719–727www.elsevier.com/locate/bone

Bone 42 (2008)

Genetic variation in the PTH pathway and bone phenotypes in elderlywomen: Evaluation of PTH, PTHLH, PTHR1 and PTHR2 genes☆

M. Tenne a, F. McGuigan a, L. Jansson a, P. Gerdhem a,c, K.J. Obrant a, H. Luthman b, K. Åkesson a,⁎

a Clinical and Molecular Osteoporosis Research Unit, Lund University, Swedenb Medical Genetics Unit, Department of Clinical Sciences, Lund University, Sweden

c Department of Orthopaedics, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden

Received 25 June 2007; revised 11 September 2007; accepted 4 December 2007Available online 23 December 2007

Abstract

Introduction: Parathyroid hormone (PTH) is a key regulator of calcium metabolism. Parathyroid hormone-like hormone (PTHrP) contributes toskeletal development through regulation of chondrocyte proliferation and differentiation during early bone growth. Both PTH and PTHrP act throughthe same receptor (PTHR1). A second receptor, PTHR2, has been identified although its function is comparatively unknown. PTH hyper-secretioninduces bone resorption, whereas intermittent injection of PTH increases bone mass. To explore the effects of genetic variation in the PTH pathway,we have analysed variations in PTH, PTHLH, PTHR1 and PTHR2 in relation to bone mass and fracture incidence in elderly women.Materials and methods: This study includes 1044 elderly women, all 75 years old, from the Malmö Osteoporosis Prospective Risk Assessment study(OPRA). Single nucleotide polymorphisms (SNPs) from 4 genes and derived haplotypes in the PTH signaling pathway were analysed in 745–1005women; 6 SNPs in thePTH gene and 3 SNPs each in thePTHLH,PTHR1 andPTHR2 genes were investigated in relation to BMD (assessed at baseline),fracture (434 prevalent fractures of all types over lifetime, self-reported and 174 incident fractures up to 7 years, X-ray verified) and serum PTH.Results and conclusion: Individually, SNPs in the 4 loci did not show any significant association with BMD. Neither were PTHLH, PTHR1 andPTHR2 polymorphisms associated with fracture. Three of 5 common haplotypes, accounting for N98% of alleles at the PTH locus, were identifiedas independent predictors of fracture. Haplotype 9 (19%) was suggestive of an association with fractures of any type sustained during lifetime(p=0.018), with carriers of one or more copies of the haplotype having the lowest incidence (p=0.006). Haplotypes 1 (13%) and 5 (37%) and 9were suggestive of an association with fractures sustained between 50 and 75 years (p=0.02, p=0.013 and p=0.034). Carriers of haplotypes 1and 5 were more likely to suffer a fracture (haplotype 1, p=0.045; haplotype 5, p=0.008).

We conclude, that while further genotyping across the gene is recommended, in this cohort of elderly Swedish women, polymorphisms in PTHmay contribute to the risk of fracture through mechanisms that are independent of BMD.© 2007 Elsevier Inc. All rights reserved.

Keywords: Genetic; Polymorphism; PTH; BMD; Fracture

Introduction

Parathyroid hormone (PTH) is a key regulator of calciummetabolism, maintaining the calcium ion levels in serum bymodulating osteoclastic bone resorption and calcium reabsorptionin the kidneys [1], whereas parathyroid hormone-like hormone(PTHrP) contributes to skeletal development through regulation of

☆ Support for the study was received from the Swedish Research Council(Project number K2006-73X-14691-04-3), the Herman Järnhardt Foundation, theKock Foundation, the Österlund Foundation and the Regional Health Council.⁎ Corresponding author. Department of Orthopaedics, Malmö University

Hospital, SE 205 02 Malmö, Sweden. Fax: +46 40 336200.E-mail address: kristina.akesson@med.lu.se (K. Åkesson).

8756-3282/$ - see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.bone.2007.12.005

chondrocyte proliferation and differentiation during early bonegrowth [2]. PTHrP thus plays a role in the growth of long bonesand potentially for stature. Secretion of PTHrP has also beenobserved in patients with tumors, causing hypercalcaemia ofmalignancy [3]. The mature PTH peptide consists of 84 aminoacids and the PTHrP peptide 141 amino acids. Of the first 13 aminoacids, 8 are identical to those of PTH and have a large degree ofstructural homology [4].

PTH and PTHrP act through the same receptor, the PTH/PTHrP receptor (PTHR1), a class B G-protein-coupled receptorwith seven trans-membrane domains mainly present in boneand kidney [4]. The physiological action induced by PTH andPTHrP is dependent on the N-terminal domain for ligand

720 M. Tenne et al. / Bone 42 (2008) 719–727

binding, where amino acid substitutions may have criticaladverse effects on high-affinity ligand binding and biologicalactivity [4]. For PTH the first 34 amino acids and for PTHrP thefirst 36, are sufficient for receptor activation, which is exploitedin recombinant pharmacological preparations. In addition, a sec-ond receptor, the PTHR2 receptor, has been identified. Thisreceptor shares more than 50 percent amino acid homology withPTHR1 and PTH binding exerts agonistic effects, howevercomparatively little is known about its biological importance [5].

Constitutive PTH hyper-secretion induces bone resorption asseen in hyperparathyroidism, whereas intermittent PTH admin-istration is a potent bone anabolic agent, stimulating osteoblastproliferation and differentiation. This leads to increased bonemass, reduced fracture risk and renders it useful in the treatmentof osteoporosis [6,7]. The importance of the PTH system inthe regulation of bone turnover through several mechanismsprompted this investigation on the influence of genetic variationin the PTH pathway on bone mineral density and fracture inelderly women.

The human gene encoding PTH is located on chromosome 11,and association studies in mainly Asian populations haveshown conclusive positive results with bone mineral density(BMD) [8–10]. The PTHR1 is located on chromosome 3. Studiesof a possibly functional tetranucleotide repeat in the P3 promoterregion suggest association with height and BMD in youngerunrelated Caucasian women [11] although not to BMD in a studyof Chinese families [12]. The parathyroid hormone-like hormone(PTHLH) gene, encoding PTHrP, located on chromosome 12 andthe PTHR2 gene, located on chromosome 2, have not previouslybeen evaluated for association with bone phenotypes.

To explore the effects of genetic variation in the PTH pathway,we evaluated variations in the PTH, PTHLH, PTHR1 and PTHR2in relation to the osteoporosis associated traits BMD and fracture,in elderly women.

