SEARCH

SEARCH BY CITATION

Keywords:

  • single nucleotide polymorphism;
  • interleukin-1-receptor-associated kinase;
  • bone mineral density;
  • association study;
  • quantitative trait

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Osteoporosis, a multifactorial common disease, is believed to result from the interplay of multiple environmental and genetic determinants, including factors that regulate bone mineral density. Interleukin-1 (IL-1) is one of the most potent bone-resorbing factors, and interleukin-1-associated kinase 1 (IRAK1) is an essential effector of the IL-1 receptor signaling cascade. In genetic studies of two independent populations of postmenopausal women (cohort A: 220 individuals and cohort T: 126 individuals) from separated geographical regions of Japan, we found that radial bone mineral density levels had similar associations with IRAK1 genotypes in both populations. Two amino acid-substituting variations in the gene, encoding Phe196Ser and Ser532Leu, were in complete linkage disequilibrium (D′ = 1.0000, r2 = 1.0000, χ2 = 192.000, p = 1.2 × 10−43), and we found two exclusive haplotypes (196F/532S, frequency 0.74; 196S/532L, frequency 0.26) of the IRAK1 gene among our test subjects. In both populations, a significant association with decreased radial bone mineral density was identified with haplotype 196F/532S (in cohort A: r = 0.21, p = 0.0017; in cohort T: r = 0.23, p = 0.011). Radial bone mineral density was lowest among 196F/532S homozygotes, highest among 196S/532L homozygotes, and intermediate among heterozygotes. Accelerated bone loss also correlated with the 196F/532S haplotype in a 5-year follow-up. These results suggest that variation of IRAK1 may be an important determinant of postmenopausal osteoporosis, in part through the mechanism of accelerated postmenopausal bone loss.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

CYTOKINE PATHWAYS INCLUDING interleukin-1 (IL1) and tumor necrosis factor α (TNFα) are considered to be among the most potent of all bone-resorbing mechanisms.(1–4) These signaling cascades participate in osteoclast differentiation and can affect bone mineral density (BMD). IL-1 receptor-associated kinase (IRAK1), one of the key regulators in the IL1 pathway, is likely to play an important role in bone metabolism by regulating c-Jun N-terminal kinase (JNK) and NF-κB-activating pathways, through stimulation of both IL-1 and TNFα.

Two IRAK1 isoforms exist because of alternative splicing: IRAK1a encodes a 713-amino acid polypeptide and IRAK1b encodes a 683-amino acid polypeptide that lacks residues 514-543.(5) In vitro cell culture experiments have revealed that the latter isoform is not only functionally active but is also highly stable and resistant to degradation; this isoform therefore would prolong signal transduction after stimulation of IL-1 ligand.(5) We speculate that amino acid variations that affect IRAK1 kinase activity or protein stability might bring about variations in bone mineral metabolism.

Osteoporosis is characterized by low BMD and by deterioration of the microarchitecture of bone tissue with a consequent increase in fragility and susceptibility to fracture.(6) BMD, an important predictor of fracture, seems to be determined by genetic as well as environmental factors.(7–11) Because people carrying genetic risk are considered more susceptible to deleterious lifestyle factors, genetic elements need to be clarified for adequate diagnosis, prevention, and early treatment of osteoporosis. If the relevant genes were identified, the pathogenesis of osteoporosis presumably could be explained by variations in those genes or in loci adjacent to them. Although several genes have already been investigated as potential risk factors for osteoporosis,(12–17) an extended panel of genes needs to be examined to elucidate in detail the genetic background of this disease, considering the polygenic nature of BMD distribution and the multiplicity of endocrine and cytokine factors known to regulate bone mass and bone turnover.

