Methylenetetrahydrofolate reductase polymorphism interacts with riboflavin intake to influence bone mineral density

Bone 35 (2004) 957 – 964 www.elsevier.com/locate/bone

Methylenetetrahydrofolate reductase polymorphism interacts with riboflavin intake to influence bone mineral density H.M. Macdonald a,*, F.E. McGuigan a, W.D. Fraser b, S.A. New c, S.H. Ralstona, D.M. Reid a a

Department of Medicine and Therapeutics, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, UK b Department of Clinical Chemistry, Royal Liverpool University Hospital, Liverpool L69 3GA, UK c Centre for Nutrition and Food Safety, School of Biomedical and Molecular Sciences, University of Surrey, Guildford, Surrey, GU2 7XH, UK Received 20 February 2004; revised 14 May 2004; accepted 25 May 2004 Available online 21 July 2004

Abstract Bone mineral density is a complex trait regulated by an interaction between genetic and environmental factors. Recent studies have identified a functional polymorphism affecting codon 677 of the methylenetetrahydrofolate reductase (MTHFR) gene that is associated with reduced bone mineral density (BMD) in Japanese and Danish postmenopausal women and increased risk of fracture in elderly Danish women. Since dietary B vitamins can influence circulating homocysteine (tHcy) levels, we examined the relationship among MTHFR genotype, B complex vitamins (folate, vitamin B12, vitamin B6 and riboflavin), BMD, and rate of change in BMD in a longitudinal study of 1241 Scottish women aged 45 – 54 years, at the time of initial study, who were followed up for a mean (SD) of 6.6 (0.7) years. There was no significant association between BMD and either MTHFR genotype or B complex vitamins when examined separately. However, we detected a significant interaction among quartile of energy-adjusted riboflavin intake, MTHFR ‘TT’ genotype, and BMD (P = 0.01 for baseline FN BMD, P = 0.02 for follow-up FN BMD). Increasing dietary riboflavin intake correlated with LS BMD and FN BMD in homozygotes for the MTHFR ‘T’ allele, which remained significant for FN after adjustment for confounders (r = 0.192, P = 0.036 for baseline; r = 0.186, P = 0.043 at follow-up) but not in the other genotypes. This raises the possibility that riboflavin intake and MTHFR genotype might interact to regulate BMD. Further work is required to determine if this association holds true for other populations and ethnic groups. D 2004 Elsevier Inc. All rights reserved. Keywords: MTHFR; Riboflavin; BMD

Introduction Osteoporosis is defined as a skeletal disorder characterized by compromised bone strength predisposing a person to an increased risk of fracture. Bone strength primarily reflects the integration of bone density, bone geometry, and bone quality. Bone density is one of the most important determinants of fracture risk, and in any given individual is determined by peak bone mass and amount of bone loss [1]. Genetic and environmental factors interact to influence bone mineral density (BMD) and bone loss, and polymorphisms in many candidate genes have been implicated in this process. These include the vitamin D receptor [2], estrogen receptor [3], apolipoprotein E [4], transforming

* Corresponding author. Osteoporosis Research Unit, Victoria Pavilion, Woolmanhill Hospital, Aberdeen, AB25 1LD, UK. Fax: +44-1224-555474. E-mail address: [email protected] (H.M. Macdonald). 8756-3282/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2004.05.018

growth factor-beta (TGFh) [5], and collagen type 1 alpha 1 (COLIA1) [6]. The methylenetetrahydrofolate reductase (MTHFR) gene has recently been identified as a possible candidate gene for regulation of BMD [7]. MTHFR is an important enzyme involved in the removal of circulating homocysteine (Fig. 1) and also lies within a linkage region for regulation of BMD on chromosome 1p36 [8 –10]. A functional polymorphism has been identified in exon 4 of the gene that results in an alanine to valine amino acid change at codon 677 (C/T), producing a heat-labile form of the enzyme that is less active than the wild type and is associated with moderately elevated tHcy levels [11]. Previous studies showed that the T variant was associated with low BMD in Japanese postmenopausal women [7] and low BMD and increased fracture incidence in Danish early postmenopausal women [12], although in a case control study of slightly older Danish women, the T variant protected against osteoporotic fracture [13]. It has also recently been reported that tHcy levels are associated with increased

