Lack of Correlation Between Start Codon Polymorphism of the Vitamin D Receptor Gene and Bone Mineral Density in Premenopausal French Women: The OFELY Study



Previous studies have demonstrated an association between bone mineral density (BMD) and a start codon polymorphism (SCP) of the vitamin D receptor (VDR) gene in pre- and postmenopausal Caucasian and Japanese women. The SCP can be determined by a restriction fragment length polymorphism defined by the FokI restriction endonuclease. VDR alleles containing the FokI site are denoted by f and alleles lacking the site by F. In this study, the association between BMD and the SCP was examined in a group of 174 premenopausal French women who previously had been studied for a relationship between BMD and the VDR BsmI polymorphism. The SCP genotypes of the French women were FF 40%, Ff 44%, and ff 16% and they were independent of the BsmI genotype. BMD was measured by dual-energy X-ray absorptiometry at the lumbar spine, proximal femur, forearm, and total body. In contrast to previous reports, there was no association of BMD with SCP genotype in this group of Caucasian women at any site. We also measured several biochemical indices of calcium homeostasis and bone turnover. We found no statistically significant associations between SCP genotype and calcium, parathyroid hormone, or vitamin D levels. There was a 33.5% higher level of the skeletal resorption marker N-telopeptides of type I collagen in the women with the ff genotype when compared with women with the FF genotype (p = 0.004). Other bone turnover markers failed to show an association with SCP genotype. In summary, the SCP genotype may not be associated with reduced BMD in all geographical or ethnic populations.


During the past 3 years, there has been substantial interest in the association of polymorphic vitamin D receptor (VDR) alleles and the risk of osteoporosis.(1) This field of research remains controversial because of the inconsistent findings in different studies.(2-4) Nonetheless, a preponderance of the evidence suggests that the VDR alleles defined by BsmI, ApaI, and TaqI restriction fragment length polymorphisms (RFLPs) in intron 8 and exon 9(1) are associated with a modest variation in bone mineral density (BMD).(5)

The nucleotide sequence of the human VDR cDNA reported by Baker et al.(6) reveals the presence of two potential translation initation (ATG) codons. Subsequent work has demonstrated that a T/C polymorphism (ATG to ACG) exists at the first ATG in humans(7,8) and we refer to this polymorphism as the start codon polymorphism (SCP).(9,10) Thus, the VDR allele with the T variant encodes two potential ATG initiation codons, and while the translation of the VDR mRNA from this allele can initiate from the first AUG, the translation of mRNA from the allele with the C variant must initiate from the second AUG. Translational initiation from the second AUG shortens the VDR by three amino acids.(11) Such a difference in VDR primary structure might contribute to altered receptor function in contrast to the silent polymorphisms in intron 8 and exon 9. Indeed, the experiments of Arai et al.(11) indicate that the shorter form of the VDR gave approximately 1.7-fold greater transcriptional activation in transfected HeLa cells than the longer form.

An RFLP using the FokI restriction endonuclease can detect the SCP polymorphism.(9,10) VDR alleles having the first ATG are denoted by f and VDR alleles lacking the first ATG by F. We have previously reported that Mexican-American Caucasian postmenopausal women from northern California with the ff genotype had a 12.8% lower BMD at the lumbar spine than women with the FF genotype. Also, over a 2-year period, the rate of loss of BMD at the femoral neck was higher for ff women.(9) In a second study, we reported that premenopausal Caucasian women from Boston, Massachusetts, U.S.A., with the ff genotype had a 12.1% lower BMD at the femoral neck and 4.3% lower total body BMD values as compared with women with the FF genotype.(10) In Japanese premenopausal women, the BMD at the lumbar spine was also 12% lower for ff genotype than for the FF genotype.(11) The current study was undertaken to determine whether the association of the SCP with BMD existed in a geographically different group of healthy well-characterized Caucasian women from France.



One hundred seventy-four premenopausal Caucasian women (age 31–56 years) were included in this study. This group of women were enrolled in the OFELY Study previously described by Garnero et al.(12) All subjects were inhabitants of the environs of Lyon, France. Exclusion criteria for these subjects included disorders of calcium metabolism, irregular menses, and elevated follicle-stimulating hormone levels. The women included in this study were all healthy and they were not taking any medications known to affect bone and mineral metabolism. Thirteen women in the initial study by Garnero et al.(12) were not included in this study because their DNA was not available for analysis. Dietary intakes of calcium were assessed by a sequential self-questionaire as previously described.(12) All subjects gave informed consent, and the study protocol was approved by the institutional review boards of Stanford University and Hopital E. Herriot.