For this study, we chose IRAK1 as the most probable candidate marker for osteoporosis that acts in the most upstream step of the IL-1 receptor signaling pathways, possibly regulating osteoclastogenesis from osteoblastic or stromal cells. We investigated a possible association between known amino acid variations in the gene product and radial BMD in samples from two independent Japanese populations of postmenopausal women.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Subjects

DNA samples for our association study were obtained from peripheral blood of 220 postmenopausal Japanese women living in Akita, Japan. In this cohort (cohort A), mean age and body mass index (BMI) were 71.9 ± 4.7 years and 23.9 ± 3.8 kg/m2, respectively. All were nonrelated volunteers and all provided informed consent before the study. No participant had medical complications or was undergoing treatment for conditions known to affect bone metabolism, such as pituitary diseases, hyperthyroidism, primary hyperparathyroidism, renal failure, adrenal diseases, or rheumatic diseases, and none was receiving estrogen replacement therapy. A second cohort (cohort T), consisting of 126 postmenopausal women living in Tokyo, was sampled independently and studied in the same manner. Values for age and BMI in cohort T were 57.0 ± 5 years and 22.9 ± 3.5 kg/m2, respectively.

Measurement of radial BMD

The BMD of radial bone (expressed in g/cm2) of each participant was measured by DXA using a DTX-200 Osteometer (Meditech Inc., Hawthorne, CA, USA). This parameter was normalized for differences in age, height, and weight using multiple regression analysis.(18,19) Mean values and SD of the measured BMD and adjusted BMD of cohort A were 0.321 ± 0.071 and 0.308 ± 0.057 g/cm2, respectively. Values for cohort T were 0.403 ± 0.072 and 0.394 ± 0.057 g/cm2, respectively. BMD in the radius was measured according to the Guidelines for Osteoporosis Screening in a health check-up program in Japan.(20) Five-year bone loss (g/cm2) was calculated in subjects from cohort T by subtracting the adjusted BMD levels recorded 5 years before the present analyses from the current values.

Genotyping for molecular variants in the IRAK1 gene

We examined two amino acid-substituting single nucleotide polymorphisms (SNPs) (rs1059702 and rs1059703) archived in the dbSNP at http://www.ncbi.nlm.nih.gov/SNP. The first of these, IRAK1 +587T>C, corresponded to Phe196Ser (F196S) and the other (IRAK1 +1595C>T) to Ser532Leu (S532L). Both SNPs were genotyped by a single-base extension assay using the ABI PRISM SNaPshot ddNTP Primer Extension Kit (Applied Biosystems, Foster City, CA, USA).(21) For detection of SNP +587T>C, genomic DNAs were amplified with primers 5′-GAAGGAATTCAGCCTTTGATGTAG-3′ and 5′-ATGAGACCCTCCAGCTACGC-3′; for SNP +1595C>T, primers were 5′-AGGCCATTCTCAGTCCTTGC-3′ and 5′-AGTCGGGACAGACACTCTGC-3′. Column-purified 1492-bp (IRAK1 +587T>C flanking) and 1203-bp (IRAK1 +1595C>T flanking) polymerase chain reaction (PCR) products used for the single-base extension assay were mixed with fluorescently labeled ddNTPs and 1 μM of appropriate primer (5′-AGAGGGGCCAGCAAAACGGA-3′ for IRAK1 +587T>C or 5′-GGATGCAGCTGGCGGCCTCC-3′ for IRAK1 +1595C>T). Genotypes were analyzed by electrophoresis on an ABI PRISM 377 DNA Sequencer (Applied Biosystems).

Statistical analysis of linkage disequilibrium and tests for association with BMD

To estimate haplotype frequencies, genotypes of two SNPs were determined among 96 subjects selected from Akita population. Haplotype frequencies among the representative 192 alleles investigated were calculated by Arlequin software (http://lgb.unige.ch/arlequin/software/). Linkage disequilibrium was investigated for all two-way comparisons of the polymorphisms according to Thompson's method (D, D′, and r2).(22,23) The coefficient of disequilibrium, D, is the difference between the observed haplotype frequency and the frequency expected under statistical independence: D = pAB × pab − pAb × paB. The normalized disequilibrium coefficient is obtained by dividing D by its maximum possible value (D′ = D/ D max); D max = min (pA × pB, pa × pb) if D < 0, and D max = min (pA × pb, pa × pB) if D > 0. r2 = D2/(A × B × a × b). Significance levels were determined by χ2 statistics for the corresponding 2 × 2 table.

