1, E. Capilla*2, C. Rosen3, J. Pessin*2, S. Judex1. 1Biomedical Engineering, State University of New York, Stony Brook, NY, USA. 2Pharmacology, State University of New York, Stony Brook, NY, USA. 3St. Joseph Hospital, Maine Center for Osteoporosis, Bangor, ME, USA

Mesenchyme can differentiate into a number of tissues, including bone, fat, cartilage, and muscle, yet the chemical/physical signals that drive this decision are relatively unknown. Considering the importance of exercise in stemming osteoporosis, diabetes and obesity, it was hypothesized that mechanical signals anabolic to bone would, in parallel, curb the production of fat in the growing animal. Forty C57BL/6J male mice, 7w of age and fed a normal diet, were randomly separated into either a mechanically stimulated (MS) or control (CO) group. For 14w, 5d/w, the MS mice were subject to 15m/d of a 90Hz, 0.2g whole body vibration induced via a vertically oscillating platform. A mechanical vibration at this magnitude and frequency is barely perceptible to human touch. Upon 12w on their respective protocols (19w of age), in vivo micro-CT scans were used to quantify subcutaneous and visceral fat of the torso (n = 12 in each group). At sacrifice (21w of age), weights of the epididymal fat pad, subcutaneous fat pad, liver, and heart were established (all animals).

Following 14w exposure to short-duration, low-level whole body vibration, food intake in MS was 7.9% lower and body mass 6.7% lower as compared to CO (p <0.05). CT measures indicated fat volume in the torso of MS mice were 27.6% lower as compared to CO (p <0.005). CT findings were directly supported by the weights of the dissected fat pads, where MS had 26.7% less epididymal and 23.5% less subcutaneous fat than CO (p <0.01). Normalized to mass, there was 22.5% less epididymal and 19.5% less subcutaneous fat in MS than CO (p <0.01). No differences in bone length, or heart and liver weights, were detected between groups.

These data support a relationship between exogenous, mechanical signals, musculoskeletal plasticity and metabolic regulation, and may help identify a systems-interdependent interaction between the benefits of mechanical loading and the consequences of a sedentary lifestyle. In hindsight, perhaps this should not be so surprising, as the musculoskeletal system is well known to be responsive to mechanical load, and muscle, adipose and bone tissues are derived from the same precursor stem cell population. Critically, our findings indicate that high frequency low amplitude vibration, inducing deformations in bone three orders of magnitude below those which arise during exercise, have marked effects not only on enhancing bone structure but on suppressing adiposity, and may ultimately play a role in the non-pharmacologic control of osteoporosis, diabetes and/or obesity.

Disclosures: C. T. Rubin, Juvent, Inc. 4, 5, 7.

This study received funding from: MIAMS.


Novel Digital Morphometric Approach to Assess Vertebral Fractures in Clinical Practice: A Multicenter Pilot Study. D. Banzer*1, R. Andresen*2. 1Osteoradiologisches Referenz-Zentrum, Evang. Krankenhaus Hubertus, D-14055 Berlin, Germany. 2Abt. für Bildgebende Diagnostik u. Interventionelle Radiologie, KMG Klinikum Güstrow, D-18273 Güstrow, Germany

Objective: To evaluate the practicability of a new digital semi-automatic morphometric method (MorphoXpress® software) for the vertebral fracture deformity assessment under clinical conditions in Germany. Patients and methods: Lateral thoracic and lumbar spine radiographs of 277 unselected patients were analyzed in 7 centers. 3878 vertebrae from TH5 to L4 were evaluated. German guidelines of standardization of radiographic techniques were followed. Qualitative x-ray evaluation was performed using a dichotomous classification of fractured or non fractured vertebra based on a 20 percent threshold value. A new semi-automatic morphometric software (MorphoXpress®) for the assessment of digitized x-rays by six-point morphometry was used for the quantitative assessment of vertebral height. Measurement are based on an active matrix contour model. Heights, height ratios and deformity indices are calculated. Results: Qualitative film evaluation recognized fractures in 109 (39.4%) patients, morphometric fractures were detected in 104 (37.6%) patients. Distribution of fractures was quite similar for both methods, at the thoracic level more fractures were found by the software program than by the radiologists, while the qualitative evaluation identified more fractures at the lumbar level. Using qualitative film evaluation as gold standard the sensitivity of the MorphoXpress® was 73.4% and the specifity 85.7% (kappa 0.6) related to the patients. Related to the single vertebra sensitivity of the software was 61% and specifity 97%. Seventy-six percent of fractured vertebrae were identified correctly by the automatic contour finding model alone. This result was improved to 98% by additional manual corrections. The whole procedure was done in six to seven minutes. Conclusions: Semiautomatic assessment of vertebral deformity using digital active matrix morphometry has a comparable sensitivity to qualitative film evaluation. MorphoXpress® software seems to be a valuable tool to support clinicians in identifying and documenting vertebral deformities and follow-ups under osteoporosis therapy, especially with the increasing budget pressure in German health care system.

Disclosures: D. Banzer, Procter & Gamble Pharmaceuticals - Germany GmbH 5.


A Role for Nitric Oxide in the Mechanical Regulation of RANKL in Bone Stromal Cells. X. Fan1, J. Rahnert*1, N. Case2, T. C. Murphy*1, F. Grassi1, M. S. Nanes1, J. Rubin2. 1Medicine, VAMC & Emory University, Atlanta, GA, USA. 2Medicine, University of North Carolina @ Chapel Hill, Chapel Hill, NC, USA

Mechanical loading is important in the maintenance of bone mass. Mechanical strain inhibits osteoclastogenesis by decreasing stromal/osteoblast cell expression of the Receptor for Activated NF-κB Ligand (RANKL) that is critical for osteoclast recruitment and differentiation. At the same time, strain induces release of nitric oxide (NO) by up-regulating expression of endothelial nitric oxide synthase (eNOS), the predominant isoform expressed in bone. NO can enhance the differentiation and function of osteoprogenitor cells, resulting in increased bone formation, another target effect of strain. Whether mechanically induced NO generated from eNOS is required for the full mechanical inhibition of RANKL, linking up the NO signal transduction pathway to strain induced activation of ERK1/2, is unclear. To address this question we here investigated whether strain would retain the ability to inhibit RANKL expression in stromal cells lacking eNOS by using cells from the eNOS null mouse (eNOS(−/-)). Overnight application of 2% strain to cultures plated in Flex-Cell dishes inhibited RANKL expression in eNOS(−/-) cells, but the inhibitory effect was lower than in wild-type cells: 37.0 ± 2% in the eNOS(−/-) compared with 47.3 ± 4% reduction in the wild type, p < 0.05. Unexpectedly, eNOS(−/-) cells continued to respond to strain with NO production of 281 ± 21% that seen in unstrained controls, a significant increase although significantly lower than strain-induced NO from WT cells (345% ± 27%, p < 0.05). Strain induction of NO from eNOS(−/-) cells was abolished in the presence of the neuronal NOS (nNOS) inhibitor, SMTC, indicating that NO synthesis was via nNOS. Since SMTC did not inhibit NO generation from WT cells this data indicated that, in the absence of eNOS, bone cells compensate by up-regulating expression of nNOS. Further investigation confirmed that strain increased nNOS mRNA levels in eNOS(−/-) but not in WT cells. In eNOS(−/-) cells treated with SMTC, strain still effectively decreased RANKL expression (34% ± 9% reduction compared to unstrained cells). These data indicate that while NO does enhance the effect of strain-activated ERK1/2 to inhibit RANKL, NO is not essential to this mechanism. Nitric oxide, and its induction by strain, however, certainly have importance in bone cell physiology, as cellular nNOS adapts to fulfill the role of the mechanically responsive NOS when eNOS is absent.

Disclosures: X. Fan, None.


Microscopic Imaging of Bone Composition and Spatial Correlation with Mechanical Strength. L. M. Miller1, W. Little*2, A. Schirmer*1, T. Feldman*3, S. Judex2. 1National Synchrotron Light Source, Brookhaven National Laboratory, Upton, NY, USA. 2Biomedical Engineering, Stony Brook University, Stony Brook, NY, USA. 3Materials Science and Engineering, Stony Brook University, Stony Brook, NY, USA

Bone diseases such as osteoporosis involve changes in bone composition and altered mechanical strength. Chemical and mechanical properties are frequently studied on the microscopic scale using Fourier transform infrared imaging (FTIRI) and nanoindentation (NI), respectively. In order to directly correlate chemical and mechanical properties, these two techniques should be performed on the same sample. Typically, however, NI is performed on thick, polished bone blocks, whereas FTIRI data are collected from microtomed thin sections. Here, we demonstrate a new method for collecting FTIRI data in a reflection geometry, such that NI and FTIRI can be performed on the same sample. This method was used to spatially correlate the chemical makeup and mechanical strength of growing mouse bone. Tibiae of female BALB mice were harvested at 8 time points (n = 4 each) distributed between 1d and 40d of age. Tibiae of 15mo old mice served as fully mineralized control specimens. FTIRI results showed that bone mineralization proceeded very rapidly at an early age; even at 1 day of age, the degree of mineralization (phosphate/ protein ratio) and mineral crystallinity had reached 36% and 87% of the adult (15mo) values, respectively. From 1–40 days of age, the mineralization level increased almost linearly by 26%, and mineral crystallinity increased by 7%. Spatially, the variability in mineralization across the mid-diaphysis was high for the early time points and declined over time. In contrast to the notable changes in mineralization, carbonate substitution into the mineral lattice and collagen cross-linking did not show any significant changes over this time period. Even though significant mineralization had occurred by 1 day of age, the elastic modulus of 1 day old bone was only 15% of the adult value and increased to 89% (of its adult value) after 40d. Between samples of different time points, significant positive correlations were observed between the elastic modulus and phosphate/protein ratio (R2 = 0.58) and crystallinity (R2 = 0.22), whereas collagen cross-linking showed a small but significant negative correlation (R2 = 0.21). These data illustrate the validity of reflection FTIRI and indicate that specific (combinations of) chemical properties modulate bone's stiffness during growth and that the initial mineral formation lags considerably behind the intrinsic stiffening of the bone, emphasizing the importance of mineral quality.

Disclosures: L.M. Miller, None.


Role of CRTAP in Prolyl 3-Hydroxylation and Recessive Osteogenesis Imperfecta. R. Morello1, T. K. Bertin*1, J. Hicks*2, P. Castagnola*3, F. H. Glorieux*4, H. Bachinger*5, P. H. Byers*6, D. R. Eyre*6, B. F. Boyce*7, B. H. Lee8. 1Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA. 2Texas Children's Hospital, Houston, TX, USA. 3Istituto Nazionale per la Ricerca sul Cancro, Genoa, Italy. 4Genetics Unit, Shriners Hospital for Children, Montreal, PQ, Canada. 5Shriners Hospital for Children, Portland, OR, USA. 6University of Washington, Seattle, WA, USA. 7University of Rochester Medical Center, Rochester, NY, USA. 8Molecular and Human Genetics, Baylor College of Medicine/Howard Hughes Medical Institute, Houston, TX, USA

Prolyl 3-hydroxylation occurs at only one position in the triple helical domain of fibrillar collagens however, unlike prolyl 4-hydroxylation, its functional, structural and biological consequences are unknown. We previously identified a novel protein, CRTAP that was differentially expressed during chondrogenesis. Crtap null mice show an osteochondrodysplasia characterized by rhizomelia, kyphosis, and severe osteopenia. The in vivo histomorphometric analyses showed significantly decreased bone mass and decreased bone formation rate, but normal numbers of osteoblasts and osteoclasts. Interestingly, the mice have very little osteoid and a decreased mineralization lag time. Together these data show that the low bone mass is due to decreased osteoid formation, but with an accelerated mineralization of available osteoid. We demonstrate that CRTAP exists in a protein complex with the recently identified prolyl-3-hydroxylase-1 protein. At a molecular level, Crtap−/− mice lack fibrillar collagen prolyl 3-hydroxylation by tandem mass spectrometric peptide mapping of cyanogen derived peptides. On EM, collagen fibrils appear thickened suggesting a defect in fibrillogenesis. Finally, we identified CRTAP mutations in two families with recessively inherited forms of OI. In the original family with OI type VII, genome-wide mapping had determined a less than 5cM critical region that included the CRTAP gene. We detected a C>G transversion in intron 1 that created a new splice donor site, activating a cryptic exon that appeared in most transcripts from both alleles. In the second, consanguineous family with recurrence in two sibships of a very severe form of OI (OI type II or OI type III), we identified a single nucleotide deletion in exon 4 that led to a premature termination codon, and virtually complete lack of mRNA and protein. Our data demonstrate the biological importance of prolyl 3-hydroxylation, a poorly understood collagen post-translational modification, for bone formation, collagen fibrillogenesis and osteoid synthesis, and that loss of function of CRTAP causes recessive OI.

Disclosures: R. Morello, None.


TGFβ1 Mutations in CED Affects Osteoprogenitor Migration: TGF-β1 as a Coupling Factor in Bone Remodeling. X. Wu*, Y. Tang*, C. Yang*, X. Peng*, M. Wan, X. Cao. Pathology, UAB, Brimingham, AL, USA

TGFβ1, the most abundant cytokine deposited in bone matrix, plays a critical role in bone remodeling, but its exact function in the bone milieu is still unclear. Recent studies of Camurati-Engelmann Disease (CED), an inherited sclerosing bone with remodeling defects, revealed that all 100 patients in 24 families carry TGFβ1 mutations. TGFβ1 mutations causing CED provides an ideal model to study the potential function of TGFβ1 in bone remodeling. Here we examined the effects of six different CED-derived TGFβ1 mutations on the activity of human primary marrow osteoprogenitors (HBMO). We found that all TGFβ1 mutants exhibit much more potent effects on migration of HBMO than the wild type TGFβ1 does. To assess whether TGFβ1 is the primary factor released from bone resorption that induces HBMO migration, we examined the effect of TGFβ1 in the osteoclastic bone resorption medium on HBMO migration in a type I collagen-coated transwell assay. Osteoclastic bone resorption medium significantly stimulates HBMO migration, whereas conditioned media without bone slices has no effect. Importantly, addition of an antibody against TGFβ1 completely inhibited the migration. Furthermore, depletion of TGFβ1 by repeating immunoprecipitation with the anti-TGFβ1 antibody inhibits conditioned medium-induced HBMO migration and reconstitution with recombinant TGFβ1 restores HBMO migration.

TGFβ1 is synthesized as a large precursor that is then cleaved into mature TGFβ1 and the latency associated protein (LAP). LAP remains non-covalently linked to mature TGFβ1, rendering it inactive. Interestingly, all of the CED mutations are at LAP region, not at mature TGFβ1. Our western blot revealed that only conditioned medium of osteoclast with bone slices contains active TGFβ1, suggesting that TGFβ1 is released and activated from bone matrix. Moreover, six TGFβ1 CED mutants were expressed individually in 293T cells. All mutants are secreted largely in active form and are hyperactive for HBMO migration, whereas wild type is primarily in latent form. These results indicate that LAP mutations of CED cause premature release of active TGFβ1. We have also generated both wild type and CED mutant TGFβ1 transgenic mice driven by 2.3 kb Col I promoter. The sizes of CED mutant mice are only one third of wild type and the bone phenotype of the mice are being analyzed. Taken together, TGFβ1 released during bone resorption induces osteoprogenitor migration. CED-derived mutations cause premature release of active TGFβ1 and therefore disrupt TGFβ1 deposition in the bone matrix and gradient formation during bone resorption which is necessary for coupling oeteoprogenitor migration.

Disclosures: X. Wu, None.


Null Mutation in TG-interacting Factor (TGIF) Causes Abnormalities in Skeletal Development and Low Bone Mass. K. S. Mohammad1, L. Bartholin*2, T. Melhuish*2, C. R. Mckenna*1, L. J. Suva3, A. Carver*3, D. Wotton*2, T. A. Guise1. 1Internal Medicine, University of Virginia, Charlottesville, VA, USA. 2Biochemistry and Molecular genetics, University of Virginia, Charlottesville, VA, USA. 3Orthopaedic surgery, University Arkansas for Medical Sciences, Little Rock, AR, USA

TGIF is a corepressor of transcriptional responses to both TGFβ and retinoic acid. Inactivating mutations in TGIF are associated with holoprosencephaly, a severe defect in craniofacial development. We generated TGIF null mice and studied the developmental and skeletal phenotypes. In mixed strain background mice, the TGIF-null mutation did not induce developmental defects. However, in a C57BL/6J (>98.4%) background, loss of TGIF caused severe abnormalities: 45% were dead at weaning, 40% of viable mice at weaning had hydrocephalus, 15% at E.18.5 had exencephaly and edema. In addition, TGIF−/− mice were growth retarded, had domed skulls, twisted noses and developmental defects in vertebrae, ribs and sternum. Loss of TGIF repressor function may enhance responses to TGFβ and retinoic acid, both of which can adversely affect bone. Therefore, we studied bone remodeling in TGIF−/− and wild-type littermate (WT) controls in the C57BL/6J background. Bone mineral density (BMD) of 10 week old mice, was significantly reduced in TGIF−/− mice at total body (p = 0.014), lumbar spine (p = 0.049), proximal tibia (p = 0.0018) and distal femur (p = 0.0005) compared with WT. Histomorphometry showed a significant reduction in trabecular bone volume (BV/TV%) at the tibia, femur and humerus in TGIF−/− mice compared to WT (17.6 ± 1.5% vs 32.0 ± 2.3%; p <0.0001), associated with increased osteoclast numbers (OC/mm-BS: 6.32 ± 0.43 vs 3.9 ± 0.36; p = 0.0002) and reduced osteoblast numbers (OB/mm-BS: 14.31 ± 1.04 vs 21.08 ± 1.49; p = 0.0005) respectively in TGIF−/− mice vs WT. Bone formation rate (BFR) at the femur was markedly reduced in 14 week old TGIF−/− mice compared to WT (0.50 ± 0.06 vs 0.89 ± 0.13; p = 0.027) as was mineral apposition rate (MAR) (1.63 ± 0.20 vs 2.53 ± 0.23; p = 0.015). In addition, evaluation of the skeletal microarchitechture by micro-CT showed a significant reduction in BV/TV%, connectivity-density, DT-Tb, N with increase in DT-Tb.Sp and TRI-SMI ratio in 18 week old TGIF−/− mice compared to WT mice. In summary, mice lacking TGIF showed growth retardation, developmental skeletal abnormalities and reduced bone mass associated with increased bone resorption, decreased bone formation and poor microarchitecture. Taken together, the data indicate that TGIF has an important role in skeletal development and bone remodeling. These effects may be due to enhanced signaling of TGFβ and/or retinoic acid caused by loss of TGIF repression on these pathways.

Disclosures: K.S. Mohammad, None.


Mutations in the BMP Type I Receptor ACVR1 in Patients with Fibrodysplasia Ossificans Progressiva (FOP). E. M. Shore1, M. Xu*1, J. M. Connor*2, F. S. Kaplan1. 1Orthopaedics, University of Pennsylvania, Philadelphia, PA, USA. 2University of Glasgow School of Medicine, Glasgow, Scotland, United Kingdom

We recently reported mapping the gene mutation for fibrodysplasia ossificans progressiva (FOP) to chromosome 2q23–24 and identified mutations in ACVR1, a gene that encodes a bone morphogenetic protein (BMP) type I receptor. The identical heterozygous mutation, c.617G>A (CGC>CAC; Arg206His), was found in ail affected individuals from seven small multigenerational families and all sporadically-affected individuals examined. This initial analysis focused on patients (familial and sporadic cases) with classic features of FOP (progressive heterotopic ossifcation and great toe malformations). A total of 60 sporadic patients with classic features of the disease have now been examined by DNA sequence analysis and, consistent with our initial data set, all show a heterozygous ACVR1 c.617G>A mutation. This mutation occurs within the GS activation domain of ACVR1, and protein structural homology modeling predicts an altered configuration affecting downstream signaling by the mutant receptor.

DNA sequence analysis of ACVR1 protein-coding exons for 11 sporadic FOP patients with one or more atypical features revealed that all 11 patients have heterozygous mutations within ACVR1. Six of 11 have mutations in the GS domain: five carry the c.617G>A mutation in codon 206 that was identified in patients with classic FOP features, and one patient (classic features plus severe growth retardation) has a unique mutation in codon 207. The remaining five patients carry mutations within the protein kinase domain of ACVR1. Four of these patients have mutations in the same codon and are distinguished by severe reduction defects of the hands and feet. The fifth patient has a mutation in a more distal codon and exhibits classic FOP features with severe hearing impairment. Similarly to ACVR1 c.617G>A, these newly identified and rare mutations occur in functional regions of the protein that are expected to affect downstream signaling by the ACVR1 receptor. Furthermore, while all identified ACVR1 mutations are associated with progressive formation of heterotopic bone, alterations in the protein kinase domain are highly correlated with more severe skeletal malformations during late stages of embryonic development.

Disclosures: E.M. Shore, None.


Impaired Interaction of Lrp6 Ringelschwanz Mutant Protein with Mesd Is Involved in Disrupted Wnt Signaling. T. Kubota*1, T. Michigami2, N. Sakaguchi*2, C. Kokubu*3, A. Suzuki*2, N. Sakai*1, S. Nakajima1, K. Imai4, K. Ozono1. 1Department of Pediatrics, Osaka University, Osaka, Japan. 2Department of Environmental Medicine, Osaka Medical Center and Research Institute for Maternal and Child Health, Osaka, Japan. 3Center for Advanced Science and Innovation, Osaka University, Osaka, Japan. 4Department of Molecular Life Science, Tokai University, Kanagawa, Japan.

Low-density lipoprotein receptor-related protein 5 and 6 (LRP5 and LRP6) function as co-receptors for Wnt ligands. Loss- and gain-of-function mutations in LRP5 gene result in low and high bone mass, respectively. Moreover, we previously identified a point mutation in Lrp6 leading to an amino acid substitution of Trp for Arg at amino acid 886 (R886W) prior to the 3rd EGF-like repeat in a spontaneous mutation ringelschwanz (rs) in the mouse. Homozygous rs/rs mice exhibited low bone mass resulting from increased bone resorption. Although these observations indicate that abnormal Wnt/beta-catenin signaling resulted from mutations in LRP5/6 affect on bone mass, the precise mechanisms by which each mutation fails to activate Wnt/beta-catenin signaling pathway have not been fully elucidated. In the current study, we performed functional analyses to determine how the rs mutation in Lrp6 affects Wnt signaling, utilizing the expression plasmids encoding wild-type (mLrp6[WT]) and the R886W mutant Lrp6 (mLrp6[rs]). First, we examined the effects of the rs mutation on T-cell Factor (TCF)-dependent transactivation using marrow-derived stromal cell line ST2. When introduced into ST2 cells, mLrp6[WT] markedly increased the transactivation of TCF reporter gene, while mLrp6[rs] failed in the transactivation despite the equivalent expression levels. The results indicate that mLrp6[rs] is a loss-of-function mutation and cannot transduce the canonical Wnt signaling appropriately. Therefore, we next examined the interactions of mLrp6[rs] protein with Wnt1, Dickkopf1 (Dkk1) that is a canonical Wnt signaling inhibitor, or Mesd (mesoderm development). Both mLrp6[WT] and mLrp6[rs] bound to Wnt1 in a similar affinity. In addition, there was no obvious difference between mLrp6[WT] and mLrp6[rs] in the affinity to Dkk1. Interestingly, regarding the interaction with Mesd, mLrp6[rs] exhibited reduced affinity compared with mLrp6[WT]. Since Mesd functions as a molecular chaperone for Lrp6 and is required for its targeting to the plasma membrane, it is very likely that the reduced affinity to Mesd leads to impaired localization of mLrp6[rs] at cell surface. Taken together, our results suggested that rs mutation in Lrp6 associated with low bone mass phenotype lost the ability to facilitate Wnt signaling due to the impaired interaction with Mesd.

Disclosures: T. Kubota, None.


The Dynamics of Alox5 and Pparγ Interactions: Delineation of “Regulon” on Mouse Chr 6 that Influences Peak Bone Acquisition. C. L. Ackert-Bicknell1, V. E. DeMambro1, J. Graber*1, B. Lecka-Czernik2, E. Canalis3, M. L. Bouxsein4, K. Shultz1, M. Horowitz5, W. G. Beamer1, C. J. Rosen1. 1The Jackson Laboratory, Bar Harbor, ME, USA. 2University of Arkansas, Little Rock, AR, USA. 3Saint Francis Hospital and Medical Center, Hartford, CT, USA. 4Beth Israel Deaconess Medical Center, Boston, MA, USA. 5Yale University, New Haven, CT, USA

Pparγ and Alox5, two genes in the osteoblast (OB) and adipocyte (Ad)differentiation pathways, are next to each other on mouse Chr 6. The product of Alox5, 5-LO, generates LTB4, an inflammatory mediator and ligand for Pparγ. The B6.C3H-6T (6T) congenic mouse was generated by backcrossing a 30 cM region of Chr 6 from C3H/HeJ (C3H) onto C57BL/6J (B6) and is C3H-like for both Pparγ and Alox5. 6T mice have low BMD and increased marrow Ads. We identified 2 polymorphisms in the C3H Pparγ2 promoter and transfection data suggest these negatively influence expression. We also found that 6T mice have decreased Pparγ and Alox5 transcripts in the femur. Yet in microarray studies, pro-adipocytic genes such as SPOT14, Srebp1, and CD36 are up-regulated in 6T bone. By contrast, Pparγ+/− knockout mice, have increased BV/TV and reduced Ads in the bone marrow. Since Pparγ2 protein production is also regulated post-transcriptionally, and its activity is ligand-dependent, we hypothesized that Pparγ activity is enhanced in 6T, but this leads to down regulation of its own transcripts. We examined 6T whole bone marrow nuclear extracts and found more Pparγ protein binding to the PPRE from the Ap2 gene vs B6 (p = 0.009). Similar differences were noted for the PEPCK and Malic Acid PPRE. We then sequenced 3′ to Alox5, and found a non-coding, highly conserved(∼99%) 1500 nt region that could influence the expression of distant genes by acting as a ‘regulon’. Our mapping data also suggested that Pparγ and Alox5 were split from this regulon by a 25cM paracentric inversion. To test the influence of this regulon, we made a B.H-6 congenic mouse from a recombinant inbred line (BxH) that is genetically like 6T; i.e. it has C3H alleles in the Chr 6 QTL, but does not contain the inversion. But unlike 6T, B.H-6 mice resemble Pparγ+/− with high cortical and trabecular bone mass(p <0.05). In summary: mice carrying C3H alleles for Pparγ and Alox 5 are similar to Pparγ +/− with increased trabecular BV/TV due to reduced Pparγ mRNA; 6T mice have increased marrow ADs and reduced BMD due to increased Pparγ binding activity; the inversion in 6T occurs in the 3′UTR region of Alox 5 and results in discordant regulation of two neighboring genes. The presence of a ‘regulon’ in the 6T QTL confirms the importance of genomic proximity and its relationship to function, while providing an in vivo platform for testing new hypotheses regarding OB and AD differentiation.

Disclosures: C.L. Ackert-Bicknell, None.

This study received funding from: NIH AR 45433.


Identification of Duffy Antigen/Receptor for Chemokine (DARC) as a Chromosome 1 BMD QTL Gene in Mice. B. Edderkaoui1, D. J. Baylink1, K. Schultz2, W. G. Beamer*2, J. E. Wergedal1, R. Porte*1, A. Chaudhuri*3, S. Mohan1. 1JLPVAMC, Loma Linda, CA, USA. 2Jackson Labs, Bar Harbor, ME, USA. 3NY Blood Ctr, NY, NY, USA

To identify the BMD QTL gene, we focused on the Chr1 QTL, as it has been identified in several different inbred strain mouse crosses and since a syntenic region in human Chr1 contains BMD QTL. We have evidence that Chr1 QTL contains four BMD loci (BMD1–1, 1–2, 1–3, & 1–4), of which BMD 1–2 QTL exhibited the highest LOD score. To narrow down the size of the BMD1–2 QTL region to permit candidate gene search, we generated additional congenic sublines C168–179 and C168–172, which carried the CAST chromosomal region underlying the BMD1–2 locus and between the BMD1–1 and 1–2 loci. Significantly higher (8.6%, P <0.001) femur vBMD was found in subline C168–179 compared with the C57BL/6J mice, while C168–172 did not show any difference with B6 progenitors (P = 0.78). After superimposing the CAST chromosomal regions carried by all of the subcongenic lines, we have narrowed down the size of the BMD1–2 locus to 172–175 Mb in Chr1. In order to screen for potential BMD candidate genes, we compared the expression levels of 48 genes and ESTs located within the BMD1–2 region between the bones of B6 control mice and the two subcongenic lines and found only the Duffy (Dfy/DARC) gene, located at 173.26 Mb, showed a 6-fold higher expression in C168–179 subline (P <0.001), but not in the C168–172 subline, compared to the B6 control. We next sequenced the Dfy gene to determine the SNPs that might be the cause of the difference in Dfy expression between the B6 and the high BMD subcongenic mice. Of the 28 SNPs that distinguished the Dfy gene B6 from the CAST strain, 6 SNPs found in the coding region led to amino acid changes. The involvement of the Dfy protein in BMD variation was confirmed by Dfy-KO mice in a B6 background that exhibited a 5% increase (P <0.001) in femur vBMD compared to B6 mice. Because Dfy binds chemokines that regulate osteoclastogenesis, we tested the possibility that Dfy is involved in regulating formation of osteoclasts (Oc1). Treatment of non-adherent bone marrow cells (NABMC) derived from Dfy KO mice with RANKL and MCSF produced 50% less TRAP-positive Oc1 in vitro compared to NABMC from wild type mice. Also, despite the increase of Dfy gene expression in C168–179, this subline showed a 70% decrease of Oc1 compared to B6 mice. To determine if decreased Oc1 formation in C168–179 line is due to reduced activity of Dfy protein caused by SNP-induced amino acid substitutions, we determined chemokine binding to NABMC and found 50% lower binding of 125I-MCP-1 to NAMBC derived from congenic mice compared to B6 mice. Based on our findings, we conclude that the Dfy gene is a BMD QTL gene that regulates Oc1 formation, bone resorption, and thereby BMD.

Disclosures: B. Edderkaoui, None.


Genetic Effects Are Associated with Increased Risk of both Osteoporosis and Obesity: A Study of RANK Gene. L. Zhao*1, D. Xiong1, Y. Guo*2, R. Recker1, H. Deng3. 1Osteoporosis Research Center, Creighton University Medical Center, Omaha, NE, USA. 2Hunan Normal University, Changsha, China. 3University of Missouri Kansas City, kansas City, MO, USA.

Background: A widely held belief is that increasing body weight (and thus higher risk of obesity) is associated with lower risk of osteoporosis. However, we previously found in quantitative genetic analyses that increasing fat mass (thus higher risk to obesity) is actually associated with decreased bone mass (thus higher risk of osteoporosis) when the latter is adjusted for body weight. RANK is a gene essential for osteoclast formation and is expressed in skeletal muscles which are import for energy metabolism.

Hypothesis: RANK is a pleiotropic gene for both osteoporosis and obesity and genetic variations in this gene may reduce/increase the risk of obesity and osteoporosis in the same direction.

Methods: We first performed a linkage analysis in a genomic region of ∼80 cM around the RANK gene in 4,087 Caucasian subjects from 482 pedigrees using 9 microsatellite markers. Then we genotyped 19 SNPs in or around the RANK gene in 405 nuclear families with 1,873 individuals. A family-based association test (FBAT) was performed for the traits of obesity (BMI, fat mass, lean mass and OB (defined as BMI>30 kg/m2)) and osteoporosis (OP, defined by lowest 10th percentile of bone mineral density (BMD) Z-score or by BMD T-score less than −2.5). Empirical global P values were derived from 10,000 permutations and reported as follows. Results: In the linkage analysis, an empirical P = .004 was achieved at the location of the RANK gene for BMI. In the family based association analysis, the allele T of marker SNP7 (rs11664594) is associated with higher fat mass (P = .008), higher lean mass (P = .003), higher BMI (P = .002), and higher risk of OP at spine (P = .0002) and hip (P = .0001). In addition, the allele G of marker SNP10 (rs4303637) is overtransmitted to both OB and OP offspring. Furthermore, haplotype GGAT in haplotype block 5, containing SNP10, is associated with higher OB (P = .001), higher risk of OP at spine (P = .004) and hip (P = .005).

Conclusion: The RANK gene is strongly associated with both osteoporosis and obesity. Our study provides the first evidence with a specific gene that the genetic variations may increase the risk of obesity and osteoporosis in the SAME direction. This result echoes our earlier quantitative genetic analyses and is contrary to the common belief widely held in the field.

Disclosures: L. Zhao, None.


Size-Corrected BMD from Adolescence to Young Adulthood: A Longitudinal Study. R. A. Faulkner, K. S. Davison*, D. A. Bailey, L. Sherar*, R. L. Mirwald*, A. P. G. Baxter-Jones*. Kinesiology, University of Saskatchewan, Saskatoon, SK, Canada

Existing data on bone development is limited because most studies have been cross-sectional, have failed to control for maturational differences in growing children, and/or have not adequately adjusted DXA derived BMD values for changing size across the growth period. The purpose of this study was to assess size-corrected BMD (BMDsc) in 42 males and 46 females who were measured longitudinally from adolescence to young adulthood. Bone mineral was measured by DXA (Hologic 2000, array mode). Peak height velocity (PHV) was first determined by fitting a cubic spline curve to the whole year height velocities for each individual subject. We then aligned each individual's bone measurements around years from PHV and fitted a sigmoidal curve to this data. BMC and BA data for each individual were then interpolated from this curve and aligned on maturational age groups as: −2, −1, 0, +1, +2, +3, and +4 years from PHV. Young adult status was defined as chronological age of 20 (which represented an average maturational age of +8 yrs from PHV for girls and +6 yrs for boys). Size-corrected BMD (BMDsc) was calculated by regressing the natural logarithms of BMC and BA (males and females separately) and then applying the power coefficient from this analysis (lambda) as: BMDsc = BMC/BAlambda[1]: For males, BMD(sc) = BMC/BA1.555; and for females, BMD(sc) = BMC/BA1.459. Previous studies have reported areal BMD to increase steadily over the growing years. However, as shown in the following figure, we found a decrease in BMD(sc) prior to age of peak linear growth (PHV), and then BMD(sc) rebounded and increased steadily until age 20. These results support the theory, that during the period of rapid growth around the age of PHV there is an associated draw on cortical bone to meet the mineral demands of the expanding skeleton [2].

  • Prentice A, Parsons TJ, Cole TJ 1994 Uncritical use of bone-mineral density in absorptiometry may lead to size-related artifacts in the identification of bone-mineral determinants. Am J Clin Nutr 60:837–842

  • Parfitt AM 1994 The 2 Faces of growth - benefits and risks to bone integrity. Osteoporos Int 4:382–398.

Thumbnail image of

Disclosures: R.A. Faulkner, None.

This study received funding from: Canadian Institutes of Health Research.


Childhood Body Mass Index Predicts Young Adult Cortical Bone Size. J. M. Kindblom, M. Lorentzon, E. Norjavaara*, A. Hellqvist*, S. Nilsson*, D. Mellstrom, C. Ohlsson. Center for Bone Research, Internal Medicine, Sahlgrenska Academy, Gothenburg, Sweden

Cortical bone geometry, such as cortical cross sectional area and outer (periosteal) and inner (endosteal) circumferences, has recently been shown to be of importance for the mechanical strength of the bone. The aim with the current study was to investigate the impact of childhood body mass index (BMI; kg/m2) for young adult bone mineral density (BMD) and cortical geometry.