958

H.M. Macdonald et al. / Bone 35 (2004) 957–964

Fig. 1. Metabolic routes for clearing homocysteine (Hcys) from the circulation MTHFR: methylenetetrahydrofolate reductase, SAM: S-adenosyl methionine, SAHcys: S-adenosyl homocysteine, THF: tetrahydrofolate, Met THF: methyltetrahydrofolate. Source: adapted from Chapter 10 in Nutritional Aspects of Osteoporosis, 2 ed. (2004) edited by P. Burckhardt, Elsevier, London.

risk of hip fracture in a Dutch postmenopausal women [14] and the Framingham elderly [15]. The MTHFR gene polymorphism has not been extensively studied in relation to BMD and other osteoporosis-related outcomes, nor have studies been performed on possible interactions with environmental factors, although riboflavin and other vitamins of the B complex have been implicated in regulating serum homocysteine levels [16]. In this study, we examined the relationship among MTHFR polymorphism, BMD, and bone loss, in relation to B-vitamin intake, in a large cohort of perimenopausal and early postmenopausal women.

Subjects and methods Subjects We studied a subset of 1241 women from the Aberdeen Osteoporosis Screening Study (APOSS). The population has been described elsewhere [17], but in brief, they were selected from a population of over 5000 women aged 45 – 54 years who came for a BMD scan in 1990 –1993 and again in 1997 – 2000. The subset included 896 mainly premenopausal women who were selected at baseline to take part in a dietary investigation [18] and approximately 350 additional women from APOSS for whom DNA had also been collected and was available at the time of analysis. At the second visit, a blood sample was taken for DNA extraction and biochemistry; a 2-h fasted urine sample was collected in the morning to analyze markers of bone

resorption, and diet was assessed by a validated food frequency questionnaire [19,20]. Bone mineral density measurements BMD was measured by dual X-ray absorptiometry primarily using a Norland XR-26 scanner (CooperSurgical Inc, Trumbull, CT). A small proportion of women were scanned using a different model of Norland Scanner (XR36) at the follow-up visit and a correction factor was used to account for the small difference (1.258%) in BMD readings between the machines detected by comparing mean phantom measurements. The coefficient of variation (CV) for measurement of BMD in our hands was 1.9% for the lumbar spine (LS) and 2.3% for the femoral neck (FN) for the XR26. The corresponding values for the XR36 were 1.2% (LS) and 2.3% (FN). Genotype analysis Genotyping was carried out by a PCR-RFLP-based method on DNA extracted from peripheral blood leukocytes using standard techniques as described by Frosst et al. [11]. The primers used were 5V-TGA AGG AGA AGG TGT CTG CGG GA-3V (forward primer) and 5V-AGG ACG GTG CGG TGA GAG TG-3V (reverse primer). In our hands for the PCR, the thermocycler was set at 94jC for 2 min, followed by 35 cycles of 94jC for 1 min, 60jC for 2 min, and 72jC for 2.5 min. Following overnight digestion at 37jC with HinfI, the digestion products (fragment sizes: CC 198 bp, CT 198 bp,

H.M. Macdonald et al. / Bone 35 (2004) 957–964

175 bp, 23 bp and TT 175 bp, 23 bp) were separated by electrophoresis on 3% agarose gels. Urinary bone resorption markers A second early morning fasted urine sample was used for analysis of free pyridinoline cross-links (fPYD) and free deoxypyridinoline cross-links (fDPD) using a modification of the high-performance liquid chromatography (HPLC) method described by Black et al. [21]. Acidified urine was applied to microgranular cellulose (CC31) in butanol (1/4) and washed before elution with heptafluorobutyric acid (0.1%) and analysis by ion-pair reverse-phase HPLC. Creatinine (Cr) was measured in urine by standard automated techniques (Roche, Lewes, UK) and results were expressed as fPYD/Cr and fDPD/Cr (nmol/mmol). The interassay CV for both cross-links methods was < 5.5% across the working concentration range for the assay [22]. Serum bone formation marker (P1NP) Serum N-terminal propeptide of type I procollagen (P1NP) was measured using an automated enzyme chemiluminescent immunoassay (ECLIA) on a ROCHE P module. The assay has a sensitivity (22%CV) of 2 Ag/l and an inter-assay CV of less than 4% across the working range of the assay.

