The Vitamin D Receptor Start Codon Polymorphism (FokI) and Bone Mineral Density in Premenopausal American Black and White Women
This study examines the association between bone mineral density (BMD) and a start codon polymorphism (SCP) at the translation initiation site of the vitamin D receptor (VDR) gene. The thymine/cytosine (T/C) polymorphism in the first of two start (ATG) codons can be detected by a restriction fragment length polymorphism (RFLP) using the endonuclease FokI, which recognizes ATG as part of its restriction site. F indicates absence of the first ATG and a VDR that is shorter by three amino acids. The FokI genotype was determined in 154 premenopausal American women (72 black and 82 white) who were 20–40 years old. BMD of the total body, femoral neck, and lumbar spine were measured by dual-energy X-ray absorptiometry. The distribution of the SCP genotypes differed significantly by race (p < 0.001): 4% of blacks versus 18% of whites were ff homozygous and 65% of blacks versus 37% of whites were FF homozygous. There was no statistically significant interaction between race and SCP genotype in analyses of BMD at any skeletal site. In the group as a whole, the ff women had femoral neck BMD that was 7.4% lower than that of the FF women. The ff white women had total body BMD values that were 4.3% lower and femoral neck values that were 12.1% lower than FF white women. Total body and femoral neck BMD did not differ significantly by genotype in black women, and spine BMD did not differ by genotype in either race. Addition of the SCP genotype to analysis of covariance models comparing BMD of the black and white women reduced estimated differences in femoral neck BMD between the two groups by about 35%. In conclusion, the SCP polymorphism, detected with the endonuclease FokI, appears to influence peak bone density, particularly at the femoral neck. Racial differences in its distribution may explain some of the racial difference in femoral neck BMD.
The vitamin D receptor (VDR) gene has two potential translation initiation or ATG start sites, three codons apart(1) and a thymine/cytosine (T/C) polymorphism has been noted in the first ATG.(2) We will refer to this as the start codon polymorphism, or SCP. Individuals who have an ACG codon instead of the first ATG codon probably initiate from the second ATG and have a VDR that is potentially three amino acids shorter. This structural difference may affect the function of the VDR and thereby influence bone remodeling and bone mineral density (BMD). A restriction fragment length polymorphism (RFLP) using the FokI restriction endonuclease allows recognition of the SCP polymorphism such that the presence of a FokI site indicates the presence of the first ATG and is designated by f, whereas the absence of the FokI site is designated by F.
Two recent studies(3,4) have examined the SCP and have described its association with BMD. One study,(3) conducted in premenopausal Japanese women and presented as an abstract, reported that BMD at an unspecified skeletal site was 10.7% lower in 16 women homozygous for the presence of the first ATG (comparable to ff) compared with 30 women homozygous for the absence of the first ATG (comparable to FF). A second study,(4) conducted in postmenopausal Caucasian Mexican-American women, showed a 12.8% lower lumbar spine BMD in 15 ff homozygous women compared with 37 FF homozygous women and intermediate values in 48 heterozygous (Ff) women. This study also demonstrated an increased 2-year rate of bone loss from the femoral neck in the ff women compared with the FF women. The percentage of the Japanese and Mexican-American women who were ff homozygous (15%) was similar, but a higher percentage of the Mexican-American women were FF homozygous (37.0%) compared with the Japanese women (27.3%).
In the present study, we have determined the SCP genotype using the FokI RFLP in 154 premenopausal American women, 72 of whom were black and 82 were white. As we reported previously,(5,6) the black women who participated in this study have substantially higher bone density than the white women at all skeletal sites measured. Distributions of the two racial groups by the BsmI genotype(7) are similar,(6) and women in the BB genotype have lower femoral neck and spine BMD than women in the bb genotype.(6) The present study had the following objectives: (1) to determine if and to what extent the SCP polymorphism is associated with BMD in premenopausal women, (2) to compare the distribution of the SCP polymorphism in black and white women, (3) to determine whether any observed associations of the SCP polymorphism with BMD are dependent on race, (4) to determine whether any racial difference in the distribution of the SCP polymorphism explains racial differences in BMD, and (5) to describe the joint distribution of the FokI and BsmI alleles in black and white American women.