Detailed growth charts, including height and weight, from birth up to 18–20 years of age were available for 668 men participating in the Gothenburg Osteoporosis and Obesity Determinants (GOOD) study. Young adult trabecular and cortical bone characteristics were analyzed using peripheral quantitative computerized tomography (pQCT) in the distal metaphyseal area (trabecular vBMD) and in the diaphyseal region (cortical analyses) of the radius and tibia. Childhood BMI at ages 3 and 7 were estimated through complex curve fitting and associated with young adult bone characteristics. Age at bone analysis, BMI at bone analysis, smoking, calcium intake and present physical activity were included as covariates in the multivariate analyses.

Linear regression analysis revealed that BMI at 3 years of age independently predicts cortical cross-sectional area (cortical CSA; β = 0.094, p = 0.016), periosteal (β = 0.19, p <0.001), and endosteal (β = 0.19, p <0.001) circumference of the tibia. Similar results were obtained for the radius. The impact of childhood BMI on these variables was even larger at 7 years of age (CSA β = 0.13, p = 0.004; periosteal circumference β = 0.22, p <0.001; endosteal circumference β = 0.21, p <0.001 for the tibia). Moreover, BMI at 3 years of age explained 8.8% and BMI at 7 years of age explains 13.6% of the variance in cortical periosteal circumference in young adult men. Cortical vBMD was not predicted by childhood BMI. Adult Trabecular vBMD in the radius, but not in the tibia, (β −0.13, p = 0.003) was predicted by BMI at 3 years.

Thus, childhood BMI predicts cortical CSA, periosteal and endosteal circumference in both the tibia and the radius. These findings indicate that low childhood BMI might be a risk factor for reduced mechanical strength and may therefore confer increased risk of fractures later in life.

Disclosures: J.M. Kindblom, None.


TGFβ Inducible Early Gene-1 Null Mice Display a Gender Specific Osteopenic Phenotype. M. Subramaniam1, J. R. Hawse1, K. D. Peters*1, D. G. Eraser*2, M. J. Oursler2, T. C. Spelsberg1. 1Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN, USA. 2Endocrine Research Unit, Mayo Clinic, Rochester, MN, USA

TGFβ inducible early gene-1 (TIEG) is a member of the kruppel-like family of transcription factors that plays an important role in TGFβ mediated Smad signaling. Previously, we have shown that TIEG overexpression in human osteoblasts mimics TGFβ action by regulating osteoblast specific marker genes. Further, we have demonstrated that estrogen, but not androgens, regulate TIEG expression in osteoblasts. To elucidate the function of TIEG in skeletal development, we have generated a TIEG-null mouse. Initial characterization of the skeletons of these mice by mechanical 3-point bending tests revealed that the femurs of TIEG-null mice are weaker when compared to wild-type littermates. Further characterization of the femurs and tibias of TIEG-null mice have revealed gender and age specific influences on this bone phenotype. Interestingly, PIXImus and pQCT analysis of 2 month old female TIEG-null mice reveal significant decreases in nearly all of the bone parameters tested to date relative to wild-type controls (Table 1). Intriguingly, no difference was observed for any of these parameters in male mice (Table 1). As the female TIEG-null mice age, these difference gradually disappear. Taken together, these data indicate that deletion of TIEG expression in mice results in an osteopenic phenotype only in females suggesting that estrogen might play a key regulatory role in this phenomenon. The fact that this phenotype is lost in old individuals could be explained either by compensation for loss of TIEG expression by related genes, or by an increased rate of bone loss in the wild-type animals with age. 

Table  . 2-month tibia and femur measurements by pQCT and PIXImus
inline image

Disclosures: M. Subramaniam, None.

This study received funding from: NIH.


Sexual Dimorphism Affects Bone Size and Shape but not Tissue-Level Mechanical Properties. S. M. Tommasini1, P. Nasser*2, K. J. Jepsen2. 1CUNY Graduate School, New York, NY, USA. 2Mt Sinai School of Medicine, New York, NY, USA

Understanding why females show a greater incidence of stress fractures early in life and fragility fractures later in life compared to males may provide important insight into the biological and mechanical mechanisms underlying fracture risk [1]. Skeletal dimorphism is one factor that may contribute to this discrepancy [2], because bone size and shape are known determinants of bone strength [3]. Fracture also depends on tissue-level mechanical properties. However, it is unclear whether sex-specific growth patterns affect the construction of bone matrix in a way that leads to differences in tissue-level mechanical properties. We tested whether females acquire similar tissue-level mechanical properties as males by the time peak bone properties are established. We also tested whether females show a correlation between cross-sectional morphology and tissue-level mechanical properties similar to that observed for males [4]. Tibiae from 17 male (age 17–46 yrs) and 14 female donors (age 22–46 yrs) were measured for bone geometry [length, cortical area (CtAr), AP and ML width, moments of inertia (IAP, IML, J)]. A slenderness index was defined as the inverse ratio of the section modulus to tibia length and body weight [5]: S = 1/[(J/width)/(L*BW)]. The diaphyses were cut into rectangular beams and, using 4-point bending, tissue-level stiffness and strength were quantified. Tissue-damageability, toughness, and ductility, were also quantified because these mechanical properties provide insight into the material response of bone when subjected to extreme load conditions such as during a fall, as well as the amount of damage accumulated within the bone when subjected to intense physical activity. Differences between females and males were determined using a Student's t-test. To determine if females and males show a similar relationship between tibial cross-sectional morphology and tissue-level mechanical properties, linear regressions were performed and differences in slopes and intercepts between female and male regressions were determined using a method equivalent to an ANCOVA. Compared to males, female tibiae were smaller (J, CtAr, width), more slender (SAP, SML, J/L), and showed a smaller J:A ratio. However, females and males showed nearly identical tissue-level mechanical properties that degraded similarly with age and correlated similarly with cross-sectional morphology. These results suggest that sex-specific growth patterns affect transverse bone size, but not tissue-level mechanical properties. [1] Winner, 1989 BMJ 298 [2] Duan, 2003 JBMR 18 [3] van der Meulen, 2001 Bone 29 [4] Tommasini, 2005 JBMR 20 [5] Selker, 1989 J Biomech 22 [6] Jepsen, 1997 J Biomech 30.

Disclosures: S.M. Tommasini, None.


Decreases in Lean Mass Are Strongly Associated with the Rate of Loss in Hip Bone Mineral Density with Aging among Afro-Caribbean Men. Y. Sheu1, J. A. Cauley*1, V. W. Wheeler*2, C. H. Bunker*1, A. L. Patrick*2, J. I. Oakley*1, J. M. Zmuda1. 1Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA. 2The Tobago Health Studies Office, Scarborough, Trinidad and Tobago

Weight loss with aging is associated with an increased rate of decline of hip bone mineral density (BMD) in men; however, less is known about the impact of body composition changes on BMD among men, particularly among those of African descent. Thus, in the present study we examined the association between changes in body composition and the rate of decline in hip BMD in an ongoing longitudinal study of Afro-Caribbean men.

A total of 2,689 men aged 40 and older were first recruited regardless of their health status for a population-based prostate cancer screening survey. BMD at the proximal femur and body composition (whole body fat and lean mass) were measured at study entry and after an average of 4.3 years with dual-energy X-ray absorptiometry (Hologic QDR-4500). A total of 1158 men (mean age, 56.1 ± 10.4; range, 40–88yrs) with complete scan data were included in the current analysis. We compared the annualized rate of change in BMD across 3 categories of change in body fat mass and lean mass using analysis of covariance. Men with a 5% or more loss of lean body mass had a substantially and significantly greater decline in femoral neck BMD compared to men with stable or increased lean mass after adjustment for initial age, BMD, fat mass, lean mass and height and change in fat mass (Table). Similar results were observed for the rate of decline in BMD at the total hip, trochanter and intertrochanter. On the other hand, changes in fat mass were not significantly associated with changes in hip BMD (Table).

Declines in lean body mass are an important contributor to bone loss with aging among men of African descent.

Table Table:. Rate of Change (%/yr) in Femoral Neck BMD by Changes in Body Composition
  1. *Values are adjusted mean (Standard Deviation)

inline image

Disclosures: Y. Sheu, None.


BMP Signaling through Type 1a Receptor (BMPR1a/ALK3) Regulates Trabecular and Cortical Bone Properties in an Age, Sex, and Skeletal Site Dependent Manner In Vivo. N. Kamiya1, T. Ward*1, L. Ye2, T. Kobayashi3, P. Atsawasuwan*4, D. M. Joiner*5, E. I. Waldorff*5, H. M. KRONENBERG3, M. Yamauchi4, S. Goldstein5, J. O. Feng2, Y. Mishina*1. 1LRDT, NIEHS/NIH, Research Triangle Park, NC, USA. 2Oral Biology, University of Missouri-Kansas City, School of Dentistry, Kansas City, MO, USA. 3Endocrine Unit, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA. 4Dental Research Center, University of North Carolina, Chapel Hill, NC, USA. 5Orthopaedic Surgery, The University of Michigan, Ann Arbor, MI, USA

In vivo function of BMP signaling during bone development and remodeling remains unknown due to the embryonic lethality of mice from conventional knockout mice of BMP ligands/receptors. We rescued the lethality of BMPR1a using a tamoxifen (TM) inducible Cre-loxP system with 3.2 kb type I collagen promoter, which realize osteoblasts/osteocytes specific Cre activity. When conditional knockout (cKO) of BMPR1a was induced for newborn stages by injecting TM into nursing mother until weaning, P25 cKO pups showed more trabecular bone volume but less bone mineral density compared to wild type (WT) littermates without a gender difference. When cKO was induced for adult stages (6 wks to 20 wks old), the aged cKO mice (34 wks old) showed hunchback phenotype with obviously hyper-calcified rib cage and spine measured by x-ray. For cKO femur, micro CT of trabecular bone areas revealed significantly higher bone volume fraction (trabecular BFR: cKO 36.2%, WT 23.9%) and tissue density. For cKO spine, the significant difference of BFR was also detected in both cortical and trabecular bone areas (trabecular BFR: cKO 50.2%, WT 22.5%) with a 72% increase in trabecular bone thickness. Bone histomorphometry of cKO confirmed the significant increase of bone volume and trabecular thickness with thicker osteoid in femur than that of WT (Osteoid Surface: cKO 6.74%, WT 2.25%). Although osteoblast surface was not significantly different in the cKO bones, TRAP positive cells were dramatically decreased and bone turn over was delayed in 34 wks old cKO. Furthermore, the content of a mature collagen cross-link, pyridinoline, was significantly higher in cKO femur when compared to WT, indicating a lower collagen turnover in cKO. Interestingly, micro CT evaluation of volumetric mineral density revealed that male cKO mice had a larger proportion of bone that was at a higher mineral density than WT, but female cKO had a larger proportion of bone that was of a lower mineral density. These data suggest that the BMP signaling regulates the distinct properties of trabecular and cortical bone independently, most likely by influencing bone remodeling, and the regulation changes in an age, sex, and skeletal site dependent manner in vivo.

Disclosures: N. Kamiya, None.


BMP3 Is a Mediator of Age-Dependent Bone Loss in Postnatal Mice through Its Action on Osteoblast Differentiation. K. Tsuji, J. Nove*, L. Gamer, K. Cox, V. Rosen. Developmental Biology, Harvard School of Dental Medicine, Boston, MA, USA

Unlike osteogenic BMPs, BMP3 acts as an inhibitor of bone formation and null mutation of BMP3 in mice results in increased trabecular bone volume by 5 weeks after birth. To date, the mechanism by which BMP3 regulates postnatal bone mass remains unclear as no significant differences in the numbers of osteoblasts or osteoclasts, or in mineral apposition rates were observed when BMP3 nulls were compared to wt littermates. Since osteoprogenitors and osteoblasts are the targets for osteogenic BMPs, we hypothesized that BMP3 acts directly on these cells to block their response to bone forming signals. To test this hypothesis, we collected bone marrow stromal cells from BMP3 nulls and examined formation of CFU-F and CFU-OB. While CFU-F values were equivalent for BMP3 null and wt mice, absence of BMP3 resulted in an increase in CFU-OB, consistent with the idea that BMP3 regulates OB differentiation and function. We next looked at the ability of BMP3 to affect osteoblast differentiation in vivo by performing bone marrow ablation, and found that BMP3 null mice have more woven bone 9 days after surgery (p = 0.03). Finally, to examine the long-term effects of removing BMP3 from bone, we performed 3D microCT analysis on bones (BMP3 KO v. wt littermates) from young growing mice (10 week-old), and compared them to mature mice (30 week old). At 10 weeks of age, trabecular bone volume (BV/TV,%) is increased in BMP3 null mice, although the differences we observe are not statistically significant. By 30 weeks of age, when BV/TV is reduced in wt mice (p = 0.03), BMP3 nulls appear to be resistant to age-related bone loss. These data, together with the finding that BMP3 levels in bone matrix increase with age, identify BMP3 as a physiological regulator of postnatal bone mass through its ability to modulate osteoblast differentiation and function and suggest that levels of BMP3 stored in bone matrix may be responsible for the decreased bone formation that occurs with aging.

Disclosures: K. Tsuji, None.

This study received funding from: NIH 5RO1AR50174-4.


OPG Inhibits Vascular Calcification but Does not Increase Atherosclerosis in Hyperlipidemic LDL Receptor Knockout Mice. S. Morony1, Z. Zhang*2, R. Cattley*3, D. Dwyer*3, M. Stolina3, P. Kostenuik3, Y. Tintut2, L. Demer2. 1MCIP UCLA/Amgen Inc., Los Angeles, CA, USA. 2UCLA, Los Angeles, CA, USA. 3Amgen Inc., Thousand Oaks, CA, USA

RANKL is essential for osteoclast formation, activation, and survival. OPG, a soluble decoy receptor for RANKL, may have a role in vascular as well as bone function. Serum OPG levels increase modestly with clinical coronary artery disease (Jono 2002, Circulation). However, mice deficient in OPG develop vascular calcification, suggesting that OPG may actually have a protective role in vascular disease (Bucay 1998, Genes Develop). To test OPG effects on the vasculature, we used a mouse model in which serum OPG increases with vascular disease. These LDL receptor knockout (ldlr−/−) mice develop atherosclerosis spontaneously, and a high-fat diet induces vascular calcification in these mice (Towler 1998, JBC). We fed ldlr−/− mice a high fat diet and treated them with synthetic OPG (Fc-OPG; 10 mg/kg, SC, 3X/week) or vehicle (PBS, SC). Plasma levels of OPG were assessed regularly, and the thoracic aorta was harvested after 2 months (n = 10/group) and 5 months (n = 16–17/group) to assess atherosclerosis and vascular calcification. In vehicle-treated mice, endogenous serum OPG increased within 2 weeks and persisted for 5 months, with a maximum 2.4-fold increase over baseline (p <0.05). Histology of the aorta revealed progressive atherosclerotic disease in this control group, with an average of 0.07+/−0.06 at 2 months and 0.25+/−0.10 at 5 months (lesions/mm; mean+/−SD). These observations suggest that ldlr−/− mice on a high fat diet recapitulate clinical observations in which OPG is an early marker of vascular disease. Treatment of ldlr−/− mice with Fc-OPG significantly reduced atherosclerotic lesion number at 2 months (by 57%, p <0.05 versus control). By 5 months, there were no significant changes in atherosclerotic lesion number between Fc-OPG and vehicle groups. Vascular calcification was not detectable in either group at month 2. At 5 months, the incidence of vascular calcification was significantly reduced in mice treated with Fc-OPG (29% vs. 65%, p <0.05).

In control mice, levels of osteocalcin protein in aortic tissue increased 22-fold from 2 to 5 months (p <0.05), consistent with the increase in aortic calcification. In mice treated with Fc-OPG, aortic osteocalcin also increased but the levels were significantly lower than controls at both time points (38% less at 2 months; 75% less at 5 months, p <0.05 vs. control). Thus, administration of Fc-OPG inhibited calcification without affecting atherosclerosis in ldlr−/− mice. These results suggest that an increase in endogenous OPG may be a response to rather than a mediator of atherosclerosis, possibly limiting calcification.

Disclosures: S. Morony, Amgen Inc. 1, 3.

This study received funding from: Amgen, Inc. and NIH/NHLBI.


Skeletal Overgrowth in Mice Lacking Chondrocyte Expression of the Pten Tumor Suppressor. A. F. Ford-Hutchinson*1, Z. Ali*1, S. E. Lines*1, B. Hallgrímsson2, S. K. Boyd3, F. R. Jirik*1. 1Biochemistry and Molecular Biology, University of Calgary, Calgary, AB, Canada. 2Cell Biology and Anatomy, University of Calgary, Calgary, AB, Canada. 3Mechanical and Manufacturing Engineering, University of Calgary, Calgary, AB, Canada

Little is known about the role of cellular signal transduction pathways lying downstream of phosphatidylinositol 3′-kinase (PI3K) in the regulation of chondrocyte development and endochondral ossification. Chondrocyte growth, differentiation, and control of endochondral ossification represents a complex process depending on an intricate set of events, involving endocrine, paracrine, autocrine, and extracellular matrix interactions. The in vivo role of the phosphatidylinositol 3′-kinase (PI3K) signaling pathway in cartilage development and endochondral bone formation remains largely unstudied, despite the fact that many of the factors that control chondrocyte function can lead to PI3K activation. This is important since signaling pathways downstream of PI3K regulate a host of cellular events, including: metabolism, cell size, protein translation, susceptibility to apoptosis, proliferation, migration and differentiation. A key regulator of the product of PI3K, PI(3, 4, 5)P3, is the 3′ phosphoinositide phosphatase and tumor suppressor, Pten. To investigate the role of this molecule in skeletogenesis, we crossed floxed Pten mice with mice expressing the Cre recombinase under the control of the Type II collagen promoter to generate Pten-deficient chondrocytes. These mice yielded a phenotype characterized primarily by increased length, with bones of endochondral origin being enlarged and demonstrating increased trabeculae and cortical thickness, as determined by microcomputed tomography. Histologically, the growth plates of the cartilage Pten-deficient group appear disorganized and were not expanded as compared to controls although there was increased cartilage matrix deposition. Trabecular growth in the metaphyseal region (sub-growth plate) as well as perichondrial bone collar formation were augmented in the Pten-deficient mice, in keeping with the resulting bone phenotype. Mesenchymal cell-derived neoplasia was absent. Our results demonstrate that the intensity of PI3K signaling in cartilage might be a key factor in determining the quality of the adult bony skeleton in mice.

Disclosures: A.F. Ford-Hutchinson, None.


Wnt/Beta-Catenin Signaling Controls Chondrocyte Hypertryphy by Inhibiting PTHrP Signaling. X. Guo, Y. Yang. Genetic Disease Research Branch, NIH/National Human Genome Research Institute, Bethesda, MD, USA

Both Wnt/beta-catenin signaling and PTHrP signaling play important roles in controlling chondrocyte hypertrophy. PTHrP signaling inhibits chondrocyte hypertrophy whereas Wnt/beta-catenin signaling promotes it. To test whether PTHrP signaling is required for the activity of Wnt/beta-catenin signaling in regulating chondrocyte hypertrophy, we have used molecular genetic approaches to investigate chondrocyte hypertrophy and maturation in mice in which Wnt/beta-catenin and PTHrP signaling has been inactivated in all chondrocytes. We have used a floxed allele of beta-catenin and Col2al-Cre mice to block the canonical Wnt signaling in the developing cartilage. We then crossed the beta-catenin conditional mice with a PTHrP+/− mutant mouse to eventually generate the double knockout mice Catnbc/c; PTHrP−/−; Col2al-Cre.

We found previously that progression to chondrocyte hypertrophy is delayed in Catnbc/c; Col2al-Cre single mutant mice, but the expression of PTHrP was hardly affected. In the Catnbc/c; PTHrP−/−; Col2al-Cre double mutant mice, we found that the pace of chondrocyte hypertrophy was the same as that in the PTHrP−/− single mutant mice according to both histological and molecular analysis. The initiation of chondrocyte hypertrophy indicated by the expression of Ihh, PTHR1 and ColX, was the same as that in the PTHrP−/− mutant mice. These results indicate that Wnt/beta-catenin signaling acts upstream of PTHrP signaling in controlling chondrocyte hypertrophy. Wnt/beta-catenin signaling promotes chondrocyte hypertrophy by inhibiting PTHrP signaling activity but not PTHrP expression itself.

We have also found previously that the final maturation of hypertrophic chondrocytes was delayed in the Catnbc/c; Col2al-Cre mice. In the Catnbc/c; PTHrP−/−; Col2al-Cre double mutant mice, final maturation of the hypertrophic chondrocyte was still delayed as compared to that in the PTHrP−/− single mutant mice. The expression domain of Ihh was expanded and the expression of VEGF and MMP13, which mark terminally matured chondrocytes, was still delayed compared to that in the PTHrP−/− single mutant mice. As a result, the hypertrophic zone was greatly expanded and trabecular bone formation was significantly delayed in the Catnbc/c; PTHrP−/−; Col2al-Cre mice compared to those in the PTHrP−/− single mutant mice. These results indicate that Wnt/beta-catenin signaling is required for terminal chondrocyte maturation independently of PTHrP signaling.

Disclosures: X. Guo, None.


Role of RAR Signaling in Skeletal Development and Growth. J. A. Williams*, N. Takeshita*, Y. Tamamura*, M. Iwamoto*, M. Pacifici. Orthopaedic Surgery, Thomas Jefferson University, Philadelphia, PA, USA

Retinoid (RA) signaling plays key roles in the development of various structures including the skeleton. Both over-expression of dominant negative forms of RA receptors (RAR) or excess uptake of RA cause skeletal deformity, indicating that appropriate control of RA signaling is essential for normal skeletal formation. To determine more specifically the physiological roles of RA signaling in cartilage development, we first analyzed the spatial and temporal gene expression patterns of RARs and RA-related molecules in developing mouse embryo limbs. We then created conditional mutant mice lacking single or dual RARs in cartilage. Gene expression of RAR alpha, beta and gamma and RA dehydrogenase 2(RALDH2) was determined by RT-PCR or in situ hybridization using longitudinal sections of E12.5–E18.5 mouse hind limbs. Cartilage-specific RAR knock out mice were created by crossing floxed RAR alpha, beta and/or gamma mice (provided by Dr. P. Chambon) with Col2al-Cre mice (provided by Dr. Y. Yamada). RT-PCR detected expression of all three RARs in cartilage, but only RAR alpha and RAR gamma were detected by in situ hybridization. RAR gamma was strongly expressed from the epiphyseal end to pre-hypertrophic zone of growth-plate and its expression pattern was similar to that of type II and IX collagens. In contrast, RALDH2, an enzyme needed for the last step during active retinoid production, was detected in perichondrium and periosteum flanking the diaphysis where in fact RAR gamma was not expressed. Cartilage specific knock out of single RARs did not create noticeable phenotypes, although efficiency of gene ablation was more than 90% as determined by Southern hybridization and PCR. A similar lack of obvious defects was noted in mice lacking one receptor plus a single copy of another receptor and in mice lacking both RAR alpha and beta. This suggested a functional redundancy in these RARs that allowed some compensation to occur. In contrast, mice lacking RAR gamma and either alpha or beta exhibited a retardation of skeletal growth that was first evident at 3 weeks of age. Overall body and limb sizes and lengths were 15–35% lower than those in wild type littermates. Also, the anteroposterior diameter of the skull was shorter. The findings indicate that RAR gamma signaling plays an important role in chondrocyte function during skeletal growth, particularly postnatally. Given that RALDH2 expression was not super-imposable to that of RAR gamma, it is possible that RAR gamma-expressing chondrocytes may not be provided with active ligands. This suggests that RAR gamma might positively regulate cartilage growth not by ligand-dependent transcriptional activation but ligand-independent repressing activity.

Disclosures: M. Iwamoto, None.

This study received funding from: NIH.


Phosphate Regulates Chondrocyte Differentiation through the Sodium-Dependent Phosphate Transporter 3. A. Sugita1, S. Kawai*1, T. Hayashibara*1, D. Tamura*2, H. Yoshikawa2, T. Yoneda1. 1Biochem, Osaka Univ Grad Sch of Dent, Suita, Japan. 2Orthopedics, Osaka Univ Grad Sch of Med, Suita, Japan

Endochondral ossification is dependent on chondrocyte differentiation that is under the control of diverse factors. Occurrence of rickets and retarded skeletal development in X-linked hypophosphatemia (XLH) suggests a crucial role of phosphate in chondrocyte differentiation. In addition, previous studies have shown that phosphate levels change according to the advancement of cartilage zone in growth plate, suggesting an involvement of phosphate in chondrocyte differentiation. Here we studied Hyp mice, the murine homologue of human XLH, to examine the role of phosphate in chondrocyte differentiation. Hyp mice showed growth retardation with short long bones and disturbed endochondral ossification compared with wild-type (WT) mice. Of interest, Hyp mice fed with high phosphate diet displayed improved endochondral ossification, suggesting that the extracellular phosphate influences chondrocyte differentiation. Since the extracellular phosphate is known to regulate the cellular metabolism through the sodium-dependent phosphate transporter (NPT), we examined the expression of NPT in chondrocytes. Hyp chondrocytes showed decreased mRNA expression of NPT3 compared with WT chondrocytes. In parallel, they also exhibited reduced 32P uptake and mineralization. Overexpression of NPT3 cDNA in Hyp chondrocytes restored mineralization. Of note, apoptosis, a prerequisite for the cartilage mineralization, and deposit of matrix vesicles in extracellular matrices were reduced in the growth plates of Hyp mice compared with WT mice. Administration of a non-selective competitive inhibitor of NPT, phosphonoformic acid (PFA), to WT mice caused hypophosphatemia and disturbed endochondral ossification with reduced apoptosis in the growth plates. To understand the mechanism by which reduced intracellular phosphate levels cause decreased apoptosis, we determined the intracellular ATP levels in Hyp chondrocytes, since ATP is required for an activation of caspase 9-mediated apoptosis. Intracellular ATP levels in Hyp chondrocytes were significantly decreased compared with WT chondrocytes and PFA reduced intracellular ATP levels in WT chondrocytes. 3-bromopyruvate, an inhibitor of ATP production, inhibited chondrocyte mineralization in culture and administration of 3-bromopyruvate reduced apoptosis in the growth plates in WT mice. In conclusion, our results suggest that phosphate uptake through the NPT3 modulates intracellular ATP levels, which in turn controls apoptosis and mineralization of chondrocytes, thereby regulates endochondral ossification.

Disclosures: A. Sugita, None.


Regulation of Cartilage Matrix Turn-Over and Degradation by Wnt/Beta-Catenin Signaling. T. Yuasa1, T. Koike*2, A. Hargett*1, M. Pacifici1, M. Iwamoto*1, M. Enomoto-Iwamoto*1. 1Orthopedic Surgery, Thomas Jefferson University, Philadelphia, PA, USA. 2Orthopedic Surgery, Osaka-city University, Osaka, Japan

A fine balance between cartilage matrix synthesis and degradation is critical for articular cartilage and for skeletal formation and growth during endochondral ossification, and imbalances in this homeostasis can lead to cartilage disease such as osteoarthritis. Recent studies have revealed that the Wnt/beta-catenin signaling pathway has important roles in chondrocyte development and function. The object of this study was to investigate whether this pathway is specifically involved in regulating cartilage matrix turn-over and may thus have roles in cartilage pathologies. Chondrocytes were isolated from the epiphyseal portions of newborn mice, treated with Wnt3A-containing medium (Wnt3A-CM) or infected with adenovirus encoding a constitutive active form of beta-catenin (CA-b-catenin). Gene expression of cartilage matrix and matrix degradation enzymes was measured by real time RT-PCR, and rates of proteoglycan degradation were determined by alcian blue staining and pulse-chase protocols. Serial tissue sections were prepared from medial plateau of knee joints of 6- to 12-month-old Hartley guinea pigs (OA animal model, N = 50) and used to determine the intracellular distribution of beta-catenin by immunohistochemistry. Treatment with Wnt3A-CM or over-expression of CA-b-catenin in primary mouse chondrocyte cultures markedly reduced proteoglycan content and gene expression of type IX collagen and aggrecan. They also stimulated release of cell-associated proteoglycans by 2 to 3 fold and gene expression of MMP-3, MMP-7, MMP-13 and ADAMTS-4 gene by 2.5-, 4-, 8- and 6.5-fold, respectively, compared to control cultures. Beta-catenin was present at relatively low levels in seemingly healthy articular chondrocytes (Grades 0∼2). However, beta-catenin was far more abundant and localized in both cytoplasm and nucleus in chondrocytes in osteoarthritic specimens (Table 1), suggesting that Wnt/beta-catenin signaling had been activated in the pathological tissues. These findings strongly suggest that the Wnt/beta-catenin signaling pathway tilts cartilage matrix remodeling toward excessive matrix degradation and may thus participate in pathogenic matrix disruption characteristic of degenerative joint diseases. 

Table  . Relationship between beta-catenin distribution and OA severity in guinea pig tibia
inline image

Disclosures: T. Yuasa, None.


Expression of Runx2 Transcription Factor in Nonskeletal Tissues. J. Jeong*1, J. Jin*1, H. Jin*1, S. Kang*1, A. van Wijnen2, J. Stein2, J. Lian2, G. Stein2, J. Choi1. 1Biochemistry and Cell Biology, Skeletal Diseases Genome Research Center, Kyungpook Natl. Univ. School of Medicine, Daegu, Republic of Korea. 2Cell Biology, UMASS Medical School, Worcester, MA, USA

Runx2 is an important transcription factor for chondrocyte differentiation as well as for osteoblast differentiation. Although Runx2 has been shown to be an essential transcription factor in skeletal development, some studies show that Runx2 is also expressed in nonskeletal tissues such as breast, T cells and testis. To better define the tissue specificity of Runx2 expression, we used transgenic mice expressing LacZ under the control of 3.0 kb or 1.0 kb of the Runx2 distal (P1) promoter. In skeletal tissues, as previous study (Lengner et al., Mech. Dev. 114:167–170, 2002), the β-galactosidase expression appeared in axial skeleton especially in periosteum of ribs and vertebral disc area of 3.0kb Runx2 distal promoter/LacZ derived transgenic mouse (3kb Tg). However, the β-galactosidase expression was not observed in any skeletal tissues of 1.0kb Runx2 distal promoter/LacZ derived transgenic mouse (1kb Tg). In nonskeletal tissues of 3.0kb Tg, the β-galactosidase expression appeared in cerebellum, breast, skin, and testis. Both 3kb and 1kb Tg showed no expression activity of Runx2 promoter in liver, kidney, spleen, and intestine. Histological analysis of 3kb Tg revealed that the β-galactosidase expression is readily detectable in seminiferous tubules of the testis and the epididymis. In a single cell level, β-galactosidase was detected in spermatids and mature sperms not in sertoli or Leydig cells. Indeed, Runx2 expression was observed in isolated mature sperms, which was confirmed by RT-PCR and western blot analysis. Collectively, these results indicate that Runx2 is expressed in several nonskeletal tissues especially in sperm of testis. It suggests that Runx2 may play an important role in nonskeletal tissues such as spermatogenesis in testis.

Disclosures: J. Choi, None.


Localization of Osteoprotegerin Gene Expression in the Human Tibial Plateau at End Stage Osteoarthritis: Correlation with Bony Adaptation by Micro-CT. M. R. Doschak1, M. Lincoln*2, T. Trinh*3, B. Hallgrimsson3, R. F. Zernicke3. 1Pharmacy & Pharmaceutical Science, University of Alberta, Edmonton, AB, Canada. 2Orthopaedic Surgery, University of Calgary, Calgary, AB, Canada. 3Joint Injury & Arthritis Research Group, University of Calgary, Calgary, AB, Canada

End-stage osteoarthritis (OA) involves joint space narrowing, osteophyte and bone cyst formation, and subchondral plate thickening. Early in the pathogenesis of OA, animal models have shown that changes in subchondral bone precede those of articular cartilage. Initially the periarticular bone becomes osteopenic, undergoing remodelling over time to produce the hypertrophied changes typical of end-stage disease. The altered joint mechanics and physiology produced by this periarticular bone remodelling may thus cause the cartilage and joint changes of OA. To date, however, factors controlling this process are poorly understood. The receptor activator of nuclear factor kappa beta (RANK), along with its soluble ligand (RANK-L), have been shown to modulate bone resorption by inducing both activation and maturation of bone-degrading osteoclasts. Activation of RANK on osteoclast cells by RANKL is opposed by another soluble factor, osteoprotegerin (OPG). Thus, local RANKL/OPG levels determine whether resorption or formation of bone will predominate.

We hypothesized that subchondral and trabecular bone densities were increased in OA bone relative to age-matched cadaveric controls, and that bone hypertrophic adaptation was associated with OPG expression. In this study, we compared the bony changes in the human tibial plateau from samples taken at total knee arthroplasty (TKA, n = 15) against those from cadaveric controls (n = 8). Micro computed tomography (micro-CT) was used to define and quantify bone changes, and these data were coupled with the molecular expression (by “in-situ” hybridization) of peptide factors known to mediate bone remodeling. In-situ hybridization of digoxygenase (DIG)-labelled OPG riboprobes showed selective uptake in osteoblasts, but not osteocytes or osteoclasts in TKA bone. Staining of DIG-labelled OPG was particularly evident in activated osteoblasts involved in bone adaptive/sclerotic processes. Using micro-CT, subchondral OA bone samples were significantly different from cadaveric controls for subchondral bone thickness, bone volume, and trabecular thickness. In conclusion, these data suggest that OPG expression by osteoblasts may precipitate the bony hypertrophy of end-stage OA. Thus, early modulation of the RANKL/OPG axis may provide hope for halting the sclerotic bone changes leading to end-stage OA.

Disclosures: M.R. Doschak, None.


Progression of Cartilage Degradation in Osteoarthritis May Involve Autocrine Induction of IL-1β and Loss of DNA Methylation. M. B. Gibson*, H. I. Roach. Bone & Joint Research Group, University of Southampton, Southampton, United Kingdom

Osteoarthritis (OA) is characterised by progressive erosion of articular cartilage, which, once started, cannot be halted. The breakdown of cartilage is mediated by proteases, including MMP-3 and −13. These are derived from the synovium and certain OA chondrocytes. In normal articular chondrocytes, the proteases are thought to be silenced by epigenetic mechanisms. In OA, chondrocytes of the superficial zone change to a “degradative” phenotype, which produce the enzymes and also proliferate so that the abnormal gene expression is stably transmitted to daughter cells. This aberrant expression has been shown to be associated with an epigenetic “un-silencing” via demethylation of specific CpG sites within the promoter regions (Arthritis & Rheumatism 52:3110–24). Why this demethylation takes place is largely unknown. The pro-inflammatory cytokine IL-1β is of potential importance in OA, especially when synovial inflammation could provide IL-1β. The cytokine upregulates MMPs in chondrocytes in vitro. We investigated whether the IL-1 mediated induction of MMPs was associated with DNA demethylation, which might explain the progressive nature of OA.

