FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |   Search Term  Advanced Search

Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe

Published at www.nejm.org April 29, 2008 (10.1056/NEJMoa0801197)

Abstract PDF Supplementary Material Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited PubMed Citation

ABSTRACT

Background Bone mineral density influences the risk of osteoporosis later in life and is useful in the evaluation of the risk of fracture. We aimed to identify sequence variants associated with bone mineral density and fracture.

Methods We performed a quantitative trait analysis of data from 5861 Icelandic subjects (the discovery set), testing for an association between 301,019 single-nucleotide polymorphisms (SNPs) and bone mineral density of the hip and lumbar spine. We then tested for an association between 74 SNPs (most of which were implicated in the discovery set) at 32 loci in replication sets of Icelandic, Danish, and Australian subjects (4165, 2269, and 1491 subjects, respectively).

Results Sequence variants in five genomic regions were significantly associated with bone mineral density in the discovery set and were confirmed in the replication sets (combined P values, 1.2x10^–7 to 2.0x10^–21). Three regions are close to or within genes previously shown to be important to the biologic characteristics of bone: the receptor activator of nuclear factor-B ligand gene (RANKL) (chromosomal location, 13q14), the osteoprotegerin gene (OPG) (8q24), and the estrogen receptor 1 gene (ESR1) (6q25). The two other regions are close to the zinc finger and BTB domain containing 40 gene (ZBTB40) (1p36) and the major histocompatibility complex region (6p21). The 1p36, 8q24, and 6p21 loci were also associated with osteoporotic fractures, as were loci at 18q21, close to the receptor activator of the nuclear factor-B gene (RANK), and loci at 2p16 and 11p11.

Conclusions We have discovered common sequence variants that are consistently associated with bone mineral density and with low-trauma fractures in three populations of European descent. Although these variants alone are not clinically useful in the prediction of risk to the individual person, they provide insight into the biochemical pathways underlying osteoporosis.

Methods

Study Populations

We studied four populations: a discovery set of 5861 Icelandic persons, 87% of whom are women, and three replication sets — 4165 other Icelandic persons (74% of whom are women), 2269 postmenopausal Danish women,^9,10 and the Australian Dubbo Osteoporosis Epidemiology Study (DOES) cohort¹¹ of 1491 persons (61% of whom are women). Dual-energy x-ray absorptiometry measurements at the lumbar spine and at the hip (a total hip or femoral neck measurement) were obtained in all patients. Standardized bone mineral density was calculated and corrected for sex, age, and weight (see the Methods section in Supplementary Appendix 1, available with the full text of this article at www.nejm.org). For example, a standardized value of X corresponds to X standard deviations above or below the population average after adjustment for sex, age, and weight. Most low-impact fractures were self-reported by persons in the Icelandic sets; reports of hip fractures in these sets were obtained from hospital records. All fractures in the Danish set were self-reported, and all fractures in the Australian cohort were ascertained through review of radiography reports. Further description of characteristics of the study populations is provided in the Methods section and Table 1 of Supplementary Appendix 1.

All participants provided written informed consent. The study was approved by the Data Protection Commission of Iceland, the National Bioethics Committee of Iceland, the Ethical Committee of Copenhagen County (in Denmark), and the St. Vincent's Ethics Review Committee (in Australia).

The study was funded by deCODE Genetics, which holds the data; the authors in Denmark and Australia also hold the data for those sets. Seven industry authors and one academic author designed the study, six industry authors analyzed the data, and six prepared the manuscript. Six industry authors and two academic authors vouch for the completeness and accuracy of the data.

Genotyping

We used either the Infinium HumanHap300 or the HumanCNV370 SNP chip (Illumina) to genotype a total of 317,503 SNPs in DNA samples from 5861 persons; 5858 had measurements of spine bone mineral density and 5715 had measurements of hip bone mineral density. Subsequent analysis was restricted to 301,019 SNPs that were of sufficient quality (see the Methods section in Supplementary Appendix 1). We genotyped SNPs in the replication sets at deCODE Genetics, using the Centaurus platform (Nanogen).¹²

Association Analyses

Age- and weight-adjusted standardized values of bone mineral density were computed for each sex and population separately (see the Methods section in Supplementary Appendix 1). For each SNP, a linear regression analysis, with the genotype as an additive covariate and standardized bone mineral density as the response variable, was fitted to test for association. Each SNP was tested separately for its association with bone mineral density of the hip and for its association with bone mineral density of the lumbar spine. The method of genomic control¹³ was used to adjust the P values and standard errors for the relatedness of Icelandic subjects in the discovery set. When the discovery set was combined with the Icelandic follow-up set, the effect of relatedness was estimated by simulating the founder gene types through the Icelandic genealogy (see the Methods section in Supplementary Appendix 1). The threshold for genomewide significance was set at a P value of less than 1.7x10^–7 (0.05÷301,019 SNPs) for each phenotype and the threshold for replication was set at a P value of less than 0.05 in the replication sets.

