Bone 36 (2005) 599 – 606 www.elsevier.com/locate/bone
LRP5 gene polymorphisms predict bone mass and incident fractures in elderly Australian women J. Bollerslev a,d,T, S.G. Wilsona,b,c, I.M. Dick a,b,c, F.M.A. Islama,c, T. Ueland d,e, L. Palmer a,c, A. Devine a,b,c, R.L. Prince a,b,c a School of Medicine and Pharmacology, University of Western Australia, Nedlands, WA, Australia Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Nedlands, WA, Australia c Western Australian Institute of Medical Research, Sir Charles Gairdner Hospital, Nedlands, WA, Australia d Section of Endocrinology, Rikshospitalet University Hospital, Oslo, Norway e Research Institute for Internal Medicine, Rikshospitalet University Hospital, Oslo, Norway b
Received 6 October 2004; revised 16 December 2004; accepted 14 January 2005 Available online 14 March 2005
Abstract Postmenopausal osteoporosis and bone mass are influenced by multiple factors including genetic variation. The importance of LDL receptor-related protein 5 (LRP5) for the regulation of bone mass has recently been established, where loss of function mutations is followed by severe osteoporosis and gain of function is related to increased bone mass. The aim of this study was to evaluate the role of polymorphisms in the LRP5 gene in regulating bone mass and influencing prospective fracture frequency in a well-described, large cohort of normal, ambulatory Australian women. A total of 1301 women were genotyped for seven different single nucleotide polymorphisms (SNPs) within the LRP5 gene of which five were potentially informative. The effects of these gene polymorphisms on calcaneal quantitative ultrasound measurements (QUS), osteodensitometry of the hip and bone-related biochemistry was examined. One SNP located in exon 15 was found to be associated with fracture rate and bone mineral density. Homozygosity for the less frequent allele of c.3357A N G was associated with significant reduction in bone mass at most femoral sites. The subjects with the GG genotype, compared to the AA/AG genotypes showed a significant reduction in BUA and total hip, femoral neck and trochanter BMD (1.5% P = 0.032; 2.7% P = 0.047; 3.6% P = 0.008; 3.1% P = 0.050, respectively). In the 5-year follow-up period, 227 subjects experienced a total of 290 radiologically confirmed fractures. The incident fracture rate was significantly increased in subjects homozygous for the GG polymorphism (RR of fracture = 1.61, 95% CI [1.06–2.45], P = 0.027). After adjusting for total hip BMD, the fracture rate was still increased (RR = 1.67 [1.02–2.78], P = 0.045), indicating factors other than bone mass are of importance for bone strength. In conclusion, genetic variation in LRP5 seems to be of importance for regulation of bone mass and osteoporotic fractures. D 2005 Elsevier Inc. All rights reserved. Keywords: Elderly women; Bone mineral density; Osteoporotic fracture; LRP5 locus
Introduction Postmenopausal osteoporosis is a heterogeneous disorder and a major health problem in modern societies. Bone mass, which peaks in early adult life, is the single most important risk factor for osteoporotic fractures in the elderly. For T Corresponding author. Section of Endocrinology, National University Hospital, N-0027 Oslo, Norway. Fax: +47 2307 1910. E-mail address:
[email protected] (J. Bollerslev). 8756-3282/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2005.01.006
postmenopausal women it is generally accepted that a reduction in bone mass of one SD is followed by a subsequent doubling of fracture rate [1]. In addition to bone mass, postmenopausal osteoporosis is influenced by multiple factors including genetic variation [2–4]. The importance of the Wnt signaling pathway for bone mass and metabolism was recently demonstrated, when the osteoporosis–pseudoglioma syndrome (OPPG) was shown to be related to a homozygous non-sense or frame-shift mutations within the LDL receptor-related protein 5 (LRP5)
600
J. Bollerslev et al. / Bone 36 (2005) 599–606
gene [5]. This resulted in loss of function of the transmembrane protein involved in the Wnt pathway. Heterozygosity for the mutation was associated with lower bone mass. Gain of function mutations in the LRP5 gene have been found in several autosomal dominantly inherited conditions characterized by increased bone mass [6–8], among these the HBM phenotypes [6,7] and autosomal dominant osteopetrosis type 1 [8]. Both conditions are associated with decreased fracture frequency and increased trabecular bone strength. Most recently, LRP5 has been shown to be a determinant for bone mass in Japanese women [9]. This study was undertaken in order to evaluate the role of polymorphisms in the LRP5 gene in regulating bone mass and in determining prospective fracture frequency in a welldescribed, large cohort of normal, ambulatory Australian women.