959

energy intake according to the residual method of Willett [24] before analysis. For a subset of 898 women from the APOSS population who had also completed the same FFQ at the baseline visit, it was found that for most nutrients, mean intake was similar and that only 15% women had changed quartile classification by more than one quartile [18] with the exception of vitamin B12 where 20% women were misclassified by more than one quartile. The mean percentage change between the two visits for dietary intakes of riboflavin, vitamin B6, and folate intake was 1% or less, whereas for vitamin B12, the decrease in mean intake was 7.5%. Statistical analysis Differences in markers of bone health (BMD and markers of bone turnover) between genotypes or quartiles of vitamin intake were analyzed by one-way analysis of variance (ANOVA) and analysis of covariance (ANCOVA) adjusting for confounders. Variables were log transformed if they were not normally distributed before determining associations between BMD and nutrient intake (Pearson’s correlations). Adjustments were made for confounders including age, height, weight, smoking, hormone replacement therapy (HRT) use, menopausal status, and socioeconomic status.

Results Dietary B vitamin intake Women completed a food frequency questionnaire, which was analyzed using a UK database of food compositions [23]. Dietary B vitamin intake was adjusted for

The average age of the study population was 47.6 F 1.4 years at baseline, and BMD measurements were repeated an average of 6.6 F 0.7 years later. The distribution of groups according to menopausal status and HRT use (Fig. 2) was

Fig. 2. Percentage of women at baseline and follow-up visits in mutually exclusive groups according to menopausal status and HRT use: premenopausal, perimenopausal, postmenopausal (none taking HRT), past HRT users, and present HRT users.

960

H.M. Macdonald et al. / Bone 35 (2004) 957–964

Table 1 Subject characteristics according to genotype Pa

CC

CT

TT

N

499

617

125

Age (years) (BL) Age (years) (V2) Weight (kg) (BL) Weight (kg) (V2) Height (cm) (BL) Height (cm) (V2) Folate intake (Ag/day) (V2)b Vitamin B12 intake (Ag/day) (V2)b Vitamin B6 intake (mg/day) (V2)b Riboflavin intake (mg/day) (V2)b Vitamin D intake (Ag/day) (V2)b Calcium intake (mg/day) (V2) Energy intake (MJ/day) (V2) Physical activity level (V2) Number of smokers (V2) Socioeconomic status (% in category 1/2/3/4/5/6) Menopausal status BL (% pre-/peri-/postmenopausal) Menopausal status V2 (% pre-/peri-/postmenopausal) HRT use BL (% none, past, present) HRT use V2 (% none, past, present)

47.6 F 1.4 54.1 F 0.7 64.5 F 11.6 67.4 F 12.3 161.2 F 5.6 160.4 F 5.7 305 F 104 6.8 F 3.8 2.1 F 0.6 2.0 F 0.7 4.3 F 2.7 1053 F 358 8.0 F 2.5 1.88 F 0.35 93 (19%) 24.1/40.3/9.7/16.4/6.7/2.8 61.6/7.8/30.6 5.1/8.3/86.7 85.7/4.4/9.9 44.6/17.8/37.6

47.7 F 1.5 54.3 F 0.7 64.7 F 11.1 67.8 F 11.7 161.2 F 8.8 160.7 F 6.1 300 F 91 6.7 F 3.5 2.1 F 0.6 2.0 F 0.6 4.2 F 2.8 1034 F 341 8.0 F 2.4 1.84 F 0.33 114 (19%) 23.9/41.7/8.6/16.9/7.5/1.3 61.9/4.9/33.2 3.7/6.5/89.7 85.7/4.1/10.3 44.8/18.7/36.5

47.7 F 1.4 54.2 F 0.7 64.3 F 10.5 66.9 F 10.6 161.4 F 5.7 160.9 F 5.6 315 F133 6.7 F 4.0 2.1 F 0.8 2.1 F 0.7 4.1 F 2.5 1025 F 309 8.1 F 2.6 1.87 F 0.31 26 (21%) 20.8/49.6/8.8/15.2/4.0/1.6 68.0/6.4/25.6 2.4/7.2/90.4 87.2/4.8/8.0 48.8/14.4/36.8