MATERIALS AND METHODS
One-hundred and fifty-four premenopausal women between the ages of 20 and 40 years were recruited from the Boston area for participation in two previous studies. The eligibility requirements for those studies have been reported in detail previously(5,8) and included good general health, black or white race, no current pregnancy or lactation, and weight under 113 kg. Written informed consent was obtained from each participant, and the study protocols were approved by the Tufts University Human Investigation Review Committee.
SCP genotyping with FokI:
Blood was drawn into Vacutainer CPT cell separation tubes (Becton-Dickinson, Franklin Lakes, NJ, U.S.A.) containing sodium citrate, and white blood cells were frozen at −70°C for later DNA extraction. As described previously,(6) genomic DNA was isolated by one of two methods: (1) a modified phenol:chloroform extraction procedure(9) or (2) with TurboGen genomic DNA isolation kits (Invitrogen, San Diego, CA, U.S.A.). The primers and polymerase chain reaction (PCR) conditions for amplifying exon 2 of the VDR gene were described previously.(4,10) Briefly, the isolated DNA was dissolved in a total volume of 25 μl TE (10 mM Tris, pH 7.8, 1 mM EDTA) and diluted 1:5 before use. The primers VDR 2a, 5′-AGCT GGCCCTGGCACTGACTCTGCTCT-3′, and VDR 2b, 5′-ATGGAAACACCTTGCTTCTTCTCCCTC-3′, in a buffer containing 1.5 mM MgCl2, 60 mM Tris HCl, pH 9.0, 15 mM NH4SO4, 10% dimethylsulfoxide (DMSO) and deoxyribonucleotide triphosphates (dNTPs) (200 μM each), were used to amplify exon 2 with Taq DNA polymerase (Stratagene, La Jolla, CA, U.S.A.). The PCR conditions were 94°C for 30 s, 60°C for 30 s, 72°C for 30 s for 35 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 Tris-EDTA (TAE) buffer and ethidium bromide. Individuals were scored as FF homozygotes, Ff heterozygotes, or ff homozygotes according to the digestion pattern.(4)
The BsmI genotypes were determined as previously described(6) by PCR amplication followed by digestion of the PCR product with the restriction enzyme BsmI (New England BioLabs).
BMD of the total body, femoral neck, and lumbar spine (L2–L4) were measured with a model DPX dual-energy X-ray absorptiometer (Lunar Radiation Corp., Madison, WI, U.S.A.). The precision estimates (percent coefficient of variation, %CV) for this scanner are 0.6% for the total body, 2.0% for the femoral neck, and 1.0% for the lumbar spine.(11) The sample size for femoral neck and lumbar spine measurements is smaller than that for the total body because these sites were measured 3 months after the initial study visit, and nine women did not attend the later visit.
Plasma 1,25-dihydroxyvitamin D (1,25(OH)2D) was measured by the competitive protein-binding method of Reinhardt et al.,(12) which has intra- and interassay %CVs of 4.9 and 7.7%, respectively. Serum intact parathyroid hormone (PTH) was measured with Allegro intact PTH radioimmunoassay kits from the Nichols Institute (San Juan Capistrano, CA, U.S.A.) that have intra- and interassay %CVs of 5.6 and 6.6%, respectively. Serum osteocalcin was measured with two-site immunoradiometric assay kits from the Nichols Institute, and the intra- and interassay %CVs were 4.0 and 7.3%, respectively.
Other subject determinations:
Standing height was measured with a wall-mounted stadiometer, and body weight was measured with a digital scale. Information about smoking history was recorded as part of a medical history interview. Dietary calcium intake over the past year was estimated by a self-administered food frequency questionnaire.(13)
Subject characteristics, bone density, and laboratory values across SCP genotypes were compared by analysis of variance (ANOVA) and analysis of covariance (ANCOVA). The chi-squared test was used to compare observed genotype frequencies with those expected under Hardy-Weinberg equilibrium.(14) Adjusted mean BMD and laboratory values (and the standard errors associated with them) were calculated and compared with the SAS General Linear Models procedure least squares means option.(15) Interactions of race and calcium intake with SCP genotype were investigated by adding interaction terms to the ANCOVA models for BMD at each skeletal site. The distribution of the genotypes by levels of other categorical variables were tested with Fisher's exact test. Analyses were conducted with SPSS,(16) SAS,(15) and StatXact (version 2.05, Cytel Software, Cambridge, MA, U.S.A.). p values less than 0.05 were considered to indicate statistical significance. Given the sample sizes of the two racial groups and an alpha level of 0.05, this study had 80% statistical power to detect BMD differences between homozygote groups in either race of about 4% at the total body and 9% at the femoral neck and spine if subjects were evenly distributed by genotype. If the distributions were uneven, the detectable differences would be greater (e.g., for a 3:2:1 distribution, about 5% at the total body and 10% at the femoral neck).
