Abstract
Context: Vitamin D deficiency is associated with many adverse health outcomes, yet little is known about the genetic epidemiology of vitamin D or its metabolites.
Objective: Our objective was to examine the relationship among three vitamin D-related genes and levels of 25-hydroxyvitamin D [25(OH)D] and 1,25-dihydroxyvitamin D [1,25(OH)2D] in Hispanics (HAs) and African Americans (AAs).
Design and Setting: The cross-sectional Insulin Resistance Atherosclerosis Family Study recruited and examined subjects in: Los Angeles, California (AAs; 513 individuals from 42 families); San Luis Valley (SLV), Colorado (HAs; 513 individuals from 30 families); and San Antonio (SA), Texas (HAs; 504 individuals from 58 families).
Main Outcome Measures: Plasma levels of 25(OH)D and 1,25(OH)2D were measured.
Results: Levels of 25(OH)D were highest in SLV-HAs [18.3 ± 7.7 ng/ml (45.7 ± 19.2 nmol/liter)], lower in SA-HAs [14.6 ± 6.4 ng/ml (36.4 ± 16.0 nmol/liter)], and lowest in AAs [11.0 ± 5.4 ng/ml (27.5 ± 13.5 nmol/liter)]. Levels of 1,25(OH)2D were similar in AAs [43.5 ± 13.9 pg/ml (113.1 ± 36.1 pmol/liter)] and SLV-HAs [43.2 ± 13.3 pg/ml (112.3 ± 34.6 pmol/liter)], but higher in SA-HAs [48.6 ± 17.0 pg/ml (126.4 ± 44.2 pmol/liter)]. After adjusting for gender and age within the site, two single nucleotide polymorphisms (SNPs) in the vitamin D binding protein gene (DBP), rs4588 and rs7041, were associated with 25(OH)D, and one SNP in the DBP, rs4588, was associated with 1,25(OH)2D at all three study centers.
Conclusions: SNPs in the DBP are associated with levels of 25(OH)D and 1,25(OH)2D in HA and AA participants in the Insulin Resistance Atherosclerosis Family Study.
Examination of three vitamin D-related genes and 25- and 1,25-dihydroxyvitamin D (1,25[OH]2D) levels in Hispanic and African Americans reveals a single nucleotide polymorphism in the vitamin D binding protein gene (DBP), rs4588, to be associated with both metabolites in both ethnic groups. No polymorphisms in the vitamin D receptor (VDR) or the 1-a hydroxylase (CYP27B1) genes are significantly associated with 25- or 1,25-[OH]2D levels.
Vitamin D deficiency is associated with many adverse health outcomes, including several bone diseases, more than a dozen types of cancer, multiple autoimmune diseases, and type 2 diabetes. Two forms of vitamin D are commonly measured in blood. The first, 25-hydroxyvitamin D [25(OH)D], is the major circulating metabolite and has clinical relevance as a marker of vitamin D status. Hydroxylation of this metabolite by 1-α hydroxylase (CYP27B1) yields 1,25-dihydroxyvitamin D [1,25(OH)2D], the more biologically active metabolite. Many studies have examined the association between levels of 25(OH)D and 1,25(OH)2D and health outcomes. Other studies have examined the role of genes in the vitamin D metabolic pathway in health outcomes (1). In general, neither the former nor the latter approaches have yielded definite results. Few studies have examined the association between vitamin D-related genes and levels of 25(OH)D and 1,25(OH)2D. Ogunkolade et al. (2) examined four polymorphisms in the vitamin D receptor gene (VDR), which encodes the vitamin D receptor that binds to 1,25(OH)2D to form a regulatory factor that modulates the transcription of genes in many tissues in the body. No association was observed between any of the studied VDR polymorphisms and 25(OH)D or 1,25(OH)2D levels in 143 and 41, respectively, healthy Bangladeshi adults. However, Ramos-Lopez et al. (3) examined these four VDR polymorphisms and an additional seven VDR polymorphisms in 158 German type 1 diabetes patients. They found an association between polymorphism rs3847987 and levels of 25(OH)D, but no association between any of the VDR polymorphisms and 1,25(OH)2D. Kurylowicz et al. (4) examined three polymorphisms in the vitamin D binding protein gene (DBP) (also known as GC), which encodes the vitamin D binding protein that binds to vitamin D metabolites and transports them to target tissues. They found an association between codon 420 (Thr→Lys) and vitamin D deficiency [defined as serum concentrations of 25(OH)D <20 ng/ml (<49.9 nmol/liter)] in 110 Polish patients with Graves’ disease. Wjst et al. (5) examined 903 individuals from 201 families with at least two asthmatic children in Germany and Sweden, and estimated an 80% heritability for 25(OH)D and 30% heritability for 1,25(OH)2D.
A few studies of 25(OH)D and 1,25(OH)2D levels have been performed in Blacks and non-Hispanic whites (6,7,8,9). Zadshir et al. (10) examined 25(OH)D levels in non-Hispanic whites, Hispanics, and Blacks from the Third National Health and Nutrition Examination Survey (NHANES III), and observed that 25(OH)D levels were highest in non-Hispanic whites, intermediate in Hispanics, and lowest in Blacks. Levels of 1,25(OH)2D were not measured. Studying populations with varying degrees of skin pigmentation is important because the primary source of vitamin D is UV-B radiation from the sun, which produces provitamin D3, the initial vitamin D metabolite, in the skin. Melanin pigmentation that makes skin darker in color also acts as a natural sunscreen by absorbing UV-B radiation from the sun. However, this protective quality also means that less UV-B radiation is available for the production of provitamin D3. Although 25(OH)D levels decrease with darker skin color, levels of the more biologically active metabolite, 1,25(OH)2D, appear to be nearly equivalent in the two extremes of skin color (6,7,8,9), but the level of 1,25(OH)2D in individuals with intermediate skin coloring is unknown.