To test for association between implicated SNPs and osteoporotic fractures, odds ratios for any low-trauma osteoporotic fracture among case patients as compared with controls were computed. A standard likelihood ratio statistic was used to calculate two-sided P values (see the Methods section in Supplementary Appendix 1). Results from multiple groups of case subjects with any low-impact osteoporotic fracture and controls were combined with the use of a Mantel–Haenszel model.¹⁴ P values of less than 0.05 were considered to indicate statistical significance.

Results

Genomewide Association Analysis

We observed a substantial excess of SNPs associated with bone mineral density, especially with spine bone mineral density, even after adjustment for relatedness of the subjects and for possible stratification by means of genomic control (adjustment factor for chi-square statistics, =1.18 for the hip and =1.22 for the spine) (Figure 1). A total of 77 SNPs (which capture at least 48 independent association signals) were significantly associated with spine bone mineral density, with adjusted P values of less than 0.0001 (adjusted chi-square statistics >15.14); only 30.1 SNPs were expected to satisfy this criterion on the basis of chance alone. These results indicate that a substantial fraction of the most strongly associated SNPs could have true associations with bone mineral density. The 50 SNPs most strongly associated with each phenotype, ranked according to P value, are listed in Table 2 in Supplementary Appendix 1; the rankings of all SNPs tested are provided in Supplementary Appendixes 2 and 3.

View larger version (17K): [in this window] [in a new window]   Figure 1. Quantile–Quantile Plots of Test Statistics (P Values) for the Associations of Single-Nucleotide Polymorphisms with Bone Mineral Density. Panel A shows data for hip bone mineral density; Panel B shows data for spine bone mineral density. The P values were corrected by means of the genomic control method (adjustment factor for chi-square statistics, =1.18 for hip bone mineral density and =1.22 for spine bone mineral density). Under the null hypothesis of no association at any locus, the data points would be expected to follow the black line.

 
Three SNPs — rs9594759 (located on 13q14) and rs2504063 and rs851982 (on 6q25) — achieved genomewide significance (P<1.7x10^–7) for spine bone mineral density, whereas no SNP reached genomewide significance for hip bone mineral density (Table 2 in Supplementary Appendix 1). We then genotyped these SNPs in the three replication sets, in addition to SNPs with P values less than 3.0x10^–5 for either skeletal site and a few SNPs with weaker associations that clustered around interesting candidate genes. In all, we genotyped 74 SNPs representing 32 chromosomal loci (Table 3 in Supplementary Appendix 1).

Twelve SNPs were both nominally significant (P<0.05) in the Danish and Australian sets combined and had genomewide significance (P<1.7x10^–7) for at least one phenotype when all sets (including the Icelandic discovery set) were combined. Eleven SNPs from five chromosomal regions — 1p36, 6p21, 6q25, 8q24 and 13q14 — were associated with spine bone mineral density, meeting the criteria of nominal significance in the Danish and Australian sets combined and genomewide significance in all sets combined (Table 1). On the basis of these same criteria, six SNPs from three chromosomal regions (1p36, 6q25, and 8q24) were associated with hip bone mineral density (Table 2). These three loci were also associated with spine bone mineral density.

View this table: [in this window] [in a new window]   Table 1. Estimated Effects on Bone Mineral Density of the Spine of Single-Nucleotide Polymorphisms (SNPs) at the Five Regions with Genomewide Significance for Either the Spine or the Hip.

 

View this table: [in this window] [in a new window]   Table 2. Estimated Effects on Bone Mineral Density of the Hip of Single-Nucleotide Polymorphisms (SNPs) at the Five Regions with Genomewide Significance for Either the Spine or the Hip.

 
We estimated that the alleles associated with low bone mineral density resulted in a decrease of 0.05 to 0.17 standardized unit. We did not observe sex-specific associations at any of the five loci, apart from 8q24, where a significantly stronger effect was observed in men than in women (P=0.03 to 0.050) (Table 4 in Supplementary Appendix 1).

The 13q14 (RANKL), 8q24 (OPG), and 18q21 (RANK) Regions

The SNP most strongly associated with spine bone mineral density in the discovery set, rs9594759 on 13q14, is located 113 kb upstream of the receptor activator of nuclear factor-B ligand gene (RANKL) and in a linkage-disequilibrium block adjacent to that gene (Fig. 1A in Supplementary Appendix 1). The association between rs9594759 and spine bone mineral density was consistent in each replication set. Another SNP, rs9594738, selected on the basis of HapMap data, showed the strongest association with spine bone mineral density when genotyped in the replication sets (combined P=2.0x10^–21) (Table 1). We estimate that the T allele of rs9594738 lowers spine bone mineral density by 0.17 standardized value per allele. With respect to its association with hip bone mineral density, this SNP reached genomewide significance in the combined Icelandic samples (Table 2) and had a nominally significant association in the Danish set but not in the Australian set.