Material and methods

Subjects

Study subjects were participants of the Osteoporosis Prospective RiskAssessment study, (OPRA), a cohort of 1604 women aged 75 years at the timeof invitation, whowere randomly selected from the city files inMalmö, Sweden.Ofthose initially invited (between December 1995 and May 1999) 1044 (65%)accepted to attend the baseline investigation. The OPRA study has previously beendescribed in detail [13].All participatingwomen in this studywere self-ambulatory,independently living and of Caucasian origin. At the baseline visit all participantsanswered a questionnaire regarding their general health including medication andhistory of falls and fracture. Bone mineral density (BMD) was assessed by DXA.Blood samples were collected for DNA analysis and serum concentrations of PTH,vitamin D and calcium. The cohort was followed prospectively and fracture datahas been collected continuously. At 5 years a comprehensive follow-up, similar tothe baseline investigation, was attended by 711 women or 76% of those still alive.This study was approved by the Ethics Committee of Lund University.

Bone mineral density

Areal bone mineral density (aBMD) of the lumbar spine and femoral neck wasassessed at baseline and follow-up using a Lunar®DPX-LDXA scanner (Madison,WI, USA). Precision coefficients were 0.5% at the spine and 1.6% at the femoralneck [14]. Calibration was performed 3 times per week using the manufacturer'sphantom.

Fracture ascertainment

At the baseline visit, data on self-reported fractures was collected. This dataincludes all fractures sustaineduntil 75 years of age and could be further categorisedaccording to whether fractures occurred after age 50 and if they were fracturescommonly associated with osteoporosis (hip, vertebral, proximal humerus andwrist fractures). The vastmajority of fractures (97%) occurring in those after the ageof 50 and in almost all after the age of 70 were related to low energy trauma.Fracture incidence, reported in this paper, is based on those women for whom asuccessful PTH genotype result was available (N=1005). Prospective fracture datawas also collected, with a mean follow-up time of 7.0 years (range 5.4–9.0 years)[15]. The prospective fracture data was verified with hospital X-ray files.

Biochemistry

Serum PTH concentration was analysed using an Elecsys PTH immunoas-say (Roche Diagnostics, Mannheim Germany). The intra- and inter assaycoefficient of variation (CV) were 1.6% and 5.7% respectively.

The serum concentration of 25(OH) vitamin D was analysed with thechemiluminescence Nichols Advantage assay (Nichols Institute Diagnostics, CA,USA). Themanufacturer's reference rangewas 10–68 ng/mLwith a less than 10%intra-assay CV for 25(OH) vitamin D levels between 8.4 and 75.1 ng/mL. Theintra-assay CV, as determined by duplicate measurements in 37 individuals, was8.7% in the OPRA cohort.

Serum calcium concentrationwasmeasured using a BeckmanSynchronLX20.The reference rangewas 0.5–5.0 mmol/L with an intra-assay CVof less than 2.4%.

DNA genotyping

Total genomic DNA was isolated from whole blood using the QIAamp 96DNA blood kit (Qiagen, Valencia, CA, USA) according to the manufacturer'sinstructions and quantified using PicoGreen. Genotyping, using 2 ng of DNAper reaction, was performed using allelic discrimination on amplified DNAwithTaqman technology using an ABI Prism 7900HT (Applied Biosystems, FosterCity, CA, USA) in a 384-well reaction plate under standard conditions (50 °C for2 min, 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at60 °C). (Details of the primers and probes are given in Appendix 1). Wholeblood was available from 1008 women and genotyping was successful for 745–1005 cases (Table 2).

Polymorphisms

For PTH, polymorphisms were selected from ensemble to encompass thegene and its surroundings and to represent the most commonly occurringhaplotypes according to Haploview. For the PTHLH, PTHR1 and PTHR2 genesSNPs were selected on the following criteria: relative position to the gene, minorallele frequency in Caucasian populations, and distance between thepolymorphisms, in order to obtain maximum genetic information.

We studied 6 single nucleotide polymorphisms (SNPs) in and around theregion of PTH. These were: rs307253 located ~4 kb downstream; rs307247located ~100 bp downstream; rs6254 in intron 2 (also known as BstBI);rs1459015 which lies ~5 kb upstream, rs10500783 located 56 kb upstream andfinally rs10500784 located 56.2 kb upstream (Fig. 1 and Table 2).

We studied 3 SNPs in the PTHLH region as follows: rs805512 located inintron 3; rs10492364 in intron 4 and rs1268693 ~2 kb downstream of the gene.

In PTHR1 we studied 3 SNPs: rs6442037 located in the first intron;rs724449 in intron 4 and rs7652849 located 3 kb downstream of the gene.

Three SNPs were also studied in PTHR2. These were: rs9288393 located~2 kb upstream of the gene; rs10497900 in intron 2 and rs897083 located inintron 9. Data on the SNPs are given in Table 2.

Statistics

Statistical analyses were performed using SPSS version 13.0 (SPSS Inc.,Chicago, 2000). Hardy–Weinberg equilibrium (HWE) was calculated by theChi2 test. Linkage disequilibrium (LD) between pairwise SNPs was tested usingthe standardised LD measure Lewontin's D′, calculated using the EH and 2BY2

Fig. 1. PTH gene structure and position of polymorphisms included in the study. The figure illustrates the structure of the PTH gene and the positions of the SNPs studiedbased on information fromHapMap. Also illustrated are the 5most common haplotypes derived from these SNPs. Circled, is the rs10500783 allele shared by haplotypes5, 9 and 1 with which an association with fracture has been observed. Only SNP rs6254 has been previously included in other studies [9,10,12,18,19,21,25–27].

Table 1Clinical characteristics of the OPRA cohort measured at baseline

Variable OPRA (n=1044)

Age (years) 75.2±0.1

721M. Tenne et al. / Bone 42 (2008) 719–727

programs and verified by the Haploview program which also provided thesquare of the correlation coefficient (r2) values [16]. PTH haplotypes defined bythe six SNPs in LD were determined by indirect haplotyping using the programPHASE v2.02 [17]. Haplotype analyses were performed, according to whetherindividuals possessed 0, 1 or 2 copies of the haplotype allele. Genotype andhaplotype specific differences in BMD and PTH levels were investigated usinggeneral linear model ANOVA and univariate analysis, correcting for confound-ing factors as appropriate including BMI, s-PTH, s-calcium and smoking, butnot age, since all individuals within the cohort were of the same age (75.2 years,range 75.03–75.99). PTH levels were further analysed using recessive anddominant models to explain the observed results. Predictors of fracture, oddsratios and 95% confidence intervals (CI), in each fracture category wereidentified by logistic regression analysis correcting for BMI, spine BMD,smoking, s-PTH and s-calcium.

The primary power analysis for the studywas calculated with regard to BMD.Based on a rare genotype frequency of 4%, and significance at the p=0.05 level,the study has an estimated 80% power to detect differences of 0.063 BMD units,assuming a standard deviation of 0.14 g/cm2. We performed multiple statisticaltests (15 polymorphisms, s-PTH concentration, bone mass measurements andfracture) however these factors are not independent of each other, for example,the SNPs are in linkage disequilibrium, BMD at the hip and spine is correlated, asis fracture and BMD. Consequently the adjusted level of significance equivalentto p=0.05 was estimated to be b0.009.