lyzer). Urine deoxypyridinoline (Dpd) was measured by high-performance liquid chromatography (HPLC) [14] and corrected for creatinine excretion. Quantitative ultrasound measurements (QUS) and osteodensitometry QUS of the left calcaneum was measured twice in all subjects using a Lunar Achilles ultrasound machine (Lunar Corp., Madison, WI). The manufacturer’s quality assurance methods were employed. The average measurements of speed-of-sound (SOS), broadband ultrasound attenuation (BUA) and stiffness were given. The CV for SOS and BUA were 0.43% and 1.59%, respectively, using the manufacturer’s standards. Bone mineral content (BMC) and bone mineral density (BMD) of the hip region were measured in 1065 subjects using a Hologic Acclaim 4500A detector fanbeam densitometer (Hologic, Waltham, MA). CV at the total hip site for BMD was 1.0%.
Materials and methods Fracture status Subjects The basis for this study consisted of a population of 1499 randomly selected women between 70 and 85 years of age, as described previously [10,11]. The women were recruited by letter, from the population aged N70 years using the Western Australian electoral roll. The population is of Caucasian origin. Participants were included if they were not taking medication known to affect bone metabolism and did not have an existing medical condition that suggested they were unlikely to survive the study period. Subjects were enrolled in a 5-year randomized study of the effects of calcium supplementation on fracture outcome. The subjects were allocated to 1.2 g of calcium carbonate or matched placebo daily. Of these, 1387 gave consent to having blood samples taken for further genetic analyses. Analyses of LRP5 polymorphisms were completed on 1301 subjects. Informed consent was obtained from each individual, and the study was approved by the Human Rights Committee of the University of Western Australia. Demographic, anthropometric and lifestyle factors At baseline, all subjects completed a lifestyle questionnaire that included age, age at menopause and smoking history. A quantitative food frequency questionnaire [12] was used to estimate the calcium intake. Weight and height were measured at baseline. Biochemistry Blood and urine samples were obtained after an overnight fast. Serum osteocalcin was measured as described previously [13]. Urine samples were analyzed for creatinine using standard laboratory methods (BM/Hitachi 747 Ana-
Prevalent fractures were estimated by obtaining a fracture history from each subject including age at time of fracture, the site and how the fracture was sustained. A prevalent fracture was included, if it occurred after the age of 50 years and was due to a minimal trauma, as defined by falling from a height of less than 1 meter. Fractures of the face, skull, fingers and toes were not included. Incident clinical osteoporotic fractures during the 60month of follow-up were available and confirmed by X-ray reports. Genotyping Genomic DNA was extracted and purified from EDTA whole blood for each subject. We examined seven different single nucleotide polymorphisms (SNPs) in the LRP5 gene using nucleotide extension reactions analyzed by MassArray MALDI-ToF mass spectrometry (Sequenom, San Diego, CA). The SNPs, designated according to dbSNP as rs312016, rs314776, rs4988321, rs667126, rs556442, rs3736228, rs901823, were located in intron 1 (IVS1 + 2130c N t), intron 4 (IVS5-4t N c), exon 9 (c.1999G N A), intron 10 (IVS10 + 120T N C), exon15 (c.3357A N G), exon 18 (c.3989C N T) and intron 19 (IVS20-345T N C), respectively. Sequencespecific oligonucleotide primers flanking each SNP were designed using SpectroDesigner software (Sequenom Inc.) and used in PCR for amplification of the target sequence. Additional sequence-specific genotyping primers were designed adjacent to the SNP. Extension of the genotyping primers across the SNP sites occurred by performing sequencing polymerase reactions. Including dideoxynucleotides in the reaction mixture terminated this extension reaction. Reaction products for each SNP differed from the genotyping primer by 1–3 nucleotides. The resultant mass
J. Bollerslev et al. / Bone 36 (2005) 599–606