0.73 0.16 0.93 0.73 0.93 0.63 0.31 0.89 0.57 0.51 0.71 0.57 0.80 0.26 0.83 0.59 0.15 0.45 0.95 0.81

Data shown as mean F SD for continuous variables. BL = baseline visit, V2 = visit 2. a One-way ANOVA or chi-squared test for categorical variables. b There was no significant difference in vitamin intake across genotype for 726 women who completed the dietary questionnaire at baseline.

significantly different at the two visits by chi-squared test (P < 0.001). At baseline, 62% were still menstruating (premenopausal – perimenopausal), and 14% had tried hormone replacement therapy (HRT), whereas at follow-up, 11% were still menstruating and 55% had tried HRT. The genotype frequencies (CC: 499 [40%] CT: 617 [50%] TT: 125 [10%]) were similar to those found in other

studies of Caucasians and were in Hardy Weinberg equilibrium. There were no significant differences in relevant baseline characteristics such as weight, height, age, menopausal status, or HRT use according to MTHFR genotype (Table 1), nor was there an association overall among MTHFR genotype baseline BMD, rates of change of BMD, or biochemical markers of bone turnover (Table 2).

Table 2 Relationship between MTHFR genotype, BMD, and markers of bone turnover

N Baseline BMD FN BMD (g/cm2) TR BMD (g/cm2) WA BMD (g/cm2) LS BMD (g/cm2) Change in BMD Change FN BMD (%/year) Change LS BMD (%/year) Change TR BMD (%/year) Change WA BMD (%/year) Bone turnover markers N fPYD/Cr (nmol/mmol) fDPD/Cr (nmol/mmol) Ratio PYD/DPD N P1NP (Ag/l)

CC

CT

TT

499

617

125

Pa

0.880 0.717 0.822 1.057

F F F F

0.12 0.12 0.16 0.15

0.888 0.720 0.831 1.066

F F F F

0.12 0.11 0.16 0.16

0.893 0.721 0.824 1.064

F F F F

0.12 0.12 0.16 0.14

0.39 0.89 0.71 0.67


0.85
0.81
0.31
0.86

F F F F

0.97 1.11 1.21 1.71


0.84
0.82
0.25
0.76

F F F F

1.04 1.11 1.43 1.83


0.96
0.81
0.29
0.96

F F F F

0.85 1.17 1.25 1.53

0.40 0.98 0.73 0.39

F 5.8 F 1.71 F 0.65

0.52 0.59 0.40

F 17.9

0.21

464 19.9 5.40 3.75 449 36.7

F 8.1 F 2.38 F 0.63 F 20.6

568 19.1 5.23 3.74 551 33.7

F 5.7 F 1.74 F 0.64 F 16.1

113 19.4 5.20 3.82 112 35.5

Data shown as mean F SD for continuous variables. FN = femoral neck, LS = lumbar spine, TR = trochanter, WA = Ward’s area, fPYD/Cr = free pyridinoline/ creatinine, fDPD = free deoxypyridinolone, P1NP = N-terminal propeptide of type I procollagen. a One-way ANOVA (variable log transformed if required).

H.M. Macdonald et al. / Bone 35 (2004) 957–964 Table 3 Relationship between intake of vitamins of the B complex and BMD Site and visit

Baseline FN BMDa LS BMDa Follow-up FN BMDa LS BMDa

Pearson’s correlation with B Vitamin (energy adjusted) Folate

Riboflavin

Vitamin B12

Vitamin B6

Unadjusted Adjustedb Unadjusted Adjustedb

0.070* 0.027 0.040
0.010

0.040 0.022 0.028 0.015

0.032 0.007 0.064* 0.045

0.073* 0.049 0.058* 0.033

Unadjusted Adjustedb Unadjusted Adjustedb

0.052y 0.006 0.029
0.014

0.055y 0.041 0.044 0.025

0.057* 0.030 0.092** 0.064*

0.065* 0.040 0.055y 0.030

a

Log transformed. Adjusted for age, weight, height, smoking, socioeconomic status, menopausal status, and HRT use. * P < 0.05. ** P < 0.01. y P < 0.07. b