Two correlated SNPs on 8q24, rs6469804 and rs6993813, reached genomewide significance for both spine and hip bone mineral density and were significantly associated in the replication sets (Table 1 and Table 2). These two SNPs are in strong linkage disequilibrium (pairwise coefficient of determination [r²]=0.94) and are in a linkage-disequilibrium block including the osteoprotegerin gene (OPG) (Fig. 1B in Supplementary Appendix 1). This gene has been tested for association with bone mineral density and osteoporosis, with inconsistent results.^{15,16,17,18,19,20,21,22,23,24,25} The two most studied SNPs in this gene, G1181C (rs2073618) and T950C (rs2073617), are in linkage disequilibrium with both rs6469804 (r²=0.57 and r²=0.77, respectively) and rs6993813 (r²=0.44 and r²=0.59, respectively) in the HapMap–CEU data set (consisting of Centre d'Etude du Polymorphisme Humain [CEPH] subjects who are Utah residents of northern and western European ancestry). Our results are consistent with those from previous studies reporting associations with bone mineral density.^{16,18,22,23,24}

RANKL and OPG encode known regulators of bone remodeling which, together with the receptor activator of the nuclear factor-B protein RANK, are central to osteoclastogenesis and activation of bone resorption. A SNP (rs3018362) located 27 kb downstream of RANK and within the same linkage disequilibrium block, on 18q21 (Fig. 1C in Supplementary Appendix 1), showed a consistent association with hip bone mineral density, although the association did not meet genomewide significance (Table 3). We did not find significant interaction between the associated SNPs at these three genes, despite their close involvement at the protein level (see the Methods section in Supplementary Appendix 1).

View this table: [in this window] [in a new window]   Table 3. Estimated Effects on Bone Mineral Density of the Spine or the Hip of SNPs at Three Regions with Suggested Associations with Bone Mineral Density and Fracture.

 
The 6q25 ESR1 Region

The genomewide association study uncovered a complex pattern of association in the 6q25 region. SNPs in this region showed an association with bone mineral density of both the hip and spine (Table 1 and Table 2, and Table 3 in Supplementary Appendix 1), and many had modest or minimal linkage disequilibrium (with respect to one another). None of these SNPs could fully account for the associations. The SNPs spanned 273 kb and were distributed over several linkage-disequilibrium blocks (Fig. 1D in Supplementary Appendix 1). Neither of the two significantly associated SNPs identified in the genomewide association study, rs2504063 and rs851982, showed associations with bone mineral density in the non-Icelandic replication sets. In total, associations between eight SNPs and spine bone mineral density were consistently replicated (Table 3A in Supplementary Appendix 1), with four (rs4870044, rs1038304, rs6929137, and rs1999805) reaching genomewide significance in the combined analysis (Table 1). Five of the eight SNPs were also associated with hip bone mineral density; two of these five, rs9479055 and rs4870044, reached nominal significance in the Danish and Australian replication sets and genomewide significance in the combined analysis (Table 2). The five SNPs have pairwise r² values ranging from 0.001 to 0.51 and are dispersed across three linkage-disequilibrium blocks (Fig. 1B in Supplementary Appendix 1).

No single SNP could fully explain the association with 6q25; at least three SNPs (e.g., rs4870044, rs1038304, and rs1999805) seemed to be required to account for the overall association (see the Methods section in Supplementary Appendix 1). The SNP rs1999805 is within an intron of the U68068 [GenBank] splice variant of the estrogen receptor 1 gene (ESR1). The SNPs rs4870044 and rs1038304 are located 44 kb to 75 kb upstream of the gene, possibly affecting its regulation of transcription. However, these two SNPs lie within C6orf97, the chromosome 6 open reading frame 97 gene of unknown function, now a candidate gene for osteoporosis (Fig. 1D in Supplementary Appendix 1).

The 1p36 Region

Two SNPs in linkage disequilibrium on chromosome 1p36, rs7524102 and rs6696981 (r²=0.73), showed an association with both hip bone mineral density and spine bone mineral density (Table 1 and Table 2). The marker showing the stronger association, rs7524102, was significant in all study populations, with similar effects on hip bone mineral density among all populations. These SNPs are located in a linkage-disequilibrium block that does not contain any known gene (Fig. 1E in Supplementary Appendix 1). The closest gene is the zinc finger and BTB domain containing 40 gene (ZBTB40), located in an adjacent linkage-disequilibrium block, 80 kb downstream from the signal. This gene, of unknown function, is expressed in bone (UniGene accession number, Hs.418966), indicating a potentially unidentified role in the biologic characteristics of bone. Farther from the signal, 250 kb in the opposite direction, lies the wingless-type MMTV integration site family member 4 gene (WNT4).