Mean age, height, weight, and calcium intake of the study population are shown in Table 1, and, except for weight in white women, the values did not differ significantly by SCP genotype. Allele distributions in both the black and white women were in Hardy-Weinberg equilibrium (given expected frequencies for the FF, Ff, and ff genotypes, respectively, of 66, 31, and 4% in the black women and 35, 48, and 17% in the white women). The distribution of the genotypes differed significantly by race (p < 0.001), with the highest percentage of black women having the FF genotype and the highest percentage of white women being heterozygous (Table 1). Only 4% of the black women had the ff genotype compared with 18% of the white women.
Table Table 1. Characteristics of the Study Population of 154 Women by SCP Genotype*
There was no statistically significant interaction between race and SCP genotype in ANCOVA analyses of BMD at any skeletal site. The p values for the race by SCP genotype interaction term in age- and weight-adjusted models were 0.406 at the total body, 0.187 at the femoral neck, and 0.691 at the spine. The results are presented both for the pooled sample (adjusted for race) and separately for black and white women. In the total group, the differences in BMD across SCP genotype, adjusted for race, weight, and age, were statistically significant only at the femoral neck, where the ff genotype had values that were 2.9 and 7.4% lower, respectively, than those of the Ff and FF groups (Table 2). Mean BMD increased in the same order at the total body and lumbar spine, but the differences at those sites were smaller and not statistically significant.
Table Table 2. Adjusted Bone Mineral Density (g/cm2) by SCP Genotype*
When the races were examined separately, BMD differences across SCP genotypes were not statistically significant in the black women (Table 2). White women in the ff genotype had total body BMD values that were 4.3% lower than those in the FF genotype, and their femoral neck values were 4.1 and 12.1% lower than those in the Ff and FF groups, respectively. Differences at the spine were not statistically significant.
There was no significant interaction between calcium intake and SCP genotype at any skeletal site, and there was no evidence that BMD differences across the genotypes were more pronounced in the small number of women (n = 25) who had dietary calcium intakes under 400 mg/day. Adjustments for calcium intake and smoking had almost no effect on the estimates of BMD across SCP genotypes. There were no statistically significant differences in plasma 1,25(OH)2D, serum PTH, or serum osteocalcin concentrations across SCP genotypes (Table 3).
Table Table 3. Laboratory Values by SCP Genotype*
Addition of the SCP genotype to ANCOVA models comparing weight- and age-adjusted BMD of the black and white women reduced the estimated difference in femoral neck BMD between the two groups from 0.069 ± 0.023 (SEM) to 0.045 ± 0.024, p = 0.015. The genotype was not statistically significant when added to similar models for the total body and lumbar spine.
The joint distribution of the SCP and BsmI genotypes is shown in Table 4. In the study population as a whole, and in the white women analyzed separately, the two distributions were not independent. In particular, a higher than expected number of FF women were Bb heterozygotes. While this pattern was fairly similar in the black women, the statistical test for independence of their SCP and BsmI distributions was not significant.
Table Table 4. Joint Distribution of the SCP (FokI) andBsmI Genotypes: The Number of Subjects Observed and, in Parentheses, the Number Expected Under Independence
The possibility that polymorphisms of the VDR gene are associated with differences in BMD, and, therefore, may predict osteoporosis risk, has received great attention since the publication by Morrison et al.(7) Subsequently, a number of studies supported this association while others found no association of VDR genotype with BMD.(17) These studies were all directed at polymorphisms in intron 8 (BsmI and ApaI) or exon 9 (TaqI) that would not change the amino acid sequence of the VDR. In contrast, the SCP polymorphism at the translation initiation site in exon 2 of the VDR gene, examined in two previous publications(3,4) and in this study, should result in a change in the VDR structure. Since the F allele lacks the first ATG, the mRNA translation would probably initiate from the second ATG downstream, thus making the F allele VDR three amino acids shorter than the f allele VDR. However, it should be noted that the N-terminal amino acids of the VDR alleles have never been determined from purified VDR. In a previous study, Gross et al. hypothesized that the ff genotype might be associated with low BMD because the f allele VDR functioned at decreased efficiency.(4) In preliminary studies, Miyamoto et al. showed that the longer SCP allele (f) may be less active in a transfected cell system.(3) The differential activity of F compared with f alleles remains to be proven, but the change in VDR amino acid sequence does provide a rational basis from which to consider how the VDR SCP polymorphism might affect BMD.