Little is known about the genetic epidemiology of vitamin D and its metabolites, compounds that appear to be vitally important to our health. Therefore, the goals of the current study were to: 1) measure the levels of 25(OH)D and 1,25(OH)2D in Hispanics and Blacks; 2) estimate the heritability of levels of 25(OH)D and 1,25(OH)2D; and 3) examine the relationship among three vitamin D-related genes, DBP, VDR, and CYP27B1, and levels of 25(OH)D and 1,25(OH)2D.
Subjects and Methods
The Insulin Resistance Atherosclerosis (IRAS) Family Study
The IRAS Family Study is a multicenter study designed to identify the genetic basis of insulin resistance and adiposity in a multiethnic population. Details of the study design and methods have been published (11). Probands for the IRAS Family Study were generally recruited from the original IRAS cohort (12). Additional probands were ascertained from the general population. Criteria for proband selection were based on large family size, not on disease status or extreme phenotypes. Participants in the IRAS Family Study were recruited by clinical centers in Los Angeles, CA (African Americans), San Luis Valley, CO (Hispanics), and San Antonio, TX (Hispanics). Race/ethnicity was determined by self-report. As part of the clinical examination, plasma from the 2-min sample of the frequently sampled iv glucose tolerance test after at least an 8-h fast was used for the measurement of vitamin D levels. Informed consent was obtained from all subjects.
Measurement of vitamin D levels
Levels of 25(OH)D were measured by a two-step process involving rapid extraction of 25(OH)D and other hydroxylated metabolites from plasma and RIA with a 25(OH)D-specific antibody (DiaSorin Inc., Stillwater, MN) with interassay coefficients of variation less than 8%. Levels of 1,25(OH)2D were measured by a two-step process involving extraction and purification of vitamin D metabolites from plasma and RIA with a 1,25(OH)2D-specific antibody (DiaSorin) with interassay coefficients of variation less than 19%.
DNA collection, single nucleotide polymorphism (SNP) selection, and genotyping
DNA was extracted from whole blood samples using the PUREGENE DNA Isolation Kit (Gentra, Minneapolis, MN). DNA was quantitated using standardized fluorometric readings on a Hoefer DyNa Quant 200 Fluorometer (Hoefer, Inc., San Francisco, CA).
The Ensembl, dbSNP, and HapMap databases were searched to identify known SNPs within DBP, VDR, and CYP27B1. There were 32 SNPs selected based upon: 1) variants reported to be associated with adverse health outcomes or SNPs very close to these variants, 2) coding SNPs, 3) tagging SNPs from HapMap, and 4) additional SNPs with a minor allele frequency more than or equal to 0.1 to provide uniform coverage of the genes.
Amplified and purified DNA samples were analyzed using the MassARRAY genotyping system (Sequenom, San Diego, CA) (13). Primer sequences are available on request. Duplicate quality control samples from 70 individuals were placed randomly throughout the plates. Concordance between the blind duplicate samples and the identified samples ranged from 95% for one of the SNPs to 100% for 23 of the SNPs.
Pedigree and genotype error detection
All pedigrees had previously been checked for pedigree errors using Pedigree RElationship Statistical Test (PREST) software (available at http://fisher.outstat.toronto.edu/sun/Software/Prest) (14). Each SNP was examined for Mendelian inconsistencies using PEDCHECK software (15). Any genotypes that were inconsistent with Mendelian inheritance were set to missing. In addition, each SNP was tested for departure from Hardy-Weinberg equilibrium (HWE). Six of the 32 SNPs had a very low success rate or were not in HWE; therefore, four replacement SNPs in approximately the same chromosomal location were genotyped to minimize gaps in coverage [average distance between SNPs = 4.5 kilobases (kb); maximum = 6.8 kb]. None of the SNPs reported in this paper deviated significantly from HWE assumptions.
Measurement of covariates
As previously reported, physical activity was assessed by a 1-yr recall using a modification of a validated instrument (16) that incorporated activities common among IRAS Family Study participants, including ranching and homemaking activities. These activities were queried in groups according to home, work, or leisure time, and according to intensity of activities (light, moderate, or vigorous) based on metabolic equivalent (MET) values. For each activity group, usual frequency and duration of participation were recorded, from which estimated energy expenditure (EEE) was determined. Total energy expended (in kcal/kg) per year was calculated by summing across all activity groups, plus the EEE from sleep (MET value of 1.0), plus the EEE from light activities (e.g. sitting MET value of 1.5). Audiotapes of the physical activity interviews were monitored centrally on a quarterly basis throughout the data collection period. Gender-specific quartiles of physical activity were calculated.
Data on solar UV-B radiation, which is the primary source of vitamin D, were obtained from the National Solar Radiation Data Base produced by the National Renewable Energy Laboratory under the U.S. Department of Energy Office of Energy Efficiency and Renewable Energy. The AVerage daily total GLObal solar radiation, defined as the total amount of direct and diffuse solar radiation in Wh/m2 received on a horizontal surface, for the month before the blood draw was determined for each individual based on the county of the study center.