Height (cm) 160.6±5.8Weight (kg) 67.7±11.7BMI (kg/m2) 26.2±4.2Menopausal age (years) 49.1±4.9Time since menopause (years) 25.9±4.9No of current smokers 145 (13.9%)Spine BMD (g/cm2) 0.993±0.19Femoral neck BMD (g/cm2) 0.748±0.13Serum PTH (pmol/L) 4.66±2.07Serum calcium (mmol/L) 2.37±0.09Serum vitamin D (ng/mL) 38.04±12.0

Fracture a

Any lifetime fracture (≤75 years) 434 (43.3%)Fracture at an osteoporotic site 227 (22.7%)Fracture occurring after age 50 361 (36.0%)Prospective fractures (N75 years) 174 (17.4%)

Values are mean±SD or number (%).a Figures refer to number of women sustaining a fracture.

Results

The clinical characteristics of the OPRA study population,includingBMDand serum biochemistrymeasurements are shownin Table 1. The genotype distribution of the SNPs studied in thePTH, PTHLH, PTHR1 and PTHR2 genes is given in Table 2. Allthe genotype frequencies were in Hardy–Weinberg equilibrium.There was strong and significant linkage disequilibrium betweenthe 6 PTH SNPs studied (D′=0.977–1.0, r2 =0.04–0.46).Reflecting the strong LD, PHASE identified five commonhaplotypes accounting for N98% of alleles at the PTH locus.These were as follows (reported in the order rs307253 (C/T) –rs307247 (C/T) – rs6254 (A/G)– rs1459015 (G/A) – rs10500783(C/T) – rs10500784 (A/C)): haplotype 5, CCAACC (36.7%),

haplotype 9, TTGGCA (19.3%), haplotype 2, CCGATA (15.3%),haplotype 8, TTGACA (14.8%), haplotype 1, CCGACA (13.1%).

Bone mineral density

Analysis of individual polymorphisms of the PTH geneidentified a general trend such that homozygotes for all of thevariant alleles had marginally lower BMD values, however itwas not incremental with allele number (Table 3). The strongesttrend was with rs10500783 at the lumbar spine (p=0.013)however this diminished to p=0.038 after correction for fac-tors with potential direct or indirect influence (smoking, BMI,s-PTH, vitamin D and calcium concentrations) (Table 4). We

Table 2Polymorphisms of the PTH gene pathway and genotype frequencies

Basechange

Location Homozygous(major allele) No (%)

HeterozygoteNo (%)

Homozygous(minor allele) No (%)

Minor allelefrequency

Total P-value(HWE)

PTHrs307253 C/T 3′UTR 332 (44.2) 326 (43.4) 93 (12.4) 0.34 751 0.35rs307247 C/T 3′UTR 332 (44.1) 327 (43.5) 93 (12.4) 0.34 752 0.37rs6254 G/A Intron 2 305 (41.0) 328 (44) 112 (15.0) 0.36 745 0.12rs1459015 A/G 5′UTR 492 (65.7) 226 (30.2) 31 (4.1) 0.19 749 0.44rs10500783 C/T 5′UTR 725 (72.1) 260 (25.9) 20 (2.0) 0.15 1005 0.55rs10500784 A/C 5′UTR 408 (40.7) 448 (44.7) 146 (14.6) 0.36 1002 0.20

PTHRLHrs805512 T/G Intron 3 383 (38.4) 478 (47.9) 136 (13.6) 0.38 997 0.49rs10492364 G/A Intron 4 533 (53.1) 397 (39.5) 74 (7.4) 0.27 1004 0.99rs1268693 C/T 3′UTR 295 (29.5) 523 (52.3) 182 (18.2) 0.45 1000 0.06

PTHR1rs6442037 A/G Intron 3 440 (43.8) 445 (44.3) 120 (11.9) 0.34 1005 0.65rs724449 A/G Intron 4 377 (37.5) 478 (47.6) 150 (14.9) 0.38 1005 0.94rs7652849 C/T 3′UTR 801 (79.7) 187 (18.6) 17 (1.7) 0.11 1005 0.12

PTHR2rs9288393 C/G 5′UTR 556 (55.5) 390 (38.) 55 (54.9) 0.25 1001 0.21rs10497900 G/T Intron 2 313 (31.1) 479 (47.7) 213 (21.2) 0.44 1005 0.24rs897083 G/A Intron 9 662 (65.9) 315 (31.4) 27 (2.7) 0.19 1004 0.15

722 M. Tenne et al. / Bone 42 (2008) 719–727

were unable to identify any association between BMD at eitherthe hip or spine and PTH haplotypes, neither did we observeany associations between bone density and individual SNPs ofthe PTHLH, PTHR1 or PTHR2 genes.

Fracture

At baseline, 434 women (43.4%) had sustained a fractureduring their lifetime. None of the individual polymorphisms inthe PTH, PTHLH, PTHR1 or PTHR2 genes was found to beassociated with fracture overall.

However, regression analysis, which also includedBMD in themodel, identified the commonPTH haplotype 9 (TTGGCA) as an

Table 3Association between clinical variables and PTH genotype

rs6254 Ad

Homozygous(major allele)

Heterozygote Homozygous(minor allele)

P-

LS BMD 0.984±0.20 0.999±0.19 0.974±0.18 0.6FN BMD 0.749±0.13 0.747±0.12 0.739±0.13 0.3s-vitamin D 37.45±11.4 38.7±11.7 38.0±11.4 0.3Calcium (mmol/L) 2.373±0.09 2.381±0.10 2.376±0.10 0.3

rs307253

LS BMD 0.988±0.18 0.995±0.20 0.976±0.20 0.5FN BMD 0.746±0.13 0.749±0.13 0.745±0.13 0.6s-vitamin D 38.79±11.9 37.5±11.1 37.4±11.9 0.3Calcium (mmol/L) 2.379±0.10 2.370±0.09 2.379±0.09 0.4

rs10500783

LS BMD 0.989±0.19 1.013±0.21 0.888±0.13 0.0FN BMD 0.744±0.13 0.762±0.14 0.734±0.13 0.3s-vitamin D 37.8±12.2 38.6±11.4 41.5±9.1 0.3Calcium (mmol/L) 2.37±0.093 2.37±0.089 2.37±0.060 0.5

Values are mean±SD.a p-values adjusted for: (BMD: BMI, smoking, years since menopause, s-PTH, v

independent predictor of fracture (overall adjusted p-value=0.018). The results indicated a decreased incidence of fracturefrom 0 to 2 copies of the haplotype, although this was onlysignificant for heterozygotes (0 v 1 copy p=0.010; odds ratio(95% CI) 1.5 (1.1–2.2) and 0 v 2 copies, p=0.17; 1.9 (0.8–4.3)(Table 5). Chi-squared analysis, indicated that individuals car-rying one or two copies of the haplotype fractured less (36% and33% respectively) than non-carriers (48%) (χ2=10.1; p=0.006).