Chondrocytes were isolated from non-OA articular cartilage and grown in monolayer culture. The cells from each patient were divided into no-treatment versus IL-1β treated groups. When confluent, simultaneous RNA and DNA extraction was carried out. Expression levels were analysed by RT-PCR and the methylation status of specific CpG sites within the promoters of MMP-3, −13, and IL-1β was determined in the same samples, using the methylation-sensitive restriction enzymes and PCR. The pre-culture control only expressed type II collagen and low levels of MMP-3. Culture per se increased expression of MMP-3 and, to a lesser extent, MMP-13. IL-1β further increased expression of the MMPs and, surprisingly, induced expression of IL-1β itself. The large increases in MMP-3 and IL-1β were associated with loss of methylation at specific CpG sites in the promoter of these mediators. The strongest correlation between expression and promoter demethylation was seen for IL-1β. IL-1β thus induced its own expression, which corresponded to promoter demethylation at one specific CpG site. If the same applies in vivo, this data suggests that an initial inflammatory episode could set up an autocrine stimulatory loop as well as epigenetic changes. This vicious circle of unregulated IL-1β expression by degradative chondrocytes combined with the heritable nature of DNA demethylation may explain, in part, the unremitting progression of OA.

Disclosures: M.B. Gibson, Wessex Medical Trust 2.

This study received funding from: Wessex Medical Trust.


Estrogen Receptor Related Receptor (ERR) α Is a Transcriptional Activator of the mSox9 Promoter. R. A. Zirngibl*1, J. M. Chan*1, T. M. Underhill*2, J. E. Aubin1. 1Molecular and Medical Genetics, University of Toronto, Toronto, ON, Canada. 2Department of Anatomy, Cell Biology and Physiology, University of British Columbia, Vancouver, BC, Canada

Estrogen Receptor Related Receptor (ERR) α is an orphan nuclear receptor identified by its similarity in the DNA-binding region to estrogen receptor (ER) α. While ERRα is unable to bind or be activated by estrogen, it can modulate the actions of the ERs. We have previously shown that antisense oligonucleotide-mediated knockdown of ERRα expression in the chondrogenic C5.18 cell line reduces chondrogenesis and expression of Sox9, the master regulator of chondrogenesis, raising the possibility that ERRα regulates chondrogenesis through its effect on Sox9 expression. When we cloned 3kb of the mouse Sox9 promoter into a luciferase reporter vector and co-transfected it with an ERRα expression plasmid into either C5.18 cells or HeLa cells, we found up-regulation of the Sox9 promoter luciferase construct. Surprisingly, mutation of the highly conserved helix 12 activation function 2 (AF2) domain (ERRαmAF2), which we predicted would severely reduce or abolish the transcriptional activity of ERRα, increased its transcriptional activity on the Sox9 promoter. Deletion of helix 11 and 12 (ERRαΔ11/12), on the other hand, abolished Sox9 promoter activity as expected. Through a series of promoter truncations of the minimal Sox9 promoter, we have now narrowed down the putative ERRα regulatory site to be between −78bp and +15bp relative to the transcription start site. Our results demonstrate that the mouse Sox9 promoter is regulated by ERRα in a cell context independent manner, an activity that may mediate the role of ERRα in chondrogenesis. In addition, we describe a novel putative silencing mutation in ERRα that renders ERRα a better transactivator of Sox9.

Disclosures: R.A. Zirngibl, None.


Circulating Cells Contribute to the Fracture Repair in the Mouse Fibula Fracture Model. K. Kumagai*1, A. Vasanji*1, J. Drazba*2, G. Muschler1. 1Biomedical Engineering and Orthopaedic Research Center, The Cleveland Clinic Foundation, Cleveland, OH, USA. 2Imaging Core, The Cleveland Clinic Foundation, Cleveland, OH, USA

Bone, bone marrow, fat and muscle tissue contain connective tissue progenitors (CTPs) that can be induced to generate progeny expressing osteoblastic markers in vitro and to contribute to new bone tissue in vivo. However, the concentration, prevalence and intrinsic biological potential of the CTP population from these sources vary widely. Several studies have also suggested that under some conditions osteogenic CTPs may transit through systemic circulation contributed to new bone formation and fracture repair. However, the magnitude of the contribution of circulating osteogenic cells to normal fracture repair is not known, and the variables modulating the mobilization, transit time and homing of circulating osteogenic cells is not characterized. This study was designed to identify and measure the contribution of circulating osteogenic CTPs to fracture repair in normal mice using a green fluorescent protein (GFP) reporter, the murine parabiosis model system and percutaneous fibula fracture model.

Transgenic GFP C57BL/6-TgN and wild type C57BL/6 mice were surgically conjoined for parabiosis at 4–5 weeks old. Three weeks after parabiosis surgery, a fibula fracture was created in the contralateral hind limb of the GFP negative partner by cutting diaphysial fibula, 3–4mm from fibular head with dissection scissors. The fracture site was harvested 1, 2 and 4 weeks after fracture and evaluated using quantitative histomorphometric methods using confocal microscopy. Cell counting was performed using automated counting of DAPI stained nuclei. Osteogenic differentiation among cells contributing to fracture repair was identified based on co-localization of GFP and alkaline phosphatase (AP) staining.

To confirm the establishment of the shared circulation, peripheral blood samples were obtained before fracture surgery. Flow cytometry analysis revealed 40–50% chimerism in the systemic circulation of the parabiosis partners. The shared circulation between parabiosis partners was maintained until the time of euthanasia. GFP positive cells gaining access to the fracture site via systemic circulation were observed in the fracture callus of all wild type partners at each time point. The % contribution of GFP+ cells at 1, 2 and 4 weeks was 1.9, 5.8, and 4.0, respectively. Of the GFP+ cells present, the percent of GFP + cells that were also colocalized with AP activity was 23.3, 88.6, and 14.9, respectively.

This study demonstrated that circulating cells physiologically home to the fracture healing region, and appear to contribute to new bone formation.

Disclosures: K. Kumagai, None.


Characterization of Sh3bp2 Knock-In Mice - A Model for Cherubism. P. Mukherjee*1, T. Jafarov*1, Y. Ueki*2, A. L. Boskey3, B. R. Olset2, E. J. Reichenberger*1. 1Center for Reconstructive Medicine and Skeletal Development, UCHC, Farmington, CT, USA. 2Department of Oral and Developmental Biology, Harvard School of Dental Medicine, Boston, MA, USA. 3Hospital for Special Surgery-Weill Medical College, Cornell University, New York, NY, USA

Cherubism is a fibro-osseous disorder characterized by bilateral symmetrical bone resorption of the jaws. Mutations in the Sh3bp2 gene have been identified as the cause for cherubism. A knock-in mouse model for cherubism carrying a Pro416Arg mutation in Sh3bp2 develops an osteoporotic phenotype. Here we characterize the bone phenotype in knock-in mice and the effects of mutant Sh3bp2 on osteoblasts.

We performed skeletal analysis by Alcian blue and Alizarin red staining and utilized MicroCT, DEXA and histomorphometry to analyze bone architecture, bone mineral density (BMD), bone mineral content (BMC), and numbers of osteoblasts and osteoclasts in long bones. We analyzed bone formation rate (BFR) and mineral apposition rate (MAR) by dynamic histomorphometry. We employed static histomorphometry to detect the number of osteoclasts and osteoblasts and used bone marrow and calvarial cell cultures to study changes in differentiation and mineralization by alkaline phosphatase and Von Kossa staining. We utilized Fourier Transformed Infrared Spectroscopy (FTIR) for a more detailed matrix analysis. Real time PCR (QPCR) was performed using RNA isolated from calvarial cultures at different time points to evaluate changes in gene expression between wild type and homozygous mice.

Homozygous mice are smaller than wild type littermates and there appears to be a delay in skeletal development. Long bones exhibit abnormal growth plates and decreased numbers of trabeculae. MicroCT and DEXA analysis showed a significant decrease in trabecular number and cortical thickness with decreased BMD and BMC. Static histomorphometry showed a marked increase in osteoblast and osteoclast numbers. Dynamic histomorphometry revealed a decrease in BFR and MAR. Differentiation and mineralization rates were decreased in homozygous calvarial osteoblast cultures. FTIR of long bones and calvarial osteoblast cultures suggested increased collagen cross linking in cortical bone of homozygous mice which may be due to faster deposition of a defective matrix. Our QPCR results demonstrated a decreased expression of osteoblast differentiation markers like Colla, BSP and OC.

These cherubism mice display excessive bone resorption, an overall decreased bone mineralization, impaired osteoblast differentiation, and increased numbers of osteoclasts and osteoblasts. Our results suggest an important role of Sh3bp2 in bone homeostasis.

Disclosures: P. Mukherjee, None.


FLJ: A Novel Collagen-Binding Small Leucine-Rich Repeat Proteoglycan in Bone. Y. Mochida*, D. Parisuthiman*, M. Katafuchi*, P. Atsawasuwan*, M. Kaku*, M. Yamauchi. Dental Research Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Fibrillar type I collagen serves as a stable template for mineral deposition and growth in bone, thus, critical for determining the bone mechanical properties. It is now clear that the process of collagen fibrillogenesis is modulated by a number of molecules including specific small leucine-rich repeat proteoglycans (SLRPs). The purposes of this study were to identify new members of collagen-binding SLRPs expressed in osteoblasts, and to partially characterize its function in matrix mineralization in vitro. By employing a bioinformatics approach, we have identified a gene encoding a novel member of SLRP family, designated as FLJ. Using cDNA derived from MC3T3-E1 osteoblastic cells (MC), the coding sequence of FLJ was cloned by PCR and the expression pattern during matrix mineralization was analyzed by quantitative real time PCR. In order to investigate the potential roles of FLJ in osteoblast biology, MC derived clones overexpressing FLJ (S clones) were established and characterized by analyzing cell proliferation, osteoblast differentiation, in vitro mineralization, and collagen fibrillogenesis. The phenotypes of S clones were evaluated by comparing to those of two control groups, i.e. MC and a clone transfected with an empty vector (EV clone). Furthermore, collagen binding assay was performed using FLJ protein and purified bone type I collagen. During matrix mineralization, the expression of FLJ was increased. Cell proliferation and osteoblast differentiation were not affected in S clones when compared to MC and EV clone. However, in S clones, in vitro mineralization was accelerated and the diameter of collagen fibrils was significantly larger in comparison to the controls. Furthermore, FLJ was directly bound to bone collagen in a dose dependent manner. These results indicate that FLJ is a novel SLRP member expressed in osteoblasts, and that it facilitates collagen fibrillogenesis and mineralization in vitro through its direct interaction with type I collagen.

Disclosures: Y. Mochida, None.


Heterogeneous Activation of Vertebral Trabecular Bone Surface Cells in Response to Isometric Loading. J. D. Delaney*1, X. Jung*1, D. J. Adams2, D. W. Rowe1. 1Genetics and Developmental Biology, University of Connecticut Health Center, Hartford, CT, USA. 2Orthopaedic Surgery, University of Connecticut Health Center, Hartford, CT, USA

It is generally accepted that bone initially responds to mechanical loading by activating resident mature osteoblasts and that prolonged loading results in the generation of new osteoblasts from a progenitor population. This model predicts a heterogeneous mixture of cells that are acting in a concerted fashion, but cannot be easily appreciated with current methods of identification. We have been building promoter-GFP reporter constructs designed to reflect different levels of osteoblast differentiation. For this study we crossed pOBCo13.6GFPcyan (3.6cyn) transgenic mice, whose reporter is active in preosteoblasts and early osteoblasts, with pOBCo12.3GFPemerald (2.3emd) transgenic mice, whose reporter is active in mature osteoblasts and osteocytes. 5–6 month old “2-color” mice were then subjected to an orbital loading protocol that produces an isometric load on the skeleton as the animals resist the motion of the cage. The animals were injected with xylenol orange (XO) two days prior to sacrifice after 7, 21 or 35 days of loading. Quantitation of fluorescent signals utilized histology generated via Cryojane frozen sections taken from non-demineralized bone. Images were generated on a Zeiss - Improvision workstation using a series of dichromic filters that allowed for the isolation of multiple of fluorescent signals from the same sample. Fluorescent ELF-97 stain under two different conditions was employed to visualize alkaline phosphatase (AP) and tartate resistant acid phosphatase (TRAP) activity. All results were compared to unexercised littermates. After only 7 days of the protocol (3 days of actual exercise) there was a 15-fold increase in male XO labeling and a 30-fold increase in female XO labeling that continued to be high throughout the exercise protocol. Five different cell populations on the bone surface were identified and quantitated: an immature osteoblast population (3.6cyn +, 2.3emd -, AP+, XO-), an active osteoblast population (3.6cyn ++, 2.3emd ++, AP ++, XO +), a mature osteoblast population (3.6cyn -, 2.3emd +, AP +, XO+), a flattened lining cell population (3.6cyn -, 2.3emd -, AP +, XO+) and a mononucleated preosteoclast population (3.6cyn +, 2.3emd -, TRAP+). The size of each cell population changed with continued orbital loading and appeared to exhibit gender dimorphism. With further refinements to this histological approach, it may be possible to fully appreciate the different bone lining populations and their response to perturbations affecting bone mass.

Disclosures: J.D. Delaney, None.


Calibrated Finite Element Voxel Models Based on High-Resolution CT Predict In Vitro Vertebral Strength Better than DEXA. Y. Chevalier*1, P. Varga*1, D. Pahr*1, E. Schneider2, P. K. Zysset*1. 1Institute for Lightweight Design and Structural Biomechanics, Vienna University of Technology, Vienna, Austria. 2AO Research Institute, Davos Platz, Switzerland

Predictions of vertebral strength based on bone mineral density (BMD) do not account for the densitometric heterogeneities of the vertebral body and for the actual mechanical behaviour of trabecular and compact bone. Recent studies have shown that finite element methods (FEM) based on quantitative computer tomography (QCT) images are valuable to predict vertebral strength. With technological developments in QCT, scanning resolutions improve constantly. Our study aims at providing a calibrated approach to predict in vitro vertebral stiffness and strength from FE models based on high-resolution QCT reconstructions.

Twelve cadaveric vertebral bodies from 4 male human lumbar spines (L1–L5, age 47 to 83) with known BMD from dual-energy X-ray absorptiometry (DEXA) were scanned with an XtremeCT system (SCANCO Medical AG, Switzerland) at 82 microns isotropic resolution to provide a detailed 3D map of BMD across the vertebral body. These specimens were then tested mechanically in axial compression to measure their in vitro stiffness and strength. The QCT voxels were resampled at a 656-micron resolution and directly converted to finite elements (FE) with linear elastic transverse isotropic material properties defined with a previously calibrated relationship between trabecular bone volume fraction and QCT BMD.

Numerical FE predictions of vertebral stiffness not only correlated satisfactorily (r2 = 0.68) but corresponded quantitatively well (y = 1.05x-193 N/mm) with the experimental stiffness which in turn correlated very well with the experimental strength (r2 = 0.95). Finally, the correlations of both the FE stiffness and DEXA BMD with the experimental strength measurements were calculated. As shown in Fig. 1, our results demonstrate that voxel models built from calibrated QCT images are significantly better at predicting in vitro compressive vertebral strength (r2 = 0.82) than BMD measured by DEXA (r2 = 0.62). Assessing vertebral strength from our FEM based on calibrated QCT images is thus an important improvement over the use of DEXA, and can be adapted to clinical resolutions for the efficient study of bone diseases and their treatments.

Thumbnail image of

Disclosures: P.K. Zysset, None.

This study received funding from: AO Research Institute, Lilly Deutschland GmbH.


Impaired Bone Healing in Matrix Metalloproteinase 13 Deficient Mice. N. Kosaki*1, S. Kamekura2, H. Kawaguchi2, T. Kimura*3, Y. Okada*3, L. Minqi*4, N. Amizuka*4, Y. Toyama*1, J. D'Armiento*5, H. Takaishi*1. 1Orthopaedic surgery, Keio University, Tokyo, Japan. 2University of Tokyo, Tokyo, Japan. 3Pathology, Keio University, Tokyo, Japan. 4Niigata University, Niigata, Japan. 5Columbia University College of Physicians and Surgeons, New York, NY, USA

Vascular and cellular invasion into cartilage are essential processes for endochondral ossification not only in bone growth, but also in bone healing. Matrix metalloproteinase 13 (MMP-13), a collagenase that plays essential roles in the remodeling of extracellular matrix, is known to be expressed in hypertrophic chondrocytes and osteoblasts. To investigate the possible involvement of MMP-13 in bone healing, we generated a bone fracture in MMP-13-deficient (MMP-13−/−) mice and compared the healing process with that of the wild-type (WT) littermates. A transverse osteotomy was created at the midshaft of the tibia in 8-week-old male mice of both genotypes (n = 30/genotype). In WT mice, callus formation was observed by day 7 and bone union by day 21 post-fracture. In contrast, fracture healing was incomplete in MMP-13−/− mice even at day 21, showing a radiolucent area remained within the fracture gap. Histologically, the ratio of the soft cartilaginous area to the total callus area was significantly greater in MMP-13−/− mice compared to that in WT mice during days 14–21, indicating that cartilage ossification was impaired by the MMP-13 deficiency. Although aggressive vascular invasion with substantial chondroclasts that attached directly to the matrix was observed in the WT cartilaginous callus, it was severely suppressed, and fewer chondroclasts were located unattached to the cartilage in the MMP-13−/− callus. To learn the molecular mechanism underlying the defect of endochondral ossification by the MMP-13 deficiency, we compared the cultures of primary chondrocytes derived from rib cartilages of WT and MMP-13−/− littermates. MMP-13−/− chondrocytes cultured in an atelocollagen gel were shown to exhibit decreases in differentiation and matrix synthesis compared to WT chondrocytes by histological analyses. When the WT chondrocyte pellets were co-cultured with human endothelial primary cell HUVEC in a collagen gel, numerous vascular sproutings were observed, in contrast to the few seen in the MMP-13−/− pellets. We therefore conclude that MMP-13 contributes to endochondral ossification in the bone healing process via vascular and cellular invasion into the callus cartilage.

Disclosures: N. Kosaki, None.


An Osteoprotegerin-like Peptidomimetic (OP3–4) Inhibits Osteoclastic Bone Resorption and Osteolytic Bone Disease in Multiple Myeloma. D. J. Heath*1, K. Vanderkerken*2, X. Cheng*3, O. Gallagher*1, M. Prideaux*1, R. Murali*3, P. Croucher1. 1Academic Unit of Bone Biology, University of Sheffield, Sheffield, United Kingdom. 2Department of Haematology and Immunology, Free University Brussels (VUB), Brussels, Belgium. 3Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA, USA

Multiple myeloma is a B-cell malignancy characterised by the uncontrolled growth of plasma cells in the bone marrow, which results in increased osteoclastic bone resorption and bone destruction. Myeloma cells express the ligand for receptor activator of NFkappa-B (RANKL), induce RANKL expression in the bone marrow and down-regulate expression of the decoy receptor, osteoprotegerin (OPG), thereby promoting bone resorption. Targeting this system in myeloma has clear therapeutic potential. However, OPG also binds tumour necrosis factor-related apoptosis inducing ligand (TRAIL) and acts as a survival factor for myeloma cells by preventing TRAIL-induced apoptosis. Whether OPG can bind TRAIL and prevent apoptosis in vivo and the relative importance of OPG binding TRAIL and RANKL is unclear. In the present study we have investigated the ability of an OPG-like peptidomimetic (OP3-4), designed to block the RANKL/RANK interaction, to inhibit osteoclastic bone resorption and TRAIL-induced apoptosis in vitro and myeloma bone disease in vivo.

OP3–4 inhibited osteoclast formation (% TRAP positive cells, 0.0 ± 0.1% vs 67.7 ± 17.0%, p <0.01) and bone resorption (% dentine resorbed, 3.1 ± 5.6% vs 90.0 ± 36.9%, p <0.01) in a dose-dependent manner in vitro, as compared to vehicle control. Unlike OPG, OP3–4 had no effect on TRAIL-induced apoptosis of RPMI-8226 myeloma cells in vitro, as compared to vehicle control (% apoptotic cells, 94.7 ± 4.0% vs 92.2 ± 3.4%). Injection of 5T2MM murine myeloma cells into C57BL/KaLwRij mice resulted in the appearance of a serum paraprotein and the development of a bone disease characterised by increased osteoclastic resorption, a decrease in cancellous bone and the development of bone lesions on x-ray. Treatment of 5T2MM bearing mice with OP3–4, from the time of paraprotein detection, decreased the surface covered by osteoclasts (17 ± 8.4% vs 24.9 ± 8.7%, p <0.05), increased cancellous bone area (1.4 ± 1.9% vs 0.5 ± 0.6%) and prevented the development of lytic lesions (3.7 ± 2.4 vs 7.5 ± 4.1, p <0.05), when compared to vehicle control. OP3–4 also reduced tumour burden (59.0 ± 31.9% vs 86.1 ± 23.0%, p <0.05) compared to vehicle control. These data suggest that OP3–4, and the selective inhibition of RANKL but not TRAIL activity, is effective in preventing myeloma bone disease and offers a novel therapeutic approach to treating this aspect of myeloma.

Disclosures: P. Croucher, None.


Lithium Inhibits the Development of Myeloma Bone Disease In Vivo. C. M. Edwards*, J. R. Edwards*, J. Esparza*, B. O. Oyajobi, B. McCluskey*, S. Munoz*, B. Grubbs*, G. R. Mundy. Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

Multiple myeloma (MM) is characterized by uncontrolled proliferation of myeloma cells within the bone marrow and the development of a unique and severe osteolytic bone disease. In addition to a well characterized increase in osteoclastic bone resorption, myeloma bone disease is also associated with a reduction in bone formation. Bone formation is regulated in vivo by canonical Wnt signaling, and MM is associated with an increase in Dkk1, a soluble antagonist of Wnt signaling. Our overall hypothesis is that increasing Wnt signaling in the bone microenvironment in MM may prevent the development of myeloma bone disease. Activation of the canonical Wnt signaling pathway can be achieved with lithium chloride (LiC1), which inhibits glycogen synthase kinase 3β and consequently activates β-catenin. The aim of this study was to test this hypothesis by determining the effect of LiCl in MM in vivo, using a murine model of MM. C57BIKalwRij mice were inoculated with 5TGM1 MM cells or PBS by intravenous injection. Mice were treated from time of tumor cell inoculation with 200mg/kg/day LiCl or vehicle control (d.H20) by oral gavage for 34 days. Intravenous injection of 5TGM1 MM cells was associated with an increase in serum IgQ and a decrease in bone mineral density (BMD) and trabecular bone volume. LiCl completely prevented the decrease in BMD associated with 5TGM1 MM-bearing mice (p <0.05). Treatment with LiCl also significantly increased trabecular bone volume in the hind limbs of MM-bearing mice (p <0.05). Furthermore, treatment with LiCl resulted in a significant decrease in serum IgG and tumor burden in bone (p <0.05), as determined by ELISA and histomorphometric analysis respectively. LiCl had no significant effect on BMD or bone volume in non-tumor bearing animals. To determine the effect of LiCl on MM growth independent from the bone microenvironment, C57B1Ka1wRij mice were inoculated with 5TGM1 MM cells or PBS subcutaneously on the right flank and treated from time of tumor cell inoculation with 200mg/kg/day LiCl or vehicle control (d.H20) by oral gavage for 28 days. Treatment with LiCl resulted in a 3-fold increase in tumor volume and tumor weight as compared to vehicle control (p <0.05). Our data demonstrate that LiCl can prevent the development of MM bone disease in vivo. However, the effect of LiCl to increase MM growth outside bone suggests that the reduction in MM tumor burden in bone is an indirect effect of LiCl treatment, and highlights the importance of the bone marrow microenvironment in regulating MM growth and survival.

Disclosures: CM. Edwards, None.


G-CSF Treatment Increases Osteoclast Formation and Promotes Tumor Growth in Bone. A. Hirbe*1, Ö. Uluçkan*1, M. Eagleton*1, E. Morgan*1, J. Prior*2, M. Kramer*1, H. Deng*1, D. Piwnica-Worms*2, R. Aft*1, M. Naughton*1, K. Weilbaecher1. 1Department of Medicine, Division of Oncology, Washington University School of Medicine, St Louis, MO, USA. 2Molecular Imaging Center, Department of Molecular Biology and Pharmacology, Washington University School of Medicine, St Louis, MO, USA

Granulocyte colony-stimulating factor (G-CSF) treatment has been associated with increased osteoclast (OC) number, decreased osteoblast number, and decreased bone mineral density (BMD). Because increased OC activity and bone turnover have been implicated in the release of tumor growth factors in the bone environment, we investigated the impact of G-CSF treatment on OC activity and tumor growth in bone in osteolytic tumor models. An eight-day course of G-CSF induced significant decreases in BMD as measured by dual x-ray absorptiometry (DEXA) compared to vehicle treatment. Quantitative histologic analyses of tibial bones demonstrated an increase in OC perimeter and a decrease in trabecular bone area in G-CSF-treated mice compared to vehicle-treated controls. We found that the G-CSF receptor was expressed on murine bone marrow macrophages and OCs and that G-CSF treatment of cultured OC precursors increased mature OC numbers in vitro. Luciferase-expressing, G-CSF receptor-negative, B161uc murine melanoma cells or 4T1LucGFP murine breast carcinoma cells were directly injected into the tibial bones of mice treated with an eight-day course of G-CSF or vehicle. G-CSF-treated mice demonstrated significantly increased tumor growth as quantitated by in vivo bioluminescence imaging and histologic tumor quantitation compared to vehicle-treated controls. To determine if the effect of G-CSF on OC formation and bone loss played a role in the observed increase in tumor growth in bone, tumors were implanted into the bones of OC-defective osteoprotegerin transgenic (OPGTg) mice. While G-CSF induced neutrophil mobilization, tumor growth was equivalent in bones from G-CSF-treated OPGTg mice compared to vehicle-treated OPGTg mice, strongly implicating the OC as a key player in G-CSF-mediated enhancement of tumor growth. Finally, we found increased markers of bone resorption in pre- and post-menopausal women with stage II and III locally advanced breast cancer treated with four cycles of G-CSF and chemotherapy. Taken together, these data suggest that G-CSF-induced OC activity and bone loss facilitate enhanced tumor growth within the bone marrow cavity and that OC inhibition suppresses this enhancement of tumor growth. Future studies are recommended to evaluate bone loss and subsequent bone metastases in cancer patients receiving G-CSF support with chemotherapy as well as the impact of OC inhibition on these parameters.

Disclosures: A. Hirbe, None.


Runx2 Levels Are Correlated with Osteolytic Disease in Prostate Cancer Cells. J. Akech1, M. van der Deen*1, J. J. Wixted*2, S. Hussain*, K. Bedard*1, A. J. van Wijnen1, J. L. Stein*1, G. S. Stein1, J. Pratap*1, J. B. Lian1. 1Department of Cell Biology and Cancer Center, University of Massachusetts Medical School, Worcester, MA, USA. 2Department of Orthopedics, University of Massachusetts Medical School, Worcester, MA, USA

The Runx2 transcription factor has been shown by several groups to be highly expressed in breast cancer cells metastatic to bone. Importantly, inhibition of Runx2 activity prevents osteolytic disease. The aim of this study is to understand the role of Runx2 in prostate cancer cells metastatic to bone to form either osteoblastic or osteolytic lesions and its role in tumor growth and invasion. Different prostate cell lines (PC3, LNCaP, C4–2B and RWPE) were phenotyped for Runx2, androgen receptor, cell growth and metastatic-related gene expression. The functional role of Runx2 was examined by performing cell migration, invasion and adhesion assays. To determine the effect of Runx2 levels on lesion formation, tibial injections were performed on mice using prostate cancer cell lines. Runx2 and other cancer cell markers (Ki67, cytokeratin) were analyzed by immunohistochemistry in mouse and human tumors.

PC3 cells expressed high levels of Runx2, while LNCaP, C4–2B and RWPE expressed very low Runx2 levels. Two sublines of PC3 cells showed differences in Runx2 levels. The PC3 line expressing high Runx2 (PC3-H) expressed low levels of cyclin D1 and the cell growth inhibitor p57, whereas the PC3 cell line expressing lower Runx2 (PC3-L) expressed high levels of p57 and cyclin Dl. We also observed increased histone H4 promoter activity in PC3-H cells, reflecting stimulated growth in response to forced-Runx2 expression. In vivo, PC3-H cells formed aggressive lytic lesions in 4 weeks, while the PC3-L subline developed into less aggressive lytic lesions. C4-2B cells, which express very low levels of Runx2, showed no osteolytic lesions even at 8 weeks. Thus the more rapid tumor growth in PC3-H cells and the extensive osteolysis correlated with the higher Runx2 expression levels. To assess the clinical significance of these findings, we examined sections of both human primary prostate tumor and secondary tumor in bone and find high Runx2 levels in the Ki67 positive cancer cells. To address Runx2-related mechanisms of tumor growth, we treated PC3-H cells with siRNA to Runx2 and observed downregulation of metastasis-related genes MMP2, MMP13, VEOF and osteocalcin. Runx2 knock-down in PC3 cells consistently showed a decrease in both cell migration and invasion through Matrigel and an increase in cell adhesion on fibronectin. We suggest that Runx2 is a key regulator of genes causing osteolytic disease and prostate tumor growth in the bone microenvironment.

Disclosures: J. Akech, None.


2-Methoxyestradiol (2ME2) Suppresses Bone Metastasis and Osteolysis by Activating TGF-beta Signaling. M. Cicek1, U. T. Iwaniec*2, M. Goblirsch*3, D. R. Clohisy3, R. Turner2, M. J. Pursier1. 1Endocrine Research Unit, Mayo Clinic, Rochester, MN, USA. 2College of Health and Human Sciences, Oregon State University, Corvallis, OR, USA. 3Department of Orthopedic Surgery, University of Minneapolis, Minneapolis, MN, USA

2-Methoxyestradiol (2ME2), a physiological metabolite of estrogen, has recently emerged as a promising cancer therapy. Using a mouse model of osteolytic tumor progression, we sought to determine 2ME2 effects on tumor growth in bone and osteolysis. We generated a 4T1 murine metastatic breast cancer cell line that stably expressed Far Red fluorescence protein (4T1/Red) to visualize tumor development. One thousand 4T1/Red cells were injected into bone marrow of the left femur and the mammary pad of 6 to 7 week-old Balb/c mice. Tumor progression and bone metastasis were monitored using a KODAK IS4000 imaging system. 2ME2 treatment began on the same day as surgery and was administered subcutaneously at 10, 25, and 50 mg/kg doses daily. We observed a dose-dependent suppression of tumor growth with maximum effects at 50 mg/kg/d (see figure; arrows indicate most intense fluorescence). Bone mass was determined by micro-CT, which documented that 2ME2 significantly repressed osteolysis. Interestingly, in addition to osteolytic tumor growth in the injected limb, osteolytic tumors were also observed in the contra-lateral limb. To determine the origins of these metastatic tumors, we injected tumor cells into the bone marrow or the mammary pad of separate mice and found that metastasis to the non-injected limb arose from both bone and soft tissue injections. Since 2ME2 induces tumor cell apoptosis, we examined its effects on tumor cell signaling and observed rapid increases in phosphorylation of MKK4 and INK activation. Moreover, 2ME2 mimicked TGF-beta and activated Smad2/3 translocation to the nucleus and increased TGF-beta Inducible Early Gene (TIEG) expression. TIEG induces tumor cell apoptosis, leading us to hypothesize that 2ME2 stimulates apoptosis by activation of JNK and Smad2/ 3 nuclear translocation and increasing TIEG expression in breast cancer cells. The data suggest that overexpression of TIEG in breast cells would slow tumor development, and tumor dissemination, as well as repress tumor-induced osteolysis.

Thumbnail image of

Disclosures: M. Cicek, None.


Aromatase Inhibition Results in Gain of Bone Mineral Density in the Spine and Femur in Female Nude Mice. W. Kozlow1, K. Mohammad1, C. R. McKenna*1, H. Walton*1, M. Niewolna*1, J. D. Dilley*2, L. J. Suva2, T. A. Guise1. 1Internal Medicine, University of Virginia, Charlottesville, VA, USA. 2Orthopaedic Surgery, University of Arkansas for Medical Sciences, Little Rock, AR, USA

Aromatase inhibitors (AIs), effective treatment for breast cancer, block the conversion of androstenedione and testosterone into estrone and estradiol. Suppression of estrogen leads to increased bone resorption and the loss of bone mineral density (BMD). Therefore, cancer treatment-induced bone loss will likely become one of the most common skeletal complications of malignancy. We hypothesized that the AI letrozole would result in loss of BMD in female nude mice.

Four-week-old female nude mice were treated with letrozole (10 mcg/d) or control. BMD was assessed at baseline and every 2 weeks thereafter. Surprisingly, mice treated with letrozole had increased BMD compared to control at the mid femur (p = 0.0030) and spine (p = 0.0002). There was no difference in BMD between control and letrozole-treated mice at the total body, proximal femur or proximal tibia. MicroCT analysis of the femur after 4 weeks of treatment did not show a significant difference in trabecular bone volume (BV/TV), although a trend toward increased BV/TV in the letrozole-treated mice was observed (p = 0.0659). However, 4 weeks of treatment with letrozole induced marked increases in skeletal microarchitecture. Significant increases in connectivity density (p = 0.0012) and trabecular number (p = 0.0538), thickness (p = 0.0280) and separation (p = 0.0348) were observed in the femurs of letrozole-treated mice, but not in the tibias. Interestingly, these data differ from published data using immunocompetent aromatase null mice, suggesting that differences in T-cell populations in nude mice may account for these distinct effects on bone density and architecture.

In a separate experiment, 4-week-old female nude mice were treated with the bisphosphonate zoledronic acid (ZA) (5 mcg/kg) twice weekly +/− letrozole (10 mcg/d) for 14 weeks. Mice treated with letrozole + ZA had increased BMD at the proximal femur (p <0.0001) and total body (p = 0.0003) compared to ZA alone but, by histomorphometric analysis, bone formation rates were not increased. Similarly, letrozole did not stimulate or inhibit osteoblast number or bone formation in ex-vivo cultures of neonatal mouse calvariae.

In conclusion, letrozole increased BMD at the spine and mid femur and increased trabecular architecture in the femur. This effect, pronounced in the presence of bisphosphonate treatment, was not due to a direct effect of letrozole on bone formation. Unlike in intact immunocompetent mice, letrozole appears to have site-specific effects on the skeletons of nude mice.

Disclosures: W. Kozlow, None.

This study received funding from: DOD W81XWH-05-1-0311 and the V Foundation.


Bone Targeting Peptide Binds to Nucleolin, Inhibits Rac-GTPase Signaling and Prostate Cancer Cell Adhesion. V. Madhu*1, G. Beck*1, P. Huang*1, M. Kagey*1, Q. Cui*1, S. Khosla*1, R. Sikes*2, J. W. Fox*3, G. Balian4. 1Orthopedics Research Lab, University of Virginia, Charlottesville, VA, USA. 2Biological Sciences, University of Delaware, Newark, DE, USA. 3Microbiology, University of Virginia, Charlottesville, VA, USA. 4Orthopaedics, Biochemistry & Molecular Genetics, University of Virginia, Charlottesville, VA, USA

Prostate cancer cells have the propensity to metastasize to bone. We examined the effect of bone targeting peptides isolated from a phage display library and have identified a peptide that binds to bone and inhibits prostate cancer cell adhesion. A peptide, termed L13, was identified from in vivo biopanning. This peptide was synthesized, labeled with biotin, and binding to bone and marrow cryosections was detected with Avidin-FITC. Similarly, L13 binding to the cell surface of endothelial cells was detected. Peptide L13 inhibits the adhesion of prostate cancer cells (LNCaP and C4–2) to human bone marrow endothelial cells in vitro. Cell lysates were fractionated by affinity chromatography on L13 bound to agarose, followed by gel electrophoresis, western blots and sequence analysis. Nucleolin was identified as the molecular target for the peptide. Direct binding of L13 was demonstrated using purified nucleolin. A peptide with similar composition to L13, but with a scrambled sequence, did not inhibit the interaction between L13 and nucleolin. The specificity of this interaction may reside in a five amino acid residue sequence within L13 that corresponds to a sequence near the N-terminal of Hepatocyte Growth Factor/Scatter Factor. Since HGF/SF may be responsible for Rac-GTPase signaling in cancer cells, we tested the effect of L13 on Rac-GTPase and found that L13 specifically inhibits the activation of Rac-GTPase. Our data suggest that nucleolin-HGF/SF interaction and subsequent cell signaling through Rac-GTPase may mediate the adhesion of cancer cells to bone marrow endothelium. L13 interferes with this interaction, thereby preventing adhesion of prostate cancer cells. Structural and functional mimics could serve to identify the potential players in prostate cancer cell adhesion and invasion and may serve as inhibitors of metastasis.