A brief questionnaire, which asked about sun exposure and regular multivitamin supplement use, was completed by participants in San Luis Valley. Sun exposure questions included hours spent outside between 1100 and 1500 h in warm and cool weather months, and the frequency of protection from sun exposure by wearing a hat with a brim and clothing covering most of the body or sunscreen in warm and cool weather months (Always = 3, Frequently = 2, Seldom = 1, and Never = 0).
Data transformation
Distributions of 25(OH)D and 1,25(OH)2D levels were positively skewed within each recruitment center. Therefore, a square root transformation was used to approximate better a normal distribution.
Association analysis
A variance component model was used that enabled us to account for the correlations among family members in pedigrees of arbitrary size and complexity. Statistical details of the application of this approach are described by Kammerer et al. (17). For the environmental association analysis, covariates reported to be associated with 25(OH)D and 1,25(OH)2D levels were included in the initial models [gender, age, solar radiation in the month before the blood draw, body mass index (BMI), physical activity quartile, and 25(OH)D (for the 1,25(OH)2D model only)]. In addition, hours spent outside and frequency of protection from sun exposure during the season of blood draw and regular multivitamin use were included in the initial model for participants from San Luis Valley, in which the vitamin D exposure questionnaire was administered. For the genetic association analysis, each SNP was tested for association under an additive model adjusting for gender and age. The significance of each independent variable was assessed by the likelihood ratio test. Due to the differences in allele frequency and linkage disequilibrium structure among the centers, all analyses were stratified by center. Analyses were conducted in the Sequential Oligogenic Linkage Analysis Routines (SOLAR) software (available at http://www.sfbr.org/solar/index.htm) (18).
Heritability analyses
The residual heritability was estimated, using the SOLAR software (18), for each phenotype within each clinical center, reflecting the residual proportion of variance of the phenotype due to familial (genetic) effects after the effects of significant environmental covariates were removed.
Results
Descriptive statistics
Descriptive statistics for the family members with genotype and phenotype data are shown in Table 1. There were no significant differences in percentage of female participants among the three centers. There was a significant difference in solar radiation in the month before the blood draw among the three centers, with Hispanics from San Luis Valley [latitude 37.27° north, altitude 2298 m (7539 feet)] having more solar radiation than African Americans from Los Angeles [latitude 33.56° north, altitude 30 m (98 feet)] or Hispanics from San Antonio [latitude 29.32° north, altitude 242 m (794 feet)]. There were small, but significant differences in age and BMI among the three centers. Hispanics from San Luis Valley were the most active followed by Hispanics from San Antonio and African Americans from Los Angeles. Levels of 25(OH)D were highest in Hispanics from San Luis Valley, lower in Hispanics from San Antonio, and lowest in African Americans from Los Angeles. Levels of 1,25(OH)2D were similar in African Americans from Los Angeles and Hispanics from San Luis Valley, but higher in Hispanics from San Antonio. As expected, the levels of 25(OH)D and 1,25(OH)2D were significantly correlated in all three centers (r = 0.43, 0.33, and 0.38 in African Americans, Hispanics from San Antonio, and Hispanics from San Luis Valley, respectively).
Table 1.
Unadjusted phenotypical characteristics of participants by study center
| Variable | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans | P values |
|---|---|---|---|---|
| No. | 504 | 513 | 513 | NA |
| No. of families | 58 | 30 | 42 | NA |
| Mean family size (range) | 8.7 (2–26) | 17.1 (4–30) | 12.2 (2–49) | NA |
| Female (%) | 59 | 56 | 58 | 0.733 |
| Age (yr)a | 39.1 ± 13.9 (38.0) | 40.2 ± 13.0 (39.0) | 41.0 ± 13.7 (40.0) | 0.002 |
| Solar radiation in month prior to blood drawa | 4,123 ± 1,445 (3,855) | 5,139 ± 1,772 (5,182) | 4,344 ± 1,512 (3,739) | <0.001 |
| BMI (kg/m2)a | 29.8 ± 6.3 (29.2) | 27.2 ± 5.3 (26.4) | 29.8 ± 6.9 (28.8) | <0.001 |
| Total energy expenditure (kcal/kg·yr)a | 16,367 ± 3,792 (15,261) | 16,903 ± 4,398 (15,558) | 15,289 ± 3,515 (14,073) | <0.001 |
| 25(OH)D (ng/mlb)a | 14.6 ± 6.4 (14.0) | 18.3 ± 7.7 (18.0) | 11.0 ± 5.4 (9.8) | <0.001 |
| 1,25(OH)2D (pg/mlc)a | 48.6 ± 17.0 (46.0) | 43.2 ± 13.3 (42.0) | 43.5 ± 13.9 (42.0) | <0.001 |
| Hours spent outside between 1100 and 1500 ha | ||||
| Warm weather months | NA | 2.4 ± 1.4 (2.0) | NA | NA |
| Cool weather months | 1.7 ± 1.3 (1.0) | |||
| Frequency of sun exposure protection | ||||
| Warm weather months | ||||
| Always | NA | 31% | NA | NA |
| Frequently | 22% | |||
| Seldom | 24% | |||
| Never | 21% | |||
| Cool weather months | ||||
| Always | NA | 37% | NA | NA |
| Frequently | 20% | |||
| Seldom | 16% | |||
| Never | 21% | |||
| Regular multivitamin use | NA | 31% | NA | NA |
NA, Not applicable.
Data are means ± sd (median).