Three hundred and sixty one women (36.1%) sustained afracture between the ages of 50 and 75 years. Individually, none ofthe SNPs in PTHLH, PTHR1 or PTHR2 was associated withfracture in this category. PTH SNP rs307253 was identified as apredictor of fracture risk with an overall (adjusted) p-value=0.03

justed a rs307247 Adjusted

value Homozygous(major allele)

Heterozygote Homozygous(minor allele)

P-value

5 0.988±0.18 0.996±0.20 0.976±0.20 0.529 0.746±0.13 0.750±0.13 0.745±0.13 0.659 38.79±11.9 37.5±11.1 37.4±11.9 0.32

2.379±0.10 2.370±0.09 2.379±0.09 0.4

rs1459015

4 0.984±0.19 1.002±0.20 0.980±0.22 0.597 0.746±0.13 0.750±0.13 0.741±0.14 0.482 38.2±11.4 37.6±11.7 38.3±11.8 0.77

2.372±0.09 2.379±0.09 2.406±0.11 0.1

rs10500784

38 0.987±0.20 0.998±0.20 0.988±0.18 0.871 0.751±0.13 0.749±0.13 0.748±0.13 0.182 37.5±12.2 38.6±11.9 38.1±11.4 0.43

2.374±0.09 2.367±0.09 2.370±0.06 0.5

itamin D and calcium) or (calcium, vitamin D).

Table 4BMD values for lumbar spine and femoral neck with regard to rs10500783 areshown as crude and adjusted for serum calcium and vitamin D separately sincethey are possible mediators

Variable PTH rs10500783 P-value

Homozygous(major allele)

Heterozygous Homozygous(minor allele)

LS BMD Unadjusted 0.989±0.19 1.013±0.21 0.888±0.13 0.013Adjusted a 0.988±0.19 1.012±0.21 0.888±0.13 0.022Adjusted b 0.988±0.19 1.012±0.21 0.888±0.13 0.038

FN BMD Unadjusted 0.744±0.13 0.762±0.14 0.734±0.13 0.178Adjusted a 0.743±0.13 0.761±0.14 0.734±0.13 0.324Adjusted b 0.743±0.13 0.762±0.14 0.738±0.13 0.284

a Adjusted for BMI and smoking.b Additionally adjusted for calcium and vitamin D.

723M. Tenne et al. / Bone 42 (2008) 719–727

(genotype CC v CT, p=0.01, 2.0 (1.2–3.5) and genotype CC vTT, p=0.23, 1.8 (0.7–4.9) (Table 5). Chi-squared analysis,however, did not indicate any significant genotype relateddifferences in the number of women who fractured (p=0.11).

Three PTH haplotypes were identified by regression analysisas predictors of fracture in women over age 50 (Table 5).Haplotype 5 (CCAACC) overall adjusted p-value=0.013; 0 v 1copy p=0.06; 1.3 (0.9–1.9) and 0 v 2 copies, p=0.005; 2.0 (1.3–3.2). Furthermore, individuals carrying 1 and 2 copies of thismost

Table 5Number of women who were fractured in each category according to PTH genotyp

rs307253 rs307247 rs6254 rs10500783

Lifetimefracture a

323 324 321 434154/138/31 155/138/31 122/144/55 304/122/8

Odds ratio[95% CI]

0.8 [0.5–1.2] 0.8 [0.5–1.2] 0.7 [0.1–3.8] 0.7 [0.5–1.1]0.9 [0.4–2.0] 0.9 [0.4–2.0] 0.5 [0.04–7.3] 0.7 [0.2–2.3]

bAdjusted OR[95% CI]

0.7 [0.4–1.2] 0.7 [0.4–1.2] 0.5 [0.07–2.9] 0.7 [0.4–1.1]0.9 [0.4–2.2] 0.9 [0.4–2.2] 0.2 [0.01–3.6] 0.9 [0.3–3.1]

P-value N/S N/S N/S N/SFracture

50–75 years a265 266 264 361123/118/24 124/118/24 93/120/51 259/96/6

Odds ratio[95% CI]

1.9 [1.2–3.3] 1.9 [1.2–3.3] 1.9 [0.4–10.7 1.6 [1.0–2.5]1.7 [0.7–4.5] 1.7 [0.7–4.5] 1.2 [0.1–19.0] 1.7 [0.5–5.9]

bAdjusted OR[95% CI]

2.0 [1.2–3.5] 2.0 [1.2–3.5] 2.9 [0.4–20.0] 1.7 [1.0–2.7]1.8 [0.7–4.9] 1.8 [0.7–4.9] 3.5 [0.2–79.0] 1.3 [0.3–4.9]

P-value 0.03 N/S N/S N/SOsteoporotic

fracture a170 170 167 22784/71/15 84/71/15 59/81/27 156/65/6

Odds ratio[95% CI]

1.4 [0.8–2.6] 1.4 [0.8–2.6] 3.7 [0.7–21.0] 1.6 [0.9–2.7]1.6 [0.5–4.8] 1.6 [0.5–4.8] 3.5 [0.2–59.7] 3.7 [1.1–13.3]

bAdjusted OR[95% CI]

1.4 [0.8–2.5] 1.4 [0.8–2.6] 3.7 [0.7–21.0] 1.6 [0.9–2.7]1.6 [0.5–4.5] 1.6 [0.5–4.8] 5.1 [0.3–92.6] 3.8 [1.1–13.7]

P-value N/S N/S N/S N/SFracture

50–82 years a350 351 348 461163/154/33 164/154/33 136/143/69 323/128/10

Odds ratio[95% CI]

2.2 [1.3–3.6] 2.2 [1.3–3.6] 0.6 [0.1–3.4] 1.7 [1.1–2.6]1.6 [0.7–3.9] 1.6 [0.7–3.9] 0.7 [0.1–9.1] 2.0 [0.6–6.1]

bAdjusted OR[95% CI]

2.3 [1.4–3.8] 2.3 [1.4–3.8] 0.8 [0.1–5.3] 1.8 [1.1–2.8]1.8 [0.7–4.4] 1.8 [0.7–4.4] 0.7 [0.03–16.0] 1.6 [0.5–5.2]

P-value 0.003 N/S N/S N/S

Odds ratios with 95% CI are reported as crude and adjusted.a Number of women who were fractured, and fracture number according to geno

order: major allele homozygotes contrasted with heterozygotes; major allele homozb Odds ratios adjusted for LS BMD, BMI, smoking. Additional inclusion of calc

common haplotype sustained, in a dose dependant manner, morethan the expected number of fractures (30%, 36%, 47%; χ2=9.7,p=0.008). Haplotype 1 (CCGACA) p=0.02 overall; 0 v 1 copyp=0.011, 0.6 (0.4–0.9); and 0 v 2 copies, p=0.29; 1.7 (0.3–8.3). Although women carrying two copies of the haplotypesustained more fractures, there was no evidence of a dose re-sponse (χ2=6.2, p=0.045). Haplotype 9 p=0.034 overall; 0 v1 copy p=0.01; 0.7 (0.5–0.9) and 0 v 2 copies, p=0.29; 0.7 (0.3–1.6). Those carrying one or two copies of the haplotype had lessfractures compared to those not carrying it, but not incrementally(χ2=6.9, p=0.032). Results shown in Table 5.