The 6p21 Region

One SNP, rs3130340, in the major histocompatibility complex region was associated with spine bone mineral density in the replication sets and in the combined overall analysis (Table 1). This SNP is located downstream of the uncharacterized chromosome 6 open reading frame 10 gene C6orf10 (Fig. 1F in Supplementary Appendix 1). This association is weaker than those of SNPs at the other loci.

Risk of Osteoporotic Fracture

We tested for associations between low-trauma fracture and the SNPs associated with bone mineral density (Table 4). The 1p36 locus was the most strongly associated with fractures. It was significantly associated with not only the broadly defined group of fractures of any kind but also individually with forearm fractures, fractures of the hip, and vertebral fractures (Table 4, and Table 5 in Supplementary Appendix 1). The 8q24 locus also was associated with all categories of fracture, albeit not as strongly as 1p36. The 6p21 MHC locus was significantly associated with the broadly defined group of fractures as well as with hip and forearm fractures individually (Table 5 in Supplementary Appendix 1). The 18q21 RANK locus was modestly associated with fractures; the 6q25 ESR1 and 13q14 RANKL loci were not.

View this table: [in this window] [in a new window]   Table 4. Association of Single-Nucleotide Polymorphisms (SNPs) with Any Low-Trauma Osteoporotic Fracture.

 
Some loci that did not reach genomewide significance in our combined analysis of association with bone mineral density did show evidence of an association with osteoporotic fractures (Table 3 and Table 4). These are rs11898505, in the spectrin β nonerythrocytic 1 gene (SPTBN1) on chromosome 2p16, which was associated with fractures in all sample sets, most strongly with vertebral fractures (Table 5 and Fig. 1G in Supplementary Appendix 1), and SNPs within and close to the gene encoding the low-density lipoprotein receptor–related protein 4 (LRP4) on chromosome 11p11 (Tables 3 and 5 and Fig. 1H in Supplementary Appendix 1).

Osteoporosis Candidate Genes

Numerous candidate genes have been studied for association with bone mineral density and with osteoporotic fractures.^26,27 We looked at the ranking, in our genomewide association study data set, of 77 candidate genes previously reported to be associated with bone mineral density or fractures and their neighboring regions (Table 6 in Supplementary Appendix 1). Apart from ESR1, RANKL, and RANK, only two of these genes had two or more markers in the list of the 500 SNPs most significantly associated with either hip or spine bone mineral density: the sclerosteosis gene (SOST) and the nuclear receptor subfamily 3, group C, member 1 (glucocorticoid receptor) gene (NR3C1). SNPs in and around the more intensively studied vitamin D (1,25-dihydroxyvitamin D3) receptor gene VDR and the low-density lipoprotein receptor–related protein 5 gene LRP5 were among the 1000 SNPs most significantly associated with bone mineral density, whereas SNPs in the collagen type I1 gene COL1A1 (which was also intensively studied) did not even feature in the top 5000 SNPs.

Discussion

We report here the results from our large genomewide association study of bone mineral density: associations with five genomic regions were replicated consistently in three replication sets of subjects of European descent (Danish, Australian, and Icelandic subjects). The four loci 1p36 (ZBTB40–WNT4), 6q25 (ESR1–C6orf97), 8q24 (OPG), and 13q14 (RANKL) influence bone mineral density at both the spine and the hip, whereas the 6p21 locus is associated with spine bone mineral density only. We identified loci that warrant further investigation in additional samples. The 1p36, 6p21, and 8q24 loci had a significant association with low-trauma fractures. Three other loci for which genomewide significance for bone mineral density was not reached but that had an association with low-trauma fractures were 2p16 (the location of the SPTBN1 gene), 11p11 (LRP4), and 18q21 (RANK). That some loci were found to show association with bone mineral density but not fracture may reflect differences in pathologic characteristics. However, other issues such as the power to detect an association with a specific type of fracture could also affect these results.