Multiply by 2.496 to convert to the International System of Units.
Multiply by 2.6 to convert to the International System of Units.
Environmental association
In the San Luis Valley Hispanics, self-reported frequency of protection from sun exposure during the season of blood draw was not significantly associated with levels of 25(OH)D (P = 0.701). Self-reported hours spent outside during the season of blood draw were significantly associated with 25(OH)D in a univariate model (P < 0.001), but not after adjusting for solar radiation in the month before the blood draw and gender (P = 0.149). All other variables included in the initial model were significantly associated with 25(OH)D and are shown in Table 2.
Table 2.
Coefficients of the effects ± se (P values) of environmental determinants on 25(OH)D levels by study center
| Variable | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans |
|---|---|---|---|
| Female gender | −0.297 ± 0.067 (<0.001) | −0.324 ± 0.073 (<0.001) | −0.217 ± 0.068 (0.001) |
| Age | 0.002 ± 0.002 (0.465) | −0.001 ± 0.003 (0.727) | 0.006 ± 0.003 (0.014) |
| Solar radiation in month before blood draw | 0.0002 ± 0.00002 (<0.001) | 0.0002 ± 0.00002 (<0.001) | 0.0001 ± 0.00003 (<0.001) |
| BMI | −0.032 ± 0.005 (<0.001) | −0.020 ± 0.007 (0.004) | −0.017 ± 0.005 (0.001) |
| Physical activity | |||
| 4th quartile (most active) | Reference | Reference | Reference |
| 3rd quartile | −0.015 ± 0.086 (0.866) | −0.083 ± 0.094 (0.381) | −0.079 ± 0.107 (0.462) |
| 2nd quartile | 0.058 ± 0.089 (0.520) | −0.069 ± 0.097 (0.482) | 0.021 ± 0.098 (0.835) |
| 1st quartile (least active) | −0.256 ± 0.097 (0.009) | −0.205 ± 0.106 (0.054) | −0.121 ± 0.099 (0.224) |
| Regular multivitamin use | |||
| Yes | NA | Reference | NA |
| No | −0.351 ± 0.078 (<0.001) | ||
| R2a | 0.233 | 0.243 | 0.139 |
NA, Not applicable.
Percentage of variance in 25(OH)D explained by all the covariates.
After adjusting for 25(OH)D, the precursor vitamin D metabolite, only age (San Antonio Hispanics only) and BMI (Los Angeles African Americans only) were significantly associated with 1,25(OH)2D (Table 3).
Table 3.
Coefficients of the effects ± se (P values) of environmental determinants on 1,25(OH)2D levels by study center
| Variable | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans |
|---|---|---|---|
| Female gender | −0.041 ± 0.099 (<0.677) | −0.019 ± 0.083 (0.818) | 0.007 ± 0.082 (0.929) |
| Age | −0.008 ± 0.004 (0.020) | −0.004 ± 0.003 (0.170) | 0.00002 ± 0.003 (0.996) |
| BMI | −0.011 ± 0.008 (0.178) | −0.011 ± 0.008 (0.158) | −0.017 ± 0.006 (0.006) |
| 25(OH)D | 0.420 ± 0.062 (<0.001) | 0.421 ± 0.049 (<0.001) | 0.472 ± 0.054 (<0.001) |
| r2a | 0.129 | 0.148 | 0.180 |
Percentage of variance in 1,25(OH)2D explained by all the covariates.
Heritability
For 25(OH)D, the heritability analysis for each center included adjustment for gender, age, solar radiation, BMI, and physical activity quartile (final model from Table 2). In addition, in San Luis Valley Hispanics, the heritability analysis included adjustment for regular multivitamin use. For 1,25(OH)2D, the heritability analysis for each center included adjustment for gender, age, BMI, and 25(OH)D (final model from Table 3). The residual heritability of 25(OH)D ranged from 0.227–0.413, and 1,25(OH)2D ranged from 0.162–0.484, indicating a moderate contribution of genetic variation to variation in both 25(OH)D and 1,25(OH)2D (Table 4).
Table 4.
Residual heritability estimates ± se (P values) of plasma vitamin D metabolites by study center
| Variable | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans |
|---|---|---|---|
| 25(OH)Da | 0.227 ± 0.105 (0.005) | 0.413 ± 0.100 (<0.001) | 0.283 ± 0.097 (<0.001) |
| 1,25(OH)2Db | 0.196 ± 0.089 (0.004) | 0.162 ± 0.083 (0.006) | 0.484 ± 0.099 (<0.001) |
Adjusted for gender, age, solar radiation, BMI, physical activity quartile, and regular multivitamin use (San Luis Valley Hispanics only).
Adjusted for gender, age, BMI, and 25(OH)D.
Genetic association
Minor allele frequencies of the 30 SNPs based on all founders in each center are shown in Table 5. After adjusting for gender and age, two SNPs in the DBP [linkage disequilibrium (r2) = 0.31, 0.26, and 0.03 in San Antonio Hispanics, San Luis Valley Hispanics, and Los Angeles African Americans, respectively; P < 0.001 for all] and one SNP in the VDR were associated with 25(OH)D, and one SNP in the DBP was associated with 1,25(OH)2D at a significance level of P < 0.01 (Table 6). No significant (P < 0.01) interactions between DBP SNPs rs4588 and rs7041 and the other SNPs were observed (data not shown). No significant (P < 0.01) interactions between vitamin D status and the SNPs on 1,25(OH)2D were found (data not shown).
Table 5.