difference due to coding variation at the SNP site led to significantly different desorption ionization during mass spectrometry on the MassArray system. Reaction products were transferred onto silicon microchips, containing crystalline matrices. Each matrix element of the chip was sequentially irradiated with ultraviolet laser pulses leading to ionization of the analyte molecules. The ions were then accelerated through a detection region at a velocity that was inversely proportional to their mass-to-charge ratio. After calibration, the resulting time-resolved spectrum was translated into a mass spectrum and analyzed by applying baseline correction, peak identification and peak area calculations using Spectro-Typer Software (Sequenom, San Diego, CA). Large-scale typing was preceded by a screening assay on a panel of 92 DNAs. Genotyping of the test panel samples for the HBM locus confirmed the locus is non-polymorphic in this study cohort. Random duplicate genotyping was routinely undertaken throughout the study and indicated an estimated genotyping error rate of 0.54%. We excluded further evaluation of IVS20-345T N C because of the absence of Hardy Weinberg equilibrium in our data and c.1999G N A because of a genotype frequency below 6%. This left five single nucleotide polymorphisms (SNPs) in the LRP5 gene (Fig. 1 and Table 2) that form the basis of this report. Study design Various types of inheritance of disorders of the LRP5 gene have been described. HBM and the osteosclerotic disorders are due to autosomal dominant gain of function mutations at the 5V end of the gene [6–8]. Loss of function mutation based on frameshift or non-sense mutations leading to the autosomal recessive syndrome of severe osteoporosis and pseudoglioma have also been demonstrated [5]. These recessive mutations have been described throughout the gene. Polymorphisms might also result in qualitative changes of the function of the protein, but these might be compensated by the other allele. Thus, effects would only be detected in the homozygous cases. Therefore, a priori we favored a recessive model for analysis of the data from this study.
601
Statistical analysis Statistical analyses were performed using SPSS for Windows version 11.5 (SPSS Inc., Chicago, USA). The chi-square test was used to verify that genotype data were in Hardy Weinberg equilibrium as implemented in the Genepop software [15]. Differences between genotype groups were examined using unpaired Student’s t test or Mann–Whitney rank sum test, when appropriate according to normal distribution of the data. BMD and QUS data were adjusted for age, weight, calcium supplementation and smoking habits by univariate analysis of variance (UNIANOVA). Fracture rates and relative risk of fractures were calculated using logistic regression analysis and adjusted as mentioned. Linkage disequilibrium (DV) between different polymorphisms was evaluated by Fischer’s exact test. Assessment of LD within the LRP5 gene was done with the HaploView program using the block definition of Gabriel et al. [16], which specifies a bstrong LDQ if the one-sided upper 95% confidence bound on pair wise DV is greater than 0.98 and the lower bound is above 0.7. Haplotype analysis was undertaken by using UNPHASED software [17]. UNPHASED is a suite of programs for association analysis of multilocus haplotypes from unphased genotype data. QTPHASE, a program for quantitative trait analysis within UNPHASED, was used to investigate the effect of haplotypes on quantitative traits. The program employed likelihood ratio tests in a log-linear model [17]. The comparison of phenotypic means for individual haplotypes with the means for all other haplotypes together was investigated by using unpaired t test. Throughout, two-tailed P values are reported, and values P = 0.05 considered significant.
Results Demographics, bone density, QUS and biochemistry The baseline characteristics including age, weight, calcium intake, QUS and biochemistry parameters, BMD parameters at 12 months are detailed in Table 1.
Fig. 1. Schematic drawing of SNP localization in LRP5. Vertical bars represent the 23 exons of LRP5, and arrows indicate the positions of the five validated SNPs with a minimum allele frequency of 16%. Percentages indicate the frequency of the rare allele.