None of the loci we describe here were reported to have an association that reached genomewide significance with bone mineral density in the Framingham Heart Study, which is, to our knowledge, the only previous publication describing a genomewide association study of bone-related phenotypes.²⁸ Genes at three of the bone mineral density loci have been previously tested for association with bone mineral density or osteoporosis. Of these, ESR1 is one of the most intensively studied candidate genes for osteoporosis.²⁹ A recent meta-analysis of more than 18,000 subjects indicated that the most intensively studied polymorphisms in ESR1 had no effect on bone mineral density but were associated with a risk of fracture.³⁰

The SNPs at the ESR1 locus that we describe here have not been previously reported to be associated with bone mineral density, nor are they in substantive linkage disequilibrium with two of the most widely studied polymorphisms in ESR1: rs2234693 (T397C) and rs9340799 (C351G) (r², 0.001 to 0.09).^29,31 Moreover, the SNPs we have identified in and around ESR1 are generally not in substantive linkage disequilibrium with one another, indicating the existence of more than one associated signal in this region and demonstrating a complex pattern of association, perhaps analogous to that observed between prostate cancer and variants in the 8q24 region.³² Furthermore, some of the associated SNPs are not only within and upstream of ESR1 but are within another uncharacterized gene, C6orf97. However, given the complexity of ESR1 transcripts and tissue-specific regulation of its different messenger RNA variants,³³ it is possible that all associated SNPs in this region affect bone mineral density by modulating ESR1 expression levels or function.

The RANKL, RANK, and OPG gene pathway has been extensively studied since the discovery of OPG^34,35 and the subsequent discovery that it interacts with RANKL^36,37,38 and RANK.³⁹ The critical involvement of this pathway in the cellular regulation of bone remodeling is evident. Furthermore, there are ongoing phase 3 clinical trials investigating the effect of a human anti-RANKL antibody (denosumab) on bone loss in patients with osteoporosis and other bone-related diseases.⁴⁰ Studies testing for association between markers at these genes and osteoporosis are few, and most of them are underpowered.^{15,16,17,18,19,20,21,22,23,24,25} Hitherto, there has been no report of consistent association between markers in these genes and bone mineral density across different populations.

The variants listed in Table 1 and Table 2 are estimated to account for approximately 3% of the total variation in hip and spine bone mineral density (calculated from the replication sets). This reinforces our belief that there are many more sequence variants relevant to osteoporosis to be identified. The variants that we have identified provide insights into the biologic pathways that influence osteoporosis. Some of them are common in the population and therefore have a greater influence on population attributable risk than would otherwise be the case. (Population attributable risk is the fraction of cases that would not occur in the population if the adverse effect of the at-risk variants could be eliminated.) For example, the odds ratio of any low-trauma fracture associated with the A allele of rs7524102 in the 1p36 region is modest (1.12) and by itself not clinically useful for purposes of prediction. Since it has a frequency of over 80% in the population, however, its population attributable risk is 17%. Follow-up studies performed with multiple and large sample sets are needed before the effect of these variants can be fully and accurately evaluated.

Supported by deCODE Genetics.

Drs. Styrkarsdottir, Halldorsson, Gretarsdottir, Gudbjartsson, Gulcher, Kong, Thorsteinsdottir, and Stefansson and Mr. Walters, Ms. Jonsdottir, and Ms. Saemundsdottir report being employees of deCODE Genetics and having equity interest in the company; Dr. Stefansson is chief executive officer and Dr. Gulcher is chief scientific officer. Dr. Eisman reports receiving consulting fees, lecture fees, or advisory fees from Amgen, deCODE Genetics, Eli Lilly, GE Lunar, Merck Sharp & Dohme, Novartis, Organon, Roche–GlaxoSmithKline, Sanofi-Aventis, Servier, and Wyeth Australia. He also reports holding a patent on the vitamin D receptor in relation to bone density and bone turnover. No other potential conflict of interest relevant to this article was reported.

We thank the subjects of the Icelandic deCODE study, the Danish Prospective Epidemiological Risk Factor (PERF) study, and the Australian Dubbo Osteoporosis Epidemiology Study (DOES) cohort for their valuable participation. We also thank the staff of the deCODE core facilities for their important contributions to this work.

Source Information

From deCODE Genetics (U.S., B.V.H., S.G., D.F.G., G.B.W., T.J., J.S., J.R.G., A.K., U.T., K.S.), Reykjavik University (B.V.H.), and University Hospital (G.S.) — all in Reykjavik, Iceland; Fjordungssjukrahusid Akureyri University Hospital, Akureyri, Iceland (T.I.); Garvan Institute of Medical Research, Sydney (J.R.C., T.V.N., J.A.E.); and Center for Clinical and Basic Research, Ballerup (Y.B.), and Nordic Bioscience, Herlev (C.C.) — both in Denmark.

This article (10.1056/NEJMoa0801197) was published at www.nejm.org on April 29, 2008. It will appear in the May 29 issue of the Journal.

Address reprint requests to Dr. Stefansson at deCODE Genetics, Sturlugata 8, IS-101, Reykjavik, Iceland, or at kari.stefansson{at}decode.is.