SNP minor allele frequencies based on all founders in each center
| Gene and SNP (also known as) | Minor Allele Frequency
|
|||||
|---|---|---|---|---|---|---|
| Region (function) | Alleles | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans | P valuesa | |
| DBP (4q12-q13) | ||||||
| rs705117 | Intron 13 | A/Gb | 0.325 | 0.306 | 0.816 | <0.001 |
| rs1491709 | Intron 11 | C/Tb | 0.035 | 0.140 | 0.079 | 0.663 |
| rs4588 (Thr420Lys) | Exon 11 (nonsyn) | Ab/C | 0.246 | 0.172 | 0.056 | 0.031 |
| rs7041 (Asp416Glu) | Exon 11 (nonsyn) | Gb/T | 0.412 | 0.533 | 0.105 | <0.001 |
| rs4752 | Exon 8 (syn) | Cb/T | 0.035 | 0.022 | 0.25 | <0.001 |
| rs222003 | Intron 6 | Cb/G | 0.107 | 0.081 | 0.556 | <0.001 |
| rs222020 | Intron 1 | Cb/T | 0.250 | 0.247 | 0.500 | 0.016 |
| rs1352843 | Intron 1 | Cb/T | 0.052 | 0.108 | 0.029 | 0.093 |
| rs222029 | Intron 1 | A/Gb | 0.197 | 0.231 | 0.421 | 0.315 |
| rs2298849 | Intron 1 | Cb/T | 0.233 | 0.234 | 0.316 | 0.454 |
| rs10011000 | 5′ UTR | C/Gb | 0.197 | 0.214 | 0.20 | 0.930 |
| VDR (12q13.11) | ||||||
| rs9729 | Exon 10 (3′ UTR) | Ab/C | 0.386 | 0.446 | 0.647 | 0.007 |
| rs1544410 (BsmI) | Intron 9 | Ab/G | 0.223 | 0.308 | 0.361 | 0.149 |
| rs2239185 | Intron 7 | C/Tb | 0.364 | 0.446 | 0.605 | 0.374 |
| rs7305032 | Intron 6 | Ab/G | 0.418 | 0.494 | 0.675 | 0.020 |
| rs987849 | Intron 4 | C/Tb | 0.371 | 0.488 | 0.695 | 0.001 |
| rs2239179 | Intron 4 | A/Gb | 0.276 | 0.362 | 0.417 | 0.841 |
| rs886441 | Intron 3 | Cb/T | 0.308 | 0.378 | 0.472 | 0.076 |
| rs3782905 | Intron 3 | C/Gb | 0.212 | 0.277 | 0.211 | 0.714 |
| rs10735810 (FokI) | Exon 3 (nonsyn) | Ab/G | 0.358 | 0.462 | 0.211 | 0.003 |
| rs2853564 | Intron 0 | Cb/T | 0.225 | 0.305 | 0.184 | 0.128 |
| rs4760648 | Intron 0 | Cb/T | 0.348 | 0.488 | 0.382 | 0.029 |
| rs4334089 | Intron 0 | Ab/G | 0.233 | 0.140 | 0.684 | <0.001 |
| rs7302235 | Intron 0 | Cb/T | 0.242 | 0.125 | 0.605 | <0.001 |
| rs10783219 | Intron 0 | A/Tb | 0.434 | 0.533 | 0.100 | 0.013 |
| CYP27B1 (12q13.1-q13.3) | ||||||
| rs1048691 | 3′ UTR | C/Tb | 0.317 | 0.327 | 0.333 | 0.909 |
| rs4646536 (+2838) | Intron 6 | Cb/T | 0.317 | 0.368 | 0.316 | 0.673 |
| rs8176344 (Lys166Val) | Exon 3 (nonsyn) | Cb/G | 0.017 | 0.006 | 0.0 | 0.535 |
| rs10877012 | 5′ UTR | G/Tb | 0.295 | 0.35 | 0.175 | 0.460 |
| rs703842 | 5′ UTR | Cb/T | 0.375 | 0.426 | 0.464 | 0.714 |
SNPs are listed in order of chromosomal position (smallest to largest); all genes were reverse transcripts so 5′ end has the largest chromosomal location. nonsyn, Nonsynonymous; syn, synonymous; UTR, untranslated region.
P values are for differences among the sites when tested using a χ2 statistic.
Minor allele.
Table 6.
Coefficients of the effects ± se (P values) of selected SNPs on vitamin D metabolites by study center
| Vitamin D metabolite | Gene: SNP | San Antonio Hispanics | San Luis Valley Hispanics | Los Angeles African Americans |
|---|---|---|---|---|
| 25(OH)D | DBP: rs4588 [C allele (Thr)] | 0.199 ± 0.068 (0.004) | 0.285 ± 0.075 (<0.001) | 0.234 ± 0.087 (0.007) |
| DBP: rs7041 [T allele (Asp)] | −0.181 ± 0.061 (0.003) | −0.220 ± 0.063 (<0.001) | −0.159 ± 0.071 (0.025) | |
| VDR: rs10783219 (T allele) | −0.164 ± 0.056 (0.004) | 0.043 ± 0.063 (0.498) | 0.111 ± 0.085 (0.194) | |
| 1,25(OH)2D | DBP: rs4588 [C allele (Thr)] | 0.255 ± 0.093 (0.007) | 0.182 ± 0.085 (0.032) | 0.203 ± 0.116 (0.080) |
Additive genetic model adjusting for gender and age.