602
J. Bollerslev et al. / Bone 36 (2005) 599–606
Table 1 Demographics, hip DXA bone density, heel quantitative ultrasound and bone turnover biochemistry of the study population
Table 3 Hip DXA bone density, heel quantitative ultrasound and bone turnover biochemistry in relation to the allele distribution of c.3357A N G
Demographic variables (n = 1301) Age (years) Weight (kg) Dietary calcium (mg/day) Ever smoked (%) Prevalent fractures (%) Incident fractures (%) Hip DXA bone density (n = 1065) Total hip BMD (mg/cm2) Femoral neck BMD (mg/cm2) Trochanter BMD (mg/cm2) Intertrochanter BMD (mg/cm2) Heel quantitative ultrasound (n = 1252) BUA (Db/MHz) SOS (m/s) Bone turnover biochemistry (n = 260) Osteocalcin (Ag/l)1 U-Dpd/Creat (Amol/mol)1
DXA bone density Total hip (mg/cm2) Femoral neck (mg/cm2) Trochanter (mg/cm2) Intertrochanter (mg/cm2) Heel quantitative ultrasound BUA (Db/MHz) SOS (m/s) Bone turnover biochemistry Osteocalcin (Ag/l) U-Dpd/Creat (Amol/mol)
75.2 F 2.7 68.6 F 12.3 864 F 330 36.4 26.8 17.5 810 690 636 951
F 124 F 103 F 106 F 157
100.5 F 7.9 1513 F 26 4.70 F 4.48 27.5 F 12.5
Results are given as mean F SD with number of subjects in brackets.
LRP5 allele distribution and the phenotypic variables considered The chromosomal positions and allele distributions of the five SNPs used in this study are given in Fig. 1 and Table 2. The distribution of each of the genotypes for the SNPs was consistent with the Hardy Weinberg distribution. The genotype effects on hip DXA, heel QUS and bone biochemistry data
AA/AG genotype
GG genotype
P value
n = 934 812 F 106 692 F 91
n = 110 791 F 106 668 F 91
0.047 0.008
638 F 94 952 F 137
619 F 94 929 F 137
0.050 0.090
n = 1092
n = 133
100.6 F 7.3 1513 F 26 n = 214
99.1 F 7.3 1509 F 26 n = 31
0.032 0.145
4.13 F 2.06 27.7 F 1.4
4.64 F 2.07 30.0 F 1.4
0.409 0.252
Results are given as mean F SD with number of subjects in brackets. The bone morphometric data are adjusted for age, weight, treatment, calcium supplementation and smoking habits. The logarithmic values of biochemistry data are adjusted for age, weight, treatment, calcium supplementation and smoking habit.
tively. Heel QUS BUA was 1.5% lower in subjects homozygous for the GG polymorphism. There was no effect of genotype on bone related biochemistry. No significant association between the genotypes and age, weight, calcium supplementation and smoking habits was found. Effects of c.3357A N G polymorphisms on fractures
The genotype effects on hip DXA, heel QUS and bone biochemistry data were investigated by using one-way ANOVA for each SNPs individually. Significant associations were observed for c.3357A N G only, which was recessive in nature. Thus, the recessive model was used to further explore the relationships. The c.3357A N G polymorphism is a synonymous base change for amino acid 1119 (Val/Val) within exon 15. The effects of c.3357A N G on osteodensitometric and QUS data are given in Table 3. Bone mineral density was significantly lower in subjects homozygous for the GG polymorphism at the total hip, femoral neck and trochanter site, the differences were 2.7%, 3.6% and 3.1%, respec-
In the 60-month follow-up period, 227 subjects experienced an osteoporotic fracture confirmed by X-ray. The relative risk of incident fractures is shown in Fig. 2. The incident fracture rate was significantly higher in subjects homozygous for the c.3357A N G GG polymorphism as 24.8% of the homozygous group had a new fracture, compared to 16.8% in the rest of the population. Further, after adjusting also for total hip BMD, the incident fracture rate was still significantly higher (RR 1.67 CI 1.02–2.78). The incident fractures were mostly related to vertebrae and forearm. There was no effect of genotype on prevalent fracture (data not shown).
Table 2 Nomenclature, position and allele distribution of the analyzed SNPs on chromosome 11 and within the LRP5 gene SNP
IVS1 + 2130C N T
IVS5-4T N C
IVS10 + 120T N C
c.3357A N G
c.3989C N T
rs number Chromosome position Site Allele distribution (%)
312016 67857763 Intron 1 CC 48.1 TC 43.1 TT 8.8
314776 67908395 Intron 4 TT 35.4 TC 49.4 CC 15.2
667126 67953088 Intron 10 TT 56.0 TC 37.6 CC 6.4
556442 67968050 Exon 15 AA 46.5 GA 42.5 GG 11.0
3736228 67976655 Exon 18 CC 72.6 TC 24.8 TT 2.6
rs number refers to dbSNP reference number. Allele distributions are given as percentages.