References

1. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359:1761-1767. [CrossRef][Web of Science][Medline]
2. Kanis JA, Melton LJ III, Christiansen C, Johnston CC, Khaltaev N. The diagnosis of osteoporosis. J Bone Miner Res 1994;9:1137-1141. [Web of Science][Medline]
3. Johnell O, Kanis JA, Oden A, et al. Predictive value of BMD for hip and other fractures. J Bone Miner Res 2005;20:1185-1194. [Erratum, J Bone Miner Res 2007;22:774.] [CrossRef][Web of Science][Medline]
4. Kanis JA, Oden A, Johnell O, et al. The use of clinical risk factors enhances the performance of BMD in the prediction of hip and osteoporotic fractures in men and women. Osteoporos Int 2007;18:1033-1046. [CrossRef][Web of Science][Medline]
5. Peacock M, Turner CH, Econs MJ, Foroud T. Genetics of osteoporosis. Endocr Rev 2002;23:303-326. [Free Full Text]
6. Albagha OME, Ralston SH. Genetics and osteoporosis. Rheum Dis Clin North Am 2006;32:659-680. [CrossRef][Web of Science][Medline]
7. Lei SF, Jiang H, Deng FY, Deng HW. Searching for genes underlying susceptibility to osteoporotic fracture: current progress and future prospect. Osteoporos Int 2007;18:1157-1175. [CrossRef][Web of Science][Medline]
8. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG. Meta-analysis of genome-wide linkage studies for bone mineral density. J Hum Genet 2006;51:480-486. [CrossRef][Web of Science][Medline]
9. Tankó LB, Bagger YZ, Nielsen SB, Christiansen C. Does serum cholesterol contribute to vertebral bone loss in postmenopausal women? Bone 2003;32:8-14. [Medline]
10. Bagger YZ, Rasmussen HB, Alexandersen P, et al. Links between cardiovascular disease and osteoporosis in postmenopausal women: serum lipids or atherosclerosis per se? Osteoporos Int 2007;18:505-512. [CrossRef][Web of Science][Medline]
11. Nguyen TV, Sambrook PN, Eisman JA. Sources of variability in bone mineral density measurements: implications for study design and analysis of bone loss. J Bone Miner Res 1997;12:124-135. [CrossRef][Web of Science][Medline]
12. Kutyavin IV, Milesi D, Belousov Y, et al. A novel endonuclease IV post-PCR genotyping system. Nucleic Acids Res 2006;34:e128-e128. [Free Full Text]
13. Devlin B, Roeder K. Genomic control for association studies. Biometrics 1999;55:997-1004. [CrossRef][Web of Science][Medline]
14. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719-748. [Web of Science][Medline]
15. Wynne F, Drummond F, O'Sullivan K, et al. Investigation of the genetic influence of the OPG, VDR (Fok1), and COLIA1 Sp1 polymorphisms on BMD in the Irish population. Calcif Tissue Int 2002;71:26-35. [CrossRef][Web of Science][Medline]
16. Langdahl BL, Carstens M, Stenkjaer L, Eriksen EF. Polymorphisms in the osteoprotegerin gene are associated with osteoporotic fractures. J Bone Miner Res 2002;17:1245-1255. [CrossRef][Web of Science][Medline]
17. Ohmori H, Makita Y, Funamizu M, et al. Linkage and association analyses of the osteoprotegerin gene locus with human osteoporosis. J Hum Genet 2002;47:400-406. [CrossRef][Web of Science][Medline]
18. Yamada T, Yamazaki H, Yamane T, et al. Regulation of osteoclast development by Notch signaling directed to osteoclast precursors and through stromal cells. Blood 2003;101:2227-2234. [Free Full Text]
19. Brändström H, Gerdhem P, Stiger F, et al. Single nucleotide polymorphisms in the human gene for osteoprotegerin are not related to bone mineral density or fracture in elderly women. Calcif Tissue Int 2004;74:18-24. [CrossRef][Web of Science][Medline]
20. Jørgensen HL, Kusk P, Madsen B, Fenger M, Lauritzen JB. Serum osteoprotegerin (OPG) and the A163G polymorphism in the OPG promoter region are related to peripheral measures of bone mass and fracture odds ratios. J Bone Miner Metab 2004;22:132-138. [CrossRef][Web of Science][Medline]