Within this group, 58% (n=211) of the women sustained afracture of an osteoporotic nature. Only PTH haplotype 5 showeda non-significant trend associated with this sub-category offracture (p=0.09) and none of the individual SNPs in any of thegenes was associated with fragility fractures.

During the 7-year follow-up period an additional 100 womensuffered a fracture. None of the SNPs individually in any ofthe genes or haplotypes of PTH was associated with fracturessustained during this period.

Including these prospective fractures, the number of womenwho fractured after the age of 50 years increased to 461 (46.1%).As before, PTH haplotype 5 predicted fracture (overall p=0.011)although in this instance there was no dose effect of the haplotypeincrease (χ2=5.6; p=0.061). By comparison, haplotype 9 (overall

e and haplotype

rs10500784 rs1459015 Hap 9 Hap 5 Hap 1

433 326 324 324 324165/197/71 232/84/10 231/83/10 123/147/54 247/72/51.2 [0.2–5.9] 1.5 [1.1–2.1] 1.6 [1.2–2.25] 0.8 [0.6–1.1] 1.3 [0.9–1.8]1.1 [0.1–15.0] 1.5 [1.0–2.4] 1.9 [0.9–2.03] 0.7 [0.4–1.0] 0.3 [0.1–1.7]1.7 [0.3–11.0] 1.5 [0.9–2.3] 1.5 [1.1–2.2] 0.8 [0.5–1.0] 1.29 [0.9–1.8]3.7 [0.2–82.0] 1.4 [0.5–3.9] 1.9 [0.8–4.3] 0.7 [0.4–1.1] 0.40 [0.1–1.2]N/S N/S 0.018 N/S N/S360 268 266 266 266129/165/66 191/68/9 190/67/9 94/121/51 212/51/30.8 [0.2–4.3] 0.6 [0.4–1.0] 0.7 [0.5–0.9] 1.3 [0.9–1.8] 0.6 [0.4–0.9]2.8 [0.2–44.9] 0.9 [0.3–2.5] 0.7 [0.3–1.5] 2.0 [1.3–3.2] 1.2 [0.3–5.6]0.6 [0.1–3.7] 0.7 [0.4–1.0] 0.7 [0.5–0.9] 1.3 [0.9–1.9] 0.6 [0.4–0.9]1.0 [0.1–22.3] 0.9 [0.3–2.8] 0.7 [0.3–1.6] 2.0 [1.3–3.2] 1.7 [0.3–8.3]N/S N/S 0.034 0.013 0.02226 171 170 170 17083/108/35 124/42/5 124/41/5 61/82/27 134/35/10.4 [0.1–2.4] 0.7 [0.4–1.1] 0.6 [0.4–1.0] 1.3 [0.9–1.9] 0.8 [0.5–1.1]0.6 [0.03–9.5] 0.8 [0.2–2.6] 0.6 [0.2–1.6] 0.3 [0.8–2.3] 0.5 [0.1–4.4]0.4 [0.1–2.3] 0.7 [0.4–1.2] 0.7 [0.5–1.0] 1.4 [0.9–2.1] 0.8 [0.5–1.2]0.4 [0.02–6.8] 0.8 [0.2–2.6] 0.5 [0.2–1.5] 1.3 [0.8–2.2] 0.7 [0.1–5.7]N/S N/S N/S N/S N/S460 353 351 351 351177/198/85 251/90/12 250/89/12 137/145/69 275/72/41.9 [0.3–10.9] 0.6 [0.4–0.9] 0.6 [0.4–0.8] 1.0 [0.7–1.3] 0.7 [0.5–0.9]4.8 [0.4–60.6] 0.8 [0.3–2.2] 0.6 [0.3–1.4] 2.2 [1.4–3.4] 1.4 [0.3–6.2]1.5 [0.2–9.5] 0.6 [0.4–0.9] 0.7 [0.5–0.9] 0.9 [0.7–1.3] 0.7 [0.5–1.0]4.7 [0.2–104] 0.8 [0.3–2.1] 0.6 [0.3–1.3] 2.0 [1.3–3.2] 2.2 [0.4–12.0]N/S N/S 0.012 0.011 N/S

type and haplotype copy number. Odds ratios and 95% CI are reported in theygotes contrasted with minor allele homozygotes.ium and vitamin D did not change results appreciably. P-values are adjusted.

Table 6aSerum PTH levels by genotype and haplotype with and without adjustment under a co-dominant model

Co-dominant model Co-dominant model

PTH SNP Homozygous(major allele)

Heterozygote Homozygous(minor allele)

P a P b PTHhaplotype

Homozygous(major allele)

Heterozygote Homozygous(minor allele)

P a P b

rs307253 4.502±1.98 4.933±2.00 4.575±1.96 0.020 0.018 Haplotype 5 4.741±1.92 4.712±2.10 4.514±1.85 0.59 0.42rs307247 4.500±1.98 4.926±2.00 4.575±1.96 0.020 0.019 Haplotype 9 4.619±1.99 4.876±1.95 4.565±2.21 0.26 0.25rs6254 4.747±1.93 4.711±2.11 4.489±1.86 0.70 0.43 Haplotype 2 4.736±2.01 4.578±1.99 4.896±1.49 0.60 0.66rs10500783 4.658±2.05 4.584±2.11 4.690±1.47 0.95 0.91 Haplotype 8 4.663±1.99 4.779±2.03 4.805±1.55 0.77 0.89rs10500784 4.696±2.07 4.641±2.10 4.475±1.85 0.56 0.51 Haplotype 1 4.707±1.99 4.622±1.93 4.797±3.21 0.30 0.63rs1459015 4.620±2.00 4.886±1.95 4.543±2.17 0.18 0.23a Unadjusted p-value.b Adjusted for calcium and vitamin D.

724 M. Tenne et al. / Bone 42 (2008) 719–727

p=0.012) was associated with reduced fracture incidence, butwithout dosage effects (χ2=10.1; p=0.006) (Table 5).

