Jan 22, 2007 - sured by a Hologic QDR4500 scanner (software version. 9.03D; Hologic .... effect in the ANCOVA analysis: the best model also re- vealed a highly ..... Chung HW, Seo J-S, Hur SE, Kim HL, Kim JY, Jung JH, Kim. LH, Park BL ...
JOURNAL OF BONE AND MINERAL RESEARCH Volume 22, Number 4, 2007 Published online on January 22, 2007; doi: 10.1359/JBMR.070114 © 2007 American Society for Bone and Mineral Research
Leptin Receptor Genotype at Gln223Arg Is Associated With Body Composition, BMD, and Vertebral Fracture in Postmenopausal Danish Women Una L Fairbrother,1 László B Tankó,2 Andrew J Walley,1 Claus Christiansen,2 Philippe Froguel,1,3 and Alexandra IF Blakemore1
ABSTRACT: Leptin is emerging as a key regulator of bone remodeling. In a population-based study of 1306 postmenopausal Danish women, nonsynonymous LEPR SNPs were associated with risk of adiposity, BMD, and vertebral fracture. Smoking exacerbates this LEPR-associated fracture risk. Introduction: Nonsynonymous single nucleotide polymorphisms (SNPs) in the human LEPR gene have been associated with adiposity in a number of studies, but there have been no large-scale studies of their implications for BMD and osteoporotic fracture risk in postmenopausal women. Materials and Methods: We carried out a population-based study of 1430 women. Three well-known nonsynonymous leptin receptor (LEPR) SNPs (Lys109Arg, Gln223Arg, and Lys656Asn) were genotyped for qualitative and quantitative association analysis. Phenotype characteristics of main interest were DXA measures of body fat and lean tissue mass, BMD, and radiographic vertebral fractures. Results: Gln223Arg associated with risk of vertebral fracture (overall OR ⳱ 1.76; OR in smokers ⳱ 2.31; p ⳱ 0.0004), in addition to BMD of the femoral neck and total hip (p ⳱ 0.036 and 0.008, respectively). Heterozygote carriers showed lower BMD at both sites. Gln223Arg was also associated with adiposity (p ⳱ 0.001 for total fat mass). For adiposity, the at-risk allele was G (resulting in an arginine at position 223). Conclusions: Variation in LEPR seemed to contribute to the variation in BMD and fracture risk in Danish postmenopausal women; the heterozygous genotype was associated with increased risk of manifest osteoporosis. Further studies are needed to replicate these data and to clarify the mechanisms involved. J Bone Miner Res 2007;22:544–550. Published online on January 22, 2007; doi: 10.1359/JBMR.070114 Key words: osteoporosis, genetic research, population study, single nucleotide polymorphisms/polymorphisms, neural factors/leptin
INTRODUCTION
G
ENETIC FACTORS MAY account for 60–80% of the variation in adult bone mass.(1) A range of genes have been reported to be associated with BMD and osteoporotic fracture risk.(2) These include genes encoding cytokines and their receptors including IL-6(3) and TGFB-1,(4) genes involved in bone development and metabolism (VDR,(5) COL1A1,(6,7) CALCR,(8) RIL,(9) SOST,(10) BMP2,(11) TCIRG1,(12) LRP5(13)), or genes with ovarian function (ESR1(14)). Several of these genes have pleiotropic effects and are also associated with metabolic conditions, such as obesity and insulin resistance.(15,16) Additionally, an obesity-related gene, part of the leptin signaling pathway, the cocaine and amphetamine-regulated transcript (CART), has been recently associated with osteoporosis.(17) Taken
The authors state that they have no conflicts of interest.
together, these data suggest a partially shared genetic background between obesity and osteoporosis. There is increasing evidence that, in addition to its roles in regulation of adiposity, inflammation, and reproductive function, the cytokine/hormone leptin is involved in regulation of bone growth and turnover. This concept is also corroborated by the observation that the concentration of umbilical cord serum leptin is significantly correlated with skeletal size and BMD at birth.(18–20) The relationship between circulating leptin levels and BMD is complex, with a number of studies reporting contradictory results,(21) including a suggestion that correlation between serum leptin levels and BMD was influenced by leptin resistance.(22) The actions of leptin are mediated by the leptin receptor, nonsynonymous genetic variants of which have been associated with a range of phenotypes including adiposity,(23) hyperlipidemia,(24) type 2 diabetes,(25,26) resting metabolic rate,(27) breast cancer,(28) and circulating levels of inflammatory markers.(29)
1
Genomic Medicine, Imperial College London, Hammersmith Hospital, London, United Kingdom; 2Centre for Clinical and Basic Research, Ballerup, Denmark; 3CNRS 8090, Institute of Biology, Pasteur Institute, Lille, France.