21. Zhao HY, Liu JM, Ning G, et al. The influence of Lys3Asn polymorphism in the osteoprotegerin gene on bone mineral density in Chinese postmenopausal women. Osteoporos Int 2005;16:1519-1524. [CrossRef][Web of Science][Medline]
22. Choi JY, Shin A, Park SK, et al. Genetic polymorphisms of OPG, RANK, and ESR1 and bone mineral density in Korean postmenopausal women. Calcif Tissue Int 2005;77:152-159. [CrossRef][Web of Science][Medline]
23. Hsu YH, Niu T, Terwedow HA, et al. Variation in genes involved in the RANKL/RANK/OPG bone remodeling pathway are associated with bone mineral density at different skeletal sites in men. Hum Genet 2006;118:568-577. [CrossRef][Web of Science][Medline]
24. Kim JG, Kim JH, Kim JY, et al. Association between osteoprotegerin (OPG), receptor activator of nuclear factor-kappaB (RANK), and RANK ligand (RANKL) gene polymorphisms and circulating OPG, soluble RANKL levels, and bone mineral density in Korean postmenopausal women. Menopause 2007;14:913-918. [CrossRef][Web of Science][Medline]
25. Ueland T, Bollerslev J, Wilson SG, et al. No associations between OPG gene polymorphisms or serum levels and measures of osteoporosis in elderly Australian women. Bone 2007;40:175-181. [Medline]
26. Kung AW, Huang QY. Genetic and environmental determinants of osteoporosis. J Musculoskelet Neuronal Interact 2007;7:26-32. [Medline]
27. Ralston SH, de Crombrugghe B. Genetic regulation of bone mass and susceptibility to osteoporosis. Genes Dev 2006;20:2492-2506. [Free Full Text]
28. Kiel DP, Demissie S, Dupuis J, Lunetta KL, Murabito JM, Karasik D. Genome-wide association with bone mass and geometry in the Framingham Heart Study. BMC Med Genet 2007;8:Suppl 1:S14-S14. [CrossRef][Medline]
29. Albagha OM, Pettersson U, Stewart A, et al. Association of oestrogen receptor alpha gene polymorphisms with postmenopausal bone loss, bone mass, and quantitative ultrasound properties of bone. J Med Genet 2005;42:240-246. [Free Full Text]
30. Ioannidis JP, Ralston SH, Bennett ST, et al. Differential genetic effects of ESR1 gene polymorphisms on osteoporosis outcomes. JAMA 2004;292:2105-2114. [Free Full Text]
31. Gennari L, Merlotti D, De Paola V, et al. Estrogen receptor gene polymorphisms and the genetics of osteoporosis: a HuGE review. Am J Epidemiol 2005;161:307-320. [Free Full Text]
32. Witte JS. Multiple prostate cancer risk variants on 8q24. Nat Genet 2007;39:579-580. [CrossRef][Web of Science][Medline]
33. Kos M, Reid G, Denger S, Gannon F. Genomic organization of the human ERalpha gene promoter region. Mol Endocrinol 2001;15:2057-2063. [Free Full Text]
34. Simonet WS, Lacey DL, Dunstan CR, et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 1997;89:309-319. [CrossRef][Web of Science][Medline]
35. Tsuda E, Goto M, Mochizuki S, et al. Isolation of a novel cytokine from human fibroblasts that specifically inhibits osteoclastogenesis. Biochem Biophys Res Commun 1997;234:137-142. [CrossRef][Web of Science][Medline]
36. Lacey DL, Timms E, Tan HL, et al. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 1998;93:165-176. [CrossRef][Web of Science][Medline]
37. Yasuda H, Shima N, Nakagawa N, et al. Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL. Proc Natl Acad Sci U S A 1998;95:3597-3602. [Free Full Text]
38. Schneeweis LA, Willard D, Milla ME. Functional dissection of osteoprotegerin and its interaction with receptor activator of NF-kappaB ligand. J Biol Chem 2005;280:41155-41164. [Free Full Text]
39. Nakagawa N, Kinosaki M, Yamaguchi K, et al. RANK is the essential signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem Biophys Res Commun 1998;253:395-400. [CrossRef][Web of Science][Medline]
40. Kearns AE, Khosla S, Kostenuik PJ. Receptor activator of nuclear factor {kappa}B ligand and osteoprotegerin regulation of bone remodeling in health and disease. Endocr Rev 2008;29:155-192. [Free Full Text]