J. Bollerslev et al. / Bone 36 (2005) 599–606
603
Using imputed haplotypes a 3-marker haplotype analysis (IVS1 + 2130C N T, IVS5-4T N C and c.3357A N G) suggested that 94% of the population is represented by five common haplotypes (Table 4). No significant associations of the haplotypes on the hip bone density or heel QUS was observed based on a likelihood ratio test in a log-linear model.
Discussion
Fig. 2. The relative risk of incident fractures over 5 years. Variables entered were age, weight, treatment (calcium or placebo), baseline calcium supplementation, pack years of smoking and c.3357A N G using a recessive (AA/AG, GG genotype) model. Treatment category, calcium supplementation and pack years of smoking were not significant.
Linkage disequilibrium and haplotype analysis Pair wise linkage disequilibrium DV and r 2 values for the five SNPs with genotype frequencies studied are given on the upper right and lower left corner of Fig. 3. These data demonstrated an LD block involving SNPs IVS10 + 120T N C, c.3357A N G and c.3989C N T. SNP c.3357A N G4 was selected to represent this block due to its high frequency.
This study demonstrates the impact of genetic variation in the LRP5 gene on bone mass and prospective fractures in a large and homogenous cohort of ambulatory elderly Australian women. The haplotype analyses supported by the single SNP data suggest important variation at the 3V end of the gene. Indeed subjects homozygous for the c.3357A N G polymorphism had reduced bone density at the hip, decreased calcaneal ultrasound measurements and an increase in 5 years incident fractures. The difference in BMD in relation to genetic variation was most pronounced in the femoral neck, where it corresponded to a third of an SD. Based on the general acceptance in postmenopausal women that a reduction in bone mass of one SD doubles fracture rate [1], and assuming linearity between reduction in bone mass and
Fig. 3. Linkage disequilibrium (LD) for five SNP pairs in the LRP5 gene, DV (upper right) and r 2 (lower left). Gray-shaded coding represents the strength of LD and r 2, according to the scale shown on the top and the right.
604
J. Bollerslev et al. / Bone 36 (2005) 599–606
Table 4 The effects of haplotypes of IVS1 + 2130C N T, IVS5-4T N C and c.3357A N G on hip DXA bone density and heel quantitative ultrasound Haplotypes
C-C-A
C-C-G
C-T-A
T-T-A
T-T-G
Haplotype frequency (%) Hip DXA bone density Total hip (mg/cm2) Femoral neck (mg/cm2) Trochanter (mg/cm2) Intertrochanter (mg/cm2) Heel quantitative ultrasound BUA (Db/MHz) SOS (m/s)
13
24
29
24
4
824 700 654 966
F F F F
8 6 7 10
101.5 F 0.5 1517 F 1.5
802 680 630 942
F F F F
6 5 5 7
99.7 F 0.3 1511 F 1.1
815 693 639 938
F F F F
5 4 4 6
100.7 F 0.3 1514 F 1.0
810 693 637 949
F F F F
6 5 5 7
100.7 F 0.3 1512 F 1.1
825 709 639 974
P1
14 12 12 18
0.28 0.11 0.12 0.34
100.5 F 0.8 1515 F 2.6
0.12 0.11
F F F F
1 Probability from a likelihood ratio test in a log-linear model of an overall association of the haplotypes on the dependent variable. Haplotypes less frequent than 3% were excluded.