Disclosures: V. Madhu, None.

This study received funding from: Department of Defense.


Administration of the RANK Ligand Inhibitor (OPG-Fc) Enhances the Inhibitory Effect of Docetaxel on the Growth of Prostate Cancer PC-3 Tumors in the Mouse Skeleton. R. Miller*1, M. Roudier1, J. Jones*1, M. Tometsko*1, A. Armstrong*1, J. Canon*2, W. Dougall1. 1Amgen Inc., Seattle, WA, USA. 2Amgen Inc., Thousand Oaks, CA, USA

Bone metastases are a frequent complication of breast and prostate cancer. Tumor cells interact with the bone microenvironment to induce osteoclastogenesis, leading to bone destruction. RANKL is essential for osteoclast formation, function, and survival. RANKL inhibition in models of breast and prostate cancer metastases inhibits tumor-induced osteolysis and reduces the progression of skeletal tumor burden. The chemotherapeutic agent docetaxel has recently been shown to increase survival, reduce pain, and improve quality of life for patients with advanced prostate cancer. Combination therapy using chemotherapies to control tumor growth and RANKL inhibition to suppress tumor-induced osteolysis may provide greater benefit than chemotherapy alone.

We examined the in vivo efficacy of RANKL inhibition by osteoprotegerin-Fc (OPG-Fc) in combination with docetaxel on prevention of tumor-induced osteolysis and reduction in skeletal tumor burden in a mouse model of prostate carcinoma metastasis. Human prostate carcinoma (PC-3) cells were engineered to express firefly luciferase to allow a noninvasive means of tumor burden assessment by bioluminescence. PC-3 cells were injected into the left cardiac ventricle. Beginning 10 days after tumor challenge, mice were treated therapeutically with either OPG-Fc (SC, 3mg/kg, 3x/week) or docetaxel (SC, 5 and 10 mg/kg 1x/week) as single agents, in combination, or with PBS. Tumor progression was monitored twice weekly by bioluminescence, and tumor volume and area were assessed by histology.

OPG-Fc treatment had little effect on the growth rate of PC-3 cells in hind limbs, but it prevented the development of osteolytic bone lesions and reduced tumor volume in bone. Docetaxel significantly reduced the growth rate of PC-3 cells in vivo resulting in a significant decrease in tumor burden by the end of the study (day 27, P <0.0001); the reduction in tumor burden additionally delayed the development of osteolytic bone lesions. The combination of docetaxel and OPG-Fc resulted in a significantly greater suppression of skeletal tumor growth than docetaxel alone (day 27, P = 0.0005). Treatment with either docetaxel or OPG-Fc alone resulted in an approximately 80% reduction in the tumor volume in hind limb bones, whereas combination treatment resulted in tumor reduction of 97.5%.

In summary RANKL inhibition not only effectively inhibits pathologic osteolysis induced by human prostate carcinoma PC-3 cells in animals with established tumors, but also enhances the effectiveness of docetaxel to reduce skeletal tumor burden in vivo.

Disclosures: R. Miller, Amgen Inc. 3.

This study received funding from: Amgen Inc.


Host-Derived RANKL Is Responsible for Osteolysis in C4–2 Human Prostate Cancer Xenograft Model of Experimental Bone Metastases. C. Morrissey1, P. J. Kostenuik*2, L. G. Brown*1, R. L. Vessella*1, E. Corey1. 1Urology, University of Washington, Seattle, WA, USA. 2Amgen, Thousand Oaks, CA, USA

Prostate cancer (CaP) bone metastases are of osteoblastic character, but osteolysis is also increased in patients with advanced CaP, and increased bone resorption is a prognostic factor for skeletal-related events in metastatic CaP. Recently, RANKL has been identified as a potential mediator of cancer-related bone destruction in humans. Numerous animal models of experimental bone metastasis showed that RANKL inhibition prevented osteolysis and reduced skeletal tumor burden. Certain tumors, including CaP, express RANKL. Tumors may also cause osteolysis by inducing host production of RANKL. The relative contribution of tumor- vs. host-derived RANKL has been difficult to establish, even with human xenografts, because murine and human RANKL both stimulate osteolysis in mice. Furthermore, the RANKL inhibitors used to date (OPG and RANK-Fc) inhibit both human and murine RANKL. The objective of our study was to investigate whether RANKL expressed by C4-2 cells is involved in the stimulation of osteolysis associated with growth of these cells in bone or whether the increased osteolysis is caused by murine RANKL.

We have reported that OPG inhibited osteolysis and tumor growth that resulted from the intratibial injection of C4–2 cells in SCID mice. In the current study, we used this model and observed an increase in osteoclast numbers and osteolysis. Immunohistochemistry revealed immunoreactivity of RANKL on C4–2 cells within these experimental bone metastases. The effects of administration of a fully human monoclonal IgG1 antibody against human RANKL (huRANKL MAb) on tumor and bone were evaluated by measurements of serum prostate specific antigen, bone mineral density, bone histomorphometry, serum TRACP 5b and calcium.

The bioactivity of huRANKL MAb was confirmed by its ability to fully inhibit the hypercalcemia associated with the injection of human RANKL in mice. The human specificity of the huRANKL MAb was confirmed by its inability to suppress bone resorption (serum TRAP-5b) in mice, while the pan-species RANKL inhibitor OPG-Fc suppressed serum TRAP-5b by 80%. Administration of huRANKL MAb did not affect the establishment or growth of the C4–2 tumors, nor did it inhibit the osteolytic response of the bone to these cells.

The inability of huRANKL MAb to suppress osteolysis and tumor growth in the bones of mice injected with RANKL-expressing C4–2 human tumor cells suggests that these cells may not express enough RANKL to directly promote osteolysis, and that C4–2 cells elaborate other factors that induce host production of RANKL, which ultimately drives tumor-associated osteolysis.

Disclosures: C. Morrissey, None.


What Do Forearm Ultrasound and DXA Measure in Children? Interpretation Relative to Peripheral QCT (pQCT). D. Buckley*, R. Dosani*, G. Uetrecht*, K. Wosje, H. Kalkwarf. Children's Hospital Med Ctr, Cincinnati, OH, USA

Peripheral QCT is useful for characterizing multiple aspects of bone including separate measures for trabecular and cortical bone compartments, volumetric density and bone geometry. Previous studies comparing quantitative ultrasound measures with DXA measures have found that speed of sound (SOS) is correlated with areal bone mineral density (BMD) at skeletal sites rich in trabecular bone. The relationships between SOS and other characteristics of trabecular and cortical bone, such as volumetric bone density, bone geometry and cortical thickness, are unknown. DXA measurements at the forearm have not previously been compared to pQCT in children. The usefulness of quantitative ultrasound and DXA as surrogates for pQCT-derived measures presumed to be related to overall forearm bone strength is unknown. The objectives of this study were to determine the association between SOS and DXA measures of bone mass (BMC) and density (BMD) with measures of BMC, BMD and bone geometry at the distal radius measured by pQCT. Subjects included 195 children (102 female, 93 male) ages 6 to 16 years; 35 were black, 158 were white. SOS was measured at the distal radius by the Sunlight OmnisenseTM. PQCT measurements were obtained at the 4% and 20% distal radius sites with a Stratec XCT2000. A forearm DXA scan (Hologic QDR 4500) was obtained, and BMC and BMD were determined at the standard ultradistal (UD) and at a 20% distal radius site. Among all pQCT measures, SOS was most strongly related to cortical density, with weaker associations with cortical thickness and cortical BMC. Forearm DXA measures of BMC were highly correlated with pQCT measures of BMC. DXA measures of BMD at the UD site were strongly correlated with total and trabecular BMC and BMD at the 4% pQCT site. DXA BMC at the 20% site was strongly related to SSI and cortical BMC by pQCT. DXA BMD at the 20% site was more strongly related to cortical BMC, cortical thickness, and SSI than to BMD. In conclusion, both ultrasound and DXA performed at the forearm provide a moderately good reflection of cortical density in children. DXA provides a good measure of BMD at the UD site, but weaker measure of BMD at the 20% site. DXA BMC at the 20% is the best single measure of bone strength assessed by pQCT. These findings can be used to identify the best modality for assessing bone health in children. 

Table  . Correlation Coefficients between SOS, DXA and pQCT Measures
  1. *p <0.05, **p <0.001

inline image

Disclosures: D. Buckley, None.


Low BMD Is Associated with Fractures in Patients with Renal Failure: A Meta-Analysis. S. A. Jamal1, J. A. Hayden*2, J. Beyene*3. 1Medicine, University of Toronto, Toronto, ON, Canada. 2Institute for Work and Health, Toronto, ON, Canada. 3Population Health Sciences, Hospital for Sick Children, Toronto, ON, Canada

The association between dual energy X-ray absorptiometry (DXA) measures and fracture in patients with dialysis dependent renal failure (DDRF) is unclear. Studies to date have been limited by small numbers of patients, different methods of classifying fractures, different sites of bone mineral density (BMD) measurement, and failure to consider other important factors (such as age and weight). To determine if DXA measurements at the lumbar spine, femoral neck, or radius were associated with fractures in patients with DDRF we performed a meta-analysis of observational studies. We conducted an electronic search of databases, hand searched relevant journals and screened cited references of studies and reviews.

We included published reports of completed observational studies that enrolled adult hemodialysis patients and measured BMD (reported as absolute values or as T scores) by DXA at the spine, femoral neck, or radius. We assessed outcomes as fractures defined clinically or by morphometry. We synthesized the study results using random effects metaanalysis of standardized mean difference (SMD) in BMD between the fracture and non fracture groups. We tested for heterogeneity using the Cochran Q test. We identified and screened 1774 potentially relevant articles. We retrieved 105 papers for detailed evaluation and included 7 observational studies (N = 869 subjects) in our metaanalysis. The studies reported on 115 morphometric spine fractures 79 clinical fractures. BMD was measured at the femoral neck, lumbar spine, ultradistal radius, 1/3 radius, and midradius. Across the studies the mean patient age ranged from 54.2 ± 8.8 years to 69.4 ± 9.1 years and duration of dialysis ranged from 36.8 ± 3.1 months to 144.4 ± 80.2 months. We found that for all BMD sites, except the femoral neck, subjects with fractures had significantly lower BMD than subjects without fractures. For example, 5 studies examined association between lumbar spine BMD and fracture; the SMD was −0.45 (95% CI: −0.81 to −0.08). Two studies assessed the association between mid-radius BMD and fracture [-0.69 (95% CI: −1.10 to −0.29)]. There was important heterogeneity in the association between BMD and fracture at most measurement sites.

Our meta-analysis suggests that BMD is lower at the spine, and radius in patients with DDRF who have fractures compared to those who do not. However, given the observed heterogeneity, we recommend larger studies with more consistent methods.

Disclosures: S.A. Jamal, None.

This study received funding from: Canadian Institutes of Health Research.


Standardization of Pediatric Spine, Femur, and Total Body Bone Mineral Density for Lunar Prodigy and QDR Discovery Densitometers. M. A. Levine1, H. S. Barden2, B. Richmond*3, W. K. Wacker*2, L. S. Weynand*2. 1Pediatric Endocrinology, Cleveland Clinic Children's Hospital, Cleveland, OH, USA. 2GE Healthcare, Madison, WI, USA. 3Radiology, Cleveland Clinic, Cleveland, OH, USA

Bone mineral density (BMD) measurements using dual-energy x-ray absorptiometry (DXA) are used increasingly to assess bone status in pediatric subjects. Results may vary among manufacturers due to differences in regions-of-interest and analysis algorithms. Several reports describe standardization formulae for spine and femur BMD in adults (1–2), but we are not aware of standardization formulae for BMD specifically derived from pediatric data. In this study we measured children age 3 to 15 years on both the Lunar Prodigy (GE Healthcare) and the QDR Discovery A (Hologic) densitometers to establish standardization of BMD at the spine (L1-L4), femur neck, total femur, and total body. We report preliminary results for 44 subjects (23 girls, 21 boys) measured on both instruments. Subjects averaged (SD) 11.7 (2.7) yrs of age, 150.5 (17.5) cm in height and 48.3 (18.5) kg in weight. Densitometry results were compared using Deming regression, taking into account the precision errors of both systems. Average absolute differences of actual Prodigy vs. predicted Prodigy BMD values for L1-L4, femur neck, total femur and total body ranged from 0.021 g/cm2 to 0.030 g/cm2 Preliminary results show strong correlations and adequate fits between results from Prodigy and QDR densitometers for lumbar spine, hip (right figure) and total body (left figure). 

Table  . Prodigy and QDR Cross-calibration for Spine, Femur and Total Body
inline image
Thumbnail image of
  • Genant HK et al., JBMR, 1994; 10:1503–1514

  • Lu Y et al., Osteoporos Int, 2001; 12:438–444

Disclosures: M.A. Levine, GE Healthcare 2.

This study received funding from: GE Healthcare.


Predictors of Repeat BMD Testing: Results from the National Osteoporosis Risk Assessment Program (NORA). P. D. Miller1, C. A. McHorney2, E. Barrett-Connor3, E. S. Siris4. 1University of Colorado Health Sciences, Colorado Center for Bone Research, Lakewood, CO, USA. 2Outcomes Research & Management, Merck & Co., Inc., West Point, PA, USA. 3Family and Preventive Medicine, UC at San Diego, La Jolla, CA, USA. 4Columbia University, College of Physicians and Surgeons, New York, NY, USA

Guidelines for repeat BMD testing are lacking. We study patterns of repeat BMDs among a cohort of postmenopausal women (PMW). From 1997–1999, NORA enrolled over 200,000 women from 49 US states and DC in a longitudinal registry of PMW. Eligibility criteria were: 50 or older, at least six months post last menses, no prior diagnosis of osteoporosis, no BMD testing in the preceding year, and no specific osteoporosis medication use. Baseline BMD testing at a peripheral site was performed. At the Year 1, 2, and 5 surveys, women reported repeat BMD testing. Stepwise logistic regression modeled predictors of repeat BMD within two and five years of the baseline BMD. We used baseline demographic and osteoporosis risk variables as independent variables. The unadjusted rate of repeat BMD testing was 29.5% up to two years and 58.4% up to five years post-baseline. The adjusted five-year rate of repeat BMD testing for patients with baseline T-score >-1.0 (normal) was 54.4% vs. 63.4% for patients with baseline T-score <−1.0 to −2.49 (“osteopenia”) vs. 62.2% for patients with baseline T-score <−2.5 (“osteoporosis”). Because two-year results were virtually identical to the five-year results, we report only the latter below. The top-five variables which increased the odds of repeat BMD testing were: (1) taking osteoporosis medications at anytime during the five-year follow-up (OR:3.22, 0:3.07–3.37); (2) talking with one's physician about baseline BMD results (OR:1.41, CI:1.36–1.47); (3) taking thyroid medications at anytime during NORA (OR:1.16, CI: 1.10–1.22); (4) T-score <−1.0 to −2.49 (OR:1.12, CI: 1.08–1.17); and (5) low body weight (OR:l.l4, 0:1.08–1.20). The top-five variables which decreased the odds of repeat BMD testing were: (1) poor/fair (vs. excellent) self-rated health (OR: 0.69,0:0.64–0.74); (2) no history of HRT therapy (OR:0.74, 0:0.71–0.77); (3) age 85+ (vs. < 65) (OR:0.28, 0:0.22–0.36); (4) less than high school education (vs. college graduate) (OR:0.71, CI:0.66–0.75); and (5) current smoker (OR:0.78, CI: 0.74–0.82). Among NORA participants, 58% had a repeat BMD within five years of a baseline BMD. Rates for women with baseline T-scores <−2.5 were only 62%. These data underscore the need for clinical guidelines for repeat BMD testing in order to help both PMW and their doctors make appropriate medical decisions about osteoporosis prevention and management.

Disclosures: P.D. Miller, Merck &Co., Inc. 5; F. Hoffman-La-Roche 2, 5;

GlaxcoSmithKline 2, 5.

This study received funding from: Merck & Co., Inc.


EFFECT - The European Femur Fracture Study Using Finite Element Analysis and 3D Computed Tomography. K. Engelke1, V. Bousson*2, L. Duchemin*3, C. Fuchs*1, D. Mitton*3, A. Mastmeyer*1, J. Adams4, W. Kalender*1, W. Skalli*3, J. D. Laredo*2. 1Inst. of Med. Physics, University of Erlangen, Erlangen, Germany. 2Hǒp. Lariboisière, AP-HP, CNRS-UMR 7052, Paris, France. 3Lab. de Biomécanique, ENSAM-CNRS UMR 8005, Paris, France. 4Clin. Rad., Manchester University, Manchester, United Kingdom

This project is an integrated framework supported by the European Community to improve fracture prediction using finite element analysis (FEA) and 3D quantitative computed tomography (QCT) of the proximal femur.

The framework includes optimization of CT scan protocols, automated 3D segmentation of the proximal femur, and definition of analysis volumes to determine bone mineral density (BMD), -content (BMC), cortical thickness, and parameters characterizing the geometry of the proximal femur. It further includes an interface with FEA to generate models for the individual prediction of fracture risk. Validation is performed using the semi-anthropometric European proximal femur phantom (EPFP) designed within this study, in-vitro scans of 80 excised femora (subject age 82 ± 11) that were mechanically tested until fracture in stance and lateral configurations, and a controlled ongoing in-vivo study in 120 elderly females to determine fracture discrimination.

The following CT protocol is a good compromise between radiation exposure and image quality: 120 kV, 170 mAs, pitch 1, slice thickness 1 mm. Range: from above the femoral head to 1 cm below the lesser trochanter. A 150 mm field of view using B40s (Siemens) or standard (GE) kernels was reconstructed. Accuracy errors of the QCT analysis between nominal and measured values of the EPFP are displayed in the table. The errors for cortical BMD in the neck and the trochanter are probably due to partial volume artifacts caused by the thin cortex.

Mean fracture load of 40 excised femora tested in stance (lateral) configuration was 9032 ± 3432 N (2568 ± 1146 N). Correlation between fracture load in stance and lateral configuration was r = 0.87. FEA analysis was completed on 40 cadavers tested in stance configuration using individual QCT BMD values. FE-computed fracture loads were significantly correlated with experimental results (slope 1.006, r2 = 0.87, p <0.05). The mean prediction error was not significantly different from 0. The FE model predicted bone strength with a SEE of 1220 N, resulting in a precision of ± 2600N, with 95% confidence. 

Table  .  
inline image

Disclosures: K. Engelke, None.


Assessment of Intra-individual Variability of 3D Bone Microarchitecture by Conventional Micro-Computed Tomography in Paired Iliac Crest Biopsies. C. Chappard, A. Marchadier*, A. Basillais, L. Benhamou. U 658 Inserm, Inserm, Orleans, France

Bone microarchitecture of the iliac bone was assessed to characterize the pathogenesis of osteoporosis in patients. Paired biopsies were used to examine the changes over time under treatments. New techniques to analyse bone tri-dimensional (3D) microarchitecture based on micro-computed tomography (μCT) are now available.

Aim of this study was to measure intra-individual variability of 3D microarchitectural parameters of trabecular bone in paired iliac crest biopsies.

Paired bone biopsies were carried out using a Bordier needle trephine system in female human cadaver. Specimens were sampled in 18 women aged 59 to 93 years (mean age 76.2 ± 10.1 years). All the samples were chemically defatted. A μ-CT Skyscan® 1072 with a conventional x-ray source was used. The system was set at 80kV and 100μA with 400 radiographic projections with an exposure time of 5.8 s and a frame averaging of 2 images. Voxel size was 10.77 μm.

For each pair, we used an adjusted global threshold set at the minimum value between peaks corresponding to bone and background. A Volume of Interest (VOI) located in the middle of each sample corresponding to 75% of the iliac crest height was analyzed. The same VOI was analyzed for each pair: mean ± SD: 100.3 ± 18.6 mm3, (min-max: 67–131). Bone Volume/Tissue Volume (BV/TV %) corresponding to the bone proportion was measured. Trabecular Thickness and Trabecular Spacing (Tb.Th, Tb.Sp μm) were evaluated with the technique described by Hildebrand. Bone Surface (BS) was calculated using surface triangulation of the binary data based on the marching cubes method. The Trabecular Number (Tb.N, mm−1) was derived from BV/TV and Tb.Th according to the plate model.

Comparison of the morphological bone parameters between right and left side were made with Pearson correlation coefficients, paired t test. The average absolute percentages of deviation were also calculated. 

Table  .  
inline image

An important intra-individual variability emerged for most microarchitectural parameters of iliac bone crest which makes it difficult to show significant differences in longitudinal studies based on small sample sizes. These results indicate that before performing comparison, high stringent quality criteria of μ-CT acquisition are necessary to obtain the best signal to noise ratio and superimposed bimodal histograms.

Disclosures: C. Chappard, None.


Relationship between Teriparatide Effects on Trabecular Bone Architecture and Bone Strength in Ovariectomized Monkeys. P. Chen*1, C. Jerome2, D. Burr3, C. Turner3, M. Sato1. 1Lilly Research Laboratories, Indianapolis, IN, USA. 2Skeletech, Inc., Bothell, WA, USA. 3Indiana University School of Medicine, Indianapolis, IN, USA

Bone strength is determined by bone mineral density (BMD) and other elements of bone quality, including bone spatial architecture. Teriparatide (TPTD) treatment increases BMD and improves both trabecular and cortical bone architecture. Increases in lumbar spine BMD were shown to account for approximately 30 to 41% of the vertebral fracture risk reduction with TPTD treatment of women with osteoporosis. The advantage of monkeys is that we can directly quantify bone strength; therefore we sought to clarify the relationship between trabecular bone spatial architecture and bone strength in clinically important sites. Ovariectomized monkeys were administered vehicle (n = 20) or TPTD at 1.0 μg/kg/day (n = 19) and 5.0 μg/kg/day (n = 21) for 18 months. Iliac crest biopsies were obtained at 15 months following the initiation of treatment. Animals were euthanized after 18 months' treatment and adjacent or contralateral bones were processed for biomechanical or histomorphometric analyses for each monkey. Pearson correlation analyses were performed to assess the relationship between biomechanical and static histomorphometric parameters of lumbar vertebra, femoral neck and iliac crest. Significant correlations between improvements in trabecular architecture and bone strength were seen at both lumbar vertebra and femoral neck after 18 months. Interestingly, trabecular bone architecture of the iliac crest correlated as well with vertebral and femoral neck strength, as these strengths did with their own histomorphometric parameters. 

Table  .  
inline image

These results show that improvements in the trabecular bone architecture are significantly correlated with increased bone strength observed with TPTD treatment for vertebra and femoral neck. Additionally, trabecular architecture measured in the iliac crest at 15 months is a good surrogate to estimate strength of the vertebra and femoral neck.

Disclosures: P. Chen, Lilly Research Laboratories 1, 3.

This study received funding from: Eli Lilly and Company.


Disagreement in Hip Strength Analysis Results between GE Lunar and Hologic DXA Systems. K. G. Faulkner1, J. A. Shepherd2, K. W. Gaither*1, B. Fan*2, E. M. Lewiecki3, P. D. Miller4, H. K. Genant2. 1Synarc, San Francisco, CA, USA. 2University of California, San Francisco, CA, USA. 3New Mexico Clinical Research & Osteoporosis Center, Albuquerque, NM, USA. 4Colorado Center for Bone Research, Denver, CO, USA

Hip Strength Analysis (HSA, Johns Hopkins University) estimates biomechanical parameters of the proximal femur from a DXA image. These parameters are determined from bone size and density data by post processing DXA scan files. HSA analysis software can be used with both GE Lunar and Hologic systems. However, GE Lunar and Hologic have known differences in calibration and image magnification, which can influence the HSA results. In this study, we compared HSA measurements obtained from the same patients on different DXA systems.

DXA scans of both hips were obtained from 91 women. Subjects were measured on both Lunar Prodigy (GE Healthcare) and Delphi (Hologic) systems and analyzed with software versions 9.3 and 12.1, respectively. All scans were acquired by experienced technologists using manufacturer recommended procedures. HSA measures of cross sectional area, cross sectional moment of inertia, section modulus and endosteal diameter were determined using version 1.3 of the HSA program. Values were compared across DXA systems using linear regression analysis at the femoral neck, trochanter, and femoral shaft. Poor to moderate correlations were found between GE Lunar and Hologic results. R-squared (R2) values ranged from 0.20 to 0.45 at the neck, 0.63 to 0.72 at the trochanter, and 0.35 to 0.77 at the shaft. Regression slopes were significantly different from unity for all HSA variables, ranging from 0.34 to 1.38. Significant regression intercepts were found for all neck variables and for the right shaft measurements.

We conclude that HSA measurements obtained on GE Lunar and Hologic DXA systems do not agree and are at best moderately correlated. Differences may be related to calibration, variable magnification effects, or variations in how the HSA program analyzes data from different DXA systems. Since bone density results from the Prodigy and Delphi are highly correlated, differences in the HSA conversion routines are the likely cause for the discrepancy. At this point, HSA measurements obtained on different DXA systems should not be compared. Additional validation studies are needed to determine the accuracy and precision of the HSA program. 

Table  . Correlations (r-square) between GE and Hologic HSA measurements
inline image

Disclosures: K.G. Faulkner, None.


Bisphosphonates Improve Bone Microarchitecture in Middle-Aged Males with Osteoporosis by Reducing Bone Turnover: A Paired Biopsy Micro-CT Analysis over 38 Months. B. Jobke*1, B. Kulle*2, J. Semler3, G. Delling1. 1Bone Pathology, University Hospital Hamburg-Eppendorf, Hamburg, Germany. 2Biostatistics, Univ. of Oslo, Oslo, Norway. 3Osteology, Immanuel-Hospital, Berlin, Germany

A reduction in the fracture rate and serum-markers of bone turnover in patients with osteoporosis (OP) receiving antiresorptive bisphosphonate (BP) therapy have been clinically proven in several studies. The aim of our retrospective analysis of paired iliac crest bone biopsy specimens from patients with OP was to investigate changes in trabecular microarchitecture and bone turnover with and without BP treatment. Retrospectively we examined 110 iliac crest biopsies from 73 patients divided into three groups. 3-D micro-CT (m-CT 20, Scanco) analysis was performed on biopsies from 23 OP patients (average age 52.6y; 18 male, 5 female) before and after 38 months of BP treatment. We further investigated microarchitecture and histological bone turnover in paired biopsies of 14 OP patients (average age 47.5y) without treatment (mean time 79 months, SD 49.6 months) in comparison to 36 control patients with “normal” bone volume (average age 49.5y).

M-CT analysis showed a mean increase in BV/TV from baseline by 1.7% per year in the treatment group. After 38 months, BV/TV had increased by 5.2% (p <0.0001). This was accompanied by an improvement in structural parameters (Conn.D, SMI, Tb.N., Tb.Th., Tb.Sp., TBPF). By comparison, 71% of the patients in the untreated group had a decrease in all trabecular bone parameters and a mean BV/TV loss of 0.7% per year. Annual changes were independent of age and gender in both groups. All patients in the “normal” control group had a BV/TV above 20%. Mean BW/TV was 26.5%, 60.8% (14/23) of the patients with a high turnover (HTO) in the treatment group had a reduction in bone turnover after 38 months. Only 35.7% (5/14) in the untreated group show a reduction of bone turnover after 79 months from baseline. 74% in the “normal” control group had a normal and balanced bone turnover.

Antiresorptive therapy with BP over more than 3 years in osteoporotic male patients contributed to a significant increase in bone volume and trabecular architecture as measured by m-CT. The results demonstrate that male OP patients equally benefit from antiresorptive therapy consistent with clinical data from DEXA measurements. Bone turnover is reduced in the majority of the patients with a HTO at baseline. Untreated patients who experience no reduction in bone turnover lose 0.7% of their bone mass per year. Despite the fact that normal BV/TV up to 26% are not restored by antiresorptive therapy, a critical bone mass is built up, preventing insufficient biomechanical stability that lead to fractures.

Disclosures: B. Jobke, None.


Load Distribution and the Relation to Morphological Indices in the Distal Radius by High Resolution Peripheral Quantitative Computed Tomography. J. A. MacNeil*, S. K. Boyd. Mechanical and Manufacturing Engineering, Schulich School of Engineering, University of Calgary, Calgary, AB, Canada

High resolution peripheral quantitative computed tomography (HR-pQCT) has recently been developed for patient measurements of the distal radius and tibia (XtremeCT, Scanco Medical, Bassersdorf, Switzerland). The nominal isotropic resolution is 82 μm which provides sufficient resolution to distinguish individual trabecular structures, thus allowing morphological analyses and estimates of mechanical strength. The purpose of this study was to determine the relative load carried by the trabecular and cortical bone in the distal radius, and the correlation between force and bone morphological indices.

10 cadaver forearms and 15 human subjects were scanned with HR-pQCT using the standard patient protocol (82 μm). The radii were segmented from the images using a semi-automated contouring method and the cortical bone was segmented from the trabecular bone using an automatic segmentation method. The voxels were automatically converted to finite elements using the voxel conversion approach. The proximal and distal surface element ids of the cortical and trabecular bone were obtained from the automatic segmentation. The boundary conditions for the finite element analyses were uniaxial compression to 1% strain. A homogeneous, isotropic material definition used for all elements. A standard patient protocol was used for the morphological analysis.

The finite element analysis and morphological analyses have been completed. It was found the trabecular bone on average carries 73% of the load at the distal end of the sample (27% carried by the cortical bone), and the cortical carries on average 74% of the load at the proximal end of the sample (26% carried by the trabecular bone). Initial results also indicate there is a range of 48 to 82 percent correlation between morphological indices such as bone volume ratio, trabecular thickness, trabecular separation and trabecular number and force to induce 1% strain. The highest correlation was found between load carried by the cortical bone and mean cortical thickness (90% to 92%).

In conclusion, the relative load carried by the trabecular and cortical bone regions, determined for a standard radius sample, differed from the distal to proximal ends. This variation is likely due to the changing cross-sectional geometry of the cortical and trabecular compartments. Furthermore, high correlations between load carrying capacity and morphometric parameters provide an optimistic possibility to assess strength changes in clinical research investigating the development of diseases such as osteoporosis, and the efficacy of interventions to arrest or reverse bone loss.

Disclosures: J.A. MacNeil, None.


Long Term Use of 1–84 PTH/7–84 PTH Ratio Avoids Adynamic Bone Disease in African American Hemodialysis Patients. T. L. Cantor1, H. Fehmi*2, O. Daramola*3, Y. Osman*3, K. Nguyen*3, J. Yee*4. 1President, Scantibodies Laboratory, Inc., Santee, CA, USA. 2Division of Nephrology, Henry Ford Hospital., Detroit, MI, USA. 3Nephrology, Henry Ford Hospital, Detroit, MI, USA. 4Div of Nephrology, Henry Ford Hospital, Detroit, MI, USA

Secondary hyperparathyroidism most often results in one of two bone disorders in patients with chronic renal insufficiency (CRI) which are high bone turnover disease (aka hyperparathyroid bone) or adynamic bone disease (ABD). Using bone biopsies and arterial calcification scoring, London et al. found that arterial calcification (and not hyperparathyroid bone) is currently associated with adynamic bone disease in renal patients. Calcification is strongly associated with mortality for these patients. ABD is iatrogenic, resulting from over-suppression with I.V. vitamin D analogs and calcium supplements. Prior to 1987, ABD was likely never observed in CRI. Today, about 50% of dialysis patients have ABD. Therefore, it is of great interest to have PTH tests that will result in a decrease in ABD when used to guide I.V. vitamin D analog therapy. Sawaya, et al. have demonstrated that iPTH results have no correlation with bone status in African American (AA) hemodialysis (HD) patients. AA patients make up 30% of all HD patients in the US.

In our study, two groups of AA HD patients had their I.V. vitamin D analog therapy guided by either the Bio-Intact PTH assay (Bi-PTH) or the 1–84 PTH/7–84 PTH ratio. We bone biopsied AA HD patients with iPTH > 100 pgm/mL) from these two groups following three years of using the Bi-PTH assay to guide therapy in group 1 (17 pts) and the 1–84 PTH/7–84 PTH ratio in group 2 (20 pts). The patients were comparable for age, gender, I.V. vitamin D therapy (90-95% of pts in both groups), vintage and diabetic state. Both groups used the same I.V. vitamin D analog (Abbott's Zemplar®).

Use of Bi-PTH for three years resulted in an average dose of 5.5 micrograms per HD session; and use of the 1–84 PTH/7–84 PTH ratio resulted in an average dose of 2.5 micrograms per HD session. The group guided by the Bi-PTH assay had 7 pts with ABD (41%), whereas, the group guided by the ratio had NO patient with adynamic bone disease (0%).

These findings suggest that long term use of the 1–84 PTH/7–84 PTH ratio can avoid ABD resulting from overuse of I.V. vitamin D analogs.

Disclosures: T.L. Cantor, Scantibodies Laboratory, Inc. 4.


High-Throughput DEXA and MicroCT Analyses of Bone in Gene Knockout Mice to Identify Novel Osteoporosis Drug Targets. R. Brommage1, Z. Z. Shi*2, J. Liu*1, L. Minze*2, L. Freay*2, J. Trackey*2, S. Wray*2, K. Combs*2, D. R. Powell*'. 1Endocrinology, Lexicon Genetics, The Woodlands, TX, USA. 2Target Validation, Lexicon Genetics, The Woodlands, TX, USA

Screening gene function in vivo is a powerful approach to discover potentially novel drug targets in the human genome (Zambrowicz and Sands, Nature Review Drug Discovery 2:38-51; 2003). We present data on the first 2000 distinct gene knockout mouse lines with viable adult homozygous mice generated at Lexicon using both gene-trapping and homologous recombination technologies. Knockout of additional genes results in embryonic or postnatal lethality. Mice are F2 crosses from C57BL/6J and 129 SvEv parental strains. Peak bone mass is determined from PIXImus DEXA scans of male and female mice (except for 42 lines with X-linked genes in which only male knockouts are generated) at ñ14 weeks of age and by microCT of bones from male mice at ñ16 weeks of age.

For most lines, initial DEXA scans are performed on 4 knockout and 2 wild-type (WT) control mice of each gender. Body BMC, body areal BMD, body volumetric BMD, body BMC/LBM ratio, femur BMD and spine BMD are monitored. Bone parameters are normally distributed, with a slight excess of low bone mass values resulting from small mice with poor health. Areal BMD in knockouts averages 99.1% of control values with a standard deviation (SD) of 4.9%. Volumetric BMD in knockouts averages 99.5% of control values with a SD of 3.8%.