Discussion
Levels of 25(OH)D in the IRAS Family Study were highest in Hispanics from San Luis Valley, lower in Hispanics from San Antonio, and lowest in African Americans from Los Angeles. The higher levels of 25(OH)D in the San Luis Valley Hispanics compared with those in the San Antonio Hispanics are not surprising considering that admixture analysis indicates that the San Luis Valley Hispanics have more European ancestry (62.7%) and less African ancestry (3.2%) than many other Hispanic groups (19). Using NHANES III data, Zadshir et al. (10) found levels of 25(OH)D to be highest in non-Hispanic whites [33.2 ng/ml (82.9 nmol/liter) in men and 30.4 ng/ml (75.9 nmol/liter) in women], lower in Hispanics [27.4 ng/ml (68.4 nmol/liter) in men and 22.7 ng/ml (56.7 nmol/liter) in women], and lowest in Blacks [20.9 ng/ml (52.2 nmol/liter) in men and 18.2 ng/ml (45.4 nmol/liter) in women]. Overall, levels of 25(OH)D were slightly higher in NHANES III participants, where northern samples were collected during the warmer months with plentiful sunshine to produce vitamin D, than in IRAS Family Study participants, where approximately half the sample collection occurred in the summer/fall and half in the winter/spring (data not shown).
Previous research has shown that despite the correlation between skin color and levels of 25(OH)D, where higher levels of 25(OH)D are seen in those with lighter skin color, levels of the subsequent metabolite, 1,25(OH)2D, are similar or even lower in non-Hispanic whites compared with Blacks (5,6,7,8). In our study, levels of 1,25(OH)2D were similar in Los Angeles AfricanAmericans and San Luis Valley Hispanics, but higher in San Antonio Hispanics. This difference between the centers persisted even after adjustment for gender, age, BMI, and 25(OH)D levels, and may be due to random variation because there is no known biological explanation.
Individuals in the least active quartile had lower 25(OH)D levels than those in the most active quartile (Table 2). This relationship has been recently reported (20), however, it was unknown whether the measure of physical activity was a surrogate for time spent outdoors or other healthy behaviors, such as multivitamin supplement use. In the current study, physical activity maintained the same level of statistical significance after adjustment for self-reported hours spent outside during the season of blood draw and multivitamin supplement use (P = 0.069; data only available in San Luis Valley). However, both hours spent outside and physical activity were self-reported, so we cannot exclude the possibility that the relationship between physical activity and 25(OH)D levels is due to physical activity that took place outside. Further research with prospective and objective measures of physical activity and UV-B radiation exposure is necessary to elucidate this relationship.
The primary source of vitamin D is UV-B radiation from the sun, which is affected by the amount of time spent outdoors. In San Luis Valley Hispanics, self-reported hours spent outside during the season of blood draw were significantly associated with 25(OH)D in a univariate model (P < 0.001), but not after adjusting for solar radiation in the month before the blood draw and gender (P = 0.149). This can be explained by the correlation between time spent outside and solar radiation and gender in the San Luis Valley. Hours spent outside were positively correlated with solar radiation (P < 0.001), indicating that participants reported spending more time outdoors in times of higher solar radiation (warmer months). Women reported fewer hours spent outside than men (P < 0.001). This finding may be specific to the San Luis Valley where cold winters make spending time outside unappealing and where farming, an occupation dominated by men that requires spending time outside, is prevalent.
After adjustment for gender, age, and 25(OH)D, BMI was independently and negatively associated with levels of 1,25(OH)2D (Table 3). This association was only significant in Los Angeles African Americans, but a similar trend was seen in San Antonio and San Luis Valley Hispanics. This finding lends support to results reported by Parikh et al. (21), in which 1,25(OH)2D was associated with BMI in both African Americans and Caucasians, but the relationship was substantially stronger in the African Americans. The negative association between BMI and 25(OH)D is well established, but that between BMI and 1,25(OH)2D is still quite controversial, and the mechanism of action is unknown. However, these results, which were adjusted for levels of 25(OH)D, indicate that the decrease in 1,25(OH)2D with an increase in BMI is not strictly due to decreased levels of the precursor metabolite, 25(OH)D, as has been previously speculated.
The residual heritability of 25(OH)D in our study is much lower than that reported by Wjst et al. (5) (0.803). The high heritability estimate in their study may be because they adjusted for age and sex, but not for solar radiation, BMI, or physical activity, environmental factors that are likely to be shared in these pediatric siblings with clinical asthma; because they are children, it is likely that their parents brought them in for their study visit and blood draw on the same day (similar solar radiation exposure). The residual heritability of 1,25(OH)2D in our study is similar to that reported by Wjst et al. (5) (0.300). This similarity, vs. the disparity in results for 25(OH)D, may be because 1,25(OH)2D is more biologically regulated and less affected by environmental factors.
One SNP in exon 11 of the DBP, rs4588, that results in an amino acid change in codon 420 Thr→Lys (ACG→AAG, resulting in a protein change from Gc-1 to Gc-2), was associated with both 25(OH)D and 1,25(OH)2D in all three centers. The A base pair leading to the lysine (Lys) amino acid and the Gc-2 protein was associated with lower levels of both 25(OH)D and 1,25(OH)2D (Table 6). This supports a recent finding by Kurylowicz et al. (4) that the A/Lys allele is associated with lower 25(OH)D [1,25(OH)2D was not measured]. In addition, Lauridsen et al. (22) found levels of Gc protein, 25(OH)D, and 1,25(OH)2D to be lowest in individuals with the Gc2–2 phenotype (corresponding to the A/A genotype at rs4588), intermediate in Gc1–2 individuals (A/C genotype), and highest in Gc1–1 individuals (C/C genotype). Biological support for these findings is provided by Arnaud and Constans (23), in which the affinity for both 25(OH)D and 1,25(OH)2D is higher for Gc-1 than Gc-2.