Abstract PDF Supplementary Material Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited PubMed Citation

This article has been cited by other articles:

• Grundberg, E., Kwan, T., Ge, B., Lam, K. C.L., Koka, V., Kindmark, A., Mallmin, H., Dias, J., Verlaan, D. J., Ouimet, M., Sinnett, D., Rivadeneira, F., Estrada, K., Hofman, A., van Meurs, J. M., Uitterlinden, A., Beaulieu, P., Graziani, A., Harmsen, E., Ljunggren, O., Ohlsson, C., Mellstrom, D., Karlsson, M. K., Nilsson, O., Pastinen, T. (2009). Population genomics in a disease targeted primary cell model. Genome Res 19: 1942-1952 [Abstract] [Full Text]  
• Richards, J. B., Kavvoura, F. K., Rivadeneira, F., Styrkarsdottir, U., Estrada, K., Halldorsson, B. V., Hsu, Y.-H., Zillikens, M. C., Wilson, S. G., Mullin, B. H., Amin, N., Aulchenko, Y. S., Cupples, L. A., Deloukas, P., Demissie, S., Hofman, A., Kong, A., Karasik, D., van Meurs, J. B., Oostra, B. A., Pols, H. A.P., Sigurdsson, G., Thorsteinsdottir, U., Soranzo, N., Williams, F. M.K., Zhou, Y., Ralston, S. H., Thorleifsson, G., van Duijn, C. M., Kiel, D. P., Stefansson, K., Uitterlinden, A. G., Ioannidis, J. P.A., Spector, T. D., for the GEFOS (Genetic Factors for Osteoporosis) C, (2009). Collaborative Meta-analysis: Associations of 150 Candidate Genes With Osteoporosis and Osteoporotic Fracture. ANN INTERN MED 151: 528-537 [Abstract] [Full Text]  
• Pitocco, D., Zelano, G., Gioffre, G., Di Stasio, E., Zaccardi, F., Martini, F., Musella, T., Scavone, G., Galli, M., Caputo, S., Mancini, L., Ghirlanda, G. (2009). Association Between Osteoprotegerin G1181C and T245G Polymorphisms and Diabetic Charcot Neuroarthropathy: A case-control study. Diabetes Care 32: 1694-1697 [Abstract] [Full Text]  
• Jin, H., van't Hof, R. J., Albagha, O. M.E., Ralston, S. H. (2009). Promoter and intron 1 polymorphisms of COL1A1 interact to regulate transcription and susceptibility to osteoporosis. Hum Mol Genet 18: 2729-2738 [Abstract] [Full Text]  
• DENNISON, E. M., SYDDALL, H. E., JAMESON, K. A., SAYER, A. A., GAUNT, T. R., RODRIGUEZ, S., DAY, I. N.M., COOPER, C., LIPS, M. A., the Hertfordshire Cohort Study Group, (2009). A Study of Relationships Between Single Nucleotide Polymorphisms from the Growth Hormone-Insulin-like Growth Factor Axis and Bone Mass: the Hertfordshire Cohort Study. The Journal of Rheumatology 36: 1520-1526 [Abstract] [Full Text]  
• Andrews, N. A. (2009). SNPs in the Clinic for Fracture Risk Prediction: A Foreseeable Reality or a Distant Dream?. IBMS BoneKEy 6: 163-168 [Full Text]  
• Hirschhorn, J. N. (2009). Genomewide Association Studies -- Illuminating Biologic Pathways. NEJM 360: 1699-1701 [Full Text]  
• Timpson, N. J., Tobias, J. H., Richards, J. B., Soranzo, N., Duncan, E. L., Sims, A.-M., Whittaker, P., Kumanduri, V., Zhai, G., Glaser, B., Eisman, J., Jones, G., Nicholson, G., Prince, R., Seeman, E., Spector, T. D., Brown, M. A., Peltonen, L., Smith, G. D., Deloukas, P., Evans, D. M. (2009). Common variants in the region around Osterix are associated with bone mineral density and growth in childhood. Hum Mol Genet 18: 1510-1517 [Abstract] [Full Text]  
• Mencej-Bedrac, S., Prezelj, J., Kocjan, T., Teskac, K., Ostanek, B., Smelcer, M., Marc, J. (2009). The combinations of polymorphisms in vitamin D receptor, osteoprotegerin and tumour necrosis factor superfamily member 11 genes are associated with bone mineral density. J Mol Endocrinol 42: 239-247 [Abstract] [Full Text]  
• Cheung, C.-L., Chan, B. Y.Y., Chan, V., Ikegawa, S., Kou, I., Ngai, H., Smith, D., Luk, K. D.K., Huang, Q.-Y., Mori, S., Sham, P.-C., Kung, A. W.C. (2009). Pre-B-cell leukemia homeobox 1 (PBX1) shows functional and possible genetic association with bone mineral density variation. Hum Mol Genet 18: 679-687 [Abstract] [Full Text]  
• Hirschhorn, J. N., Gennari, L. (2008). Bona Fide Genetic Associations with Bone Mineral Density. NEJM 358: 2403-2405 [Full Text]  
• Xiong, D., Ferrari, S. L., Deng, H.-W. (2008). Genomewide Association Studies in Osteoporosis: Have We Reached the Bottom of the Ocean or the Tip of the Iceberg?. IBMS BoneKEy 5: 182-185 [Full Text]