Bone densitometry

BMD was measured at the lumbar spine in the frontal view (L1–L4), at the proximal femur, distal radius, and the whole body (bone mineral content [BMC]) using dual-energy X-ray absorptiometry (DXA) on a Hologic QDR 2000 instrument (Hologic, Waltham, MA, U.S.A.). The in vivo precision of DXA measurements was 0.9, 1.2, 1.7, 2.2, 1.0, 0.6, and 1.0% at the frontal lumbar spine, femoral neck, trochanter, Ward's triangle, total hip, distal radius, and whole body BMC, respectively.


DNA was isolated from peripheral white blood cells using methods previously described.(12) The primers and PCR conditions for amplifying exon 2 of the VDR gene were also described previously.(9,13) Briefly, genomic DNA served as a template for amplification of exon 2 with the primers VDR2a: 5′-AGCTGGCCCTGGCACTGCTCTGCTCT-3′ and VDR2b: 5′-ATGGAAACACCTTGCTTCTTCTCCC TC-3′ (each 0.49 μM final concentration), in a buffer containing 1.5 mM MgCl2, 60 mM Tris HCl, pH 9.0, 15 mM (NH4)2SO4, 10% dimethylsulfoxide (DMSO), 200 μM each of dATP, dCTP, dGTP and dTTP, and 1.5 U of Taq DNA Polymerase (Stratagene, La Jolla, CA, U.S.A.). The PCR conditions were 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s for 30 cycles. PCR products were digested with FokI (New England Biolabs, Beverly, MA, U.S.A.) at 37°C for 3 h and then electrophoresed through a 3% agarose gel containing ethidium bromide (0.5 μg/ml). Alleles having the FokI site in exon 2 are denoted by f and alleles lacking the restriction site denoted by F. The subjects had been genotyped previously at the BsmI, ApaI, and TaqI sites.(12)

Biochemical measurements

Serum levels of calcium, phosphate, creatinine, albumin, and urinary levels of creatinine were measured by standard clinical laboratory techniques. Measurements of the levels of intact parathyroid hormone (PTH), 25-hydroxyvitamin D (25(OH)D), the bone formation markers, osteocalcin, bone-specific alkaline phosphatase (BAP), and procollagen type I peptide (PICP), and the bone resorption markers, urinary type I collagen cross-linked N-telopeptide (NTX, Osteomark, Ostex International, Inc., Seattle, WA, U.S.A.) and C-telopeptide of an α1 chain of type I collagen (CTX, Crosslaps ELISA, Osteometer A/S, Ballerup, Denmark), were as previously described.(12)

Statistical analysis

Data were evaluated by analysis of variance with Fisher's protected least significant difference post hoc test using Statview 4.5 software (Abacus Concepts, Berkeley, CA, U.S.A.) and presented as means ± SD, except where noted. This study has a 90% power to detect an 8.5% difference in BMD, assuming an α of 0.05.


Of the 174 subjects in this study, the SCP genotype frequencies were 40% FF, 44% Ff, and 16% ff. These values are consistent with previously reported values for Caucasian women from northern California(9) and Boston.(10) A summary of published frequencies for the SCP genotypes is given in Table 1. When the relationship between the SCP genotype and the BsmI genotype in the French women was determined by contingency table analysis using Statview 4.5 (Abacus Concepts), the genotypes appeared to be independent of each other (chi-square p-value of 0.63; Table 2).

Table Table 1. Summary of the Prevalence of SCP Genotypes Published to Date
original image
Table Table 2. Joint Distribution of the SCP (FokI) andBsmI Genotypes
original image

The clinical characteristics of the study population are shown in Table 3. No differences were found between the three genotype groups. The study subjects were calcium replete with a mean intake of 808 ± 300 mg/day. Serum levels of calcium, PTH and 25(OH)D were not different between SCP genotypes (Table 4). The differences in the levels of phosphate between the genotypes approached statistical significance but they were considered not to be biologically significant (Table 4).