increase in fracture rate, the increased fracture rate found was more pronounced than could be expected from the difference in bone mass. This finding is substantiated by our finding of an increased incident fracture rate in the presence of the GG genotype, also after adjustment for bone mass. This may be related to the micro-architecture of bone and the three dimensional lattice of trabeculae. In accordance with this, ex vivo analyses of bone from patients with a mutation in the first propeller of the LRP5 protein followed by increased bone mass [8] have previously shown increased trabecular bone strength even after adjustment for bone mass (ash weight) [18]. However, extra-skeletal effects of the genetic variation in LRP5 should also be considered, as LRP5 is expressed in several other tissues [7,19–21], although loss of function mutation so far only has been related to phenotypic consequences for bone and the eye [5]. In total, we investigated five different SNPs of the LRP5 gene based on the fact that missense mutations related to the first propeller of the protein, originally believed to be the primary binding site for the inhibitory protein Dickkopf (Dkk) [7], are followed by increased bone mass [6–8], whereas truncation of the protein as a consequence nonsense or frame-shift mutations is followed by severe osteoporosis in childhood [5]. However, recent studies of the HBM mutation (G171V) in transfected cell cultures have shown the primary binding site for Dkk to be related to the third propeller, whereas the chaperon protein (Mesd) binds the first propeller [22]. The G171V mutation was thus found to disrupt the LRP5–Mesd interaction followed by a reduced LRP5 expression at the cell membrane and a decrease in Wnt signaling in the paracrine, however not the autocrine paradigm [22]. The significance of these results for the understanding of bone metabolism in the osteosclerotic LRP5–related mutations is still a matter of debate and further research [23]. The SNP c.3357A N G is a synonymous polymorphism within exon 15 and corresponds to the fourth highly conserved YWTD motif predicted to form propeller structures. Since no amino acid substitution occurs due to the variant allele of SNP 4, this suggests that some other polymorphism within the highest LD region (IVS10 + 120T N C, c.3357A N G and c.3989C N T) may be
responsible for the observed effects on bone density, ultrasound measurements and fracture. The distal part of the gene (the LD-block in this study spanning from intron 10 to exon 18) has been found to be of particular importance for variation in bone mass by others. In the study by Mizuguchi et al. [9], a strong LD block was identified between intron 7 and exon 18. Within this block, three individual SNPs, one related to exon 10, were found to be significantly associated with bone mass in healthy Japanese women. In a study of normal individuals, a polymorphism in exon 9 (c.2047G N A) was significantly associated with bone mass, area and also personal height, mainly in men [24]. A SNP at the intron-exon boundary of exon 21 was found to be significantly associated with bone mass, accounting for 3.5% of the population variance in hip BMD. Again, the association was strongest in men. Moreover, linkage was found between a SNP in exon 18 and BMD at the hip and spine [24]. Polymorphisms in exon 18 have in other studies been associated with bone mass in both sexes [25]. In our study, haplotypes were built by using three SNPs, namely IVS1 + 2130C N T, IVS5-4T N C and c.3357A N G, and when attempting to characterize genetic variation within the gene, no associations with either BMD parameters or heel QUS were observed. Ferrari et al. [26] examined the relationship between LRP5 haplotypes and gain in vertebral bone mass and size over one year in 386 prepubertal children finding significant associations for males, but not females. In the OPPG syndrome, children have very low bone mass and are prone to developing fractures and deformations. However, no defect in collagen synthesis, calcium homeostasis or bone metabolism have been found [27] and histological investigations have revealed normal density and appearance of bone forming and resorbing cells on bone surfaces [5]. Concerning the human HBM phenotype, no systematic histomorphometric analyses have been published so far, whereas biochemical analyses have indicated increased bone formation and normal or decreased bone resorption [6]. In mice expressing the HBM LRP5 gene [28] an increased bone mass and strength was found and related to an increased number of osteoblasts, possibly caused by increased osteoblastic lifespan. However, in the human
J. Bollerslev et al. / Bone 36 (2005) 599–606
osteosclerotic phenotypes related to missense mutations in the same area of the gene [8], no indication of increased bone formation has hitherto been demonstrated. We thus performed analyses of calciotropic hormones and biochemical markers of bone turnover in a subset of subjects, and basal biochemistry and plasma OPG in all persons, at baseline. However, although the turnover markers tended to be higher in subjects homozygous for the c.3357A N G polymorphism, no significant differences were found. It could be speculated that an altered function of the LRP5 protein in relation to polymorphisms in the gene could alter downstream function in Wnt signaling and eventually affect osteoblast–osteoclast interaction by regulating the OPG/RANKL system [29]. Coupling of the Wnt signaling pathway to this system and thereby osteoclast recruitment has recently been illustrated by the fact that a decoy receptor to Frizzle (the transmembrane receptor that binds Wnt and LRP5), secreted frizzle-related protein-1 inhibits RANKL-dependent osteoclast formation [30]. Thus, alteration in the function of LRP5 and thereby Wnt signaling might affect bone formation as well as resorption. By measuring OPG in the circulation (and RANKL, data not shown), we were not able to contribute to this discussion, as no significant differences were demonstrated. As expected [31], OPG levels were positively correlated with age also in our total population, even with this relatively narrow age span (data not shown). In conclusion, genetic variation in LRP5 seems to be important for regulation of bone mass and subsequent fractures in elderly women. In particular, in relation to low bone mass, a region within the strong LD region distal in the gene is of particular interest and should be investigated further.