Using a Scanco μCT40, trabecular bone parameters in the fifth lumbar vertebrae (LV5) have been analyzed in 1622 lines and midfemur cortical thickness (CT) in 1593 lines. LV5s are scanned at 16 μm voxel size and femurs at 20 μm voxel size. Specially-designed plastic inserts hold 48 LV5s for overnight scanning and allow 4 LV5s to be scanned simultaneously. Separate inserts hold 18 femurs, allowing 6 femurs to be scanned simultaneously in 10 minutes. LV5 trabecular bone volume (TBV) in knockouts averages 17.2 % with a SD of 3.9%. Femoral CT averages 247 μ with a SD of 16 μm. Knockout lines with high bone mass are subjected to additional analyses employing additional mice. Confirmed high bone mass knockout lines without side-effects provide in vivo-validated targets for osteoporosis drug development.

These high throughput screens successfully identified two genes known to be important in bone metabolism. LRP5 knockouts have low bone mass (vBMD = 89% of WT, CT = 226 μ) and klotho knockouts have high vBMD (119% of WT) and high TBV (43%) but low CT (165 μm). Knockout of the amino-bisphosphonate target farnesyl diphosphate synthase is embryonic lethal. Mice with knockout of GPR54 do not undergo puberty, have low circulating levels of androgens and estrogens, and have low bone mass by both DEXA and microCT.

Disclosures: R. Brommage, Lexicon Genetics Incorporated 3.

This study received funding from: Lexicon Genetics.


A Gain-of-Function Mutation in the Cytochrome P450 3A4 Gene Is Associated with Low Bone Mineral Density. Y. Rang1, S. Lee*1, H. Kwak*1, M. Jung*1, C. Hwang*1, S. Park*2, C. Yim*2, H. Yoon2, K. Han*2. 1Laboratory of Endocrinology, Cheil General Hospital and Women's Healthcare Center, Sungkyunkwan University School of Medicine, seoul, Republic of Korea. 2Department of Medicine, Cheil General Hospital and Women's Healthcare Center, Sungkyunkwan University School of Medicine, seoul, Republic of Korea

Osteoporosis is a disease that is strongly influenced by genetic factors. Low bone mass and high bone turnover, the two important risk factors of osteoporosis, are highly related with sex hormones. Marked interindividual variability in catalytic activity of the human cytochrome P450 (CYP) 3As, the metabolic enzyme of a wide range of endobiotics including sex hormones, may influence the risk of osteoporosis.

To determine the presence of SNPs in the CYP3A4 gene, the 596 bp 5′-upstream and all 13 exons in CYP3A4 gene were amplified and sequenced in DNA from 225 unrelated Koreans. Our screening of whole exons detects only two SNPs located in exon 10; a T[RIGHTWARDS ARROW]C transition mutation which causes proline to be substituted for leucine at codon 293 (L293P) and L295L, a silent novel mutation. In 2,089 healthy postmenopausal women, the L293P variant was analyzed by PCR-RFLP method. BMD at lumbar spine and femur was measured by DXA using a QDR-2000. The cDNA for the variant allele was constructed by site directed mutagenesis and expressed in a baculovirus expression system. Metabolism of testosterone and estrogen by the recombinant wild type and mutant CYP3A4 alleles was assessed by using LCMS/MS method.

The allele frequency of L293P was 2.1%; 42 women were heterozygotes and only one woman was homozygote. Women with L293P were significantly lower in BMD measured at both lumbar spine (P <0.01) and total hip (P <0.001) than the age and BMI matched control women. Testosterone 6β and estrone 16α-hydroxylation by L293P variant were increased compared with wild type (P <0.05). In vitro functional studies indicate that L293P is a gain-of-function mutation, resulting in rapid metabolism of the substrates of CYP3A4.

In conclusion, our study shows that the L293P variant increases the enzyme activity of CYP3A4, the major metabolizing enzyme of sex steroids including estrogen and testosterone. These results suggest that postmenopausal women who carry a gain-of-function mutation of CYP3A4 gene may have a risk of osteoporosis.

Disclosures: Y. Kang, None.


Distal Chromosome 1 Orchestrates Gender and Genes to Regulate Bone Phenotypes. W. G. Beamer, K. L. Shultz, C. L. Ackert-Bicknell, L. G. Horton, K. M. Delahunty, H. F. Coombs*, V. E. DeMambro, L. R. Donahue, C. J. Rosen. The Jackson Laboratory, Bar Harbor, ME, USA

Skeletal bone mineral density (BMD) is a complex, heritable trait in humans and mice. Using mouse B6C3H-F2 intercross progeny and large segment B6.C3H congenic strains, we found Quantitative Trait Loci (QTL) on Chromosomes (Chrs) 1, 4, and 18 account for the largest variation in femoral volumetric (v)BMD between B6 (low vBMD) and C3H (high vBMD) mice. The purpose of this investigation was to positionally clone the vBMD QTL located on distal mouse Chr 1 - a chromosomal region with genetic homology to human 1q that also associates with human skeletal regulation. Our approach exploited a set of 9 nested B6.C3H congenic sublines carrying overlapping C3H segments in the B6 genetic background. These segments vary from 1 to 8 megabases (Mb) in size and are located between 169 and 177 Mb. Isolated femurs from groups (n = 15–40) of 16 wk females and males from congenics and B6 controls were measured by pQCT and by Micro CT40 densitometry. Density, microstructure and size phenotypes were tested for association with C3H alleles present in the 9 congenics. We found females from 6 of 9 congenics increased vBMD (2.43 to 6.25%; p <0.0001), as well as mid diaphyseal periosteal circumference and cortical thickness - thus mapping the VBMD phenotype (QTL-1) to a 0.26 Mb region between genetic markers for Dfy and Ifi202b. These pQCT phenotypes were unchanged or significantly decreased (p <0.0001) in congenic males. In contrast, the distal femoral Micro CT40 data for both females and males showed increased BV/TV (11 to 55%; p <0.0001) in the same 6 congenics showing increased female vBMD. The Micro CT40 data showed a second QTL for trabecular BV/TV and thickness specific for males flanked by SNP rs3675022 and D1 Mit149. This 2.83 Mb QTL-2 is closely linked with and proximal to QTL-1.

Expression studies for QTL-1 (1 known, 2 novel; Ensembl Bldg 37) and QTL-2 (56 known, 9 novel) have been undertaken by Real Time PCR to search for candidate genes. Whole bone RNA from 8 wk B6.C3H-1–11 and B6 females and males was chosen to cast a broader net than evaluating marrow or bone lining cells alone. QTL-1 genes have yet to reveal significant expression differences by genotype, whereas QTL-2 has shown 1 of 42 genes examined with significant genotype specific expression in congenic male, but not in female, bone. Thus, congenic sublines revealed two distinct bone QTL with compartment and gender specific roles: a) QTL-1 increased vBMD in females only, whereas BV/TV increased in both sexes; and b) QTL-2 did not affect vBMD in either sex, but increased BV/TV in males. Complex gene expression profiles for these QTL have not readily matched biological phenotypic differences observed in the congenic strains.

Disclosures: W.G. Beamer, None.

This study received funding from: NIH AR43618, NIH AR45433.


Collaborative Genome Search Meta-Analysis for Bone Mineral Density. S. H. Ralston1, J. Ioannidis*2, M. Ng*3, P. Sham*3, E. Zintzaras*4, H. Deng5, E. J. Michael6, D. Karasik7, M. Devoto*8, C. Kammerer*9, T. Spector10, A. Cupples*11, T. Foroud6, D. Kiel7, D. L. Koller6, L. Bente12, B. Mitchell13, M. Peacock6, R. Recker5, H. Shen*13, K. Sol-Church*14, L. Spotila15, A. G. Uitterlinden16, S. G. Wilson*17, A. W. C. Kung3. 1Rheumatic Diseases Unit, University of Edinburgh, Edinburgh, United Kingdom. 2University of Ioannina, Ioannina, Greece. 3University of Hong Kong, Hong Kong, Hong Kong 4University of Thessaly, Thessaly, Greece. 5Creighton University, Omaha, NE, USA. 6University of Indiana, Indianapolis, IN, USA. 7Harvard University, Boston, MA, USA. 8CHOP, Pittsburgh, PA, USA. 9University of Pittsburgh, Pittsburgh, PA, USA. 10St Thomas Hospital, London, United Kingdom. 11Boston University, Boston, MA, USA. 12Aarhus University Hospital, Aarhus, Denmark. 13University of Maryland, Baltimore, MD, USA. 14CHOP, Philadelphia, PA, USA. 15Sciencescribe, Haddonfield, NJ, USA. 16Erasmus University, Rotterdam, The Netherlands. 17Sir Charles Gairdner Hospital, Nedlands, Australia.

Bone mineral density (BMD) is an important predictor of osteoporotic fracture risk. Several genome wide linkage scans (GWS) have been performed to detect loci that regulate BMD but there has been limited replication between studies. In an attempt to resolve these inconsistencies we conducted a meta-analysis of GWS in which femoral neck (FN) or lumbar spine (LS) BMD had been studied. Data were accumulated from 9 GWS involving 11,842 subjects and were analysed separately for LS-BMD and FN-BMD and for males and females. For each study, genomic bins of 30cM were defined and ranked according to the maximum LOD score they contained. Significance for high average rank and heterogeneity was obtained through Monte Carlo testing. The strongest evidence for linkage to LS-BMD was on chromosome Ip13.3-q23.3 (p = 0.004). Other loci with significant evidence of LS-BMD linkage (p <0.05 or less) were found on chromosomes 12q24.31-qter, 3p25.3-p22.1, Hp12-q13.3, 1q32-q42.3 and 18p11-q12.3. For FN-BMD the strongest locus was on chromosome 9q31.1-q33.3 (p = 0.002) but other loci with significant evidence of linkage were detected on chromosomes 17p12-q21.33, 14q13.1-q24.I, 9q21.32-q31.1 and 5q14.3-q23.2. The loci that regulated BMD in men and women and at different sites were largely distinct. This meta-analysis has identified several loci that regulate BMD in a site- and gender-specific manner and many more regions were identified than would be expected by chance. Despite the large sample size however, no individual locus reached the threshold for genome wide significance (p <0.00042), emphasising the continuing challenge faced by investigators who seek to identify genes that predispose to complex traits such as osteoporosis.

Disclosures: S.H. Ralston, Novartis 2; Novartis 5: Proctor & Gamble 5.


The Association between Physical Activity and BMD Is Modulated by Catechol-O-Methyltransferase (COMT) Genotype - The GOOD Study. M. Lorentzon, A. Eriksson, P. Mellström, C. Ohlsson. Internal Medicine, Center for Bone Research at the Sahlgrenska Academy (CBS), Göteborg, Sweden

The gene coding for the estrogen degrading enzyme Catechol-O-Methyltransferase (COMT) contains a functional G to A polymorphism, which results in a valine to methionine substitution at codon 158. This amino acid replacement leads to a 60–75% difference in enzyme activity between the Val (high activity) and Met (low activity) variants. We have previously shown that both physical activity (PA) and the COMT polymorphism are associated with areal bone mineral density (aBMD) and trabecular volumetric BMD (vBMD) in young men. The aim of the present study was to determine if the COMT genotype modulates the impact of PA on BMD

The Gothenburg Osteoporosis and Obesity Determinants (GOOD) study is a population based study that consists of 1068 young men, age 18.9 ± 0.6 yrs (mean ± SD). aBMD of the total body, lumbar spine, trochanter, radius, and femoral neck was measured using dual X-ray absorptiometry. vBMD and bone size of the trabecular and cortical bone of the radius and tibia were measured using peripheral quantitative computerized tomography. The amount (hours/week) of PA was determined through questionnaires.

Using the general linear model (including age, height, weight, smoking, and calcium intake as covariates), significant interactions between the COMT genotype and PA were seen for aBMD at all bone sites measured (p <0.01) and for trabecular vBMD in both the radius and the tibia (p <0.01) while no interactions were seen for parameters of cortical bone size or cortical vBMD. Subjects were divided into groups according to amount of PA (low PA: <4 hrs/wk (n = 554) and high PA ≥4hrs/wk (n = 514)). The difference in aBMD and trabecular vBMD between high and low PA was greater in subjects homozygous for the Met genotype than in subjects homozygous for the Val genotype (total body aBMD: Met 5.0% vs. Val 1.7%, p = 0.001; lumbar spine aBMD: Met 7.5% vs. Val 3.5%, p = 0.005; femoral neck aBMD: Met 8.5% vs. Val 4.9%, p = 0.04; radius trabecular vBMD: Met 7.0% vs. Val 0.3% p = 0.005; tibia trabecular vBMD: Met 6.7% vs. Val 0.5%, p <0.001).

These findings demonstrate that the impact of PA on trabecular bone is clearly modulated by the functional COMT polymorphism, and suggest that young men with the Met genotype have superior ability to adapt, in terms of increased bone mass, to PA.

Disclosures: M. Lorentzon, None.


A Robust and Comprehensive Analysis of Twenty Osteoporosis Candidate Genes by Very High-Density Single-Nucleotide Polymorphism Screen among 405 Caucasian Nuclear Families Identified Significant Association and Gene-Gene Interaction. D. Xiong1, H. Shen*2, H. Deng2. 1Osteoporosis Research Center and Department of Biomedical Sciences, Creighton University, Omaha, NE, USA. 2Orthopedic Surgery and Basic Medical Sciences, University of Missouri - Kansas City, Kansas City, MO, USA

Many ‘novel’ osteoporosis candidate genes have been proposed in recent years but seldom analyzed adequately. To advance our knowledge of their roles in osteoporosis, we screened twenty such genes using a set of high-density SNPs in a large family-based study. Our efforts lead to the prioritization of those genes in relation to osteoporosis and the detection of certain gene-gene interactions.

Introduction: Considerable and complex genetic effects are important in the etiology of osteoporosis. As a precursor to genomewide association research, we performed a large-scale family-based association study by screening 20 important and novel osteoporosis candidate genes with 277 SNPs (single nucleotide polymorphisms) for the quantitative trait - BMD (bone mineral density) variation and the qualitative trait - osteoporosis (OP) at three clinically important skeletal sites - spine, hip and ultradistal radius (UD).

Materials and Methods: 1873 subjects from 405 Caucasian nuclear families were genotyped and analyzed with an average density of 1 SNP per 4 kb across the 20 genes. We conducted association analyses by SNP- and haplotype-based FBAT (family-based association test), and performed gene-gene interaction analyses using multi-analytic approaches such as MDR (multifactor-dimensionality reduction) and conditional logistic regression.

Results and Conclusions: We detected four genes (DBP, LRP5, CYP17 and RANK) that showed ‘highly suggestive’ association (10,000-permutation derived empirical global P values ≤.01) with spine BMD/OP; four genes (CYP19, RANK, RANKL and CYP17) highly suggestive for hip BMD/OP; and four genes (CYP19, BMP2, RANK, TNFR2) highly suggestive for UD BMD/OP. Some of these findings supported previous results while the others were novel discoveries. The associations between BMP2 with UD BMD, and those between RANK with OP at spine, hip and UD also met the experimentwide stringent criterion (empirical global P values ≤.0007). In addition, we identified and validated a two-locus gene-gene interaction model involving GCR and ESR2, for which prior biological evidence exists. Our results suggested the prioritization of osteoporosis candidate genes from among the many proposed in recent years and revealed the significant gene-gene interaction effects influencing osteoporosis risk.

Disclosures: D. Xiong, None.


A Large Scale Candidate Gene Screen Identifies an Association between the Sclerostin (SOST) Locus and Femoral Neck Areal BMD in Older Men. S. P. Moffett1, L. Yerges1, K. Roeder*2, C. Kammerer*1, R. Ferrell*1, L. Klei*2, J. Cauley1, K. Ensrud*3, T. Lang*4, E. Orwell*5, J. Zmuda1. 1University of Pittsburgh, Pittsburgh, PA, USA. 2Carnegie Mellon University, Pittsburgh, PA, USA. 3VA Med Center and University of Minnesota, Minneapolis, MN, USA. 4University of California, San Francisco, CA, USA. 5Oregon Health Sciences Univ, Portland, OR, USA

Many candidate genes have been proposed for osteoporosis; however it has not been possible to thoroughly investigate large panels of putative candidate genes until recently. In the current study, we comprehensively investigated the association between femoral neck areal BMD (nBMD) and single nucleotide polymorphisms (SNPs) within or flanking 374 physiologically defined candidate genes for bone metabolism in 882 Caucasian American men aged ≥65 years in the Osteoporotic Fractures in Men (MrOS) Study. SNPs in each candidate gene region were comprehensively interrogated by first creating a reference SNP panel spanning 10kb downstream and 30kb upstream of each gene using the International HapMap Phase I data and then selecting tag SNPs from the reference SNP panel with a minor allele frequency (MAF) ≥5% using a pairwise correlation method (r2 ≥ 0.80). In addition, non-synonomous coding SNPs with a reported MAF ≥1% and other potentially functional SNPs with a MAF ≥2% were also selected for genotyping. 2,722 SNPs in 242 genes genotyped using the Illumina Bead Station platform met quality control criteria and were analyzed for their association with BMD measured at the femoral neck using dual-energy x-ray absorptiometry. Analysis of variance was used to assess the association between the most optimal genotype split for each SNP and nBMD adjusted for age and study site using HelixTree genetic analysis software. We identified 33 SNPs in 24 genes that were associated with nBMD after a gene-wise alpha correction (adjusted ≤p0.05). A SNP (rs1230399) flanking sclerosteosis (SOST), a negative regulator of bone formation, showed the most significant association with nBMD (adjusted p <0.001). Men with the G/G genotype had an adjusted mean BMD 7% higher than men with the A/A genotype representing a difference of nearly half a standard deviation (mean ± SD: A/A = 0.79 ± 0.12g/ cm2, A/G = 0.80 ± 0.12g/cm2 and G/G = 0.85 ± 0.13g/cm2). Another SNP in the SOST region, rs1877632, was significantly associated with nBMD (adjusted p <0.01). Regression analysis of the 7 typed SNPs in the SOST region showed that while rs1230399 and rs1877632 are moderately correlated (R = 0.48; D' = 0.51), they were independently associated with nBMD (p = 0.0017 and p = 0.023, respectively). These results highlight the importance of the SOST gene region in the determination of femoral neck areal BMD variation among otherwise healthy men.

Disclosures: S.P. Moffett, None.


FSH Regulation of Bone Turnover Via FSH-R: Regulated Expression of FSH-R in Osteoclasts and Probable Effects on Mesenchymal Stem Cells. A. C. Sharrow*1, B. B. Yaroslavskiy1, M. Zaidi2, H. C. Blair1. 1Pathology, University of Pittsburgh, Pittsburgh, PA, USA. 2Department of Medicine, University of Pittsburgh, Mount Sinai School of Medicine, NY, USA

Osteoclast differentiation assays and measurements in vivo showed previously that FSH regulates bone mass, at least in part via direct response of osteoclasts to FSH. Glycoprotein hormones are typically subject to feedback receptor regulation, and FSH would be expected to effect other bone cells because mesenchymal stem cells (MSC) express FSH receptors (FSH-R). These possibilities had not been studied. We examined long-term changes in histomorphometry of FSH-R−/− and control animals to determine if effects related to MSC-derived cells such as osteoblasts could be detected. In addition, we compared FSH-R in MSC, osteoblasts, and osteoclasts with and without estrogen or FSH stimulation. Histomorphometry of FSH-R−/− mice relative to control animals showed no change whatever in overall bone formation rate, but there were subtle differences in trabecular pattern. This was apparent as greater average length of calcein-labeled bone in cross sections of trabecular bone. In addition, knockout animals had mild sclerotic changes in the luminal surfaces of the shafts of the long bones. These findings are in keeping with small differences in bone formation pattern dependent on FSH. However, direct labeling of FSH-R in murine bones showed labeling above background associated only with osteoclasts near the growth plate, suggesting that osteoblast FSH-R is absent or low relative to osteoclast FSH-R expression. Quantitative real-time PCR for FSH-R in human MSC, MG63 osteosarcoma cells, and CD14-derived osteoclasts was done, comparing expression of FSH-R relative to GAPDH. We found that MSC and osteoclasts from a female CD14 donor expressed FSH-R, typically amplifying 8 cycles later (relative abundance 1/256) relative to GAPDH on several replicates. The FSH-R was markedly reduced when cultures were made in the presence of 10 ng/ml FSH, 14 ± 4 cycles beyond GAPDH, suggesting that FSH-R in osteoclasts is downregulated by its ligand. Estradiol, 10 nM, had no effect. The MG63 human osteosarcoma cells made very little FSH-R, with amplification typically >15 cycles beyond GAPDH and of doubtful physiological significance. CD14 cells from a single male donor amplified 13 ± 2 cycles beyond GAPDH, and did not show a further decrement with FSH. Our findings suggest that FSH effects on MSC-derived cells such as osteoblasts probably occur, although minimal FSH-R are present in mature human osteoblasts. Expression of FSH-R itself may depend not only on cell type but on other factors including sex and exposure to continuous high levels of FSH, which may downregulate the receptor.

Disclosures: H. C. Blair, None.

This study received funding from: NIH.


Estradiol Potentiates the Inhibitory Effects of Anti-Cytokines on Bone Resorption in Naive and Arthritic Mice. D. Saintier*1, C. Schiltz*1, N. Saidenberg*2, M. de Vernejoul*1, M. E. Cohen-Solal1. 1606, INSERM, Paris, France. 2ERI 18, INSERM, Bobigny, France

Osteoporosis occurs in patients with rheumatoid arthritis (RA). Clinical evidence suggests the beneficial role of estrogen in RA and estrogen deficiency might participate in RA-related bone loss. Pro-inflammatory cytokines produced in synovial pannus are involved in osteoclast differentiation and bone resorption. We hypothezised that estrogen deficiency might contribute to bone resorption through modulation of inflammatory cytokines. Collagen-induced arthritis (CIA) was generated in DBA-1 female mice, and were sacrificed when arthritis reached a plateau. Bone marrow cells from CIA and naive mice were cultured in phenol red-free DMEM and 10%FCS with RANKL and M-CSF for 14 days in the presence of estradiol (108M), IL1ra (50μg/ml), TNF antibody (A-TNFab, 1 μg/ml) and with the combination of estradiol and anti-cytokines. Resorption pits on dentin slices and TRAP multinucleated cells were measured at the end of culture in duplicate wells. The presence of OPG completely abolished TRAP cell formation and resorption pits in all bone marrow cultures (data not shown). Estradiol alone had no effect on TRAP and pits formation in naive and CIA mice. IL1ra decreased TRAP multinucleated cells in naive but not in CIA mice. Combination of estradiol and IL1ra, A-TNFab significantly decreased the number of TRAP cells in both naive and CIA mice (p <0.01). Moreover, addition of estradiol and the 2 anti-cytokines was the only condition able to reduce significantly bone resorption in CIA mice (p <0.04), but not in naive mice. 

Table  .  
  1. *: different from controls of the same group;° : different from IL1ra In conclusion, estradiol is required to decrease ex vivo bone resorption in combination with anti-cytokines in CIA mice. Such a synergistic effect might be relevant for the prevention of postmenopausal bone loss in patients with RA.

inline image

Disclosures: M.E. Cohen-Solal, None.


Smurf1, a Ubiquitin E3 Ligase, Targets BMP and AP-1 Signaling Proteins and Mediates TNF-Induced Inhibition of Osteoblast Function in Mature Osteoblasts. R. GUO*, D. Chen., E. M. Schwarz, Y. Zhang*, B. F. Boyce, L. Xing. University of Rochester Medical Center, Rochester, NY, USA

Smad ubiquitin regulatory factor 1 (Smurf1) negatively regulates osteoblast (ob1) function by promoting proteasomal degradation of BMP signaling molecules and Runx2 protein. Recently, we reported that TNF stimulates Runx2 degradation and increases Smurf1 expression in ob1 precursor cell lines, indicating that Smurf1-medatied protein ubiquitination and degradation may play a critical role in TNF-induced ob1 inhibition and osteopenia. Here, we used Smurf1−/− mice to determine if Smurf1 mediates TNF's effects on obl function and if it has targets apart than Smads and Runx2. Since osteopenia occurs in long bones of TNF-transgenic (TNF-Tg) mice, we used mature obls derived from bone marrow (BM) as a cell model system. We cultured wt murine BM cells in obl differentiation medium for various times up to 20 days and found that treatment with TNF decreased ALP mRNA expression (by real time RT-PCR) maximally at 2 weeks and significantly decreased protein levels of pSmad1/5, total Smad1 and Runx2, and increased Smurf1, but not Smurf2 expression at this time. Similarly, metaphyseal bone samples from TNF-Tg arthritic mice had decreased BMP-Smads and Runx2 and increased Smurf1 proteins levels, suggesting that our in vitro data reflect in vivo effects of TNF. To examine the effect of Smurf1 blockade on TNF-induced obl inhibition, we injected TNF over the calvariae of Smurf1−/− and wt mice. TNF decreased ALP (0.4 ± 0.1 vs 1 ± 0.2) and OC (0.4 ± 0.02 vs 1.1 ± 0.1) mRNA in calvarial samples from wt, but not from Smurf1−/− mice (ALP 3 ± 0.6 vs 1.9 ± 0.4; OC 2.4 ± 0.5 vs 1.7 ± 0.2). Accordingly, TNF-induced reduction of Smad1 and Runx2 proteins was attenuated in the Smurf1−/− mice. To identify new targets of Smurf1, we focused on AP-1 transcriptional factor proteins because they play important roles in various cell functions and both JunB and c-Jun contain PY motifs, recognition sites for Smurf1-mediated ubiquitination. Over-expression of Smurf1 reduced nuclear JunB protein levels, but had no effect on c-Jun and c-Fos in 2T3 preosteoblast cells. The proteasome inhibitor, MG132 (10 μM, 4 h), partially reversed Smurf1-reduced JunB protein. Smurf1−/− obls expressed high levels of JunB, suggesting that JunB is a likely physiologic target for Smurf1. In summary, Smurf1 mediates the inhibitory effect of TNF on BMP signaling proteins in vitro and in vivo and the AP-1 protein, JunB, is a new target for Smurf1 in obls. We propose that Smurf1-mediated ubiquitination and proteasomal degradation of BMP and AP-1 signaling proteins is a novel mechanism whereby TNF induces osteopenia in general and inhibits bone formation in inflammatory bone diseases.

Disclosures: R. Guo, None.


Skeletal Integrity in Low Serum IGF-1 States Is Determined by the Ligand Delivery System of Binary and Ternary Complexes. S. Yakar1, M. L. Bouxsein2, V. Glatt*2, H. Sun*1, D. Clemmons3, C. J. Rosen4. 1The Mount Sinai School of Medicine, New York, NY, USA. 2Beth Israel Deaconess Medical Center, Boston, MA, USA. 3University of North Carolina, Chapel Hill, NC, USA. 4The Jackson Laboratory, Bar Harbor, ME, USA

Serum IGF-1 is secreted mainly by the liver and circulates bound to IGF-binding proteins (IGFBPs) in two ways: 1) as binary complexes to IGFBP-3 or IGFBP-5, or 2) as ternary complexes with the acid labile subunit (ALS) plus IGFBP-3 or IGFBP-5. These complexes stabilize IGF-I and prolong its half life. The goal of this study was to delineate the role of these circulatory complexes in skeletal integrity. We employed three mouse lines of serum IGF-I deficiency that reflect differences in complex composition: 1) the liver-specific IGF-I deficient (LID) mice with 80% reduction in serum IGF-1 but intact binary and ternary complexes; 2) the ALSKO (ALS deficient) mice, with 60% reduction in IGF-I and binary, but no ternary complexes, and 3) the BP3KO (IGFBP-3 deficient) mice, with a 40% reduction in serum IGF-I and no IGFBP-3 binary complex, but alternative ternary complexes using IGFBP-5. We analyzed femorae of 8 week-old male mice by μCT. Trabecular bone volume/total volume (BV/TV%) was significantly decreased in all serum IGF-1 deficient models (Table). However, BP3KO mice lacking the binary complex exhibited the most profound decline in BV/TV (−25%). Cortical thickness was decreased in LID, but not in ALSKO or BP3KO. Mid-femoral cross-sectional area was decreased significantly in LID and ALSKO, but not in BP3KO. In summary, femoral trabecular BV/TV is markedly decreased in the IGF-1 deficiency state and is largely dependent on the presence of the IGFBP-3 binary complex. Cortical thickness however is more sensitive to total circulating IGF-1, and is markedly decreased in LID mice. Interestingly, BP3KO mice have increased body weight and normal bone size, despite low serum IGF-1 suggesting that the alternative IGFBP-5 binary complex affects bone acquisition. We conclude that absolute serum IGF-1 concentrations do not predict BMD or morphology primarily because of differential skeletal responsiveness to alterations in the complexes that affect IGF-I delivery.

*p <0.05 vs controls 

Table  .  
inline image

Disclosures: S. Yakar, None.


O-Linked Glycosylation by ppGaNTase-T3 Prevents Processing of Fibroblast Growth Factor (FGF) 23. N. Ito*1, S. Fukumoto1, T. Yamashita2, Y. Frishberg*3, T. Fujita*1. 1Division of Nephrology & Endocrinology, Department of Medicine, University of Tokyo Hospital, Tokyo, Japan. 2Pharmaceutical Research Laboratories, KIRFN Brewery Co., Ltd., Takasaki, Japan. 3Division of Pediatric Nephrology, Shaare Zedek Medical Center, Jerusalem, Israel

FGF23 with phosphaturic activity is known to be partly processed into inactive fragments. Tumoral calcinosis (TC) is characterized by hyperphosphatemia due to enhanced phosphate reabsorption and relatively high serum level of 1,25-dihydroxyvitamin D. This disease is a mirror image of hypophosphatemic diseases caused by excess FGF23 activity. Two genes, FGF23 and GALNT3 encoding a peptide (ppGaNTase-T3) involved in the initiation of mucin-type O-glycosylation, were identified to be responsible for TC. Patients with TC caused by mutations of either gene have rather low levels of biologically active full-length FGF23 while the amount of inactive C-terminal fragments is markedly elevated. Therefore, the processing and production of FGF23 seem to be enhanced in patients with TC. Although FGF23 is known to have several O-linked sugar chains, it has been unknown whether mutations in GALNT3 actually affect the processing of FGF23. We therefore investigated the effect of silencing of GALNT3 expression. HOS-TE85 cells that were found to express ppGaNTase-T3 were treated with siRNA for GALNT3 or control siRNA, and then transfected with expression vector for FGF23. Immunoblotting confirmed that siRNA for GALNT3 decreased the expression of ppGaNTase-T3. Western blotting demonstrated that the amount of full-length FGF23 decreased while the processed C-terminal fragment increased in conditioned media by treatment with siRNA for GALNT3. In agreement with these findings by Western blotting, measurement of FGF23 by full-length assay indicated that me amount of FGF23 in the conditioned media of cells with siRNA for GALNT3 decreased to 60.9 +/− 6.1% of cells with control siRNA. In contrast, C-terminal assay, which detects both full-length and processed C-terminal fragment, demonstrated that FGF23 produced by cells with siRNA for GALNT3 was 96.2 +/− 17.5% of control cells. These results indicate that the inhibition of ppGaNTase-T3 expression enhances the processing of FGF23. Therefore, ppGaNTase-T3 activity is necessary for maintenance of full-length FGF23. TC caused by mutations in either FGF23 or GALNT3 seems to be caused by deficient actions of full-length FGF23. These results indicate that FGF23 activity is physiologically essential for normal phosphate and vitamin D metabolism.

Disclosures: N. Ito, None.


Increased Cb1-Medlated PI3K Degradation Mediates Osteoblast Apoptosis in Response to FGFR-2 Activation. C. Dufour*1, K. Kaabeche*1, D. Bouvard*2, P. J. Marie1. 1Laboratory of Osteoblast Biology and Pathology, Inserm U606. Paris, France. 2LEDAC, UMR CNRS/UJF 5538, La Tronche, France

Fibroblast growth factor receptor (FGFR) signaling is an important mechanism controling osteoblast proliferation, differentiation and apoptosis. To gain further insight into the regulatory role of FGFR-2 signaling in osteoblastogenesis, we investigated the molecular mechanisms involved in apoptosis in human calvarial osteoblasts expressing an activated FGFR-2 induced by the S252W mutation in Apert craniosynostosis. Western blot analysis revealed decreased levels of total and phosphorylated phosphatidylinositol 3-kinase (PI-3K), a key player in the control of cell survival, in confluent FGFR-2 mutant osteoblasts compared to wild type cells. Additionally, phosphorylated levels of the downstream survival kinase Akt were decreased in FGFR-2 mutant osteoblasts compared to wild type cells. In contrast, PI3K and Akt mRNA levels were identical in mutant and wild type cells. Biochemical analyses showed that PI3K co-immunoprecipitates with the E3 ubiquitin ligase Cb1, ubiquitin and FGFR-2, suggesting interactions between these three molecules. Treatment with lactacystin, a potent inhibitor of proteasome, restored PI3K levels in FGFR-2 mutant osteoblasts, suggesting that FGFR-2 activation induces PI3K proteasome degradation. Transfection with 70Z-Cb1 mutant which lacks the RING domain required for Cbl-ubiquitin interaction, or with the G306E Cb1 mutant that abolishes the binding ability of Cb1 phosphotyrosine-binding domain restored normal total and phophorylated PI3K levels, showing that Cb1-mediated ubiquitination is involved in PI3K proteasome degradation induced by FGFR-2 activation. Because lipid rafts are dynamic membrane microdomains involved in the regulation of cell growth and apoptosis, we investigated whether FGFR-2, Cb1 and PI3K interactions occur in lipid rafts. Biochemical analysis showed that FGFR-2 was mainly expressed with caveolin in clustered lipid rafts. Immunocytochemical analysis confirmed that PI3K co-localised with FGFR2 and Cb1 in lipid rafts clusters detected by cholera toxin (CTx) staining. The results show that Cbl and PI3K recruitment in lipid rafts induced by FGFR-2 activation triggers Cbl-dependent PI3K degradation by the proteasome, resulting in diminished PI3K/Akt signaling and increased osteoblast apoptosis. This identifies a functional role of PI3K signaling in apoptosis triggered by FGFR-2 activation in osteoblasts, and reveals a ccordinated mechanism by which Cbl and PI3k recruitment to lipid rafts controls osteoblast apoptosis in response to FGFR-2 activation.

Disclosures: P.J. Marie, None.