Table Table 3. Clinical Characteristics of the Women in This Study
original image
Table Table 4. Serum Biochemical Characteristics of the Women in This Study
original image

There was no association between SCP genotype and BMD at the lumbar spine, proximal femur, or BMC of the whole body (Table 5). The relatively high calcium intake of the subjects precluded a meaningful analysis of SCP genotype and BMD stratified by dietary calcium. Correction for body mass index, tobacco use, and oral contraceptive use did not change these results.

Table Table 5. Bone Mass and Turnover Measurements of the Women Studied
original image

There was a statistically significant association of SCP genotype and NTX values with NTX levels in ff women 33.5% higher than in FF women (FF 19.7 ± 8.1 vs. Ff 20.7 ± 9.8 vs. ff 26.3 ± 8.3 nmol/mmol creatinine, p = 0.004; Table 5). Although not statistically significant, the same trend was observed for CTX, with a 19% increased levels in ff women compared with FF women. No association was found between the SCP genotype and the markers of bone formation.


The distribution of the SCP genotypes found in this study is consistent with that found for other Caucasian women(9,10) and with our unpublished data (Eccleshall and Feldman) (Table 1). It is slightly different from that reported for Japanese women where the genotype prevalences were 31% FF, 55% Ff and 13% ff(11) and very different from that for African-Americans whose genotype frequencies were 65% FF, 31% Ff, and 4% ff.(10) Nonetheless, in all of these populations the distribution of the genotypes approximates the Hardy-Weinberg equilibrium.

We did not find an association of SCP genotype with BMD in the group of premenopausal French women studied here. This is in contrast to previous reports that demonstrated a 12% lower BMD at either the lumbar spine or the femoral neck for women with the ff genotype.(9–11) However, in the study by Harris et al.(10) where the prevalence of the ff genotype for African-Americans was only 4%, no association with BMD was detected for African-American women. Harris et al.(10) suggested that as much as a third of the inter-racial variation in BMD between African-Americans and Caucasians might be explained by the susceptibility of Caucasians with the ff genotype to have decreased BMD. It is probable, that ethnic or environmental factors modulate any potential SCP effect on BMD. Possibly, the high calcium diet in the French women decreases the impact of the VDR polymorphism on BMD. Indeed, Krall et al.(14) found that the rate of bone loss in genetically susceptible postmenopausal women was substantially reduced when their calcium intake was supplemented with 500 mg of calcium per day.

Variability in the association of SCP genotype with BMD in different populations is reminiscent of the initial reports concerning the BsmI VDR polymorphism.(4) The results of a number of studies and a recent meta-analysis(5) found a significant association with this polymorphism and BMD. We believe that the SCP association with BMD also will prove to be an important genetic factor contributing to the determination of BMD. We chose to study this group of women from the OFELY cohort because previously they have been shown not to have an association of BMD with the BsmI polymorphism.(12) Since Harris et al.(10) did not find a statistically significant association between the BsmI genotype and the SCP genotype, we chose to see if the SCP might better correlate with BMD. While this was not found, the independence of the SCP genotypes and the BsmI genotype was confirmed.

We found increased levels of bone resorption markers in ff women as compared with FF women suggesting that SCP polymorphism could be related to difference in bone resorption rate. This is consistent with the loss in BMD at the femoral neck observed by Gross et al.(9) in Mexican-American women. However, the clinical significance of these findings in this group of premenopausal women remains unclear because no differences in BMD were detected and this effect on the markers of bone resorption needs to be confirmed in larger studies and in longitudinal studies where rates of bone loss are examined.

This study was designed to have a 90% power to detect an 8.5% difference in BMD. This was based upon previous studies that demonstrated as much as a 12% difference in BMD between genotype groups.(9,10) While the effect of the SCP polymorphism on BMD could prove to be less than originally thought, as was found in the case of the BsmI polymorphism,(1,5) it is more probable that confounding environmental and regional factors, such as high calcium diet, may be masking the relationship between the SCP genotype and BMD. Further studies will be useful in evaluating this relationship to observe how it varies in different populations.


This study was supported by National Institutes of Health grants DK 50802 (D.F.) and DK 02459-02 (C.G.) and by a contract INSERM-MSD-Chibert (OFELY Study) (P.G. and P.D.D.).