Acknowledgments The study was supported by research grants from Healthway Health Promotion Foundation of Western Australia, the Australian Menopause Society, Sir Charles Gairdner Hospital Research Fund and National Health and Medical Research Council Project Grant (254627).
References [1] Melton III LJ, Atkinson EJ, O’Fallon WM, Wahner HW, Riggs BL. Long-term fracture prediction by bone mineral assessed at different skeletal sites. J Bone Miner Res 1993;8:1227 – 33. [2] Gueguen R, Jouanny P, Guillemin F, Kuntz C, Pourel J, Siest G. Segregation analysis and variance components analysis of bone mineral density in healthy families. J Bone Miner Res 1995;10:2017 – 22. [3] Seeman E, Hopper JL, Bach LA, Cooper ME, Parkinson E, McKay J, et al. Reduced bone mass in daughters of women with osteoporosis. N Engl J Med 1989;320:554 – 8. [4] Uitterlinden AG, Burger H, Huang Q, Yue F, McGuigan FE, Grant SF, et al. Relation of alleles of the collagen type Ialpha1 gene to bone density and the risk of osteoporotic fractures in postmenopausal women. N Engl J Med 1998;338:1016 – 21.
605
[5] Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, et al. LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 2001;107:513 – 23. [6] Boyden LM, Mao J, Belsky J, Mitzner L, Farhi A, Mitnick MA, et al. High bone density due to a mutation in LDL-receptor-related protein 5. N Engl J Med 2002;346:1513 – 21. [7] Little RD, Carulli JP, Del Mastro RG, Dupuis J, Osborne M, Folz C, et al. A mutation in the LDL receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 2002;70: 11 – 9. [8] VanWesenbeeck L, Cleiren E, Gram J, Beals RK, Benichou O, Scopelliti D, et al. Six novel missense mutations in the LDL receptorrelated protein 5 (LRP5) gene in different conditions with an increased bone density. Am J Hum Genet 2003;72:763 – 71. [9] Mizuguchi T, Furuta I, Watanabe Y, Tsukamoto K, Tomita H, Tsujihata M, et al. LRP5, low-density-lipoprotein-receptor-related protein 5, is a determinant for bone mineral density. J Hum Genet; 2004;49:80 – 6. [10] Bollerslev J, Wilson SG, Dick IM, Devine A, Dhaliwal SS, Prince RL. Calcium-sensing receptor gene polymorphism A986S does not predict serum calcium level, bone mineral density, calcaneal ultrasound indices, or fracture rate in a large cohort of elderly women. Calcif Tissue Int 2004;74:12 – 7. [11] Dick IM, Devine A, Marangou A, Dhaliwal SS, Laws S, Martins RN, et al. Apolipoprotein E4 is associated with reduced calcaneal quantitative ultrasound measurements and bone mineral density in elderly women. Bone 2002;31:497 – 502. [12] Ireland P, Jolley D, Giles G, O’ Dea K, Powles JRI, Wahlqvist ML, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pac J Clin Nutr 1994;1:19 – 31. [13] Price PA, Nishimoto SK. Radioimmunoassay for the vitamin K-dependent protein of bone and its discovery in plasma. Proc Natl Acad Sci U S A 1980;77:2234 – 8. [14] Randall AG, Kent GN, Garcia-Webb P, Bhagat CI, Pearce DJ, Gutteridge DH, et al. Comparison of biochemical markers of bone turnover in Paget disease treated with pamidronate and a proposed model for the relationships between measurements of the different forms of pyridinoline cross-links. J Bone Miner Res 1996;11:1176 – 84. [15] Garnier-Gere P, Dillmann