No individual SNPs in the PTHLH, PTHR1 or PTHR2 geneswere associated with fracture, however, SNP rs307253 in thePTH gene was predictive of fractures (overall p=0.003, genotypeCC v CT, p=0.002, 2.3 (1.3–3.9) and genotype CC v TT,p=0.26, 1.7 (0.7–4.4) with a trend for a reduction in the numberof women who fractured when the rare allele was carried(p=0.055) (Table 5).

PTH serum concentration

The correlation between s-PTH level and BMD was weaklynegative at both the lumbar spine (−0.041) and the femoralneck (−0.03). An association was identified between serum PTH(s-PTH) concentration and PTH SNPs rs307247 (p=0.02) andrs307253 (p=0.02). In both instances, heterozygotes had thehighest s-PTH values (Table 6a). This could be explained undera recessivemodel, whereby individuals who possessed 1 or morecopies of the variant allele had lower s-PTH levels (p-value0.019 rs307247 and 0.018 rs307253). After correction for serumcalcium and vitamin D levels, the p-values for rs307247 andrs307253 increased to 0.025 and 0.023 respectively (Table 6b).No association between s-PTH levels and SNPs of the PTHLH,PTHR1 or PTHR2 genes was observed, nor did we find anyrelationship with haplotypes of PTH.

Discussion

In this study we have comprehensively evaluated the keygenes in the PTH and PTH receptor system and bone pheno-types in a well-defined, ethnically homogeneous and clinically

Table 6bSerum PTH levels for SNPs with p-values of b0.05 (rs307247, rs307253) under rec

Recessive model

PTH SNPs Minor allele carrier Homozygous (major allele) P a P

rs307247 4.499±1.97 4.844±1.99 0.019 0.rs307253 4.502±1.97 4.854±1.99 0.018 0.

a Unadjusted p-value.b Adjusted for calcium and vitamin D.

relevant population. Based on the clear implications of theimportance of PTH in the regulation of bone metabolism andthe recently observed effects of PTH as an anabolic agent, wetested the hypotheses that variations in genes of the PTHcomplex were associated with bone mineral density andfracture in elderly women. Contrary to our expectations, wefound no correlation with bone mass but suggestive implica-tions for a relationship between PTH polymorphisms andfracture.

The studies performed to date have investigated polymorph-isms of the PTH and PTHR1 genes in relation to bone mass, butgenerally these have been in non-Caucasian populations, studieswith small sample sizes or in cohorts with a primary diagnosis ofhypoparathyroidism or other endocrinological disorders. Directcomparisons between studies are often complicated by differ-ences in the phenotypes studied and the polymorphisms select-ed. The BstB1 RFLP (rs6254) frequently referenced in studiesof the PTH gene is however included in our study allowing acertain amount of direct comparison.

In our cohort we did not find evidence to associate poly-morphisms of the PTH, PTHLH, PTHR1 or PTHR2 genes withbone density. While there have been no previous studies in-cluding variation in the PTHR2 gene, earlier studies of PTHgene polymorphisms and BMD have shown both positive andnegative results. In Japanese studies, positive correlations withPTH SNP's have been observed in young, healthy girls [10]and postmenopausal women [18]. Similar observations havebeen made in Chinese postmenopausal women [9]. All of thesestudies have included relatively small numbers of individualshowever and the number of SNPs studied has, in general, beenvery restricted. In unrelated Caucasian women [11] polymorph-isms of the PTHR1 gene are associated with BMD and evidence

essive and dominant models

Dominant modelb Major Allele Carrier Homozygous (minor allele) P a P b

025 4.7132±1.99 4.575±1.96 0.54 0.41023 4.7183±2.00 4.575±0.53 0.53 0.40

725M. Tenne et al. / Bone 42 (2008) 719–727

of suggestive linkage to femoral neck BMD has been observedin families [19]. Negative correlations in Chinese women [12]and men [20] suggest that there may be both ethnic and genderdifferences in the influence of variants in this gene. Significantallelic and genotypic differentiation between Caucasian andChinese populations has been demonstrated for several candi-date genes for osteoporosis, including PTH [9].

In this study we have for the first time evaluated the as-sociation between variation in genes of the PTH system andfracture. Our data suggests that PTH polymorphisms may playa role in fracture, especially those sustained in the years afterthe menopause, regardless of whether they are categorised asfragility fractures.

Since variation in PTH appears to be relatively unimportantin terms of bone density in our cohort, it suggests that themechanism of action is independent of bone mass. Bone loss,following fracture, is likely to make a negligible contributionsince the upper extremities are the predominant sites of fractureover lifetime, and are not associated with immobilisation.

One possible explanation for this observation may lie in thefact that the PTH genotypes affect bone size rather than BMD.It has been shown that variation in the PTH gene and PTH/PTHR1 genes affects bone size in non-weight bearing bonessuch as the arm and hand [21] but not the hip or spine [20]. Thisbeing the case, fracture sites such as the ankle or proximal tibia,which are not normally classified as sites for osteoporoticfracture, could also be more vulnerable to reduced bone size. Inour cohort, peripheral fractures constitute the majority of thoserecorded. Evidence also suggests that PTH exerts differentialeffects on cortical and trabecular bone [22]. Furthermore,clinical trials evaluating the effects of PTH treatment on BMDhave indicated negative effects on cortical bone at the distalradius site, while at the same time causing large increases inBMD at the mainly trabecular bone containing vertebrae [6,7].Further evidence for the fracture association being independentof BMD comes from the weak negative correlation betweenserum PTH and bone mass observed in this study and by others[23,24].

Since maintaining calcium homeostasis through the hormo-nal feed back mechanism is vital for cellular functions, it islikely that variation in the PTH gene exerts more subtle effectsthan we can detect, on serum calcium via PTH. Although twoSNPs of the PTH gene were suggestively associated with s-PTH levels, calcium or vitamin D levels did not appreciablyalter this relationship. It seems likely that since none of theSNPs we have studied has a known function, there may be otherdisease-associated polymorphisms within or close to the genethat contribute to its regulation.

The absence of association between fracture and polymorph-isms of the PTHLH gene suggests that this hormone exerts itsgreatest effect during early bone growth, becoming relatively lessimportant in adulthood. Neither PTHR1 nor PTHR2 polymorph-isms were related to fracture. While the function of the receptor 2is relatively unknownwith regard to bone function, earlier studiesin Caucasian cohorts indicate an association between PTHR1 andbone mass, which might have been expected to translate into acontribution to fracture.

One of the strengths of our study is the comprehensivecoverage of the key genes in the PTH system. Our choice of SNPscapturing genetic variation in both the 5′ and 3′ region of all thegenes has tried to ensure maximum information content for themost likely functional parts of the genes. In particular at the PTHlocus, our choice of SNPs extending more than 50 kb upstreamand 4 kb downstream of the PTH gene itself ensured that wecaptured as much as possible of the genetic variation across theregion of the gene. The 5 common haplotypes estimated from theselected SNPs, capture 89% of the genetic variation in the locus.This is comparable with the variation detectable had tag SNPsbeen selected.