544
LEPR SNPs IN OSTEOPOROSIS The aim of this study was to investigate the implications of the three validated nonsynonymous LEPR gene variants (Lys109Arg, Gln223Arg, Lys656Asn) for traits concerned with body size and composition, BMD, and fracture risk in a well-characterized, extensively phenotyped populationbased sample of Danish postmenopausal women. This study design has the advantage that multiple traits can be examined simultaneously. In addition, the available information concerning lifestyle factors, such as smoking habits, allows analysis of the joint effects of genetic and environmental influences.
MATERIALS AND METHODS Subjects The study population consisted of 1430 postmenopausal women between 60 and 84 years of age (mean age, 71.0 ± 5.2 [SD] years) living in the Aalborg area in Denmark. These women were recruited by questionnaire surveys also offering a screening examination for assessment of skeletal and cardiovascular status. All participating women signed an approved consent form, and the study was carried out in accordance with the Helsinki Declaration II and the European Standards for Good Clinical Practice. Women underwent a thorough examination. Height and weight were measured to the closest 0.1 cm and 0.1 kg, respectively. Information on lifestyle factors with potential risk implications for osteoporosis were gathered during personal interviews using a preformed questionnaire. BMD at the distal forearm was measured by a DTX200 arm scanner (Osteometer MediTech, Rødovre, Denmark). BMD at the lumbar spine L1–L4, total hip, and femoral neck was measured by a Hologic QDR4500 scanner (software version 9.03D; Hologic, Waltham, MA, USA). BMD measurements were available for 1425 women at the lumbar spine, for 1415 women at the femoral neck and total hip, and for 1351 women at the distal forearm. The Hologic scanner also provided direct measures of body fat and lean tissue mass. Total body fat mass was divided into two main compartments: (1) central fat mass (CFM), including the subcutaneous and visceral fat of the trunk, and (2) peripheral fat mass (PFM), including the subcutaneous fat mass of the four extremities. Body fat mass was also expressed as percentage of total body soft tissue (i.e., total fat mass pus total lean mass). Lateral X-rays of the thoracic and lumbar spine were obtained using standard X-ray equipment. Vertebral deformities from T4 to L4 were assessed by digital measurements of morphologic changes using the Image Pro Image Analyzer software (version 4.5 for Windows; Media Cybernetics, Silver Spring, MD, USA). The ratio of the anterior and posterior heights of each vertebral body was determined digitally, and a difference of >20% between the anterior and posterior edges was considered as a radiographic vertebral fracture. None of the fractures were caused by a traffic accident.
Genotyping Amplifiable DNA was available for 1306 samples. Single nucleotide polymorphism (SNP) analysis was carried out
545 using the Sequenom MALDI-TOF mass spectrometer. Triplex PCR reactions were carried out in 384-well plates for LEPR coding SNPs, Lys109Arg (rs1137100), Gln223Arg (rs1137101), and Lys656Asn (rs8179185), according to Sequenom protocols. Forward, reverse, and extension primers (ACGTTGGATGAGCTAATGCTTACCTATTTG, ACGTTGGATGGCTCCTTATGTGCAGACAAC, and TGTGCAGACAACATTGAAGGAA for Lys109Arg; ACGTTGGATGATGGGCTGAACTGACATTAG, ACGTTGGATGCAAACTCAACGACACTCTCC, and CTGGTGGAGTAATTTTCC for Gln223Arg; ACGTTGGATGAGTTCCTATGAGAGGACCTG, ACGTTGGATGACCTTCCAAAGTAAAGTGAC, and AAAGTAAAGTGACATTTTTCTC for Lys656Asn) were obtained from Sigma-Genosys, based on output from Sequenom assay design software. PCR was carried out using 1× Qiagen HotStar buffer, 2.5 mM Mg2+, 1 mM dNTPs, 0.1 M primer mix, and 0.1 U Qiagen HotStar Taq polymerase with 5 ng DNA in a 5-l reaction.(30,31) The overall genotyping success rate was 93.4%.