There were weak associations between B complex vitamins and BMD, but most of these disappeared after adjustment for confounders, with the exception of vitamin B12 and LS BMD at follow-up (Table 3). However, for women homozygous for the heat-labile enzyme (TT variant), there was a positive relationship between riboflavin intake (as a continuous variable) and BMD at the spine and femoral neck (baseline: LS r = 0.193, P = 0.003; FN r = 0.263, P = 0.003; follow-up LS r = 0.185, P = 0.039; FN r = 0.252, P = 0.005). This remained significant for FN BMD after adjustment for confounders

961

(r = 0.192, P = 0.036 for baseline; r = 0.186, P = 0.043 at follow-up). There was no relationship with dietary folate and BMD for TT homozygotes. Examining the association between MTHFR genotype and BMD according to quartile of energy-adjusted riboflavin intake (Fig. 3) showed lower FN BMD for the two lowest quartiles of riboflavin intake and greater BMD for the highest two quartiles of riboflavin intake, for TT homozygotes (ANOVA P = 0.001 baseline, P = 0.002 follow-up; ANCOVA P = 0.005 at baseline and followup). A similar trend, although not significant after adjustment for confounders, was observed for LS BMD (ANOVA P = 0.05 baseline, P = 0.19 follow-up, ANCOVA P = 0.30 baseline, P = 0.44 follow-up; Fig. 3). There was also a significant interaction, after adjustment for confounders, among FN BMD at baseline, category of genotype (TT vs. combined CT and CC), and quartile of energy-adjusted riboflavin (P = 0.01), although this failed to reach significance when CT and CC genotypes were included separately (P = 0.07). The same pattern was seen for FN BMD at follow-up, with a significant interaction for TT versus combined CC and CT (P = 0.02) and also for the individual genotypes (P = 0.03). Furthermore, for the other hip sites (greater trochanter [TR] and Ward’s triangle [WA]), the same trends were observed (ANOVA P = 0.03 for TR BMD and P = 0.03 for WA BMD at follow-up, which remained significant after adjustment for confounders for WA BMD, P = 0.01). There was a statistically significant interaction between genotype and MTHFR genotype for WA BMD at followup (P = 0.03 for TT vs. combined CC and CT). No significant interaction or relationships were found between

Fig. 3. Relationship between baseline FN BMD, LS BMD, and MTHFR genotype according to quartile of energy-adjusted riboflavin intake (FN BMD: ANOVA P = 0.001, ANCOVA P = 0.005. LS BMD: ANOVA P = 0.051, ANCOVA P = 0.302). The interaction between riboflavin quartile and TT versus CT and CC genotypes was statistically significant (P = 0.01 after adjustment for confounders age, weight, height, smoking, socioeconomic status, menopausal status, and HRT use) for FN but not LS. Similar trends were seen for follow-up BMD.

962

H.M. Macdonald et al. / Bone 35 (2004) 957–964

genotype and riboflavin intake on bone loss rates or bone resorption or formation markers (data not shown). We investigated potential interactions with dietary folate and MTHFR genotype at all sites but did not observe any relationship or trends.

Discussion It is suggested that the MTHFR genotype may influence BMD through elevated tHcy levels caused by a defect in the variant form of the MTHFR enzyme [7]. It is known that tHcy can be removed from the circulation either by remethylation, catalyzed by MTHFR, or by transulfuration catalyzed by beta cystathionine synthase, which requires vitamin B6 (Fig. 1). The activity of both enzymes is regulated via Sadenosyl methionine (SAM) so that a defect in one pathway automatically leads to impairment in the other, with resultant elevation in homocysteine levels [25]. In vitro studies have shown that homocysteine thiolactone, a naturally occurring metabolite of homocysteine, inhibits lysyl oxidase, the enzyme required to produce collagen cross-links [26]. In addition, in homocystinuria, the early onset osteoporosis linked with the disease is a result of impaired cross-linking not with a deficiency of collagen synthesis [27]. Although MTHFR genotype was not associated with BMD or bone turnover in this population of perimenopausal and early postmenopausal women, there was a significant interaction between MTHFR genotype and riboflavin intake in relation to BMD, such that for subjects homozygous for the less active MTHFR enzyme, riboflavin intake was positively associated with BMD. In other words, BMD was lower for TT homozygotes at low intakes of riboflavin compared to the other genotypes, but at high intakes, BMD was higher. There was no association between riboflavin intake and BMD in the other genotype groups, so that BMD remained constant between high and low intakes of riboflavin. The association between riboflavin intake and BMD for TT homozygotes was observed with BMD and not with change in BMD or bone markers. However, because the women were studied around the time of the menopause, the effects of estrogen deficiency may dominate and any influence of genotype on bone turnover could have been difficult to detect. There is a plausible mechanism behind our observations in that the flavin adenine dinucleotide (FAD) prosthetic group (for which riboflavin is a precursor) is 10 times more likely to dissociate from the less-active enzyme compared to the normal enzyme and there will be a greater requirement for this vitamin [28]. Riboflavin was found to be the limiting factor in normalizing tHcy levels in a cross-sectional study of 286 healthy subjects aged 19 – 63 years [16]. We did not find any association with dietary folate and BMD, perhaps because folate intake is adequate in our population with mean intakes of 300 Ag/day (compared to 250 Ag/day in UK women aged 19 – 64, and a UK reference nutrient intake of 200 Ag/day [29]). It appears that TT