Adjusted BMD (g/cm2) values between genotypic (haplotypic) categories were compared using ANOVA and linear regression analysis as a post-hoc test. Statistical significance was determined by ANOVA, f test. Differences were considered significant when p values were less than 5% (p < 0.05).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Complete linkage disequilibrium between F196S and S532L

Our test populations were genotyped for two SNPs encoding amino acid substitutions in the IRAK1 gene (rs1059702 and rs1059703 in the NCBI dbSNP database). These two sites, located at either end of an important catalytic kinase domain of IRAK1 at nt + 587 and nt + 1595, respectively, could conceivably affect the enzymatic activity of the gene product. The SNP at IRAK1 + 587T>C brings about substitution at amino acid residue 196 from phenylalanine to serine (F196S); the other, at IRAK1 + 1595C>T, causes substitution of leucine for serine at residue 532 (S532L). We first genotyped 220 Japanese postmenopausal women from Akita (cohort A) for the two sites; an obvious consistency led us to analyze linkage disequilibrium for possible two-way comparisons of the variations using Thompson's method (D, D′, and r2). The results confirmed complete linkage disequilibrium of the SNP F196S at position + 587 with variation S532L at position + 1595 (D′ = 1.0000, r2 = 1.0000, c2 = 192.000, p = 1.2 × 10−43). On this basis we constructed haplotypes, calculated their frequencies using the Arlequin algorithm, and identified only two distinct haplotypes among the test subjects. The major haplotype, 196F/532S or +587T/+1595C, constituted 74.0%, and the minor haplotype (196S/532L or +587C/+1595T) represented 26% of the haplotypes present in the Japanese populations examined. No haplotypic assortments of 196F/532L or 196S/532S were found.

Association of 196F/532S with low BMD in the A population

Genotypes and haplotypes for each of the two variant sites were determined and correlated with BMD among cohort A of 220 postmenopausal women living in Akita, on the western side of northern mainland Japan. A significant correlation with BMD was identified with both +587T>C SNP and +1595C>T SNPs, thus with the haplotypes of IRAK1. Haplotype 196F/532S was associated with decreased BMD in an ANOVA and a linear regression test as a post-hoc test. When BMD values were compared among the three haplotypic categories (196F/532S homozygotes, heterozygotes, and 196S/532L homozygotes), BMD was lowest among 196F/532S homozygotes (0.312 ± 0.056 g/cm2), highest among 196S/532L homozygotes (0.353 ± 0.059 g/cm2), and intermediate among heterozygotes (0.330 ± 0.055 g/cm2; r = 0.21, p = 0.0017), indicating a codominant effect of 196F532S haplotype for lowered BMD (Fig. 1A). The codominant effect of the haplotype was also supported by multiple comparison test of Tukey-Kramer (p < 0.05). Coefficient of determination value (r2) was 0.44, indicating about 4% of population variance in BMD was explained by these two SNPs in this cohort. No deviation of genotype frequencies from Hardy-Weinberg equilibrium was observed in any of the polymorphisms among our subjects (cohort A: p = 0.86, cohort T: p = 0.93).

thumbnail image

Figure FIG. 1.. Effect of IRAK1 alleles on quantitative adjusted BMD. Open circles indicate mean values; error bars indicate SD. (A) Akita population (cohort A). (B) Tokyo population (cohort T). r, correlation coefficient; p, p values, calculated by ANOVA ƒ test.

Download figure to PowerPoint

Association of 196F/532S with low BMD in the T population

To examine reproducibility of the noted association of IRAK1 variation with BMD, we examined a second cohort (cohort T) of Japanese postmenopausal women, independently collected in Tokyo, on the eastern side of central mainland Japan. ANOVA and a linear regression test in these subjects again revealed a significant correlation between the IRAK1 haplotypes and the adjusted BMD, indicating a codominant effect; that is, it was lowest among 196F/532S homozygotes (0.383 ± 0.057 g/cm2), highest among 196S/532L homozygotes (0.420 ± 0.046 g/cm2), and intermediate among heterozygotes (0.404 ± 0.058 g/cm2; r = 0.23, p = 0.011; Fig. 1B). Coefficient of determination value (r2) was 0.53, indicating about 5% of population variance in BMD was explained in this cohort.