Another SNP in exon 11, rs7041, that results in an amino acid change in codon 416 Asp→Glu [GAT→GAG (Gc-1s protein)] was associated with 25(OH)D, but not 1,25(OH)2D. However, neither Kurylowicz et al. (4) nor Lauridsen et al. (22) detected this association. This is probably because the T allele (Asp) is found in both the phenotype with the highest affinity for the two vitamin D metabolites (Gc-1f) and the phenotype with the lowest affinity (Gc-2), whereas the G allele (Glu) is found in the phenotype with intermediate affinity (Gc-1s) (23). This splitting of the T-allele effect between two opposite phenotypes could result in a diluted association or no association at all.
There are a few limitations to this study. First, solar radiation data were obtained for the county of the study center, however, some participants did not live near the center in which the study visit took place, and we do not have access to their county of residence. Therefore, some nondifferential misclassification is likely, although this measure of vitamin D exposure from the sun is more accurate than season of blood draw, which has been used by most previous studies. Second, ethnicity was self-reported, not determined by skin reflectance or ancestry informative markers (AIMs). This likely resulted in a slightly more heterogeneous population at each center with respect to ethnicity. Third, due to financial and logistical constraints, the vitamin D exposure questionnaire was only administered in San Luis Valley and was relatively brief. Therefore, the effect of self-reported hours spent outside, frequency of protection from sun exposure, and regular multivitamin supplement use on levels of 25(OH)D and 1,25(OH)2D in San Antonio Hispanics and Los Angeles African Americans is unknown. Finally, due to financial constraints, the coverage within the three vitamin D genes was not very dense by current standards (approximately one SNP every 4.5 kb). However, this is the densest coverage of the three genes with respect to vitamin D levels thus far, and the first study to examine and report all three vitamin D-related genes together. Moreover, the genetic findings were replicated in three distinct populations. These characteristics make the findings an important advancement of prior knowledge.
Because vitamin D status is being linked to more and more adverse health outcomes, understanding the genetic variants that are responsible for variation in relevant vitamin D metabolites in different populations is vital. Most of the research that has been done thus far has been on a smaller scale and/or in patients with a particular disease. This study examined environmental and genetic risk factors of two vitamin D metabolites, 25(OH)D and 1,25(OH)2D, in a large multicenter study of healthy Hispanics and African Americans. One SNP in the DBP, rs4588, was associated with 25(OH)D and 1,25(OH)2D in both Hispanics and African Americans, comparable to results in Caucasian samples (4,22). Genetic variation at or around this SNP may well play an important role in adverse health outcomes such as bone disease, cancer, autoimmune disease, and type 2 diabetes in Caucasians, Hispanics, and African Americans.
Footnotes
This research was supported by American Diabetes Association Grant 7-04-RA-83, National Institutes of Health Grants HL060894, HL060919, HL060931, HL060944, HL061019, and HL061210, and the General Clinical Research Centers Program, National Center for Research Resources Grant M01RR00069.
Disclosure Statement: The authors have nothing to disclose.
First Published Online July 1, 2008
Abbreviations: BMI, Body mass index; CYP27B1, 1-α hydroxylase; 1,25(OH)2D, 1,25- dihydroxyvitamin D; 25(OH)D, 25-hydroxyvitamin D; DBP, vitamin D binding protein gene; EEE, estimated energy expenditure; HWE, Hardy-Weinberg equilibrium; IRAS, Insulin Resistance Atherosclerosis; kb, kilobase; Lys, lysine; MET, metabolic equivalent; NHANES III, Third National Health and Nutrition Examination Survey; SNP, single nucleotide polymorphism; VDR, vitamin D receptor gene.