homozygotes have higher BMD compared to those carrying the wild-type C allele if riboflavin is sufficient, but show lower BMD when riboflavin intake is low. Differences in riboflavin intake may explain the contradictory findings so far reported in the literature: the T allele was associated with low BMD in Japanese postmenopausal women [7] and increased fracture incidence in Danish early postmenopausal women [12], whereas in a case control study of slightly older Danish women, the T allele protected against osteoporotic fracture [13]. However, none of these studies has investigated the role of dietary riboflavin, although one study reported adequate folate intake (mean intake 300 Ag/day) [12], which was similar to the mean intake of folate in our study. Milk and dairy products are important sources of riboflavin in the UK [29,30]. Denmark is one of the highest consumers of dairy produce, along with much of Northern Europe, whereas Japan is characterized by its low dairy product consumption [31]. Whether there are differences in the amount of dairy foods consumed within or between different populations that lead to major differences in riboflavin intake is not known. A relationship between riboflavin and plasma tHcy, which was influenced by MTHFR genotype, was also found in a Norwegian study in a population that was not B vitamin-deficient [32], whereas in the Framingham offspring cohort, the relationship between riboflavin and tHcy was influenced by both folate status and MTHFR status such that it was only seen in TT homozygotes whose folate intake was low [33]. These workers noted that products made from refined flour have been enriched with riboflavin in the United States since the 1940s, which may explain the differences in findings between studies carried out in the US compared to Europe. Hcy increases with age [34] and is higher for postmenopausal women compared to premenopausal women [35]. In a study in Northern Ireland, 49% of free-living elderly subjects were found to have insufficient riboflavin in their diets [36]. This raises the possibility that the bone health of TT homozygotes might become increasingly compromised with advancing age. In conclusion, our data suggest that the association between MTHFR genotype and BMD may depend upon B vitamin status, in particular riboflavin intake. The possible increased risk of osteoporosis as a result of elevated tHcy levels could easily be attenuated by changes in the diet. Further work is required to fully elucidate the role of MTHFR and its interaction with B complex vitamins on BMD regulation and fracture incidence.

Acknowledgments We are grateful for the assistance of Dr. N. Hoyle (Roche Diagnostics) who kindly supplied the materials for the P1NP assays, and to Grace Taylor and Euan McLeod for technical assistance with genotyping. We are also extremely grateful for the hard work of the radiographers and research

H.M. Macdonald et al. / Bone 35 (2004) 957–964

nurses at the Osteoporosis Research Unit and to all the women who kindly participated in the study. This work was supported in part by the Food Standards Agency, an MRC Co-operative group grant (G9823281), a grant from the Chief Scientist’s Office of the Scottish Executive (CZB/4/ 18), and by an Integrated Clinical Arthritis Centre Grant (R0609) from the Arthritis Research Campaign. Any views expressed are the authors’ own.