Significant difference in bone loss between 196F/532S and non-196F/532S

To test whether bone loss of postmenopausal women is affected by the variation of IRAK1 gene, Student's t-test was examined. Five-year bone loss of the subjects in cohort T who had been followed up longitudinally for over 5 years was examined. Five-year bone loss (g/cm2) was calculated in each individual of cohort T by subtracting the adjusted BMD value obtained in previous BMD measurement carried out 5 years ago from the BMD value measured in the present analysis. Significant difference in 5-year bone loss was identified between women with the major haplotype 196F/532S homozygotes (0.0506 ± 0.0304 g/cm2, n = 65) and other women (0.0382 ± 0.0252 g/cm2, n = 61; p = 0.014; Fig. 2). This result suggests that genetic variation of IRAK1 contribute to development of osteoporosis in part through the mechanism of accelerated postmenopausal bone loss. We hypothesized that haplotype 196F/532S is an important risk factor for decreased BMD in postmenopausal women.

thumbnail image

Figure FIG. 2.. Five-year longitudinal bone loss between subjects harboring major haplotype 196F/532S and other subjects in cohort T. Mean values of adjusted BMD are plotted on the diagram from data at 5 years ago (1995) and the present data (2000). Open circles indicate values among 196S/532L homozygous or heterozygous individuals. Open squares are for individuals with 196F/532S homozygotes. Lines are synonymous to regression lines drawn for each haplotypically classified subject group. Significant difference was indicated for inclination of two lines by Student's t-test (p = 0.01).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

In the work reported here, we studied two SNPs encoding amino acid substitutions in the IRAK1 gene among two independent populations of Japanese women, analyzing haplotypes and correlating them with forearm cancellous BMD. We demonstrated complete linkage disequilibrium between the two SNPs and determined that two (and only two) haplotypes were present in our test populations. Significant association of BMD with these variants was detected; adjusted BMD was lowest in 196F/532S homozygotes, intermediate among heterozygotes, and highest among 196S/532L homozygotes in both populations. The data implied that variation in the coding region of the IRAK1 gene might have affected bone metabolism in these women, eventually introducing variation in BMD. Lowered BMD in postmenopausal women could be a result of accelerated bone loss and/or lesser acquisition of bone mass before maturation. A trend of correlation between rate of bone loss and variation of IRAK1 was indicated by analysis of bone loss over 5 years in one of the cohorts. The trend may suggest that the main contribution of the IRAK1 variation is to increase bone turnover and bone loss.

There are several interpretations for the observed associations, given the complete linkage disequilibrium between the two SNP sites. First, one site or the other may play the predominant mechanistic role that explains the associations. Second, both amino acid changes together are required to impart the functional change that underlies the observed association (codominant effect of the two polymorphisms). Third, these markers may themselves be in linkage disequilibrium with other unmeasured and functional variants at or near IRAK1 that are the true mechanistic basis for the associations. Functional studies will be required to distinguish between these possibilities. However, the S532L variation may be more important on theoretical grounds because it is located just downstream from the C-terminal end of the essential kinase domain (IRAK1 + 1595) and also because it lies within the sequence that is variably spliced between the two IRAK1 isoforms; a 30-amino acid domain is retained in IRAK1a but spliced out in IRAK1b.(5) Analysis of the shorter isoform in cultured cells showed that IRAK1b is functionally active but highly stable and resistant to degradation after stimulated by receptor binding of IL-1 ligand, implying that signal transduction is prolonged.(5) Three serine residues and one tyrosine residue are present within the spliced-out 30 amino acid sequence, and they represent potential phosphorylation sites required for hyperphosphorylation/destabilization of the enzyme.(5) The lack of a serine residue may therefore affect efficiency of degradation. Detailed characterization of the function of IRAK1 protein will be required for testing this hypothesis, as well as examinations of both variants as to their effects on signaling pathways. Such investigations may help to clarify mechanisms of postmenopausal bone loss as well as general bone loss associated with interleukin-exerted pathways.

Along with others, we have focused on molecules acting on inflammatory cytokine pathways such as IL-1, TNFα, and their effectors, which are known to be involved in bone loss.(24–29) Among many other effectors on this pathway, IRAK1 seems to be especially important for bone metabolism. However, many other effectors and inhibitors act on this signaling cascade, and clarification of their contributions awaits for further investigation.