Despite this, we must be cautious and recognise the pos-sibility that we may not have fully delineated the complete generegion given that some regulatory regions lie in excess of 50 kbfrom the gene itself. Despite the fact that we have provided themost comprehensive coverage of the PTH locus reported to datein an osteoporosis study, there may be other as yet unidentifiedlinked or regulatory polymorphisms associated with the diseasephenotype.

Another strength is our study cohort which is ethnicallyhomogeneous, phenotypically well defined, and of particularimportance, a clinically relevant population, comprising as itdoes 75-year old women who are at increased risk of fracture.

Fracture ascertainment is a potential limitation sincefractures in the retrospective part of the study are self-reported,as compared to the radiographically verified prospectivefractures. While it is not clear if this leads to an over orunderestimation of fracture incidence, this is unlikely to beinfluenced by genotype. Nevertheless, it would have beenpreferred if a baseline spinal radiogram had also been includedin the study protocol, since vertebral fractures are thesymptomatically least obvious fracture to the patient. Anunavoidable weakness of our study, since a large number ofgenes and polymorphisms were studied, is the possibility ofobserving results purely by chance. Although our mostsignificant results fall short of the adjusted level of significance,the data would suggest that further analysis of the effect of PTHgene variation on fracture is justified.

We conclude, on the basis of our results, that polymorph-isms of the genes in the PTH system are not associatedwith bone mass in the OPRA population of elderly Caucasianwomen. However, an influence on fracture risk cannot beexcluded due to the suggestive association between PTH ha-plotypes and fracture, and since the PTH pathway is the mainsystem for maintaining the calcium homeostasis and bonemetabolism we are convinced that further investigation of theeffect of PTH gene variation on fracture in other similarcohorts is merited.

Acknowledgments

We thank the Laboratory of Clinical Chemistry, MalmöUniversity Hospital, for serum calcium measurements; Jan-ÅkeNilsson for statistics advice; the research nurses at the Clinicaland Molecular Osteoporosis Research Unit, and all the womenwho kindly participated in the study.

Appendix 1 Probe and primer details for PTH complexgenes

PTH

rs307253 Primers Forward 5´-CCTCCCCTTCTTTCATTCTTCATCTReverse 5´-GAAAGCCTAGACAGACTGAGTGATProbes VIC-CCCTCGTCAAGGTGFAM-CCCCTCATCAAGGTG

rs307247 Primers Forward 5´-AAGTATTTTTTTAGTCATGTGTTTCACATTTGGAReverse 5´-GGACAACAGTATTGAGCTAGTGAAGAProbes VIC-TTGGTAATACTAAAACGTTGATGTFAM-TTGGTAATACTAAAACATTGATGT

rs1459015 Primers Forward 5´-TTGATGTCAAATTCCCAAAGCTGTACTA Reverse 5`-GTGCCCTGTTCCCTTATGGAProbes VIC-AAGGAAAAAGTATCCATCTTAFAM-AAGGAAAAAGTATCCGTCTTA

rs6254 Primers Forward 5´-AGTACTGTTTTGCCTTGGAATTGGAReverse 5´-GGGCCACTTCCCATTAGCTProbes VIC-CCCCACTTCGAAATGFAM-CCCCACTTTGAAATG

rs10500783 Primers Forward 5´-GTGTGCTTGACCGCATTCTTReverse 5´-GTTTAGGATGATCACATTGAAAAGTTGTTGAProbes VIC-CCGTGGTGAAAGCFAM-TTCCGTGATGAAAGC

rs10500784 Primers Forward 5´-GGAACTTGCAGAAATACAATAGGAGGTA Reverse 5´-CTGGGTCAGATAAAATGAGCTGTTGProbes VIC-CAATTGTCAGATGATATCFAM-AATTGTCAGAGGATATC

PTHRP

rs805512 Primers Forward 5´-CAGAGAAAGAAATAAAAGCCCCCAATAC Reverse 5´-CATAAGCACTACAGTTTAATAAGTATCTTTTAGCAAAAProbes VIC-TCACATTAAAAAGCAGTTCTFAM-CACATTAAAAAGAAGTTCT

rs10492364 Primers Forward 5´-AGAATAATAAAACAGACATGGTGATATGAAATACTGATReverse 5´-GGATTTTTCTGTTAGCATTGGGAAAGTProbes VIC-CTTAATTATTCCCACATTTACFAM-TTATTCCCGCATTTAC

rs1268693 Primers Forward 5´-GTTTGCGAGGTGCTCAAGTGReverse 5´-CCAAACCAATGTGAGATGAGAAGGTProbes VIC-TCTCTGAGGACCGGTTCFAM-TCTCTGAGAACCGGTTC

PTHR1

rs6442037 Primers Forward 5´-CATGGCCAAATTGCCCTCTAAAAReverse 5´-CCTGGTTGGTCTTGGTGTTGTATAProbes VIC-CTTGTGCTGCTTTATACTFAM-CTTGTGCTGCTTTGTACT

rs724449 Primers Forward 5´-CTGAGCTCGAGACTTAGGACAAGReverse 5´-GTGGTGCAAGAGTGGAGTCAProbes VIC-CATGGATGCCCAAGCCFAM-CATGGATGCCCGAGCC

rs7652849 Primers Forward 5´- AGCCACCTTCCCTCTCTGAReverse 5´- AGGTCATGAAGCCTGAGCAGProbes VIC-ACCTGTGAGGAGAAGTFAM-ACCTGTGAAGAGAAGT

PTHR2

rs9288393 Primers Forward 5´-GGGCCTTATGAGGCAACTTGTTATTReverse 5´-CGAGGGTTTTTGTCTTGACTCCAA

Appendix 1 (continued)

PTH

Probes VIC-CACGAGCCACTGCACFAM-ACGAGCGACTGCAC

rs10497900 Primers Forward 5´-TCAGAGATCACCTAGGCCAACATTAReverse 5´-AGCTTACTCTGGGCACATAATAATATTTGTProbes VIC-CATCGTCAGTGTCCCFAM-ATCGTCATTGTCCC

rs897083 Primers Forward 5´-CCAGGCAAAGAAAACTCCAGGAAReverse 5´-AAGAAAACATTTCAGCCATGGAAGTGProbes VIC-CCAAGTCTCAGAGACAGFAM-CCAAGTCTCGGAGACAG

rs9288393

R2

726 M. Tenne et al. / Bone 42 (2008) 719–727

References

[1] Brommage R, Hotchkiss CE, Lees CJ, Stancill MW, Hock JM, Jerome CP.Daily treatment with human recombinant parathyroid hormone-(1–34),LY333334, for 1 year increases bone mass in ovariectomized monkeys.J Clin Endocrinol Metab 1999;84:3757–63.