In the study population as a whole, ff homozygote women had femoral neck BMD that was 7.4% lower than FF homozygote women, suggesting that the SCP polymorphism may have a clinically significant influence on peak bone density at the femoral neck. However, the appropriateness of pooling the black and white women for this analysis is uncertain. While there was no statistical evidence that the association of the genotype with femoral neck BMD differed by race, the statistical power to detect an interaction of this kind was limited. In general, for a given analysis, the power to detect an interaction between two independent variables is much lower than that to detect the main effects of the same independent variables.(18) For this reason, and to allow comparison with the only other Caucasian group studied thus far, we presented results for the two races separately.
In white women alone, femoral neck BMD of the ff genotype was 12.1% lower than that of the FF genotype. If this difference were maintained or increased after menopause, women in the ff genotype would have a substantially greater risk of osteoporotic fracture. Gross and colleagues(4) did not observe an association between the genotype and cross-sectional femoral neck BMD in postmenopausal Mexican-American Caucasian women but did observe that ff women had higher rates of femoral neck bone loss than FF women. Although speculative, it is possible that the SCP genotype effect on femoral neck BMD is altered during menopause.
Gross et al.(4) also reported that spine BMD of the postmenopausal Mexican-American women in the ff group was 12.8% lower than that of women in the FF group, whereas we observed a smaller and not significant difference (4.1%) in younger white women. There have not yet been other studies of total body BMD and the SCP genotype to corroborate our finding that white women in the ff genotype had total body BMD that was 4.3% lower than that of the FF women. Taken together, these findings suggest that effects of the SCP genotype on BMD may vary depending on race, ethnicity, age, and menopausal status and may differ by skeletal site. Other environmental factors may also influence the effects of the genotype in ways that have not yet been demonstrated. The finding that white women in the ff group were heavier than women in the other two groups is interesting but should be interpreted with caution unless it is confirmed in other populations; a similar association was not observed in postmenopausal women.(4)
This study shows major differences in the distributions of the FokI alleles in black and white women. This is in contrast to an earlier report in the same women that showed similar distributions of the BsmI genotype.(6) Although distributions of the SCP and BsmI genotypes were not independent, the higher than expected number of women who were both Bb and FF does little to clarify the relationship of the two polymorphisms. We cannot be certain that these women, a convenience sample of volunteers from the Boston area, are representative of American black and white women in general, but the distribution of the FokI alleles in the white women studied (18% ff and 37% FF) was very similar to that reported for the older group of Caucasian Mexican-American women (15% ff and 37% FF).(4) To date, the distribution of the SCP genotype has not been reported in other black women. Racial differences in BMD have long been assumed to have at least some genetic basis, and the racial difference in distribution of the SCP genotype appears to explain about a third of the racial difference in femoral neck BMD.
Biochemical measurements did not differ significantly by genotype. Elevated levels of 1,25(OH)2D and PTH would have provided evidence of resistance to the action of 1,25(OH)2D and supported the concept that the SCP genotype difference in BMD results from impaired efficiency of the VDR in the ff genotype. However, lack of such evidence in this cross-sectional study does not negate the possibility that an impairment of this kind may have adversely affected bone accrual or maintenance at earlier ages.
In conclusion, the SCP polymorphism in the VDR gene appears to influence peak femoral neck BMD in premenopausal black and white women examined together and both total body and femoral neck BMD in white women examined alone. The distribution of the SCP genotype differs by race, with a smaller percentage of black women having the unfavorable ff genotype. This difference in distribution appears to explain part of the racial difference in femoral neck BMD. These results must be interpreted with caution until they are confirmed by additional studies in larger populations.
The authors thank Dr. James Fleet of the Mineral Bioavailability Laboratory at the Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University for extracting the DNA used in these analyses and for determining the BsmI genotypes. This study was supported by National Institutes of Health grant DK50802 to D. Feldman and by the Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University (Contract No. 53–3K06–5–10). The contents of this publication do not necessarily reflect the views or policies of the U.S. Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.