In summary, we have identified a novel candidate osteoporosis-susceptibility gene IRAK1, whose two amino acid variations were significantly associated with BMD among postmenopausal Japanese women, where rate of bone loss, in part, seemed to be affected. Structural inspection indicated a possible contribution of S532L variation in IRAK1, which alters an amino acid immediately downstream of the kinase domain. The possible involvement of genetic variations in IRAK1 may explain, at least in part, the pathogenesis of postmenopausal osteoporosis, and may contribute to establishment of designs for suitable treatments and preventive strategies for this disease.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

This work was supported in part by a special grant for Strategic Advanced Research on “Cancer” and “Genome Science” from the Ministry of Education, Science, Sports and Culture of Japan; a research grant for research from the Ministry of Health and Welfare of Japan; and a Research for the Future Program Grant of The Japan Society for the Promotion of Science.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Kim GS, Kim CH, Choi CS, Park JY, Lee KU 1997 Involvement of different second messengers in parathyroid hormone and interleukin-1-induced interleukin-6 and interleukin-11 production in human bone marrow stromal cells. J Bone Miner Res 12:896902.
  • 2
    Machwate M, Zerath E, Holy X, Hott M, Godet D, Lomri A, Marie PJ 1995 Systematic administration of transforming growth factor-beta 2 prevents the impaired bone formation and osteopenia induced by unloading in rats. J Clin Invest 96:12451253.
  • 3
    Sjogren U, Mukohyama H, Roth C, Sundqvist G, Lerner UH 2002 Bone-resorbing activity from cholesterol-exposed macrophages due to enhanced expression of interleukin-1alpha. J Dent Res 81:1116.
  • 4
    Suda K, Woo JT, Takami M, Sexton PM, Nagai KJ 2002 Lipopolysaccharide supports survival and fusion of preosteoclasts independent of TNF-alpha, IL-1, and RANKL. Cell Physiol 190:101108.
  • 5
    Cao Z, Henzel W, Gao X 1996 IRAK: A kinase associated with the interleukin-1 receptor. Science 271:11281131.
  • 6
    Kanis JA, Melton LJ III, Christiansen C, Johnston CC, Khaltaev N 1994 The diagnosis of osteoporosis. J Bone Miner Res 9:11371141.
  • 7
    Pocock NA, Eisman JA, Hopper JL, Yeates MG, Sambrook PN, Eberl S 1987 Genetic determinants of bone mass in adults. A twin study. J Clin Invest 80:706710.
  • 8
    Slemenda CW, Christian JC, Williams CJ, Norton JA, Johnston CC Jr 1991 Genetic determinants of bone mass in adult women: A reevaluation of the twin model and the potential importance of gene interaction on heritability estimates. J Bone Miner Res 6:561567.
  • 9
    Kelly PJ, Nguyen T, Hopper J, Pocock N, Sambrook P, Eisman J 1993 Changes in axial bone density with age: A twin study. J Bone Miner Res 8:1117.
  • 10
    Seeman E, Hopper JL, Bach LA, Cooper ME, Parkinson E, McKay J, Jerums G 1989 Reduced bone mass in daughters of women with osteoporosis. N Engl J Med 320:554558.
  • 11
    Krall EA, Dawson-Hughes B 1993 Heritable and life-style determinants of bone mineral density. J Bone Miner Res 8:19.
  • 12
    Morrison NA, Qi JC, Tokita A, Kelly PJ, Crofts L, Nguyen TV, Sambrook PN, Eisman JA 1994 Prediction of bone density from vitamin D receptor alleles. Nature 367:284287.
  • 13
    Nelson DA, Kleerekoper M 1997 The search for the osteoporosis gene. J Clin Endocrinol Metab 82:989990.
  • 14
    Sano M, Inoue S, Hosoi T, Ouchi Y, Emi M, Shiraki M, Orimo H 1995 Association of estrogen receptor dinucleotide repeat polymorphism with osteoporosis. Biochem Biophys Res Commun 217:378383.