Statistical analysis Statistical analysis was carried out using SPSS v13. Correlation analysis was carried out using the Spearman rank test. As expected, BMD at all sites was significantly correlated with height, weight, BMI, fat mass, lean mass, and percentage fat (Table 1). Pearson’s 2 tests were used to check for departures from Hardy-Weinberg equilibrium and to compare genotype distributions between groups. There was no significant deviation from Hardy-Weinberg equilibrium for any of the three SNPs. Quantitative analysis was carried out using a general linear model univariate analysis of covariance (ANCOVA) method in SPSS version 13.0. Where appropriate (i.e., where residuals were otherwise not normally distributed), variables were natural logarithm-transformed. For analysis of ln-BMD, ln-transformed body mass index (BMI) and years since menopause were included as covariates, with genotype, current smoking status, and current treatment for osteoporosis as fixed factors: The proportion of BMD variation explained by each SNP was estimated by the adjusted r2 value for the ANCOVA. For analysis of body composition, the number of years since menopause was included as a covariate, with genotype and current smoking status as fixed factors: the best model (assessed by highest adjusted r2 value) also included an interaction term between years since menopause and genotype. To determine linkage disequilibrium (LD) relationships and to test for associations between traits and haplotypes, we used the THESIAS(32) program, which jointly estimates haplotype frequencies and haplotype effects, as well as performing likelihood ratio test statistics under various hypotheses. As expected, there was strong pairwise LD between the SNPs, shown in Table 2. Haplotype analysis was uninformative for these data. p values are presented uncorrected. p values retaining significance after modified Bonferroni correction (adjusting for LD between the SNPs and for mean correlation coeffi-
546
FAIRBROTHER ET AL.
TABLE 1. CORRELATION ANALYSIS (SPEARMAN RANK) BETWEEN BMD AT THE FEMORAL NECK, TOTAL HIP, SPINE, MEASURES OF BODY SIZE AND COMPOSITION
Femoral neck BMD Total hip BMD
Spine BMD
Arm BMD
Correlation coefficient Significance (two-tailed) N Correlation coefficient Significance (two-tailed) N Correlation coefficient Significance (two-tailed) N Correlation coefficient Significance (two-tailed) N
TABLE 2. ALLELE FREQUENCIES
AND
AND
ARM
AND
Height
Weight
BMI
Central fat mass
Total fat mass
Total lean mass
Percent fat
Percent lean
0.18 0.000001 1404 0.15 0.000001 1404 0.12 2.8 E−06 1413 0.094 0.000581 1339
0.42 0.000001 1404 0.48 0.000001 1404 0.33 0.000001 1413 0.42 0.000001 1339
0.36 0.000001 1404 0.43 0.000001 1404 0.29 0.000001 1413 0.39 0.000001 1339
0.34 0.000001 1392 0.41 0.000001 1392 0.26 0.000001 1402 0.35 0.000001 1330
0.35 0.000001 1392 0.40 0.000001 1392 0.25 0.000001 1402 0.34 0.000001 1330
0.33 0.000001 1392 0.38 0.000001 1392 0.31 0.000001 1402 0.33 0.000001 1330
0.23 0.000001 1381 0.25 0.000001 1381 0.12 3.3 E−06 1390 0.22 0.000001 1318
−0.27 0.000001 1381 −0.30 0.000001 1381 −0.18 0.000001 1390 −0.26 0.000001 1318
LINKAGE DISEQUILIBRIUM MEASURES (D⬘
SNP
Alleles
Allele frequencies
Lys109Arg Gln223Arg Lys656Asn
A/G A/G G/C
0.71/0.29 n ⳱ 1268 0.55/0.45 n ⳱ 1134 0.84/0.16 n ⳱ 1263
cient between quantitative traits) are indicated with an asterisk. We also carried out a false discovery rate (FDR) analysis for this study: cut-off p values for significance were 0.019, 0.003, and 0.0004 at FDR q-values