References [1] NIH Consensus Development Panel on Osteoporosis Prevention, Diagnosis and Therapy. Osteoporosis prevention, diagnosis, and therapy. JAMA 2001;285:785 – 95. [2] Cooper GS, Umbach DM. Are vitamin D receptor polymorphisms associated with bone mineral density? A meta-analysis. J Bone Miner Res 1996;11:1841 – 9. [3] Becherini L, Gennari L, Masi L, Mansani R, Massart F, Morelli A, et al. Evidence of a linkage disequilibrium between polymorphisms in the human estrogen receptor alpha gene and their relationship to bone mass variation in postmenopausal Italian women. Hum Mol Genet 2000;9:2043 – 50. [4] Cauley JA, Zmuda JM, Yaffe K, Kuller LH, Ferrell RE, Wisniewski SR, et al. Apolipoprotein E polymorphism: a new genetic marker of hip fracture risk—The Study of osteoporotic fractures. J Bone Miner Res 1999;14:1175 – 81. [5] Yamada Y, Miyauchi A, Goto J, Takagi Y, Okuizumi H, Kanematsu M, et al. Association of a polymorphism of the transforming growth factor-beta1 gene with genetic susceptibility to osteoporosis in postmenopausal Japanese women. J Bone Miner Res 1998;13: 1569 – 76. [6] Grant SFA, Reid DM, Blake G, Herd R, Fogelman I, Ralston SH. Reduced bone density and osteoporosis associated with a polymorphic Sp1 binding site in the collagen type I alpha1 gene. Nat Genet 1996;14:203 – 5. [7] Miyao M, Morita H, Hosoi T, Kurihara H, Inoue S, Hoshino S, et al. Association of methylenetetrahydrofolate reductase (MHTFR) polymorphism with bone mineral density in postmenopausal Japanese women. Calcif Tissue Int 2000;66:190 – 4. [8] Devoto M, Shimoya K, Caminis J, Ott J, Tenenhouse A, Whyte MP, et al. First-stage autosomal genome screen in extended pedigrees suggests genes predisposing to low bone mineral density on chromosomes 1p, 2p and 4q. Eur J Hum Genet 1998;6:151 – 7. [9] Devoto M, Specchia C, Li HH, Caminis J, Tenenhouse A, Rodriguez H, et al. Variance component linkage analysis indicates a QTL for femoral neck bone mineral density on chromosome 1p36. Hum Mol Genet 2001;10:2447 – 52. [10] Wilson SG, Reed PW, Bansal A, Chiano M, Lindersson M, Langdown M, et al. Comparison of genome screens for two independent cohorts provides replication of suggestive linkage of bone mineral density to 3p21 and 1p36. Am J Hum Genet 2003;72:144 – 55. [11] Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 1995;10: 111 – 3. [12] Abrahamsen B, Madsen JS, Tofteng CL, Stilgren L, Bladbjerg EM, Kristensen SR, et al. A common methylenetetrahydrofolate reductase (C677T) polymorphism is associated with low bone mineral density and increased fracture incidence after menopause: longitudinal data from the Danish osteoporosis prevention study. J Bone Miner Res 2003;18:723 – 9. [13] Jorgensen HL, Madsen JS, Madsen B, Saleh MM, Abrahamsen B, Fenger M, et al. Association of a common allelic polymorphism

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31] [32]