Klotho-FGFR1(IIIc) Complex Comprises FGF23 Specific Receptor. Y. Yamazaki*1, I. Urakawa*1, T. Shimada1, K. Iijima*1, H. Hasegawa*1, T. Fujita*2, S. Fukumoto2, T. Yamashita1. 1Kirin Brewery Co., Ltd, Takasaki, Japan. 2University of Tokyo School of Medicine, Tokyo, Japan

While most FGF family molecules function in an autocrine/paracrine manner, FGF23 is a humoral factor that is produced in bone and regulates phosphate reabsorption and vitamin D metabolism in kidney. Therefore, there must be a specific mechanism that accounts for the unique renotropic FGF23 function. We reported at the last ASBMR meeting that 1) the administration of FGF23 into mice led to increase expression of Early Growth Response gene-1 (Egr-1) in kidney, 2) kidney-specific expressing protein Klotho bound to FGF23, 3) FGF23 increased expression of Egr-1 in only exogenously Klotho-expressing cells, 4) most phenotypes of klotho-deficient mice resemble those of FGF23-null mice and 5) an anti-klotho antibody increased serum levels of phosphate and 1,25(OH)2 vitamin D as well as some anti-FGF23 antibodies. These findings indicated that Klotho was a key molecule for FGF23 function. Klotho is a Type I transmembrane protein that has a large extracellular domain and a very short cytoplasmic resion. Since an intracellular region of Klotho is too small to be able to evoke intracellular signaling, it seemed that an additional factor is needed for the FGF23 receptor. We suspected the involvement of FGFRs in the Klotho-dependent FGF23 signaling, because FGF23 could not induce Egr-1 expression even in the presence of Klotho in L6 rat myoblast cells, which are known to lack FGFRs. We examined several soluble FGFRs-Fc (sFGFR-Fc, an extracellular domain of FGFR fused with an IgG Fc domain) if they could affect the Klotho-dependent FGF23 signaling in vitro and found that the sFGFR1(IIIc)-Fc reduced the FGF23-induced activation of Egr-1 promoter activity. Moreover, coexpression of the full-length FGFR1(IIIc) and Klotho enabled to induce activation of Egr-1 promoter activity by FGF23 in L6 rat myoblast cells. FGFR1(IIIc) and Klotho seems to interact with each other in their extracellular portions, since sFGFR1(IIIc)-Fc coprecipitated an extracellular domain of Klotho. The Klotho and FGFR(IIIc) complex was stabilized in the presence of FGF23, whereas basic FGF inhibited formation of the Klotho and FGFR(IIIc) complex. These findings suggest that Klotho converted FGFR1(IIIc), a canonical receptor for various FGFs, into the FGF23 specific receptor, which may account for the unique biological activity of FGF23 apart from those of other FGFs.

Disclosures: Y. Yamazaki, Kirin Brewery Co., Ltd 3.


Stage-Specific Effects of TGF-β on Wnt Signaling in Chick Chondrocytes. Y. Dong, D. Soung, D. Chen, E. M. Schwarz, R. J. O'Keefe, H. Drissi. Orthopaedics, University of Rochester, Rochester, NY, USA

Here we investigate the effects of TGF-β on primary chick chondrocytes at various stages of their maturation process. We utilized immature chondrocytes from chick lower stemae, the more advanced upper sternal chondrocyte differentiation model and the hypertrophic growth plate chondrocyte model. All three chick chondrocyte models were treated with TGF-β for 3, 6, 12, 24, 48, and 72 hours, and harvested for RNA and protein isolation. Our real time RT-PCR and western blot analyses showed that while TGF-β significantly induced β-catenin and TCF-4 in immature chondrocytes, this induction was switched to a strong repression on these canonical Wnt mediators in upper sternal, and to a greater extent in hypertrophic growth plate chondrocytes. Together these results suggest a stage dependant regulation of canonical Wnt signaling by TGF-β. We further assessed the effects of TGF-β on canonical Wnt signaling using TOP-flash in all three cell models. Transfection experiments showed an increase in promoter activity by TGF-β in chick lower sternal chondrocytes, while TGF-β time dependently inhibited promoter activity in upper sternal and growth plate chondrocytes. Thus, TGF-β differentially regulates Wnt signaling in chick chondrocyte depending on their levels of maturation. To define the molecular mechanisms underlying TGF-β regulation of the canonical Wnt pathway we assessed possible interactions between Smad and β-catenin and/or TCF/LEF proteins. Our immunoprecipitation experiments showed that Smad3 strongly interacts with β-catenin, LEF1 and TCF-4 while its interaction with TCF1 and TCF3 was very weak in chick chondrocytes. Furthermore, while TGF-β inhibited Smad3 interaction with β-catenin and TCF-4, Wnt3a enhanced it. Experiments are underway to determine the effects of TGF-β on this interaction in lower sternal immature chondrocytes and hypertrophic growth plate chondrocytes. Finally we assessed the occupancy of the Runx2 promoter by these proteins as Runx2 is a direct target for Wnt signaling in chondrocytes. Our Chip assay results show that both TGF-β and Wnt3a enhanced TCF-4 and β-catenin interaction with the Runx2 promoter in MLB13MYC17 mouse prechondrocytes. Furthermore, the use of a specific antibody in our chip assays also shows that Smad3 is part of this interaction. The effects of TGF-β and Wnt3a on these interactions in our three chick chondrocyte models are currently being examined. Together, our results indicate that TGF-β effects on Wnt signaling involve direct interaction between Smad3, TCF/LEF and β-catenin. These interactions may result in activation or repression of canonical Wnt signaling depending on chondrocyte maturation stage.

Disclosures: Y. Dong, None.


TGF-β Activates Cyclin D1 Expression through a Smad3/β-Catenin-Dependent Mechanism in Chondrocytes. T. Li, M. Chen, Q. Wu, T. Sheu, H. Drissi, M. Zuscik, D. Chen, R. J. O'Keefe. Orthopaedics, University of Rochester, Rochester, NY, USA

TGF-β regulates chondrocyte proliferation but mechanisms involved remain undefined. In mouse sternal chondrocytes isolated from TOP-gal transgenic mice, TGF-β (0–5 ng/ml) dose-dependently stimulated β-Gal activity. In chondrogenic RCJ3.1C5.18 (RCJ3.1) cells, TGF-β (2 ng/ml) increased β-catenin protein levels within 15 minutes with maximal levels achieved after 2 hours. In contrast, TGF-β did not alter β-catenin mRNA expression. To determine the role of Smad3 in TGF-β-stimulated β-catenin activation, the steady-state protein levels of β-catenin were examined in Smad3−/− chondrocytes, β-catenin protein levels were significantly reduced in Smad3−/− chondrocytes. To determine if Smad3 interacts with β-catenin, we performed immunoprecipitation (IP) assays in RCJ3.1 cells. Weak Smad3/β-catenin interaction was detected in the absence of TGF-β while strong Smad3/β-catenin interaction was observed in the presence of TGF-β (30 minute treatment) in RCJ3.1 cells. In contrast, Smad2/β-catenin interactions were absent. Immunostaining showed that TGF-β stimulated Smad3 and β-catenin nuclear translocation in RCJ3.1 cells. In Smad3−/− chondrocytes, the nuclear translocation of β-catenin following TGF-β administration was completely blocked. β-TrCP is an E3 ligase in the SCF−βTrCP protein complex, which induces β-catenin degradation. Transfection of β-TrCP reduced β-catenin protein levels in a dose-dependent manner in RCJ3.1 cells. Co-transfection of Smad3 with β-TrCP significantly attenuated β-TrCP-induced β-catenin degradation in RCJ3.1 cells. The results suggest that Smad3 may prevent β-catenin degradation by β-TrCP. Since β-catenin has been associated with cell cycle regulation, we examined if β-catenin was involved in proliferation downstream of TGF-β. BrdU-positive cells in the growth plate were significantly reduced in Smad3−/− mice compared with wt littermates. Cyclin D1 reporter activity and protein expression were decreased in Smad3−/− chondrocytes. In contrast, expression of constitutively active β-catenin (β-cateninS33Y) stimulated cyclin D1 reporter activity and protein expression in RCJ3.1 cells. Primary mouse sternal chondrocytes isolated from β-catenin-loxP mice were infected with Ad-Cre recombinase to delete the β-catenin gene. This resulted in completely loss of TGF-β-induced cyclin D1 expression in chondrocytes. These results suggest that the induction of cyclin D1 expression by TGF-β/Smad3 is dependent upon β-catenin signaling in chondrocytes. Our findings provide new evidence concerning the cross-talk between TGF-β and β-catenin signaling pathways in cartilage.

Disclosures: T. Li, None.


TGFβ1 Released during Bone Resorption Couples Osteoprogenitor Migration. Y. Tang*, X. Wu*, S. Xie*, L. Zhao*, W. Yuan*, Z. Shi*, X. Feng, X. Peng*, M. Wan, X. Cao. Pathology, UAB, Brimingham, AL, USA

Bone is continuously remodeled by the balanced processes of bone resorption and concomitant bone formation. The recruitment of osteoprogenitors to the site of bone resorption is the initial events necessary for the coupling of bone resorption and bone formation. However, the factor(s) in the bone microenvironment responsible for this process is not clear. Here we demonstrate that TGFβ1 released from bone matrix during osteoclastic bone resorption induces migration of primary human bone marrow osteoproginitors (HBMO). We established a type I collagen-coated transwell assay in which HBMO migration can be examined in coupling with osteoclastic bone resorption. We found that osteoclastic bone resorption-conditioned medium stimulates HBMO migration, whereas conditioned media from monocytes/macrophage or osteoclasts without bone slices has no effect. Importantly, addition of an antibody against TGFβ1 inhibits the migration, but antibodies against TGFβ2, TGFβ3, FGF-I, FGF-II, PDGF or norggin do not affect cell migration. In addition, Western blot analysis of both mature active and latent inactive TGFβ1 in conditioned media revealed that only conditioned medium of osteoclast with bone slices contains active TGFβ1, indicating that TGF-β1 released and activated during osteoclastic bone resorption is a primary factor coupling bone resorption and formation.

To characterize the signaling mechanism that mediates TGFβ1-induced HBMO migration, we constructed a yeast two-hybrid HBMO library which was then screened using a cytoplasmic domain of TGFβ type II receptor (TRβII) as a bait. One of the clones identified was GDIα. GDIα is the central regulator of Rho GTPases and binds all three subfamily members (i.e., RhoA, Rac and Cdc42) that are required for cell migration. The interaction between endogenous TRβII and GDIα in HBMO is then confirmed in an immunoprecipitation assay. Importantly TGFβ1 activates RhoA by inducing the interaction between TRβII and GDIα as a novel Smad independent TGFβ signaling pathway. We have also established an in vitro TGFβ1 gradient migration model, in which HBMO are polarized against the gradient during migration. Immunostaining demonstrated that TGFβ1 induces formation of cell focal adhesion at proximal side of the gradient via activation of RhoA. In addition, TβRI specific inhibitor blocks TGFβ1-induced cell migration by inhibiting Smad signaling pathway, whereas a constitutive active TβRI enhances cell migration by promoting actin polymerization at the proximal side of the gradient. Taken together, TGFβ1 released and activated during bone resorption induces osteoprogenitor migration as a coupling factor by activating both classical Smad pathway and RhoA pathway.

Disclosures: Y. Tang, None.


Cartilage-Specific Over-Expression of Smurf2 Increases Callus Cartilage Formation by Degradation of Phospho-Smad3. Q. Wu. A. Naik*, K. Kim*, Y. Lim*, C. Xie*, M. J. Zuscik, D. Chen, R. J. O'Keefe, R. N. Rosier. Orthopaedics, University of Rochester, Rochester, NY, USA

Endochondral cartilage maturation during fracture healing recapitulates many of the events that occur during embryonic development. This process is tightly regulated by TGF-β and BMR During cartilage maturation, TGF-β inhibits chondrocyte proliferation and differentiation, while BMP promotes this process. Supporting this idea, over-expression of Smurf2, an E3 ubiquitin ligase, leads to reduced TGF-β signaling and accelerated chondrocyte maturation in chondrocytes in vitro. We therefore hypothesized that over-expression of Smurf2 in cartilage in vivo leads to loss of TGF-β signaling in chondrocytes, stimulates chondrocyte maturation in cartilage calluses, and accelerates fracture healing. We tested this hypothesis by examining the effect of Smurf2 on TGF-β/BMP signaling and endochondral cartilage maturation during fracture healing. Western-blot showed that the protein levels of Smad1, phosphorylated Smad1, Smad2, phosphorylated Smad2, and Smad3 as well as TGFβ type I receptor were similar in chondrocytes from col2-Smurf2 and WT mice; however, phospho-Smad3 was 4 fold lower in col2-Smurf2 mouse chondrocytes than WT. We performed immunoprecipitation, and observed the ubiquitinated phospho-Smad3 ladders in sternal chondrocytes from col2-Smurf2 mice but not from WT. This data suggests that Smurf2 specifically ubiquitinates and degrades phospho-Smad3 in chondrocytes. Consistent with this finding, only the TGF-β promoter (P3TP) activity, but not BMP promoter (12 X SBE colX-Lux) activity was inhibited in col2-Smurf2 chondrocytes. Furthermore, the mRNA levels of colX, VEGF, and MMP13, markers for hypertrophic chondrocytes, were 3 fold higher in col2-Smurf2 chondrocytes than WT. Therefore, Smurf2 blocks TGF-β signaling in chondrocytes and accelerates chondrocyte maturation by degradation of phospho-Smad3. Since the balance between TGF-β and BMP signaling is essential for maintaining a normal rate of endochondral cartilage maturation, we examined whether overexpresion of Smurf2 in chondrocytes accelerates endochondral cartilage maturation during fracture healing in femur fractures. Histology showed that 1) the area of cartilage within calluses in col2-Smurf2 is twofold and threefold of that in WT at 7 and 10 days post-fracture, respectively and 2) callus chondrocyte proliferation and differentiation in col2-Smurf2 mice is increased V.S. that in WT. Therefore, overexpression of Smurf2 in cartilage enhances endochondral cartilage maturation during fracture healing.

Disclosures: Q. Wu, None.

This study received funding from: NIH/NIAMS RO1 AR45700 (Rosier, PI).


E-Selectin Ligand 1 Negatively Regulates TGFβ in the Golgi during Skeletogenesis. T. Yang*1, R. Mendoza-Londono*1, H. Lu*1, K. Li*1, B. Keller*1, M. Jiang*1, Y. Chen*1, B. Dabovic*2, D. B. Rifkin*2, J. Hicks*3, A. L. Beaudet*1, B. Lee1. 1Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA. 2Cell Biology, New York University Medical Center, New York, NY, USA. 3Pathology, Texas Children's Hospital, Houston, TX, USA

Transforming growth factor β (TGFβ) signaling is critical for the regulation of growth and differentiation during development and disease. Its context dependent action is specified by numerous control mechanisms at the extracellular level and downstream of ligand-receptor interactions, but little is known about the regulation of its post-translational trafficking. E-Selectin Ligand-1 (ESL-1), the cysteine rich protein originally isolated as a ligand for E-Selectin, was also found to interact with FGFs and to be co-purified with TGFb1 in a large protein complex. To elucidate the in vivo function of ESL-1, we mutated Esl-1 gene in mice by conventional knock-out strategy. The newborn and adult Esl-1−/− mice are notably smaller with narrow chests and generalized shortening and thinning of the long bones, ribs and spine. By histological analysis, P1 Esl-1−/− mice showed shortening of the growth plates in both the proliferating zone and hypertrophic zone. Further molecular biology assays show that ESL-1 acts as a negative regulator of TGFβ production by binding TGFβ precursors in the Golgi in a cell autonomous fashion. In vivo, loss of ESL1 function causes increased TGFβ signaling resulting in decreased cell proliferation and delayed terminal differentiation in the cartilaginous growth plate, independent of effects on BMP and FGF signaling. Moreover, in vivo genetic models of gain vs. loss of TGFβ signaling in the growth plate confirm this effect. These data identify a novel cellular mechanism for regulating TGFβ during skeletogenesis and cartilage homeostasis.

Disclosures: T. Yang, None.


Potential Interaction of Prostaglandin and Wnt Signaling Pathways Mediating Bone Cell Responses to Fluid Flow. M. A. Kamel*, B. R. Holladay*, M. L. Johnson. Oral Biology, UMKC School of Dentistry, Kansas City, MO, USA

The canonical Wnt signaling pathway has been shown to be critical for the regulation of bone mass and the responsiveness of bone to mechanical load. Prostaglandin E2 (PGE2) is an important signaling molecule in bone and several studies have shown that PGE2 is released from bone cells immediately upon the application of load. In an effort to determine if there is a relationship between the rapid release of PGE2 from bone cells and Wnt pathway activation in response to loading, we tested the hypothesis that initial (early) activation of the Wnt signaling pathway might result as a consequence of PGE2 acting through its EP2 receptor. We used Wisp2 (Wnt-1 Induced Signaling Protein) gene expression, as a marker for Wnt pathway activity. 2T3 osteoblastic and MLO-Y4 osteocytic cells were either treated with 5 uM PGE2 for 2 hours or subjected to pulsatile fluid flow shear stress (PFFSS) at 16 dynes/cm2 with a pulse height of +/− 0.6 dynes/cm2 for 2 hours. Cells were also treated with 5 uM AH6809, an EP2 receptor antagonist, for 3 hours prior to and during fluid flow. RNA was extracted from the cells immediately after PFFSS and Wisp2 mRNA was quantitated using real-time PCR (TaqMan). In 2T3 cells, both PGE2 and PFFSS resulted in increased levels of Wisp2 mRNA. The effects of PFFSS on Wisp2 gene expression were blocked by treatment of the cells with AH6809. In MLO-Y4 cells both PGE2 and PFFSS resulted in a decrease in Wisp2 mRNA. However, in MLO-Y4 cells AH6809 did not appear to prevent the fluid flow induced decrease in Wisp2 mRNA. These results are consistent with our hypothesis that there is an interaction between the PGE2 and Wnt signaling pathways in bone cells in response to fluid flow as a model for mechanical loading. Our data are consistent with an EP2 receptor mediated interaction with the Wnt pathway in 2T3 cells and a different receptor mediating the interaction in MLO-Y4 cells. Our data suggesting a possible connection between PGE2 and the Wnt signaling pathway implies a temporal order of events that occur in response to mechanical loading. We are currently testing a model in which the initial activation of the canonical Wnt pathway by PGE2 (produced in rapid response to loading) is subsequently amplified by downstream products of the Wnt pathway acting in an autocrine fashion at the level of the Lrp5/frizzled co-receptors in order to produce a bone formation response.

Disclosures: M.A. Kamel, None.


Deletion of β1 and β2 Adrenergic Receptors Does not Prevent Bone Loss Induced by Tail Suspension in Mice. R. Levasseur1, E. Belin de Chantemele*2, M. A. Custaud*2, E. Legrand1, M. Audran1, D. Chappard3. 1Rheumatology Angers Teaching Hospital, Inserm EMI 335, Angers, France. 2Laboratoire de Physiologie, UMR-CNRS 6188, Angers, France. 3Inserm EMI 335, Angers, France

Sympathetic nervous system is implicated in the regulation of bone remodeling. β2-adrenergic receptors are expressed by bone cells and propranolol prevents bone loss induced by ovariectomy in mice. β2-adrencrgic receptors KO mice have high bone mass by uncoupling favouring bone formation and decreasing bone resorption. Propranolol also prevents bone loss induced by tail-suspension in rats but sympathetic nervous system does not mediate the load-induced cortical new bone formation in mice. To further investigate the role of beta-adrenergic receptors in bone regulation and particularly in mediating mechanical loading, we used the tail-suspension model with mice lacking β1- and β2-adrenergic receptors. Male wild-type C57/BL6 or KO mice at 2 months of age were suspended during 25 days and compared with non suspended mice. MicroCT analysis of the tibia showed, when comparing WT suspended with non suspended mice, a significant 152% decrease in bone volume (BV/TV) (p = 0.01), a 149% decrease in trabecular number (Tb.N) (p = 0.007), a 36% increase in trabecular separation (Tb.Sp) (p = 0.04), a non significant change of trabecular thickness (Tb.Th), a highly significant 32% increase of the structure model index (SMI), a 46% increase in the trabecular bone pattern factor (TBPF). When comparing KO suspended with non suspended mice, we observed a significant 195% decrease in BV/TV (p = 0.01), a 152% decrease for Tb.N (p = 0.007), a 41% increase in Tb.Sp (p = 0.03), a 17% decrease in Tb.Th (p = 0.02), a significant 13% increase of the SM1 (p = 0.02), a 30% increase in the TBPF (p = 0.008). The same results were obtained when analyzing WT and KO female mice. Thus, bone loss is significantly induced by tail-suspension in both WT and KO male and female mice. These results are against a mediation of β1 and β2 adrenergic receptors, i.e. the sympathetic nervous system, in the bone loss observed in the tail-suspension mice model.

Disclosures: R. Levasseur, None.


ATP Release Mediates Fluid Flow-induced Calcium Signaling in Human Mesenchymal Stem Cells. R. C. Riddle, A. F. Taylor*, H. J. Donahue. Department Orthopaedics and Rehabilitation, Pennsylvania State University College of Medicine, Hershey, PA, USA

The activation of intracellular signaling cascades by autocrine/paracrine factors, such as ATP and PGE2, is a vital component of the mechanism by which bone cells respond to mechanical stimulation. While we have recently reported that exposing human mesenchymal stem cells (hMSCs) to oscillatory fluid flow induces proliferation, the factor(s) that initiate the signaling events leading to this response have yet to be identified. In these studies, we examined the effect of oscillatory fluid flow on the release of ATP from hMSCs, and the effect of this signaling molecule on intracellular calcium signaling, which we previously found to be important for hMSC proliferation in response to fluid flow. Exposing hMSCs to oscillatory fluid flow inducing a shear stress of 20dynes/cm2 at 1 Hz triggered a robust increase in ATP release compared to static controls (6.3 ± 0.9nM to 0.3 ± 0.1 nM). ATP levels peaked within 1 minute of the initiation of fluid flow but remained significantly elevated through 15 minutes of flow and were not accompanied by disruption of the plasma membrane. The release of extracellular nucleotides was found to be directly responsible for fluid flow-induced calcium signaling. In control cells, a rapid increase in intracellular calcium concentrations (184.9 ± 12.14nM) was observed in response to fluid flow. However, in the presence of apyrase (10U/mL), an enzyme that degrades extracellular nucleotides, the increase in intracellular calcium concentration in response to fluid flow was significantly attenuated (65.61 ± 7.85nM). These alterations in flow-induced calcium signaling were accompanied by changes in the activation of downstream targets, such as the calcium-responsive phosphatase calcineurin. In untreated cells, 5 minutes of fluid flow exposure increased calcineurin activity from 309.6 ± 33.84 to 790.3 ± 81.61pmol of phosphate liberated in 30 minutes/mg of protein. In the presence of apyrase, calcineurin activity did not increase significantly in response to fluid flow (306.7 ± 53.55 compared to 314.9 ± 31.73pmol of phosphate liberated in 30min/mg of protein). Currently we are focused on the effects of fluid flow on calcium-regulated transcription factors, such as NFAT, and the effect of these factors on hMSC proliferation. The consistency between these data and previously published reports utilizing osteoblastic cell models imply that a common pathway exists by which mechanical signals are translated to cellular responses.

Disclosures: R.C. Riddle, None.

This study received funding from: NIH.


Molecular Mechanisms Enhancing Osteoclastogenesis and Bone Resorption under Modeled Microgravity. A. Rufo*, M. Alamanou*, A. Teti, N. Rucci. Department of Experimental Medicine, University of L'Aquila, L'Aquila, Italy

Prolonged mechanical unloading is known to be detrimental for the skeleton, but the underlying molecular mechanisms are still unclear. We investigated the osteoblast ability to regulate osteoclastogenesis using mouse calvarial cell cultures grown for 24 hours at 1xg, 0.08xg and 0.008xg in a NASA-developed Rotating Wall Vessel bioreactor. Mouse bone marrow cultures were then incubated with the osteoblast conditioned media (CM), which progressively enhanced osteoclastogenesis and bone resorption if harvested from osteoblast cultures grown under microgravity, with a maximum effect at 0.008xg. Interestingly, among the pro-osteoclastogenesis factors produced by osteoblasts, RANKL/OPG ratio was progressively increased at 0.08xg and 0.008xg, while similar levels were observed in the various gravity conditions for IL-6, IL-1beta, M-CSF, PTHrP and TNFalpha. Consistently, treatment with excess OPG blunted the higher osteoclastogenesis potential of CM from microgravity-treated osteoblasts. In contrast, microgravity failed to affect osteoblast differentiation and function in the time frame (24 h) of the experiment, as evidenced by unchanged ALP activity, and ALP, runx2, osteocalcin, osteopontin and collagen 1A2 mRNA expression. This is at variance with long-term exposure to low gravity which impairs osteoblast activity, suggesting that the effect on osteoclastogenesis precedes the outcome on the osteoblast phenotype. Various protein kinases which are major determinants of osteoblast activity, including c-Src, PKC and the MAPKs p38 and JNK, were not changed by any of the microgravity conditions. In contrast, microgravity induced an increase of ERK-1/2 phosphorylation starting within 60 min at 0.08 × g, and 30 min at 0.008 × g. The increase reached a plateau at 10 hours, still persisted at 24 hours, and could be blunted by the MEK-1/2 inhibitor PD98059. Treatment with PD98059 had no effect on the RANKL/OPG ratio, but enhanced apoptosis under microgravity, suggesting a protection role by ERK-1/2 against cell death. In conclusion, microgravity exerts and indirect stimulation of osteoclast formation and activity by affecting the secretion of crucial regulatory factors such as RANKL and OPG by osteoblasts. We hypothesize that this mechanism could contribute to bone loss in individuals subjected to unloading conditions, including the weightlessness in spaceflight.

Disclosures: A. Rufo, None.


Mechanoresponsiveness in Neonatal Rat Femur Organ Cultures. M. Saunders, A. Taylor*. Orthopaedics and Rehabilitation, Penn State College of Medicine, Hershey, PA, USA

Bone mechanotransduction is the process by which bone cells respond to a physical stimulus. While in vitro and in vivo mechanotransduction models are common, a less utilized model is the ex vivo (organ culture) system whereby bones maintained in culture are subjected to mechanical load and the effect of the stimulation assessed. Here, we developed an organ culture model for mechanotransduction.

Structurally intact rat femurs harvested from 2 day old Wistar rats were weighed and measured and placed in BGJb culture medium supplemented with 15% FBS and 2% P/S. Medium was changed every 72 Hr and bones were maintained in culture for 8 days. To verify bone viability in culture we assessed growth changes during the culture period (between day 1 and 8). We harvested femur pairs and compared mechanical strength; femur pairs harvested from the same litters were mechanically tested 24 Hr post harvest or 8 days post harvest. We also verified that diffusion was capable of providing adequate nutrition via incorporation of Procion red dye. To assess the effect of mechanical stimulation on organ culture, bones were mechanically stimulated using three-point bending 24 Hr after harvesting (day 1) and compared at day 8 to non-stimulated controls. Loaded limbs underwent a single loading bout of 350 cycles at 1Hz with a minimum of 50 and a maximum of 500 microstrain at the mid-diaphyseal shaft. Where appropriate, data were analyzed with ANOVAs and a p value of 0.05.

We found that during the 7 day culture period, bones significantly grew with respect to length. Specifically, shaft length was increased 9.7% (p = 0.0273) over the 7 day culture period and overall femur length was increased 13.5% (p = 0.0037) during this period (n = 12). Mechanically, a greater than 2-fold increase in strength was demonstrated in limbs tested at 8 days when compared to contralateral limbs tested 24 Hr post harvest. Histologically, bone was found to be viable and Procion dye labeling demonstrated that diffusion was adequate. Based upon these results, we concluded that the bones utilized in the organ culture model remained viable during the culture period. From mechanical testing on day 8, we found that mechanical stimulation increased bending stiffness 41%, increased failure load 10.7% and decreased failure displacement 44.7% over non-stimulated controls; differences in stiffness (p = .0015) and displacement (p = .0169) were significant (n = 4). We concluded that the organ culture model was mechanoresponsive.

Our interest is in the development of an ex vivo mechanotransduction model. Ex vivo systems have the advantage of eliminating systemic effects found in in vivo systems while maintaining the 3D structure, milieu and cell interactions of the physiologic system ignored in in vitro models.

Disclosures: M. Saunders, None.


LRP5 G171V Mutation and Disuse-Related Bone Structural Properties. M. P. Akhter, G. K. Alvarez*, B. L. McGuckin*, D. M. Cullen. Medicine, Creighton University, Omaha, NE, USA

Osteoporosis is a disease of low bone mass that results in fragility fractures. Lifestyle choices such as a sedentary (disuse or lack of exercise) life style and smoking may further deteriorate low bone mass associated with osteoporosis. In this study we characterized bone structural properties to determine if the LRP5 G171V mutation will protect against bone loss associated with disuse. Forty eight adult male mice representing three genotypes (WT = wild type, KO = Lrp5-knockout), HBM = High bone with LRP5 G171V mutation) were randomly divided between control and Disuse groups (Table). The mice in the disuse group were hind limb suspended in individual cages for four weeks. All mice were inspected daily to assure compliance, adequate food intake, and general health. At the time of necropsy, femurs were collected and frozen in saline for subsequent biomechanical and structural evaluations. Trabecular bone in the distal femurs was analyzed for structure (micro-CT scans). We analyzed the data using the General Linear Model for univariate analysis to test for differences due to disuse within each genotype. Bone volume fraction (BV/TV) declined in all the three genotypes (Table). While trabecular thickness was lower in the HBM and KO disuse group as compared to the controls (Table). The remaining parameters of bone structure (Trabecular number/separation, connectivity density, apparent and tissue density) were lower in the KO with disuse as compared to the controls. Disuse resulted in the same absolute loss of bone volume fraction (0.06); however the relative loss was far greater in the KO mice (67%) than in the HBM mice (14%). The bone loss resulted in at least a 20% decrease in trabecular number and thickness for KO, while the decline was less than 10% for the HBM. While the already fragile KO had a large decrease in bone structure, the HBM losses were moderate with final bone structure remaining well above normal for the C57B16 wild type mice. The mice with the LRP5 G171V mutation did show a significant response to disuse demonstrating mechanosensitivity. However, the relative advantage in bone volume between HBM and KO went from 4.8 fold in the control mice to 12.3 fold after disuse suggesting that the HBM genotype is relatively protected against disuse-related bone loss and fragility. 

Table  .  
  1. * Differences due to Disuse (P < 0.05)

inline image

Disclosures: M.P. Akhter, None.

This study received funding from: The State Nebraska (LB595) Funds.


The Climbing Exercise Enhances Osteoblast Differentiation and Inhibits Adipogenic Differentiation with Highly Expressed PTH / PTHrP Receptor in Bone Marrow Cells. K. Menuki*1, T. mori1, M. Sakuma*1, N. Okimoto*1, A. Sakai1, N. Kunugita*2, T. Nakamura1. 1Department of Orthopaedic Surgery, University of Occupational and Environmental Health, Kitakyushu, Japan. 2Department of Health Information, University of Occupational and Environmental Health, Kitakyushu, Japan

To investigate the effect of voluntary tower climbing exercise on osteogenic and adipogenic terminal differentiation of bone marrow, 56 C57BL/6J male mice, 8 weeks of age, were assigned to two groups: ground control and climbing exercise. The mice were sacrificed at 4, 7, 10, 14 and 28 days after exercise or not. Mice were housed in a 100-cm tower cage and had to climb up to drink water toward a bottle placed at the top. We obtained bilateral femurs and tibias to evaluate the population of mesenchymal stem cells (MSCs) by flow cytometric analysis, osteogenic and adipogenic potential by bone marrow cell culture, bone marrow adipocyte volume by histomorphometry, and mRNA expression in bone marrow cells by real-time RT-PCR. Flow cytometric analysis revealed that the population of CD31-CD45-CD166+ cells in bone marrow cells were significantly decreased in the climbing mice at 4 days, which has a property of MSCs. In bone marrow cell cultures, the number of alkaline phosphatase-positive colony forming units-fibroblastic (ALP+ CFU-f) and Oil red O positive cells were increased in the climbing mice at 4 days. The trabecular bone volume (BV/TV) of the proximal tibia was increased in the climbing mice at 28 days. In contrast, the marrow adipocyte volume (AV/TV) and adipocyte numbers were decreased. Quantitative RT-PCR revealed that PTH/PTHrP receptor 1 (PTHR1) mRNA expression significantly increased in the climbing mice at 4 days. In addition, flow cytometric analysis revealed that the expression of PTH1R in bone marrow cells significantly increased in the climbing mice at 7 days. Anti-PTH1R antibody suppressed the number of ALP+ CFU-f and increased the number of Oil red O+ cells in bone marrow cell culture. Our results suggested that climbing exercise enhanced MSCs differentiation, in consequence transiently decreased the population of MSCs and increased bipotential adipocyte-osteoblast precursor cell; however, inhibited the terminal differentiation of marrow adipocyte. This subsequent increase in bone formation and decrease in marrow adipocyte after climbing exercise seems to be associated with highly expressed PTHR1 in bone marrow cell.

Disclosures: K. Menuki, None.


Estrogen Receptor β's Contribution to Bone's Loading-Related Response Differs between Cortical and Cancellous Bone. L. K. Saxon*, J. S. Price*, L. E. Lanyon. Veterinary Basic Sciences, Royal Veterinary College, London, United Kingdom

The osteogenic response of the ulna cortex to artificial loading in mice lacking Estrogen Receptor a (ERα) is less than in Wild Type mice but in mice lacking ERβ it is greater. This suggests that ERα is necessary for cortical bone's adaptive response to loading but that its action is opposed by ERβ. To investigate the contributions of the ERs to the loading response in trabecular bone we axially loaded in vivo the right tibia of 7 female (F) and 6 male (M), 17 week old WT (C57/BL6) and ERβ−/− mice at 13 N (producing midshaft strains of 1500μω) for 40 cycles/day, 3 days a week for 2 weeks, using a trapezoid waveform with a 14.9 sec rest between each cycle. The left tibia acted as a normally loaded control. Differences in trabecular bone volume (BV/TV), number (Tb.N), thickness (Tb.Th), separation (T.Sp) and structure model index (SMI) in the proximal metaphysis of the loaded (right) versus control (left) tibiae were measured using microCT. Body weight was not different between male or female WT and ERβ −/− groups and there was no difference at baseline between the respective groups in BV/TV (M 17.2 vs 19.5; F 9.9 vs 9.7 %), Tb.N (M 3.4 vs 3.9; F 1.9 vs 1.8 mm) Tb.Th (M 0.05 vs. 0.5; F 0.05 vs. 0.05 μm), T.Sp (M 0.2 vs 0.2; F 0.3 vs 0.3 μm) and SMI (M 1.4 vs 1.3; F 1.9 vs 2.0). In contrast, while there was no difference in the loading-related increase between male ERβ−/− and WT mice (BV/TV −2.9% vs 3.9%; Tb.N 0.3% vs −5.5%; Tb.Th 4.4% vs. 2.9%, T.Sp 3.1% vs 1.1%; SMI −8.1% vs. 4.3%, respectively), the female ERβ−/− mice showed a significantly (p <0.05) lower response compared to the WT mice for BV/TV (9.6% vs 30.6%, respectively) and Tb.Th (8.6% vs 19.9%, respectively); no other parameter was different between these two groups. This outcome is the opposite to that reported in cortical bone where female ERβ−/− mice show a larger response to loading than WT. In conclusion, the absence of ERβ−/− does not appear to influence trabecular bone phenotype in response to normal loading in either male or female mice. However, in contrast to the situation in the cortex, the absence of ERβ in females but not males enhances the osteogenic response to loading in trabeculae.

Disclosures: L.K. Saxon, None.