[2] Kronenberg HM. Developmental regulation of the growth plate. Nature2003;423:332–6.

[3] Stewart AF, Horst R, Deftos LJ, Cadman EC, Lang R, Broadus AE. Biochem-ical evaluation of patients with cancer-associated hypercalcemia: evidence forhumoral and nonhumoral groups. N Engl J Med 1980;303: 1377–83.

[4] Gensure RC, Gardella TJ, Juppner H. Parathyroid hormone and parathyroidhormone-related peptide, and their receptors. Biochem Biophys ResCommun 2005;328:666–78.

[5] Usdin TB, Gruber C, Bonner TI. Identification and functional expression ofa receptor selectively recognizing parathyroid hormone, the PTH2 receptor.J Biol Chem 1995;270:15455–8.

[6] Neer M, Slovik DM, Daly M, Potts Jr T, Nussbaum SR. Treatment ofpostmenopausal osteoporosis with daily parathyroid hormone pluscalcitriol. Osteoporos Int 1993;3(Suppl 1):204–5.

[7] Reeve J, Meunier PJ, Parsons JA, Bernat M, Bijvoet OL, Courpron P,Edouard C, Klenerman L, Neer RM, Renier JC, Slovik D, Vismans FJ,Potts Jr JT. Anabolic effect of human parathyroid hormone fragment ontrabecular bone in involutional osteoporosis: a multicentre trial. Br Med J1980;280:1340–4.

[8] Deng HW, Shen H, Xu FH, Deng HY, Conway T, Zhang HT, Recker RR.Tests of linkage and/or association of genes for vitamin D receptor,osteocalcin, and parathyroid hormone with bone mineral density. J BoneMiner Res 2002;17:678–86.

[9] Dvornyk V, Liu XH, Shen H, Lei SF, Zhao LJ, Huang QR, Qin YJ, JiangDK, Long JR, Zhang YY, Gong G, Recker RR, Deng HW. Differentiationof Caucasians and Chinese at bone mass candidate genes: implication forethnic difference of bone mass. Ann Hum Genet 2003;67:216–27.

[10] Katsumata K, Nishizawa K, Unno A, Fujita Y, Tokita A. Association ofgene polymorphisms and bone density in Japanese girls. J Bone MinerMetab 2002;20:164–9.

[11] Scillitani A, Jang C, Wong BY, Hendy GN, Cole DE. A functional poly-morphism in the PTHR1 promoter region is associated with adult height andBMD measured at the femoral neck in a large cohort of young Caucasianwomen. Hum Genet 2006;119:416–21.

[12] Lei SF, Deng FY, Dvornyk V, LiuMY, Xiao SM, Jiang DK, Deng HW. The(GT)n polymorphism and haplotype of the COL1A2 gene, but not the(AAAG)n polymorphism of the PTHR1 gene, are associated with bonemineral density in Chinese. Hum Genet 2005;116:200–7.

[13] Gerdhem P, Brandstrom H, Stiger F, Obrant K, Melhus H, Ljunggren O,Kindmark A, Akesson K. Association of the collagen type 1 (COL1A 1) Sp1binding site polymorphism to femoral neck bone mineral density and wristfracture in 1044 elderly Swedish women. Calcif Tissue Int 2004;74: 264–9.

[14] Karlsson MK, Obrant KJ, Nilsson BE, Johnell O. Bone mineral densityassessed by quantitative ultrasound and dual energy X-ray absorptiometry.Normative data in Malmo, Sweden. Acta Orthop Scand 1998;69:189–93.

727M. Tenne et al. / Bone 42 (2008) 719–727

[15] Gerdhem P, Ivaska KK, Isaksson A, Pettersson K, Vaananen HK, Obrant KJ,AkessonK.Associationsbetweenhomocysteine, bone turnover,BMD,mortality,and fracture risk in elderly women. J Bone Miner Res 2007;22: 127–34.

[16] Terwilliger JD, Ott J. Handbook of human genetic linkage. Johns HopkinsUniversity Press; 1994.

[17] Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotypereconstruction from population data. Am J Hum Genet 2001;68:978–89.

[18] Hosoi T, Miyao M, Inoue S, Hoshino S, Shiraki M, Orimo H, Ouchi Y.Association study of parathyroid hormone gene polymorphism and bonemineral density in Japanese postmenopausal women. Calcif Tissue Int1999;64:205–8.

[19] Zhang YY, Liu PY, Lu Y, Xiao P, Liu YJ, Long JR, Shen H, Zhao LJ, ElzeL, Recker RR, Deng HW. Tests of linkage and association of PTH/PTHrPreceptor type 1 gene with bone mineral density and height in Caucasians.J Bone Miner Metab 2006;24:36–41.

[20] Lei SF, Zhang YY, Deng FY, Liu MY, Liu XH, Zhou XG, Deng HW. Bonemineral density and five prominent candidate genes inChinesemen: associations,interaction effects and their implications. Maturitas 2005;51: 199–206.

[21] Gong G, Johnson ML, Barger-Lux MJ, Heaney RP. Association of bonedimensions with a parathyroid hormone gene polymorphism in women.Osteoporos Int 1999;9:307–11.

[22] Calvi LM, Sims NA, Hunzelman JL, Knight MC, Giovannetti A, SaxtonJM, Kronenberg HM, Baron R, Schipani E. Activated parathyroidhormone/parathyroid hormone-related protein receptor in osteoblasticcells differentially affects cortical and trabecular bone. J Clin Invest2001;107:277–86.

[23] Neuprez A, Bruyere O, Collette J, Reginster JY. Vitamin D inadequacyin Belgian postmenopausal osteoporotic women. BMC Public Health2007;7:64.

[24] von Muhlen DG, Greendale GA, Garland CF, Wan L, Barrett-Connor E.Vitamin D, parathyroid hormone levels and bone mineral density incommunity-dwelling older women: the Rancho Bernardo study. Osteo-poros Int 2005;16:1721–6.

[25] Deng FY, Long JR, Lei SF, Li MX, Deng HW. Potential effect of inter-genic action on peak bone mass (PBM) in Chinese females. Yi Chuan XueBao 2005;32:1003–10.

[26] Kanzawa M, Sugimoto T, Kobayashi T, Kobayashi A, Chihara K.Parathyroid hormone gene polymorphisms in primary hyperparathyroid-ism. Clin Endocrinol (Oxf ) 1999;50:583–8.

[27] Goswami R, Mohapatra T, Gupta N, Rani R, Tomar N, Dikshit A, SharmaRK. Parathyroid hormone gene polymorphism and sporadic idiopathichypoparathyroidism. J Clin Endocrinol Metab 2004;89:4840–5.