963

(C677T) in the methylene tetrahydrofolate reductase gene with a reduced risk of osteoporotic fractures. A Case Control Study in Danish Postmenopausal Women. Calcif Tissue Int 2002;71:386 – 92. van Meurs JB, Dhonukshe-Rutten RA, Pluijm SM, van der Klift M, de Jonge R, Lindemans J, et al. Homocysteine levels and the risk of osteoporotic fracture. N Engl J Med 2004;350:2033 – 41. McLean RR, Jacques PF, Selhub J, Tucker KL, Samelson EJ, Broe KE, et al. Homocysteine as a predictive factor for hip fracture in older persons. N Engl J Med 2004;350:2042 – 9. McNulty H, McKinley MC, Wilson B, McPartlin J, Strain JJ, Weir DG, et al. Impaired functioning of thermolabile methylenetetrahydrofolate reductase is dependent on riboflavin status: implications for riboflavin requirements. Am J Clin Nutr 2002;76:436 – 41. Macdonald HM, New SA, Campbell MK, Reid DM. Longitudinal changes in weight in perimenopausal and early postmenopausal women: effects of dietary energy intake, energy expenditure, dietary calcium intake and hormone replacement therapy. Int J Obes Relat Metab Disord 2003;27:669 – 76. Macdonald HM, New SA, Golden MHN, Campbell MK, Reid DM. Nutritional associations with bone loss during the menopausal transition: evidence of a beneficial effect for calcium, alcohol, and fruit and vegetable nutrients; and a detrimental effect for fatty acids. Am J Clin Nutr 2004;79:155 – 65. New SA. An epidemiological investigation into the influence of nutritional factors on bone mineral density and bone metabolism. Scotland, UK: PhD Thesis, University of Aberdeen; 1995. Bodner CH, Soutar A, New SA, Scaife AR, Byres M, Henderson GD, et al. Validation of a food frequency questionnaire for use in a Scottish population: correlation of antioxidant vitamin intakes with biochemical measures. J Hum Nutr Diet 1998;11:373 – 80. Black D, Duncan A, Robins SP. Quantitative analysis of the pyridinium crosslinks of collagen in urine using ion-paired reversed-phase high-performance liquid chromatography. Anal Biochem 1988;169: 197 – 203. MacDonald AG, Birkinshaw G, Durham B, Bucknall RC, Fraser WD. Biochemical markers of bone turnover in seronegative spondylarthropathy: relationship to disease activity. Br J Rheumatol 1997; 36:50 – 3. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson’s the composition of foods. Cambridge, UK: Royal Society of Chemistry; 1992. Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr 1997;65:1220S – 8S. Selhub JMJW. The pathogenesis of homocysteinemia: interruption of the coordinate regulation by s-adenosylmethionine of the remethylation and transsulfuration of homocysteine. Am J Clin Nutr 1992; 55:131 – 8. Liu G, Nellaiappan K, Kagan HM. Irreversible inhibition of lysyl oxidase by homocysteine thiolactone and its selenium and oxygen analogues. J Biol Chem 1997;51:32370 – 7. Lubec B, Fang-Kircher S, Lubec T, Blom HJ, Boers GHJ. Evidence for McKuisik’s hypothesis of deficient collagen cross-linking in patients with homocystinuria. Biochim Biophys Acta 1996;1315:159 – 62. Guenther BD, Sheppard CA, Tran P, Rozen R, Matthews RG, Ludwig ML. The structure and properties of methylenetetrahydrofolate reductase from Escherichia coli suggest how folate ameliorates human hyperhomocysteinemia. Nat Struct Biol 1999;6:359 – 65. Henderson L, Irving K, Gregory J, et al. National diet and nutrition survey: adults aged 19 to 64 years. Vitamin and Mineral Intake and Urinary Analytes, vol. 3. London: TSO; 2003. Henderson L, Gregory J, Swan G. National diet and nutrition survey: adults aged 19 to 64 years. Types and Quantities of Foods Consumed, vol. 1. London: TSO; 2003. Danish Dairy Board. Diary Statistics 2002. Arhus, Denmark: Danish Dairy Board; 2003. Hustad S, Ueland PM, Vollset SE, Zhang Y, Bjorke-Monsen AL, Schneede J. Riboflavin as a determinant of plasma total homocys-

964

H.M. Macdonald et al. / Bone 35 (2004) 957–964

teine: effect modification by the methylenetetrahydrofolate reductase C677T polymorphism. Clin Chem 2000;46:1065 – 71. [33] Jacques PF, Kalmbach R, Bagley PJ, Russo GT, Rogers G, Wilson PW, et al. The relationship between riboflavin and plasma total homocysteine in the Framingham offspring cohort is influenced by folate status and the C677T transition in the methylenetetrahydrofolate reductase gene. J Nutr 2002;132:283 – 8. [34] Ganji V, Kafai MR. Demographic, health, lifestyle, and blood vitamin determinants of serum total homocysteine concentrations in the third

National health and nutrition examination survey, 1988 – 1994. Am J Clin Nutr 2003;77:826 – 33. [35] Hak AE, Polderman KH, Westendorp IC, Jakobs C, Hofman A, Witteman JC, et al. Increased plasma homocysteine after menopause. Atherosclerosis 2000;149:163 – 8. [36] Madigan SM, Tracey F, McNulty H, Eaton-Evans J, Coulter J, McCartney H, et al. Riboflavin and vitamin B-6 intakes and status and biochemical response to riboflavin supplementation in free-living elderly people. Am J Clin Nutr 1998;68:389 – 95.