PTH and PTHrP Targets Cyclin D1 and Induces Osteoblastic Cell Proliferation. N. S. Datta, G. J. Pettway*, C. Chen*, A. J. Koh*, L. K. McCauley. Periodontics and Oral Medicine, University of Michigan, Ann Arbor, MI, USA

Parathyroid hormone (PTH) and PTH related protein (PTHrP) are in clinical and investigative use for the treatment of osteoporosis; however, their underlying mechanisms of action remain unclear. Controversy has surrounded the ability of PTH to increase osteoblast proliferation versus osteoblast recruitment and/or differentiation; hence, the purpose of this study was to evaluate the impact of PTH and PTHrP in vitro and in vivo on the celt cycle regulator, cyclin D1. Mouse calvarial MC3T3-E1 clone 4 (MC-4) cells were synchronized by serum starvation and induced with 100nM PTHrP in media with 2% serum. PTHrP significantly up-regulated cyclin D1 expression and MC-4 cell proliferation when cells were seeded at low density. The rate of proliferation and cyclin D1 expression was dramatically reduced at confluent culture density. Electrophoretic mobility shift and luciferase transcription assays revealed that PTHrP regulation of cyclin D1 involved both AP-1 and CRE sites in the cyclin D1 promoter. The PTHrP mediation of cyclin D1 transcription was abolished with AP-1 and CRE double mutant promoter constructs. PTHrP mediated up-regulation of cyclin D1 was found to be PKA and MAPK dependent in proliferating MC-4 cells. Furthermore, a decrease in p27 and an increase of CDK1 proteins was noted in response to PTHrP. To corroborate these in vitro findings, ectopic ossicles were generated using bone marrow stromal cells (BMSCs) implanted in nude mice. One week following implantation two groups of animals were injected with either an anabolic dose of PTH or vehicle control, for one week (targeting proliferative stage cells) or 3 weeks (targeting proliferative stage and differentiating cells). Ossicles were collected, RNA isolated and analyzed for cyclin D1 expression with real time PCR, and histomorphometry was performed on parallel samples. In immature ossicles after one week of treatment there was negligible bone present but a dramatic up-regulation of cyclin D1 mRNA with PTH treatment. In contrast, in more mature ossicles with 3 weeks treatment, there was significantly increased bone area with PTH versus vehicle but only a slight non-significant increase in cyclin D1 expression. These studies suggest that PTH targets immature cells of the osteoblastic lineage to increase their cell cycle progression which results in increased bone, but at the later stages of osteoblast differentiation PTH does not facilitate cell cycle progression. The induction of cyclin D1 expression by PTH and PTHrP in proliferating osteoblasts, but not in more differentiated cells, may be important in bone cell turnover and anabolic actions of PTH.

Disclosures: N.S. Datta, None.

This study received funding from: NIH.


Decreased Alpha5 Betal Expression Mediates the Immediate Osteoblast Apoptosis Induced by Skeletal Unloading in Rats. C. Dufour*1, X. Holy2, P. J. Marie1. 1Laboratory of Osteoblast Biology and Pathology, Inserm U606, Paris, France. 2Dept Aerospace Physiology, IMASSA, Bretigny-sur-Orge, France

Skeletal unloading in rats results in bone loss as a consequence of alterations in bone formation. The cellular and molecular mechanisms underlying the altered osteoblastogenesis induced by unloading are not fully understood. In this study, we investigated the effect of skeletal unloading on osteoblast apoptosis and identified the underlying cellular mechanisms. Skeletal unloading induced by tail suspension in rats reduced trabecular number and thickness and induced metaphyseal bone loss at 4 and 7 days, as a result of diminished osteoblast surface in the long bone metaphysis. Unloading induced an immediate increase in the number of apoptotic osteoblasts in the tibial metaphysis compared to loaded rats, as shown by Tunel staining in situ. Osteoblast apoptosis increased by 4-fold at 2 and 4 days and by 2-fold at 7 days of unloading. Osteocyte apoptosis was unaffected at 2 days but was increased by 60% at 4 and 7 days of unloading. Surrenal gland weights and plasma corticosterone levels did not differ in loaded and unloaded rats at all times, indicating that osteoblast/osteocyte apoptosis occured independently of endogenous glucocorticoids. Western blot analysis in metaphyseal bone in vivo revealed a 50% decrease in α5 and β1 integrin protein levels in unloaded bones compared to loaded bones at 2 and 4 days. This transient decrease in α5 and β1 integrin expression was associated with a parallel decrease in the pro-survival protein Bc1–2. This led to a 2.4-fold increase in Bax/Bc12 in unleaded compared to loaded bone at 2 and 4 days. These results show for the first time in unloaded rats that skeletal unloading induces a glucocorticoid-independent, immediate increase in osteoblast apoptosis in metaphyseal long bone which is associated with reduced α5β1 expression and decreased Bc1–2 levels in metaphyseal bone. We conclude that skeletal unloading in rats leads to an immediate decrease in α5β1 integrin levels in unloaded metaphysis, which results in diminished Bc1–2 levels and subsequent osteoblast apoptosis. This provides an integrin-mediated mechanism that contributes to the reduced osteoblastogenesis and bone loss induced by unloading in rats.

Disclosures: P.J. Marie, None.


PTH/PTHrP Suppresses Cell Death in Cultured Osteoblasts by Inhibiting Akt Signaling. M. S. Schnoke*, R. J. Midura. Biomedical Engineering, ND20, Cleveland Clinic, Cleveland, OH, USA

PTH/PTHrP suppresses osteoblast apoptosis in vitro and in vivo. This decrease in cell death is thought to be one mechanism by which PTH functions to enhance bone mass in vivo. However, the mechanism of action underlying this suppression of osteoblast apoptosis is not completely understood. The aims of this study were to determine the predominant signaling mechanism responsible for this anti-apoptotic activity, and determine whether these effects are mediated by Akt/PKB signaling. UMR106–01 BSP mature osteoblastic and MC3T3-E1 pre-osteoblastic cell lines were used in this study, and the results from both of these cell lines were similar. Apoptosis was determined utilizing the Roche Cell Death Detection assay, and live cell number was confirmed with the MTT assay. All reagents tested that enhanced protein kinase A (PKA) activity provided potent anti-apoptotic effects, these included: [1] PTH1–84, PTH1–34, PTH1–31, and PTHrP1–36 peptides from 10–100 nM; [2] forskolin from 10–25 uM; and [3] 8-Br-cAMP from 0.3–0.6 mM. All reagents tested that did not enhance PKA activity failed to provide anti-apoptotic effects, these included: PTH3–34, PTH7–34, PTHrP107–111, and PTHrP107–139 peptides from 10–100 nM. The anti-apoptotic effects of PTH peptides and forskolin were blocked by the simultaneous exposure to the PKA inhibitors, H89 at 10 uM or the A-kinase anchoring protein inhibitor (InCELLectTM AKAP St-Ht31) at 10 uM. All reagents that stimulated PKA activation resulted in a decreased phosphorylation of Akt on Ser473 as determined by Western blot analysis, and a simultaneous exposure to PKA inhibitors reversed these effects on Akt phosphorylation. Blocking Akt phosphorylation using the PI3 kinase inhibitors wortmannin at 1 uM or LY294002 at 20 uM mimicked the anti-apoptotic effects of PKA activators in these two osteoblastic model systems. Western blot analysis confirmed that these 2 compounds potently attenuated the phosphorylation of Akt at Ser473. Altogether we conclude that PTH/PTHrP's activation of PKA leads to a decrease in the phosphorylation of Akt, which attenuates Akt downstream signaling and results in an inhibition of apoptosis in cultured osteoblastic cells. The findings of this study suggest that Akt can act as a pro-apoptotic factor in osteoblasts.

Disclosures: M.S. Schnoke, None.

This study received funding from: NIAMS, NIH.


Phosphatidylinositol (PI) 3 Kinase/Akt Signaling Represents a Potential Mechanism for BMP-2-Induced Expression of Colony Stimulating Factor-1 (CSF-1) which Mediates Osteoclast Differentiation. C. C. Mandal*1, P. K. Singha*1, S. L. Pandeswara*1, G. Ghosh-Choudhury*2, N. Ghosh-Choudhury3. 1Pathology, UTHSC, San Antonio, TX, USA. 2Medicine, UTHSC, STVHCS, San Antonio, TX, USA. 3Pathology, UTHSC, STVHCS, San Antonio, TX, USA

Cooperative action of RANKL and CSF-1 regulates formation of mature osteoclasts (OC). CSF-1 is produced by the osteoblasts (OB) and acts in a paracrine fashion to stimulate its receptor present in the OC progenitors. BMP-2 promotes differentiation of multipotent C2C12 cells to OB lineage. Coculturing mouse spleenocytes with BMP-2-treated C2C12 cells promoted TRAP positive multinucleated osteoclast formation with concomitant increase in expression of CSF-1 mRNA. The mechanism by which BMP-2 stimulates CSF-1 expression is not known. To elucidate the signal transduction pathway, we investigated the involvement of PI 3 kinase. In C2C12 cells, BMP-2 increased PI 3 kinase activity, with concomitant increase in Akt kinase activity. Ly294002 (Ly) inhibited BMP-2-induced PI 3 kinase activity, resulting in attenuation of CSF-1 mRNA expression. Adenoviral-mediated expression of PTEN, which blocks PI 3 kinase signaling, or dominant negative Akt inhibited CSF-1 mRNA expression by BMP-2. Using a reporter construct in which luciferase cDNA is driven by 627 bp 5 flanking sequence spanning transcription start site of CSF-1 gene (−627-Luc), we have recently shown that BMP-2 stimulates CSF-1 transcription through BMP-specific Smad 1/5. To test the involvement of PI 3 kinase in CSF-1 transcription, −627-Luc-transfected C2C12 cells were treated with Ly followed by incubation with BMP-2. Ly significantly inhibited the transcription of CSF-1 induced by BMP-2. Transfection of dominant negative PI 3 kinase or PTEN or dominant negative Akt kinase with the reporter plasmid blocked BMP-2-induced transcription of CSF-1, indicating PI 3 kinase/Akt signaling contributes to CSF-1 expression during osteoblastic differentiation of C2C12 cells. BMP-2-induced CSF-1 transcription was abolished when the BMP response element (BRE) present in the −627-Luc was mutated, suggesting the functional involvement of the BRE. Furthermore, in C2C12 cells, BMP-2 stimulated the transcription from a heterologous promoter in which the transcription of the reporter luciferase is driven by three copies of the BRE (3xBRE-Luc). Incubation of 3xBRE-Luc transfected C2C12 cells with Ly inhibited BMP-2-induced transcription of the reporter gene. Together, these data for the first time demonstrate a novel mechanism where PI 3 kinase/Akt signaling in OB contributes to BMP-2-mediated expression of CSF-1 that controls osteoclastogenesis of the progenitor cells.

Disclosures: C.C. Mandal, None.


The Liganded Versus Unliganded Estrogen Receptor Has Opposite Effects on Osteoblast Differentiation. M. Almeida, X. Chen, L. Han, M. Martin Millan, V. Lowe*, A. Warren*, S. A. Stewart*, S. Kousteni, R. S. Weinstein, C. A. O'Brien, T. Bellido, R. L. Jilka, S. C. Manolagas. Center for Osteoporosis and Metabolic Bone Diseases, University Arkansas for Medical Sciences, and Central Arkansas Veterans Healthcare System, Little Rock, AR, USA

It has been recently shown that the unliganded estrogen receptor (ER) plays an essential role in the transduction of mechanical signals that promote osteoblast and osteocyte survival. Moreover, whereas estrogens suppress osteoblastogenesis and estrogen loss increases it, ERα null mice exhibit a paradoxical decrease in osteoblast number and bone formation rate. Based on this and evidence for distinct roles of unliganded versus liganded ER in transcriptional repression of the TNF gene (Molecular Cell 21:555, 2006), we have tested the hypothesis that the ER may have two opposite effects in osteoblast differentiation: attenuation in the presence of E2 and potentiation in the absence of E3. We report that E2 at 10−8 M suppressed BMP-2-induced osteoblast progenitor commitment and differentiation in both the uncommitted mesenchymal cell line C2C12 and in primary cultures of murine (C57B1/6) bone marrow cells, as measured by alkaline phosphatase activity. Strikingly, treatment of C2C12 cell cultures with ICI 182,780 (10−7 M) also attenuated BMP-2-induced alkaline phosphatase activity. Furthermore, consistent with a suppressive effect on osteoblastogenesis, ICI also attenuated BMP-2-induced osteocalcin as well as the expression of several BMP-2 responsive genes such as Smad6, collagen I, Runx2, and Axin2. Confirming the known effect of ICI on ER degradation, Western blot analysis (using an ERα-specific antibody) of C2C12 cells treated with this compound over a 5 day period demonstrated a 50% and 70% decrease in the abundance of the ERα protein levels at day 1 and 3, respectively. However, by the 5th day the ERα protein level rebounded, coinciding with a reversal of the inhibitory effect of ICI on the alkaline phosphatase expression. Finally, we compared mice carrying two vs. one copy of the functional ERα gene (ERα+/+ vs ERα+/−). Strikingly, at 8 months of age, spinal, femoral and total BMD (measured by DEXA), were higher in the ERα+/+, as compared to ERα+/−. These results strongly suggest that the ERα can either induce or repress osteoblast differentiation and bone formation, depending on its activation by estrogens. Hence, the increased osteoblastogenesis that follows loss of estrogens at menopause may result from increased availability of unliganded ERα serving as co-activator of BMP- and perhaps Wnt-induced genes involved in osteoblastogenesis. Remarkably, the same molecular mechanism may be responsible for increased cytokine production and osteoclastogenesis.

Disclosures: M. Almeida, None.


Inducible cAMP Early Repressor Blocks ATF4 Function at its Binding Site in the Mouse Osteocalcin Promoter. T. K. Chandhoke, B. E. Kream. Medicine, University of Connecticut Health Center, Farmington, CT, USA

Inducible cAMP early repressor (ICER) is a member of the ATF/CREB superfamily of transcription factors, containing a basic leucine zipper (bZIP) domain for DNA binding and dimerization. Encoded by the P2 promoter of the CREM gene, ICER functions as a dominant negative repressor of cAMP-mediated transcription. To assess in vivo the role of ICER in bone, osteoblast-targeted ICER transgenic mice were previously generated by cloning ICER cDNA with an N-terminal FLAG epitope downstream of a 3.6 kb fragment of the rat Col1a1 promoter (pOBCol3.6-ICER). ICER transgenic mice showed a definitive low bone mass phenotype, with marked reduction in overall body size and a disruption of trabecular bone with a significant reduction in BV/TV, trabecular number and thickness by MicroCT. Interestingly, Northern blot analysis of in vivo femoral mRNA expression of bone markers bone sialoprotein (BSP) and Col1a1 from ICER mice were unchanged, while osteocalcin expression was dramatically reduced. Activating transcription factor-4 (ATF4), another member of the ATF/CREB family, was recently shown to induce osteocalcin expression at the OSE1 site within the mouse OG2 (mOG2) promoter. The purpose of the present study was to determine if the reduction in osteocalcin expression could be due to suppression of ATF4 function by ICER. To determine if ICER regulates ATF4-induced osteocalcin gene expression, MC3T3-E1 cells were transiently transfected with a promoter-reporter construct containing four multimerized OSE1 sites cloned into the pGL3-basic vector (4OSE1-Luc). Cells were co-transfected with pCR3.1-ICER and/or pCMV-ATF4 expression vectors and pRL-SV40 (Renilla luciferase) internal control vector. ATF4 robustly induced 4OSE1-Luc activity 30-fold over control. When ICER DNA was introduced, ATF4 induction at the OSE1 site was potently repressed to baseline in a dose-dependent manner. Results from this study show that ICER suppresses ATF4 function at the OSE1 site, which may be responsible for the loss of osteocalcin expression in ICER transgenic mice.

Disclosures: T.K. Chandhoke, None.




Downregulation of Ubiquitin Ligase Cb1 Induced by Twist Haploinsufficiency in Saethre-Chotzen Syndrome Results in Increased PI3K/Akt Signaling and Osteoblast Proliferation. H. Guenou*, C. Dufour*, K. Kaabeche*, H. Miraoui*, P. J. Marie. Laboratory of Osteoblast Biology and Pathology, Inserm U606, Paris, France

Genetic mutations of Twist, a bHLH transcription factor, induce premature fusion of cranial sutures (craniosynostosis) in Saethre-Chotzen syndrome (SCS). The signaling mechanisms that are involved in the altered osteoblast phenotype in SCS are not fully understood. We report here a previously undescribed mechanism implicated in the altered osteoblastogenesis in SCS. Cranial osteoblasts from an SCS patient with a Y103X Twist mutation causing deletion of the bHLH domain showed decreased expression of the E3 ubiquitin ligase Cbl compared to wild type osteoblasts. Biochemical studies revealed that the decreased Cbl expression in Twist mutant osteoblasts induced reduction of the ubiquitin-mediated degradation of phosphatidyl inositol 3 kinase (PI3K), resulting in increased PI3K/Akt protein levels and signaling. Increased PI3K immunoreactivity was also found in vivo in osteoblasts in affected cranial sutures from SCS patients with distinct Twist mutations. Transfection of Twist mutant osteoblasts with Twist or Cb1 expression vectors corrected Twist and Cb1 levels and abolished the increased PI3K/Akt signaling. Forced expression of Cbl did not correct the altered expression of osteoblast differentiation markers in Twist mutant cells. In contrast, pharmacological inhibition of PI3K signaling in Twist mutant osteoblasts corrected the increased cell growth induced by Twist haploinsufficiency. These results show that Twist haploinsufficiency in osteoblasts reduces Cb1 expression which in turn decreases Cbl-mediated PI3K degradation, causing PI3K accumulation and activation of PI3K-dependent osteoblast growth. This provides novel genetic and biochemical evidence for a role of Cb1-mediated PI3K signaling in the altered osteoblast phenotype induced by Twist haploinsufficiency in SCS.

Disclosures: P.J. Marie, None.


Inhibition of the Transcriptional Activity of Osterix by Interactions with Obelix, a Jumonji Family Chromatin Protein. K. M. Sinha*, X. Zhou*, C. Zhang*, L. Zhang*, B. de Crombrugghe. Molecular Genetics, M.D. Anderson Cancer Center, Houston, TX, USA

Osterix (Osx) is a critical osteoblast-specific transcription factor required for bone formation. Genetic inactivation of Osx in mice leads to complete arrest of osteoblast differentiation, although chondrocyte differentiation and cartilage formation are normal. However, the mechanism of Osx function during osteoblast differentiation is not well understood.

In the present study, the Jumonji (JmjC) domain containing protein, NO66 (Obelix, renamed by us), was identified as an Osx-interacting protein. This result was based on the use of an osteoblast cell line stably transfected with a doubly tagged-Osterix followed by tandem immunoaffinity purification and mass spectrometry. Co-immunoprectpitation and GST-pulldown assays showed that Obelix physically interacted with Osx. Obelix interacts through its JmjC domain with the transcription activation domain of Osx. In situ RNA hybridization experiments revealed that Obelix and Osx were co-expressed in skeletal elements of hind limbs and forelimbs, in vertebrae, ribs, and craniofacial bones during mouse embryonic development. In DNA transfection assays, Osx strongly stimulated the activity of a 1kb osteocalcin promoter and that of an osteoblast specific 2.3-kb Collal promoter. Co-transfection of Obelix resulted in a strong inhibition of the Osx-dependent activity of these promoters, whereas the activity of several other promoters was not affected by Obelix. In addition, transfection of Obelix in osteoblasts also inhibited the activity of the endogenous osteocalcin gene. Furthermore, knockdown of Obelix expression in osteoblasts by specific siRNAs resulted in increased expression of several osteoblast-specific marker genes including Collal, Osteocalcin and Bone Sialoprotein (BSP) without apparent changes in the level of Osx expression.

Overall our data suggest the hypothesis that Obelix is a negative regulator of Osx activity and, hence, of osteoblast differentiation or function. The presence of a JmjC domain in Obelix raises the intriguing possibility that the inhibitory effect of Obelix is mediated by detnethylation of specific lysine residues of histones within the chromatin environment of osteoblast-specific genes.

Disclosures: K.M. Sinha, None.


A Transgenic Parabiosis Model Does not Show the Presence of Circulating Osteoprogenitor Cells. I. Boban*, T. Barisic-Dujmovic*, S. H. Clark. Deparment of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, CT, USA

The goal of this study was to determine the presence of osteoprogenitor cells in peripheral blood. A parabiosis model was developed utilizing various collagen promoter transgenes. The Col3.6 promoter has been shown to be active in preosteoblasts and osteoblasts as well as in other collagen producing cells. However, the expression of Col2.3 transgenes is restricted to osteoblasts and osteocytes. Several experiments were conducted with the parabiosis model in which Col3.6GFP or Col2.3GFP transgenic mice were surgically joined with mice bearing a Col2.3 delta TK transgene. After a common circulation was established, parabiotic pairs were treated with ganciclovir (GCV). In individual Col2.3 delta TK transgenic mice, 15 days of GCV treatment caused an ablation of osteoblasts and destroyed bone marrow integrity. In initial experiments that we reported with Col3.6GFP/Col2.3 delta TK pairs, after 15 days of GCV treatment, GFP expression was observed in histological preparations from the Col2.3 delta TK. parabiont around trabecular bone and on endosteal and periosteal surfaces. In addition, bone marrow integrity was maintained. However, more detailed analyses demonstrated that these GFP expressing cells were of hematopoietic origin, more specifically belonging to the osteoclast lineage. In more recent experiments, Col2.3GFP transgenic mice were paired with mice bearing a Col2.3 delta TK transgene as the Col2.3GFP transgene expression is limited to osteoblasts and osteocytes. In parabionts treated for 15 days and immediately examined for GFP expression, Col2.3GFP expression was not detected in bone histological preparations or in marrow stromal cell cultures from the Col2.3 delta TK parabiont. This was also the case in pairs treated for 15 days and allowed to recover from GCV treatment for 3 months. Finally, pairs were treated continuously with GCV for 1.5 to 2.0 months. Again there was a failure to detect Col2.3GFP expressing cells in the Col2.3 delta TK parabiont, either in histological preparations or in marrow stromal cell cultures. Based on these observations, at least within the limits of this model system, we are lead to the conclusion that osteoprogenitor cells are not a component of the peripheral blood.

Disclosures: I. Boban, None.


Feasibility Clinical Trial: Early Results Show Healing of Recalcitrant Non-Union Fractures Using Autologous Tissue Repair Cells, a Bone Marrow-Derived Tissue Enriched for Adult Stem and Progenitor Cells. M. Jimenez*1, T. Lyon*2, J. Balazsy*3, G. Nowinski*3, J. Goulet*4, S. Wolff*5, J. Block*6, T. Sferra*1, R. D. Armstrong*7, J. Douville*7, G. O'Brien*7, J. M. Hock7. 1Lutheran General Hospital, Park Ridge, IL, USA. 2Lutheran Medical Center, Brooklyn, NY, USA. 3William Beaumont Hospital, Detroit, MI, USA. 4University of Michigan, Ann Arbor, MI, USA. 5Meharry Medical College, Nashville, TN, USA. 6Vanderbilt University, Nashville, TN, USA. 7Aastrom Biosciences, Inc, Ann Arbor, MI, USA

Autologous Tissue Repair Cells (TRC) are obtained from small volume bone marrow aspirate after processing for 12–13 days by proprietary single pass perfusion in an automated, computer-controlled AastromReplicell System. Flow cytometry and clonogenic and lineage fate assays show enrichment for stem and progenitor cells of mesenchymal, endothelial and hematopoietic lineages, with supporting stroma. Clinical trials of TRC in bone marrow reconstitution in >187 cancer patients showed safe successful cell engraftment equivalent to 20-fold higher volume of fresh bone marrow. We hypothesized TRC as a source of progenitor and endothelial cells would promote healing of recalcitrant non-union fractures. We assessed safety of TRC in FDA-approved prospective multi-site consecutive case series of 36 patients (pilot phase 11 + expanded phase after process improvements: 25). We report results in the first 7 patients who completed 6 months of observation, as initial proof-of-principle.

Five men and 2 women with tibial non-union that failed to heal after 1–3 standard of care bone grafting and surgical treatments in the prior 12 (range: 8–29) months, were enrolled. Previous treatment included failures in internal or external fixation to align and immobilize the fractured bone, autologous bone grafting and bone morphogenetic protein (BMP) therapy. TRC-treated patients, aged 30-73 years, underwent open reduction and internal fixation surgery in which TRC at 3 × 10E6 cells/cc matrix were applied to exposed bone surfaces, after mixing with commercial allograft matrix graft extender (MTF, NJ). TRC were characterized by flow cytometry.

Bone regeneration, as radiographic callus formation or bone bridging was observed in 4 patients by 3 months, and in 7 (100%) patients, 2 of whom were smokers, by 6 months. Post-surgical evaluations of these patients using standard clinical and radiographic evaluations of the healing fracture site will continue for a total of 12 months surveillance. There were no TRC-adverse events. This is the first report of the use of a standardized manufactured bone marrow-derived, multi-lineage cell product to promote healing of recalcitrant non-union fractures.

Disclosures: J.M. Hock. Aastrom Biosciences 1, 3.

This study received funding from: Aastrom Biosciences, Inc.


Bone Marrow Stromal Cell Transplants Recapitulate Osteoporotic Bone. M. H. Mankani1, B. Halloran2, G. W. Marshall3, R. A. Nissenson2. 1Surgery, UCSF, San Francisco, CA, USA. 2Medicine and Physiology, UCSF, San Francisco, CA, USA. 3Preventive and Restorative Dental Sciences, UCSF, San Francisco, CA, USA

Bone marrow stromal cells (BMSCs) mediate bone abnormalities in gene mutation-associated diseases such as fibrous dysplasia and McCune-Albright syndrome. To date, BMSCs have not been implicated in the formation of abnormal bone in diseases which do not have a discrete genetic etiology, such as osteoporosis arising from estrogen-deficiency. The purposes of this study were to 1) Evaluate the effects of ovariectomy on BMSC activity, and 2) Establish a model for engraftment of BMSCs in osteoporotic animal recipients using BMSC transplants.

Methods: Six-month-old retired breeder female Sprague-Dawley rats underwent either ovariectomy (OVX group) or sham surgery (control group). BMSCs were isolated from all rats two months later. BMSC CFU number, apoptotic activity (via an Annexin V-FITC/propidium iodide (PI) assay), proliferation rate (via BrDU staining), and alkaline phosphatase (AP) expression were characterized. BMSC numbers were expanded in cell culture. At passages 3–5, BMSCs were combined with hydroxyapatite-tricalcium phosphate particles and autotransplanted to the subcutis of the back. Transplants and fibulas were harvested at 2 months; bone content was evaluated by histology and mechanical properties evaluated by nanoindentation using modified atomic force microscopy (AFM).

Results: The ratio of OVX-derived BMSC colonies to control-derived BMSC colonies was 1.00 (for CFU) and 1.12 (for AP+ colonies). In contrast, the OVX/control ratio for BrDU+ cells was 0.12, and the OVX/control ratio for apoptotic cells was 22.6. Histologic evaluation of the transplants demonstrated equivalent amounts of bone from both groups. AFM of bone within the transplants demonstrated that the OVX bone had a lower elastic modulus but equivalent hardness (Figure); this was in contrast to the fibulas of the OVX rats, which had both a lower elastic modulus and hardness.

Conclusions: Ovariectomy acts to diminish proliferation and increase apoptosis in rat BMSCs, but leaves the number of CFUs and AP+ colonies unchanged. BMSCs derived from OVX donors and auto-transplanted into OVX recipients can recapitulate the mechanical properties of osteoporotic bone, suggesting that BMSCs may act as mediators in the formation of weakened bone.

Thumbnail image of
Thumbnail image of

Disclosures: M.H. Mankani, None.

This study received funding from: UCSF Research Evaluation and Allocation Committee.


Genetic Disruption of All Nitric Oxide Synthase Isoforms Enhances Bone Turnover and Bone Mineral Density in Mice In Vivo. K. Sahanai*1, M. Tsutsui*2, A. Sakai1, H. Hirasawa*1, S. Tanaka*1, E. Nakamura*1, N. Yanagihara*2, T. Nakamura1. 1Department of Orthopaedic Surgery, University of Occupational and Environmental Health, Kitakyushu, Japan. 2Department of Pharmacology, University of Occupational and Environmental Health, Kitakyushu, Japan

Nitric oxide (NO) has multiple important actions that contribute to the maintenance of bone homeostasis. NO is synthesized by three different isoforms of NO synthase (NOS), including neuronal, inducible and endothelial NOSs. Because there are substantial compensatory interactions among the NOS isoforms, the ultimate roles of endogenous NO in bone metabolism still remain to be fully elucidated. We have recently succeeded in developing mice lacking all three NOS isoforms (triply n/i/eNOS−/− mice) (PNAS 2005). Thus, in this study, we investigated the role of NO derived from the whole NOS system in those mice. Experiments were performed in 12-week-old male wild-type C57BL/6 and triply n/i/eNOS−/− mice. The kinetics of bone formation and the kinetics of bone absorption were evaluated by morphological and biochemical markers. Bone mineral density was analyzed with dual energy X-ray absorptiometry. Bone formation rate and mineral apposition rate, which are morphological markers of bone formation, and serum alkaline phosphatase concentration, which is a biochemical marker of bone formation, were all significantly enhanced in the triply n/i/eNOS−/− mice as compared with the wild-type mice. Furthermore, osteoclast number and osteoclast surface, which are morphological markers of bone absorption, and urinary deoxypyridinoline concentration, which is a biochemical marker of bone absorption, were also ail significantly more increased in the triply n/i/eNOS−/− than in the wild-type mice. In addition, significantly higher levels of bone mineral density in the spine and the femur diaphysis were noted in the triply n/i/eNOS−/− mice. These results provide the first evidence that genetic disruption of all three NOS isoforms accelerates both bone formation and bone absorption, demonstrating a novel inhibitory effect of the NOS system-derived NO against bone turnover. Since bone mineral density was significantly higher in the triply mutant mice, the NOS system-derived NO appears to affect predominantly the process of bone formation, rather than that of bone absorption, in mice in vivo.

Disclosures: K. Sabanai, None.


GSK-3β Inhibitor Stimulates Osteoblast Differentiation In Vitro and Bone Formation In Vivo. M. Chen, T. Sheu, M. Zhu, R. J. O'Keefe, P. Chen. Orthopaedics, University of Rochester, Rochester, NY, USA

Mouse and human genetic evidence indicates that canonical Wnt/β-catenin signaling plays an important role in bone formation. GSK-3β kinase forms a complex with β-catenin and regulates β-catenin function through phophorylation of β-catenin followed by the proteasomal degradation. BIO (6-biomoindirubin-3′-oxime) is derived from Tyrian Purple Indirubin of Shellfish and is a permeable and selective GSK-3β inhibitor. In the present studies, we cultured primary mouse calvarial osteoblasts with 1 μM BIO for 0–12 hours. Western blot analysis showed that non-phosphorylated β-catenin (active form) was increased within 2 hours, and was maximal at 4 hours. Nuclear localization of β-catenin was observed by immunofluorescence staining within 4 hours after BIO treatment. BIO up-regulated β-catenin signaling reporter activity by 5-fold in osteoblasts isolated from β-catenin signaling reporter TOPGAL transgenic mice. Total RNA was extracted from primary osteoblasts treated with BIO or vehicle. The expression of osteoblast marker genes was measured by real-time RT-PCR assay. The expression of osteoblast marker genes such as Runx2 and osteocalcin was up-regulated by BIO by 2.5 and 2-fold respectively. In primary osteoblasts cultured in differentiation media for 9 days, BIO (1 μM) induced alkaline phosphatase activity 5-fold. Cell viability studies demonstrated minimal toxicity in primary osteoblast cultures in vitro up to a dose of 5 μM. The results suggest that BIO mimics canonical Wnt signaling in osteoblasts and promotes osteoblast differentiation. To determine the effect of BIO on bone formation in vivo, BIO (50 μg/10g/day, x5 days) was injected subcutaneously over the surface of calvaria in 4-week-old male mice. Calcein was injected at day 7 and 14 after BIO administration to label the mineralization fronts of newly formed woven bones. Mice were sacrificed at day 15 after BIO administration. A significant amount of new woven bone was formed on the surface of calvaria in mice receiving BIO. In contrast, no new bone formation was observed in the mice receiving vehicle injections. These results demonstrate that inhibition of β-catenin phosphorylation by GSK-3β inhibitor BIO leads to osteoblast differentiation in vitro and new bone formation in vivo and suggest that selective activation of β-catenin signaling in osteoblasts may have a potential for clinical intervention for the treatment of bone loss-associated diseases.

Disclosures: M. Chen, None.

This study received funding from: NIH.


Antibody-Induced Dimerization Activates the Ror2 Tyrosine Kinase Receptor and Stimulates 14-3-3 Phosphorylation and Osteogenesis in Human Mesenchymal Stem Cells. Y. Liu1, J. Ross*2, B. S. Komm1, P. V. N. Bodine1, J. Billiard1. 1Women's Health & Musculoskeletal Biology, Wyeth, Collegeville, PA, USA. 2Protein Chemistry and Proteomics, Wyeth, Cambridge, MA, USA

We have previously shown that the Ror2 receptor tyrosine kinase promotes osteoblastogenesis by an unidentified mechanism. To elucidate the Ror2 signaling pathway, we performed mass spectroscopy analysis of the potential Ror2 binding factors. U2OS osteosarcoma cells were transiently transfected with Ror2-flag or pcDNA3.1 expression plasmids, and the whole-cell extracts were immunoprecipitated on anti-flag affinity agarose and analyzed by mass spectroscopy. This analysis identified a scaffold protein 14-3-3β as a potential Ror2 binding partner. 14-3-3 proteins modulate an enormous number of cellular processes by binding to phosphorylated sites on diverse target proteins. The interaction between Ror2 and 14-3-3β was confirmed by co-immunoprecipitation from U2OS cell lysates and by in vitro binding assay using purified GST-Ror2 and 14-3-3β. We also found that 14-3-3β was endogenously phosphorylated on tyrosine(s) in U2OS cells and that this phosphorylation was almost completely abolished by down-regulating Ror2 expression with siRNA. On the other hand, Ror2 over-expression in U2OS cells dramatically increased the levels of tyrosine phosphorylation of 14-3-3β. Furthermore, purified Ror2 phosphorylated 14-3-3β in vitro, confirming 14-3-3β as the first identified substrate of the Ror2 tyrosine kinase.

Several receptor tyrosine kinases have been shown to be activated by dimerization, and we tested if the Ror2 receptor can form homo-dimers and if this can lead to receptor activation. Using an interaction assay between flag-tagged and his-tagged Ror2 over-expressed in U2OS cells we demonstrated that Ror2 does form homo-dimers and that the extent of dimerization can be greatly enhanced by treatment with a bivalent Ror2 antibody. The antibody-induced dimerization leads to activation of the Ror2 tyrosine kinase, as measured by Ror2 autophosphorylation and phosphorylation of 14-3-3β. Furthermore, Ror2-specific antibody dose-dependently enhanced mineralization of human mesenchymal stem cells (hMSC) upon Ror2 over-expression or induction of endogenous Ror2 by dexamethasone. The antibody effect was abolished by down-regulating Ror2 expression with shRNA.

In conclusion, we have identified 14-3-3β scaffold protein as the first substrate for the Ror2 tyrosine kinase and shown that antibody-induced dimerization of Ror2 leads to receptor autophosphorylation and tyrosine phosphorylation of 14-3-3β. Activation of the Ror2 receptor through dimerization promotes osteogenesis in hMSC.

Disclosures: Y. Liu, Wyeth 3.


12/15-Lipoxygenase Knockout Mice Have Decreased Bone Mass Compared to Their Wild-type Counterparts. S. J. Hoffman1, C. A. Capriotti*1, P. Liang1, G. B. Stroup2, S. Kumar1. 1Musculoskeletal Diseases, GlaxoSmithKline, Collegeville, PA, USA. 2GlaxoSmithKline, Collegeville, PA, USA

12/15-Lipoxygenase (12/15-LO), produced from the Alox 15 gene, catalyzes formation of hydroperoxy fatty acids (HETEs, HODEs) and has been implicated in the determination of bone mass by controlling marrow precursor cell differentiation into osteoblasts and adipocytes, possibly via activation of PPARγ. The purpose of this study was to evaluate the long term effects of 12/15-LO deletion on 1) body composition and whole body bone mass using Dual-energy X-ray Absorptiometry, and 2) bone turnover in the proximal tibia using dynamic histomorphometry.