C. A computer program for testing pairwise linkage disequilibria in subdivided populations. J Hered 1992;83:239. [16] Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225 – 9. [17] Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol 2003;25:115 – 21. [18] Bollerslev J, Mosekilde L, Nielsen HK, Mosekilde L. Biomechanical competence of iliac crest trabecular bone in autosomal dominant osteopetrosis type I. Bone 1989;10:159 – 64. [19] Dong Y, Lathrop W, Weaver D, Qiu Q, Cini J, Bertolini D, et al. Molecular cloning and characterization of LR3, a novel LDL receptor family protein with mitogenic activity. Biochem Biophys Res Commun 1998;251:784 – 90. [20] Figueroa DJ, Hess JF, Ky B, Brown SD, Sandig V, HermanowskiVosatka A, et al. Expression of the type I diabetes-associated gene LRP5 in macrophages, vitamin A system cells, and the islets of Langerhans suggests multiple potential roles in diabetes. J Histochem Cytochem 2000;48:1357 – 68. [21] Hey PJ, Twells RC, Phillips MS, Yusuke N, Brown SD, Kawaguchi Y, et al. Cloning of a novel member of the low-density lipoprotein receptor family. Gene 1998;216:103 – 11. [22] Zhang Y, Wang Y, Li X, Zhang J, Mao J, Li Z, et al. The LRP5 highbone-mass G171V mutation disrupts LRP5 interaction with Mesd. Mol Cell Biol 2004;24:4677 – 84. [23] Johnson ML, Harnish K, Nusse R, VanHul W. LRP5 and Wnt signaling: a union made for bone. J Bone Miner Res 2004;19: 1749 – 57.
606
J. Bollerslev et al. / Bone 36 (2005) 599–606
[24] Koay MA, Woon PY, Zhang Y, Miles LJ, Duncan EL, Ralston SH, et al. Influence of LRP5 polymorphisms on normal variation in BMD. J Bone Miner Res 2004;19:1619 – 27. [25] van Meurs J, Hugens W, Arp P, Bartels C, Ai M, Hofman A, et al. Association of LRP5 polymorphisms with bone mineral density in elderly men and women. Calcif Tissue Int 2003;72:327. [26] Ferrari SL, Deutsch S, Choudhury U, Chevalley T, Bonjour JP, Dermitzakis ET, et al. Polymorphisms in the low-density lipoprotein receptor-related protein 5 (LRP5) gene are associated with variation in vertebral bone mass, vertebral bone size, and stature in whites. Am J Hum Genet 2004;74:866 – 75. [27] Gong Y, Vikkula M, Boon L, Liu J, Beighton P, Ramesar R, et al. Osteoporosis-pseudoglioma syndrome, a disorder affecting skeletal strength and vision, is assigned to chromosome region 11q12-13. Am J Hum Genet 1996;59:146 – 51.
[28] Babij P, Zhao W, Small C, Kharode Y, Yaworsky PJ, Bouxsein ML, et al. High bone mass in mice expressing a mutant LRP5 gene. J Bone Miner Res 2003;18:960 – 74. [29] Bollerslev J, Ueland T, Odgren PR. Serum levels of TGF-beta and fibronectin in autosomal dominant osteopetrosis in relation to underlying mutations and well-described murine counterparts. Crit Rev Eukaryot Gene Expr 2003;13:163 – 71. [30] Hausler KD, Horwood NJ, Chuman Y, Fisher JL, Ellis J, Martin TJ, et al. Secreted frizzled-related protein-1 inhibits RANKL-dependent osteoclast formation. J Bone Miner Res 2004;19:1873 – 81. [31] Yano K, Tsuda E, Washida N, Kobayashi F, Goto M, Harada A, et al. Immunological characterization of circulating osteoprotegerin/osteoclastogenesis inhibitory factor: increased serum concentrations in postmenopausal women with osteoporosis. J Bone Miner Res 1999; 14:518 – 27.