References
- Uitterlinden AG, Fang Y, van Meurs JB, van Leeuwen H, Pols HA 2004 Vitamin D receptor gene polymorphisms in relation to Vitamin D related disease states. J Steroid Biochem Mol Biol 89–90:187–193 [DOI] [PubMed] [Google Scholar]
- Ogunkolade BW, Boucher BJ, Prahl JM, Bustin SA, Burrin JM, Noonan K, North BV, Mannan N, McDermott MF, DeLuca HF, Hitman GA 2002 Vitamin D receptor (VDR) mRNA and VDR protein levels in relation to vitamin D status, insulin secretory capacity, and VDR genotype in Bangladeshi Asians. Diabetes 51:2294–2300 [DOI] [PubMed] [Google Scholar]
- Ramos-Lopez E, Jansen T, Ivaskevicius V, Kahles H, Klepzig C, Oldenburg J, Badenhoop K 2006 Protection from type 1 diabetes by vitamin D receptor haplotypes. Ann NY Acad Sci 1079:327–334 [DOI] [PubMed] [Google Scholar]
- Kurylowicz A, Ramos-Lopez E, Bednarczuk T, Badenhoop K 2006 Vitamin D-binding protein (DBP) gene polymorphism is associated with Graves’ disease and the vitamin D status in a Polish population study. Exp Clin Endocrinol Diabetes 114:329–335 [DOI] [PubMed] [Google Scholar]
- Wjst M, Altmuller J, Braig C, Bahnweg M, Andre E 2007 A genome-wide linkage scan for 25-OH-D(3) and 1,25-(OH)2-D3 serum levels in asthma families. J Steroid Biochem Mol Biol 103:799–802 [DOI] [PubMed] [Google Scholar]
- Daniels ED, Pettifor JM, Schnitzler CM, Moodley GP, Zachen D 1997 Differences in mineral homeostasis, volumetric bone mass and femoral neck axis length in black and white South African women. Osteoporos Int 7:105–112 [DOI] [PubMed] [Google Scholar]
- Parisien M, Cosman F, Morgan D, Schnitzer M, Liang X, Nieves J, Forese L, Luckey M, Meier D, Shen V, Lindsay R, Dempster DW 1997 Histomorphometric assessment of bone mass, structure, and remodeling: a comparison between healthy black and white premenopausal women. J Bone Miner Res 12:948–957 [DOI] [PubMed] [Google Scholar]
- Harris SS, Dawson-Hughes B 1998 Seasonal changes in plasma 25-hydroxyvitamin D concentrations of young American black and white women. Am J Clin Nutr 67:1232–1236 [DOI] [PubMed] [Google Scholar]
- Bikle DD, Ettinger B, Sidney S, Tekawa IS, Tolan K 1999 Differences in calcium metabolism between black and white men and women. Miner Electrolyte Metab 25:178–184 [DOI] [PubMed] [Google Scholar]
- Zadshir A, Tareen N, Pan D, Norris K, Martins D 2005 The prevalence of hypovitaminosis D among US adults: data from the NHANES III. Ethn Dis 15(Suppl 5):S5–S97–S101 [PubMed] [Google Scholar]
- Henkin L, Bergman RN, Bowden DW, Ellsworth DL, Haffner SM, Langefeld CD, Mitchell BD, Norris JM, Rewers M, Saad MF, Stamm E, Wagenknecht LE, Rich SS 2003 Genetic epidemiology of insulin resistance and visceral adiposity. The IRAS Family Study design and methods. Ann Epidemiol 13:211–217 [DOI] [PubMed] [Google Scholar]
- Wagenknecht LE, Mayer EJ, Rewers M, Haffner S, Selby J, Borok GM, Henkin L, Howard G, Savage PJ, Saad MF, Bergman RN, Hamman R 1995 The insulin resistance atherosclerosis study (IRAS) objectives, design, and recruitment results. Ann Epidemiol 5:464–472 [DOI] [PubMed] [Google Scholar]
- Buetow KH, Edmonson M, MacDonald R, Clifford R, Yip P, Kelley J, Little DP, Strausberg R, Koester H, Cantor CR, Braun A 2001 High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Proc Natl Acad Sci USA 98:581–584 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McPeek MS, Sun L 2000 Statistical tests for detection of misspecified relationships by use of genome-screen data. Am J Hum Genet 66:1076–1094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- O'Connell JR, Weeks DE 1998 PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet 63:259–266 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sidney S, Jacobs Jr DR, Haskell WL, Armstrong MA, Dimicco A, Oberman A, Savage PJ, Slattery ML, Sternfeld B, Van Horn L 1991 Comparison of two methods of assessing physical activity in the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Epidemiol 133:1231–1245 [DOI] [PubMed] [Google Scholar]
- Kammerer CM, Dualan AA, Samollow PB, Perisse AR, Bauer RL, MacCluer JW, O'Leary DH, Mitchell BD 2004 Bone mineral density, carotid artery intimal medial thickness, and the vitamin D receptor BsmI polymorphism in Mexican American women. Calcif Tissue Int 75:292–298 [DOI] [PubMed] [Google Scholar]
- Almasy L, Blangero J 1998 Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 62:1198–1211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bonilla C, Parra EJ, Pfaff CL, Dios S, Marshall JA, Hamman RF, Ferrell RE, Hoggart CL, McKeigue PM, Shriver MD 2004 Admixture in the Hispanics of the San Luis Valley, Colorado, and its implications for complex trait gene mapping. Ann Hum Genet 68(Pt 2):139–153 [DOI] [PubMed] [Google Scholar]
- Scragg R, Sowers M, Bell C 2007 Serum 25-hydroxyvitamin D, ethnicity, and blood pressure in the Third National Health and Nutrition Examination Survey. Am J Hypertens 20:713–719 [DOI] [PubMed] [Google Scholar]
- Parikh SJ, Edelman M, Uwaifo GI, Freedman RJ, Semega-Janneh M, Reynolds J, Yanovski JA 2004 The relationship between obesity and serum 1,25-dihydroxy vitamin D concentrations in healthy adults. J Clin Endocrinol Metab 89:1196–1199 [DOI] [PubMed] [Google Scholar]
- Lauridsen AL, Vestergaard P, Hermann AP, Brot C, Heickendorff L, Mosekilde L, Nexo E 2005 Plasma concentrations of 25-hydroxy-vitamin D and 1,25-dihydroxy-vitamin D are related to the phenotype of Gc (vitamin D-binding protein): a cross-sectional study on 595 early postmenopausal women. Calcif Tissue Int 77:15–22 [DOI] [PubMed] [Google Scholar]
- Arnaud J, Constans J 1993 Affinity differences for vitamin D metabolites associated with the genetic isoforms of the human serum carrier protein (DBP). Hum Genet 92:183–188 [DOI] [PubMed] [Google Scholar]