Female wild type (WT) mice were consistently heavier than their 12/15-LO knockout (KO) counterparts starting as early as one month of age. Fat and lean mass were significantly greater in WT than KO mice. In addition, whole body bone mineral density (BMD) was greater in WT mice than KO mice at 1 month of age and these differences were maintained up to 7 months of age.

As determined by histological analysis of the proximal tibia of 1 month old mice, % trabecular area and trabecular number tended to be higher for WT compared to the KO mice. Bone resorption levels (% eroded perimeter) were significantly greater in WT than KO mice but no difference in bone formation rates were observed. At 4 months of age, % trabecular area was significantly greater in WT than KO mice likely due to a greater trabecular number in WT than KO mice with no difference in thickness. No significant differences were noted between the genotypes with respect to bone formation and resorption rates. 

Table  .  
inline image

Data are expressed as mean +/− SEM for 4 month old mice. * indicates p value of <0.05 compared to WT by 2-tailed ttest.

In conclusion, female 12/15-LO KO mice have decreased whole body BMD compared to WT mice that is observed as early as one month of age and continues to at least 7 months of age. 12/15-LO KO mice also had reduced trabecular bone area and fewer trabeculi of the proximal tibia than their WT counterparts.

Disclosures: S.J. Hoffman, None.


Nephroblastoma Overexpressed (Nov) Is a Novel Bone Morphogenic Protein (BMP) and Wnt Antagonist in Cells of the Osteoblastic Lineage. S. Rydziel, E. Canalis. Research, Saint Francis Hospital and Medical Center, Hartford, CT, USA

CCN (Cyr 61, connective tissue growth factor, CTGF, and Nov) are conserved cysteine rich proteins expressed in skeletal cells. Cyr 61 and CTGF play a role in chondrogenesis, but the physiological role of Nov and the function of CCN proteins in bone are not known. To determine the actions of Nov in osteoblasts, we transduced ST-2 cells with a retroviral vector, where a cytomegalovirus promoter directs the expression of the nov gene. Nov transduced cells had Nov mRNA levels 45 fold higher than control cells. Nov overexpression inhibited osteoblastogenesis, reduced the effect of BMP and Wnt 3 on alkaline phosphatase activity and decreased osteocalcin mRNA levels. To explore mechanisms involved in the suppression of osteoblastogenesis, we examined whether Nov modified BMP or Wnt signaling. Nov impaired the effect of BMP-2 on Smad 1/5/8 phosphorylation and on the transactivation of a BMP/Smad dependent 12xSBE-Oc-pGL3 reporter construct, where 12 Smad binding repeats direct luciferase expression. BMP-2 did not induce the phosphorylation of the mitogen activated protein kinases p38 or JNK, but did increase ERK phosphorylation, an effect that was enhanced by Nov overexpression. Although, the activation of ERK 1/2 may alter Smad phosphorylation and distribution, additional experiments revealed that Nov induced the transcript levels of the BMP antagonist, gremlin by 40 fold. The induction of gremlin explains the impaired BMP signaling and activity, but other mechanisms may be operational. In fact, Nov decreased cytoplasmic β-catenin levels, the transactivation of the Wnt/β-catenin dependent 16xTCF-Luc reporter construct, where 16 Tcf4/Lef1 binding repeats direct luciferase expression, and the mRNA levels of the Wnt dependent gene, WISP 1, indicating impaired Wnt signaling. Nov decreased the effect of a transfected Wnt 3 expression construct on the 16xTCF-Luc reporter. The inhibition persisted following the transfection of a stable β-catenin mutant expression vector, and was not rescued by a glycogen synthase inhibitor 3 β that prevents β-catenin ubiquitination. This indicates that normalization of β-catenin is not sufficient to rescue the inhibitory effect of Nov, and that interactions between Nov and Wnt are beyond receptor ligand activation. These did not involve Notch, as in myoblasts, since Nov did not activate Notch signaling, and the inhibitory effect of Nov on Wnt signaling was not rescued by a γ-secretase II inhibitor. In conclusion, Nov is a newly discovered antagonist of BMP and Wnt signaling and activity in cells of the osteoblastic lineage.

Disclosures: S. Rydziel, None.


Identification of DKK1 Residues Necessary for Interaction with LRP5/6. L. Lipfert*, P. Nantermet*, R. A. Perlow-Poehnelt*, A. A. Reszka, S. Harada, H. Glantschnig*. Molecular Endocrinology, Merck & Co., Inc., West Point, PA, USA

Canonical Wnt-induced signaling depends on frizzled co-receptors LRP5/6, which can be inhibited by DKK1. The binding site for DKK1 on LRP5/6 has been roughly described to reside within the third YWTD-domain, although hypermorphic mutations in first YWTD domain of LRP5 clearly affect DKK1 interaction and render LRP5/6 refractive to DKK1 inhibition. The protein domain on DKK1 necessary and sufficient for interacting with cell surface receptors has been traced to the second cysteine-rich domain of DKK1, and disturbance by a C220A substitution results in loss of DKK1/LRP6 interaction. However, the protein surface on DKK1 that directly interacts with LRP5/6 has not been described. Here we describe a hot-spot within DKK1 crucial for its interaction with LRP5/6. Full-length rhesus DKK1 cDNA or regions encoding either the NH2-terminal (ΔC-DKK1) or the COOH-terminal (ΔN-DKK1) regions of DKK1 were fused to EGFP. Fusion proteins were expressed in HEK-293 and analyzed for binding to LRP6 by fluorescence microscopy in cell-based assays and for their effects on DKK1 function in TCF/LEF signaling assays or by measuring a marker of cell differentiation in C3H10T1/2 cells. DKK1 binding to LRP6 was indeed detectable with ΔN-DKK1 whereas ΔC-DKK1 was not sufficient for interaction. Within the ΔN-DKK1 protein we focused on protein regions based on 1) homology models of co-lipase/lipase interaction; 2) a protein region conserved within LRP5/6 binding proteins DKK1, DKK2, DKK4 but not DKK3; and 3) putative DKK1 glycosylation events. Crystal structures of DKK1 have not been reported, but structural homology models for DKK1 using co-lipase can be generated. We show that the protein-protein interacting surface of co-lipase/lipase (homology model) differs from that of DKK1/LRP6, however, since substitutions (K215A, E216A) in DKK1 did not overtly alter the DKK1/LRP6 interaction. Nonetheless, scanning alanine mutagenesis within a conserved DKK1 domain (R203-I209) clearly diminished its binding to LRP6-transfected cells, thus identifying this region on DKK1 protein surface as crucial for receptor binding. Although we show that DKK1 is clearly glycosylated, an N256Q substitution of the single predicted N-glycosylation site of ΔN-DKK1 did not alter the apparent binding to LRP6 and the construct maintained inhibitory activity in functional assays. In summary we provide first evidence for a protein-region (R203-I209) that is conserved in members of the DKK-family capable of binding to LRP5/6. This region has the potential to be involved in establishing direct protein-protein contacts with the LRPs during receptor/ligand complex formation.

Disclosures: H. Glantschnig, None.


Osteopenia Produced by Expression of a Gi-activating Receptor in Osteoblasts. J. T. Peng*1, M. Bencsik*1, W. Lu*1, A. Storer*, A. Burghardt*2, S. M. Vanegas*3, S. Majumdar*2, T. J. Wronski*3, B. R. Conklin*2, B. Halloran1, R. A. Nissenson1. 1Endocrine Research, UCSF, San Francisco, CA, USA. 2UCSF, San Francisco, CA, USA. 3University of Florida, Gainesville, FL, USA

Intracellular signaling by G proteins is thought to provide critical regulation of osteoblast differentiation and function in response to extracellular signals. However, the precise role of specific osteoblast G protein signaling pathways in initiating anabolic and catabolic responses is unclear. In the present study, we have targeted a Gi-activating engineered G protein-coupled receptor (Ro1) to osteoblasts in transgenic mice, using the tet-off system to control transgene expression. Expression of Ro1 under the control of the 2.3 kb mouse α1Col 1 promoter resulted in perinatal lethality due to respiratory failure. Lethality was fully rescued by providing the mother with doxycycline during gestation thereby suppressing transgene expression. The affected pups displayed reduced bone mineralization particularly evident in the ribs and calvaria, suggesting a negative effect of Gi signaling on osteoblast function. Expression of Ro1 from birth resulted in a marked loss of trabecular bone, assessed at 12 weeks of age. Fractional trabecular bone volume (BV/TV) was reduced by over 50% in mice expressing Ro1. This was associated with a significant decrease in the number and connectivity of trabeculae. Histomorphometric analysis indicated little if any effect of Ro1 expression on the number of osteoclasts or osteoblasts, and the expression of osteoblastic differentiation markers (alkaline phosphatase, osteocalcin, and osteopontin) was significantly reduced in Ro1-expressing mice. These results indicate that Ro1 affects bone formation primarily by reducing osteoblast differentiation and/or functional activity. The similarity in the phenotypes between mice expressing Ro1 and mice lacking osteoblast expression of Gs-alpha (Sakamoto et al., J Biol Chem 280:21369, 2005) suggest that the effects of Ro1 on osteoblast differentiation are likely mediated by constitutive activation of Gi signaling and consequent inhibition of cAMP production in osteoblasts. An increase in apoptotic osteoblasts is also observed in Ro1 expressing mice, and this effect may also be due to Gi-mediated inhibition of cAMP production. Interestingly, BrdU labeling revealed an increase in the number of proliferating osteoblasts in Ro1-expressing mice, suggesting that Gi signaling has a proliferative effect on osteoblasts. Thus, signaling by a Gi-coupled GPCR in osteoblasts produces multiple cellular effects that appear to result in the accumulation of immature osteoblasts and the loss of trabecular bone.

Disclosures: J.T. Peng, None.


G Protein-Coupled Receptor (GPCR) Kinase 2 (GRK2) Inhibits Canonical Wnt Signaling. L. Wang*, R. F. Spurney. Medicine, Duke University Medical Center, Durham, NC, USA

Belonging to the large superfamily of heptahelical receptors are the frizzled (Frz) receptor subgroup which are activated by a family of secreted glycoproteins termed Wnts. Following Wnt binding, Frz receptors cause release and stabilization of the transcription regulator B-catenin from a destruction complex composed of axin and the adenomatous polyposis coli (APC) protein (canonical signaling pathway). This signaling cascade plays a key role in osteoblast (OB) biology because genetic mutations that alter Frz receptor activity alter bone mass, predominantly by modulating OB function. In most heptahelical receptor systems, receptor activation following ligand binding is regulated by a family of seven enzymes termed GRKs by a process that generally involves direct phosphorylation of receptor proteins by GRKs. To investigate the role of GRKs in regulating canonical Wnt signaling, we monitored Wnt-induced activation of a reporter construct in HEK293 cells transfected with either GRK2 or its empty vector. We found the GRK2 inhibited activation of the reporter construct (319 +/− 26 [vector] vs 92 +/− 18 [GRK2] % increase above baseline; P <0.001). In support of this negative regulatory effect, GRK2 inhibited Wnt-induced stabilization of B-catenin by immunoblotting. GRK2 enzymatic activity was required for the inhibitory effect because a GRK2 mutant lacking enzymatic activity [GRK2 (K220R)] had no effect on Wnt-dependent activation of canonical signaling (361 +/− 22 [vector] vs 362 +/− 51 [GRK2 (K220R)] % increase above baseline; P = NS). The negative regulatory effect did not appear to be associated with phosphorylation of Frz receptor proteins, but was associated with recruitment of GRK2 to the APC-axin complex in co-immunoprecipitation studies. To determine if this GRK2-dependent effect was observed in OBs, we used protein transduction to introduce GRK2 into calvarial OBs by tagging GRK2 with the TAT HIV protein sequence. We found that treatment with the GRK2 TAT protein [GRK2-TAT (+)] potently attenuated Wnt-induced activation of the reporter construct (558 +/− 157 [control] vs 150 +/− 26[GRK2-TAT (+)] % increase above baseline; P <0.05). In contrast, a GRK2 protein lacking the TAT sequence [GRK-TAT (−)] did not significantly effect activation of the reporter construct. Taken together, these data suggest that: 1. GRK2 is a negative regulator of canonical Wnt signaling, 2. Enzymatic activity is required for the negative regulatory effect, and 3. GRK2-dependent regulation of canonical signaling is associated with recruitment of GRK2 to the APC-axin destruction complex. We speculate that GRK2 may phosphorylate a component of the APC-axin protein complex and modulate canonical Wnt signaling.

Disclosures: L. Wang, None.

This study received funding from: National Institutes of Health.


Prostaglandin E2 (PGE2) Regulates Osteoblastic Jagged1: A Putative Modulator of the Hematopoietic Stem Cell (HSC) Niche. B. J. Frisch*1, J. M. Weber1, B. J. Gigliotti*1, L. Xing2, R. J. O'Keefe2, L. M. Calvi1. 1Department of Medicine and Center for Musculoskeletal Research, University of Rochester, Rochester, NY, USA. 2Center for Musculoskeletal Research, University of Rochester, Rochester, NY, USA

In the bone marrow, osteoblastic cells are a crucial component of the HSC niche. Specifically, activation of the PTH/PTHrP receptor (PTHR1) in osteoblastic cells increases HSC frequency through Notch activation. We previously demonstrated that PTH(1–34) stimulates the Notch ligand Jagged1, known to expand human HSC in vitro. This stimulatory effect of PTH on Jagged1 is Protein Kinase A (PKA)-dependent, since it is blocked by the specific PKA inhibitors H89 and PKA Inhibitory Peptide (14–22)(PKI), and is reproduced by the direct PKA agonists 8-bromo-cAMP and dibutiry1-cAMP. To establish whether alternative PKA activators could also regulate osteoblastic Jagged1 and alter the HSC niche, we utilized PGE2, a member of the prostaglandin family known to stimulate PKA. By real-time RT-PCR analysis, Jagged1 mRNA was increased up to 5 fold at 2 hours in UMR106 cells when treated with PGE2 (10−7 M) compared to vehicle. Jagged1 protein was also increased as early as 2 hours after treatment with PGE2. Treatment with PGE2 in the presence of myristoylated PKI (200ug/ml) blocked the PGE2-dependent Jagged1 increase, demonstrating that PKA is necessary for osteoblastic Jagged1 stimulation by PGE2. Since systemic PGE2 is known to have bone anabolic effects in both humans and animal models, 8-10 week old wild-type FVB/N male mice were treated with PGE2 (6mg/Kg/day i.p.) daily for 12 and 21 days. These regimens have previously been shown to have anabolic effects in rats. At day 12, histologic analysis demonstrated an anabolic effect in the long bones of PGE2-treated mice compared to control, as was confirmed by histomorphometry (% bone volume in primary spongiosa means 41% vs 12%; cortical thickness means 210 vs 129 μm). Evaluation of HSC frequency by flow cytometric analysis of the c-Kit+, Sca1+, lin- fraction (KSL) showed a trend towards an increase in KSL in bone marrow from PGE2-treated vs control mice (0.25% vs 0.20%). By day 21, the anabolic effect was less pronounced (% bone volume in primary spongiosa means 29% vs 23%; cortical thickness means 224 vs 154 m) and there was no increase in the KSL fraction. In summary, PGE2 stimulates Jagged1 in osteoblastic cells through PKA activation and increases both trabecular and cortical bone in vivo. Ongoing studies will confirm whether the anabolic effect of PGE2 is associated with HSC expansion, and whether Jagged1 regulates this process. This study identifies PGE2 as a novel regulator of osteoblastic Jagged1, and as a potential new microenvironmental modulator of HSC.

Disclosures: B.J. Frisch, None.


Transgenic Overexpression of p20C/EBPbeta with the pOBCol3.6 Promoter Causes Expansion of Stromal Progenitor Populations and Enhanced Adipogenesis, P. Palacios*1, A. Bazos*2, P. L. Kelly*1, M. Kronenberg*3, J. R. Harrison1. 1Department of Craniofacial Sciences, University of Connecticut Health Center, Farmington, CT, USA. 2University of Connecticut Health Center, Farmington, CT, USA. 3Department of Developmental Biology and Genetics, University of Connecticut Health Center, Farmington, CT, USA

To determine the role of CCAAT enhancer binding protein (C/EBP) transcription factors in bone, we developed a novel loss-of-function transgenic (TG) mouse model with overexpression of p20C/EBPbeta, a naturally occurring dominant negative C/EBPbeta isoform. These mice, in which expression of the transgene has been targeted with the 3.6-kb Col1a1 promoter and first intron (pOBCol3.6), exhibit osteopenia due to decreased bone formation. Analysis of osteoblast differentiation in vivo and ex vivo suggests a block in terminal osteoblast differentiation characterized by reduced expression of osteocalcin and pOBCol2.3-green fluorescent protein (GFP), established markers of the mature osteoblast phenotype. However, despite this inhibition of terminal differentiation, there is evidence that commitment of progenitor cells to the osteoblast lineage may be enhanced, including increased expression of the early osteoblast markers alkaline phosphatase and bone sialoprotein. The goal of the present study was to assess the cellular mechanisms underlying the changes in early osteoblast differentiation and to determine whether adipogenic commitment was affected. In both primary calvarial and bone marrow stromal culture systems, increased cell number was observed in TG cultures. Growth of TG calvarial cells diverged from WT after day 7, at a time when the pOBCol3.6 promoter becomes widely expressed. Flow cytometric analysis did not detect differences in BrdU incorporation between WT and TG calvarial cells during divergence, but annexin V labeling was decreased in TG cells, suggesting reduced apoptosis in these cultures. Remarkably, adipogenesis in response to a standard adipogenic cocktail or rosiglitazone was enhanced in bone marrow stromal cells from TG mice, suggesting that C/EBP transcription factors may regulate the availability of multipotent stromal progenitor cells. These observations suggest that C/EBP transcription factors act at multiple stages in the stromal-osteoblast lineage, regulating both pluripotent progenitor pool size and terminal osteoblast differentiation and function.

Disclosures: J.R. Harrison, None.


A Mouse Model with Pycnodysostosis Phenotype Reveals that Cathepsin K Functions as a Regulator of Osteoclast Apoptosis and Senescence. W. Chen*1, S. Yang*1, Y. Abe*2, R. Moroi*2, M. Li*2, Y. Wang*1, J. Shao*1, Y. P. Li1. 1Cytokine Bilogy, The Forsyth Institute & Harvard School of Dental Medicine, Boston, MA, USA. 2Cytokine Bilogy, Forsyth Institute, Boston, MA, USA

Mutation of Cathepsin K causes pycnodysostosis, a condition of defective bone homeostasis characterized by short stature, wide cranial sutures, acroosteolysis, spondylolysis, and osteosclerosis. Despite numerous previous studies involving Cathepsin K mutation, the pathomechanism for pycnodysostosis remains largely unknown due to the unavailability of a suitable mouse model with the phenotypes of pycnodysostosis. We generated a novel Cathepsin K−/− mouse strain that exhibits a human pycnodysostosis-like phenotype in the 129/Sv background. The cathepsin K−/− tibiae stained with Goldner's Trichrome showed lack of the Howard's lacunas between the osteoclasts and bone and exhibited un-degraded matrixes, suggesting impaired resorptive capacity. Contrary to what might be expected, the cathepsin K−/− knockout did not exhibit the same impairment of bone resorption in the calvariae, indicating that cathepsin K has varying effects around the skeleton. Cathepsin K−/− osteoclasts in 129/Sv background lacked normal apoptosis and senescence and exhibited abnormal growth both in vitro and in vivo, resulting in an extraordinarily high number of osteoclasts. In contrast, Cathepsin K−/− osteoclasts in C57BL/6J mice without phenotypes of pycnodysostosis exhibited normal apoptosis and senescence. The expression levels of p19, p53, and p21 were significantly reduced in these 129/Sv Cathepsin K−/− osteoclasts, and forced expression of Cathepsin K in the pre-osteoclast cell line, MOCP-5, induced premature senescence and increased expression of p19, p53, and p21. Our study is the first to propose a novel Cathepsin K function as a mediator of osteoclast apoptosis and senescence, possibly through the p19, p53, and p21 pathways. In summary, we found the novel function of Cathepsin K as a regulator of apoptosis and senescence. Cathepsin K−/− mice exhibit pcynodysostosis due to the combination of the impairment of this novel function as well as of the previously established cysteine protease function. An improved understanding of the Cathepsin K-mediated apoptosis and senescence pathways may provide a new drug target for diseases involved in osteoclast-related abnormal bone homeostasis.

Disclosures: Y.P. Li, None.

This study received funding from: NIH grant AR-44741, (Y. P. Li).


Orally Active Cathepsin K Inhibitor Potently and Rapidly Improved both Bone Mineral Density and Strength of not only Lumbar Vertebra but also Femoral Neck in Ovariectomized Cynomolgus Monkeys. Y. Ochi*, H. Yamada*, N. Kawada*, Y. Nakanishi*, H. Mori*, R. Kayasuga*, A. Hatayama*, K. Ohmoto*, M. Tanaka*, K. Kishikawa*. Discovery research laboratories ii, Ono Pharmaceutical Co., LTD, OSAKA, Japan

Cathepsin K, a cysteine protease highly expressed in osteoclasts, is capable of degrading bone type I collagen. Recently, cathepsin K inhibitors have drawn much attention as novel therapeutic agents for treatment of elevated bone turnover including osteoporosis. We investigated anti-resorptive cathepsin K inhibitor, Compound A, in ovariectomized cynomolgus monkeys.

Compound A potently inhibited human cathepsin K activity (Ki = 0.10 nM) and rabbit osteoclast bone resorption (CTX release into the medium, IC50 = 6.8 nM). In ovariectomized cynomolgus monkeys (9–25 years old), Compound A (0.3, 3 or 30 mg/kg), or alendronate (0.5 mg/kg) was given orally once daily from the next day of ovariectomy for 8 months. Compound A potently and rapidly suppressed bone resorption markers, such as serum C-terminal telopeptide of type I collagen (CTX), urinary CTX, serum N-telopeptide of type I collagen (NTX), urinary NTX in dose-dependent manner without affecting bone formation markers, such as serum osteocalcin and serum bone-specific alkaline phosphatase (BAP). Compound A immediately incresed lumbar bone mineral density (BMD) dose dependently. In addition, Compound A at the dose of 30 mg/kg further increased lumbar BMD compared with alendronate-treated group. As the results of isolated bone experments, Compound A has a positive effect on lumbar and femoral neck BMD and strength in dose dependent manner. In particular, the effect of Compound A on femoral neck BMD and strength was greater compared with alendronate-treated group. In histomorphometrical study, Compound A decreased cortical bone formation less than alendronate. We found out the possibility that Compound A may have a greater effect on cortical bone than bisphosphonates.

These results suggest that an orally active cathepsin K inhibitor, Compound A specifically suppressed bone resorption and rapidly improve BMD and bone strength in ovariectomized cynomolgus monkeys. Especially, Compound A has a greater effect on femoral neck BMD and strength. We conclude that cathepsin K inhibitors have the therapeutic potential in primates, acting in not only lumbar vertebra, but also femoral neck. Compound A would become a new ideal drug for osteoporosis therapy.

Disclosures: Y. Ochi, None.


The Inositol-1,4,5-trisphosphate Receptor-1 (IP3-R1) Regulates Osteoclast Intracellular Ca2+ and Activates μ-Calpain for Motility in Response to Nitric Oxide or cGMP. B. B. Yaroslavskiy, A. C. Sharrow*, A. Wells*, L. J. Robinson, H. C. Blair. Pathology, University of Pittsburgh, Pittsburgh, PA, USA

Nitric oxide (NO) is a key regulator of Ca2+ homeostasis affecting several bone regulatory systems. Skeletal flexion causes stretch-activated NO production in osteoblasts. NO synthesis is stimulated by inflammatory cytokines. NO synthase II (eNOS) is required for normal osteoblast activity and regulation of bone mass by estrogen. In regulation of bone turnover, osteoclasts release from their degradation sites and move to new locations. An NO dependent guanylyl cyclase and the cGMP-dependent protein kinase I (PKG I) initiate motility and terminate matrix degradation reversibly. We studied how activation of the NO/PKG I pathway mediates osteoclast relocation using human osteoclasts made from CD14 cells by RANKL and CSF-1 dependent maturation in vitro. We found that μ-calpain activity, regulated by increased intracellular Ca2+, is essential to motility. The μ-calpain was identified as the active proteinase by RNA expression screening, and in osteoclasts and osteoclast lysates by protein electrophoretic mobility, Western blot, and Ca2+ or calpain inhibitor-regulated activity. Increased Ca2+ during motility, and its dependency on NO or cGMP, were demonstrated by single or dual wavelength Ca2+ indicators in situ with emission microfluorescence cinematography. Ca2+ increased from 70 nM to 6 μM with motility. RNA interference showed that calpain activation required PKG I and an attachment-related adaptor protein, VASP. These findings implicated a membrane associated protein complex in calpain activation. Pharmacological inhibitor and RNA interference studies showed that calpain activation is via the inositol-1,4,5-trisphosphate receptor-1 (IP3-R1) Ca2+ channel, the only IP3-R that was expressed in the osteoclast by RNA or Western analysis. Inhibition of Src-family protein activity, which is functions in adhesion-related complexes in the osteoclast, was also required calpain activity. However, the IP3-R1 is a novel inositol trisphosphate pathways than was not related to inositol-3,4,5-trisphosphate/PI-3-kinase activity, which is involved in alternative pathways including Akt activation also dependent on Src and activated by integrin, RANK, and Y-kinases. We conclude that Ca2+-activated μ-calpain proteinase activity is essential for NO-regulated motility of the osteoclast. The Ca2+ release is mediated by the IP3-R1. This work is the first demonstration of IP3-R-dependent motility in a macrophage-family cell. Activation of the IP3-R1 in the osteoclast requires Src, or a Src family kinase sensitive to PP2, as well as PKG I and the PKG I-activated intermediate protein VASP.

Disclosures: H.C. Blair, None.

This study received funding from: NIH.


Deletion of TNF Rescues the Increased Osteoclastogenesis in TSH Receptor Deficient Mice: Role of TNFα in Mediating the Skeletal Effects of TSH. A. Brebene*1, H. Hase*1, H. Amano2, L. Sun1, M. Zaidi1, E. Abe1. 1Medicine, Mount Sinai School of Medicine, New York, NY, USA. 2Pharmacology, School of Dentistry, Showa University, Tokyo, Japan

We have shown that TSH negatively regulates osteoclast differentiation. Both heterozygotic and homozygotic TSH receptor (TSHR) knock out mice are thus osteopenic with evidence of enhanced osteoclast differentiation. We had previously attributed this increased osteoclastogenesis to elevated TNFα levels, as it was abolished in vitro by a TNFα blocking antibody. Here, we provide genetic evidence showing that the enhanced osteoclast formation in TSHR−/− and TSHR+/− mice is abrogated by deleting the TNFα gene, i.e. in compound TSHR−/−/TNFα−/− and TSHR+/−/TNFα+/− mice, respectively. We also provide direct evidence that TSH inhibits TNFα production, reduces the number of TNFα-producing osteoclast precursors and attenuates the induction of TNFα expression by IL-1, TNFα and RANK-L. TSH also suppresses osteoclast formation in murine macrophages and RAW-C3 celts, and in RAW-C3 cells overexpressing the TSHR. In contrast, the constitutively active, ligand-independent TSHR (caTSHR) abrogates osteoclast formation even in the absence of TSH. Finally, TSHR interferes with IL-1/TNFα and RANK-L signaling. These cytokines fail to trigger AP-1 and NFκB binding to DNA in cells transfected with TSHR or caTSHR. Together, the results suggest that TNFα is a critical cytokine mediating the anti-osteoclastogenic actions of TSH on the skeleton.

Disclosures: A. Brebene, None.


Dap12 Mediates M-CSF-Induced Cytoskeletonal Reorganization in Osteoclasts. W. Zou, S. L. Teitelbaum, F. P. Ross. Pathology, Washington University School of Medicine, St. Louis, MO, USA

The actin cytoskeleton is essential for osteoclastic bone resorption. While M-CSF induces cytoskeletal reorganization in osteoclasts (OC), the detailed mechanisms remain incompletely defined. M-CSF initiates intracellular signaling by binding to its receptor, c-Fms, which dimerizes and autophosphorylates seven tyrosine residues within its cytoplasmic tail, thus recruiting adaptor proteins that mediate OC functions. We find that M-CSF induces activation of downstream molecules Syk, Vav3 and Rac1, all of which are required for osteoclastic bone resorption. Dap12 is a trans-membrane adaptor protein found in myeloid cells that transmits activation signals via two tyrosine (Y) residues in its intracellular tyrosine activation motif (ITAM). Dap12-deficient bone marrow macrophages (BMMs) differentiate into multinucleated, TRAP positive OCs, but the cells fail to spread or form actin rings. Treatment of BMMs with M-CSF leads to phosphorylation of Dap12, suggesting this adaptor may mediate cytoskeletal reorganization. While re-expression of wild type Dap12 rescues spreading and actin ring formation in OCs, mutation of either Y residue to phenylalanine (F) results in failure of cytoskeletal remodeling, suggesting that signaling through the ITAM regulates actin reorganization. M-CSF stimulation induces Dap12 association with spleen tyrosine kinase (Syk) in a Src-kinase dependent manner; furthermore activation of Syk, Vav3 and Rac1 is absent in Dap12 deficient cells. Expression in DAP12 null cells of the adaptor in which both Y residues are mutated to F or in Syk null cells of a Syk mutant mutated the C- terminal SH2 domain required for Dap12 binding blocks M-CSF induced syk phosphorylation and osteoclast spreading. Finally, Y to F mutation of c-Fms 559, the c-src binding site, blocks M-CSF-induced src and syk activation. Taken together our data suggest that Dap12 is a specific ITAM adaptor for M-CSF signaling and that M-CSF, by binding its receptor c-Fms, initiates a series phosphorylation events resulting in src/Dap12/syk/vav/rac activation, which mediates OC cytoskeletal reorganization at least in part.

Disclosures: W. Zou, None.


Phosphatases Mediate the Inhibition of Osteoclast Differentiation by Tolllike Receptor Ligand. A. Amcheslavsky*, Z. Bar-Shavit. Experimental Medicine and Cancer Research, Hebrew University Faculty of Medicine, Jerusalem, Israel

Activation of either Toll-like receptors (TLRs) or RANK induces similar responses in osteoclast precursors (OCPs), namely NF-κB nuclear mobilization, p38 and ERK phosphorylation, and cytokines synthesis and release. However, several investigators have demonstrated that activation of TLRs results in inhibition of RANKL-induced osteoclastogenesis. Studies with TLR9, the receptor for bacterial DNA, have shown that activation of these receptors results in decreased levels of M-CSF receptors (essential for osteoclastogenesis) in OCPs and in induction of IL-12 (an osteoclastogenesis inhibitor) synthesis and release. Here we examine the molecular mechanism mediating the anti-osteoclastogenic effect of TLR9 activation. Synthetic oligodeoxynucleotides containing unmethylated CpG dinucleotides (CpG-ODNs) that were shown to mimic bacterial DNA are used in our studies to activate TLR9. Careful comparisons between the effects of RANKL and CpG-ODN on murine bone marrow-derived OCPs showed that while both induce p38 and ERK phosphorylation, RANKL-RANK interactions result in more sustained/durable phosphorylation and CpC-ODN-TLR9 interactions result in transient phosphorylation. Furthermore, addition of both ligands together also results in transient phosphorylation. Duration of the signal (ERK phosphorylation) is a key factor determining induction of c-Fos, a protein critical for osteoclastogenesis, and its absence leads to osteopetrosis. Indeed, we found that CpG-ODN, in contrast to RANKL, does not induce c-Fos and inhibits RANKL induction of the protein by enhancing its degradation. Thus, sustained ERK phosphorylation is critical for osteoclast differentiation. These findings led us to examine the possibility that induction of ERK dephosphorylation by CpG-ODN mediates the anti-osteoclastogenic effect. We indeed find that CpG-ODN or CpG-ODN together with RANKL, but not RANKL alone, induce the expression of phosphatases (PP2A and DUSP2). Moreover, CpG-ODN induces sustained ERK phosphorylation in the presence of the phosphatase inhibitor, okadaic acid. In correlation with this result, the addition of okadaic acid reduces the CpG-ODN-induced c-Fos degradation. Our findings suggest that the increased rate of c-Fos degradation by TLR9 ligand mediates the inhibition of RANKL-induced osteoclast differentiation. TLR9 ligand, via induction of phosphatases, prevents the sustained ERK phosphorylation required for maintaining the high c-Fos levels that are essential for osteoclast differentiation.

Disclosures: A. Amcheslavsky, None.

This study received funding from: Israel Science Foundation.


Osteoblasts Prepare the Osteoclast Niche which Determines the Site of Osteoclast Generation. Y. Nakamichi1, T. Mizoguchi*1, Y. Yamamoto*2, N. Udagawa3, Y. Kobayashi*1, A. Hosoya*4, M. Nakamura3, H. Ozawa1, J. M. Penninger5, N. Takahashi1. 1Institute for Oral Science, Matsumoto Dental Univ., Shiojiri, Japan. 2Dept. of Oral Diagnostics, School of Dentistry at Matsudo, Nihon Univ., Matsudo, Japan. 3Biochemistry, Matsumoto Dental Univ., Shiojiri, Japan. 4Second Dept. of Anatomy, Matsumoto Dental Univ., Shiojiri, Japan. 5IMBA, Vienna, Austria

RANKL and M-CSF are osteoblast (OB)-derived cytokines essential for osteoclast (OC) formation. It is believed that the distribution of RANKL or M-CSF, or both determine the site of OC generation. Using OPG−/−, RANKL−/− and M-CSF-deficient op/op mice, we studied how the site of OC generation determines. (1) The serum RANKL level in OPG−/−mice is extremely high, suggesting that serum RANKL may stimulate OC formation outside bone tissues in OPG−/− mice. When collagen disks containing BMP-2 (BMP-disks) or vehicle were implanted into OPG−/− mice for 7 days, TRAP-positive OCs and ALP-positive OBs appeared simultaneously in BMP-disks but not in the control disks. OCs were always detected near OBs. OPG added to BMP-disks completely suppressed OC formation in the BMP-disks. (2) BMP-disks and control-disks were implanted for 7 days into RANKL−/− mice, which were intraperitoneally administered with RANKL for 7 days. OCs appeared only in BMP-disks and bone tissues but not in the control-disks or other tissues in RANKL−/− mice. (3) When op/op mice were treated with 1,25(OH)2D3 for 2 days, many OCs were induced in bone as if they were existing in WT mice. Similarly, administration of RANKL to op/op mice for 2 days successfully induced OCs only in bone tissue even in the absence of M-CSF, suggesting that M-CSF is not required for in vivo OC formation in op/op mice. (4) We previously showed that OCs are directly differentiated from postmitotic OC precursors (pOCPs), which differentiate into OCs within 2 days in the presence of RANKL without cell proliferation. Therefore, we speculated that OBs retain pOCPs even in the absence of RANKL and M-CSF, and pOCPs differentiate into OCs in response to RANKL. When RANKL−/− mice were treated with BrdU and then with RANKL, many OCs were induced in bone tissues, and nuclei of OCs induced were all negative for BrdU. Thus, distribution of neither RANKL nor M-CSF determined me site of OC generation. These results suggest that OBs prepare the OC niche, which takes care of pOCPs independent of RANKL and M-CSF signals and determines the exact sites of OC generation.

Disclosures: Y. Nakamichi, None.