  • 15
    Kobayashi S, Inoue S, Hosoi T, Ouchi Y, Shiraki M, Orimo H 1996 Association of bone mineral density with polymorphism of the estrogen receptor gene. J Bone Miner Res 11:306311.
  • 16
    Uitterlinden AG, Burger H, Huang Q, Yue F, McGuigan FE, Grant SF, Hofman A, van Leeuwen JP, Pols HA, Ralston SH 1998 Relation of alleles of the collagen type Iα1 gene to bone density and the risk of osteoporotic fractures in postmenopausal women. N Engl J Med 338:10161021.
  • 17
    Hosoi T, Miyao M, Inoue S, Hoshino S, Shiraki M, Orimo H, Ouchi Y 1999 Association study of parathyroid hormone gene polymorphism and bone mineral density in Japanese postmenopausal women. Calcif Tissue Int 64:205208.
  • 18
    Kleinbaum DG, Kupper LL, Muller KE 1998 Applied Regression Analysis and Other Multivariate Methods, 2nd ed. PKS-Kent Publishing, Boston, MA, USA, pp. 299301.
  • 19
    Tsukamoto K, Yoshida H, Watanabe S, Suzuki T, Miyao M, Hosoi T, Orimo H, Emi M 1999 Association of radial bone mineral density with CA repeat polymorphism at the interleukin 6 locus in postmenopausal Japanese women. J Hum Genet 44:148151.
  • 20
    Orimo H, Hayashi Y, Fukunaga M, Sone T, Fujiwara S, Shiraki M, Kushida K, Miyamoto S, Soen S, Nishimura J, Oh-hashi Y, Hosoi T, Gorai I, Tanaka H, Igai T, Kishimoto H 2001 Diagnostic criteria for primary osteoporosis: Year 2000 revision of the Osteoporosis Diagnostic Criteria Review Committee; Japanese Society for Bone Miner Research. J Bone Miner Metab 19:331337.
  • 21
    Iwasaki H, Ezura Y, Ishida R, Kajita M, Kodaira M, Knight J, Daniel S, Shi M, Emi M 2002 Accuracy of genotyping for single nucleotide polymorphisms by a microarray-based single nucleotide polymorphism typing method involving hybridization of short allele-specific oligonucleotides. DNA Res 9:14.
  • 22
    Thompson EA, Deeb S, Walker D, Motulsky AG 1988 The direction of linkage disequilibrium between closely linked markers: RFLPs at the AI-CIII apolipoprotein genes. Am J Hum Genet 42:113124.
  • 23
    Miller PT, Sardina IB, Saccone NL, Putzel J, Laitinen T, Cao A, Kere J, Pilia G, Rice JP, Kwok PY 2000 Juxtaposed regions of extensive and minimal linkage disequilibrium in human Xq25 and Xq28. Nat Genet 25:324328.
  • 24
    Rachon D, Mysliwska J, Suchecka-Rachon K, Wieckiewicz J, Mysliwski A 2002 Effects of oestrogen deprivation on interleukin-6 production by peripheral blood mononuclear cells of postmenopausal women. J Endocrinol 172:387395.
  • 25
    Ota N, Nakajima T, Emi M 2001 Genetic markers for prevention of osteoporosis. Nippon Ronen Igakkai Zasshi 38:498500.
  • 26
    Ota N, Nakajima T, Nakazawa I, Suzuki T, Hosoi T, Orimo H, Inoue S, Shirai Y, Emi M 2001 A nucleotide variant in the promoter region of the interleukin-6 gene associated with decreased bone mineral density. J Hum Genet 46:267272.
  • 27
    Shinohara Y, Ezura Y, Iwasaki H, Nakazawa I, Ishida R, Kodaira M, Kajita M, Shiba T, Emi M 2001 Linkage disequilibrium and haplotype analysis among ten single-nucleotide polymorphisms of interleukin 11 identified by sequencing of the gene. J Hum Genet 46:494497.
  • 28
    Ota N, Hunt SC, Nakajima T, Suzuki T, Hosoi T, Orimo H, Shirai Y, Emi M 2001 Linkage of human tumor necrosis factor-alpha to human osteoporosis by sib pair analysis. Genes Immun 1:260264.
  • 29
    Keen RW, Woodford-Richens KL, Lanchbury JS, Spector TD 1998 Allelic variation at the interleukin-1 receptor antagonist gene is associated with early postmenopausal bone loss at the spine. Bone 23:367371.