Vitamin D deficiency has been recognized as a major public health issue worldwide. Recent studies have indicated that genetic factors might play an important role in determining serum 25-hydroxyvitamin D [25(OH)D] levels in Caucasians and African Americans. However, the genes that contribute to the variation in serum 25(OH)D levels in Chinese are unknown. In this study, we screened 15 key genes within the vitamin D metabolic pathway using 96 single-nucleotide polymorphism (SNP) markers in a group of 2897 unrelated healthy Chinese subjects. Significant confounding factors that may influence the variability in serum 25(OH)D levels were used as covariates for association analyses. An association test for quantitative traits was performed to evaluate the association between candidate genes and serum 25(OH)D levels. In the present study, variants and/or haplotypes in GC, CYP2R1, and DHCR7/NADSYN1 were identified as being associated with 25(OH)D levels. Participants with three or four risk alleles of the two variants (GC-rs4588 and CYP2R1-rs10766197) had an increased chance of presenting with a 25(OH)D concentration lower than 20 ng/mL (odds ratio 2.121, 95% confidence interval 1.586–2.836, p = 6.1 × 10−8) compared with those lacking the risk alleles. Each additional copy of a risk allele was significantly associated with a 0.12-fold decrease in the log-25(OH)D concentration (p = 3.7 × 10−12). Haplotype TGA of GC rs705117-rs2282679-rs1491710, haplotype GAGTAC of GC rs842999-rs705120-rs222040-rs4588-rs7041-rs10488854, haplotype CA of GC rs1155563-rs222029, and haplotype AAGA of CYP2R1 rs7936142-rs12794714-rs2060793-rs16930609 were genetic risk factors toward a lower 25(OH)D concentration. In contrast, haplotype TGGGCCC of DHCR7/NADSYN1 rs1790349-rs7122671-rs1790329-rs11606033-rs2276360-rs1629220-rs2282618 were genetic protective factors. The results suggest that the GC, CYP2R1, and DHCR7/NADSYN1 genes might contribute to variability in the serum 25(OH)D levels in a healthy Chinese population in Shanghai. These markers could be used as tools in Mendelian randomization analyses of vitamin D, and they could potentially be drug targets in the Chinese population in Shanghai.
Vitamin D deficiency has been recognized as a major public health issue worldwide. This condition has been linked to bone disorders such as rickets and osteoporosis[1-4] as well as extraskeletal diseases that include cancer,[5, 6] cardiovascular disease, and diabetes. It has been reported that 40% to 100% of elderly U.S. and European men and women suffer from vitamin D deficiency. A high prevalence of vitamin D deficiency has also been described in Asian populations.[10-12]
The serum concentration of 25-hydroxy vitamin D [25(OH)D] is the established clinical marker for vitamin D. The impact of heritable factors on the 25(OH)D concentration is estimated to range from 29% to 80%.[13-15] Because several genes control pathways that synthesize, transport, and degrade forms of vitamin D, common genetic variants may play a role in individual (and potentially population) differences in vitamin D status; thus, a more complete understanding of the determinants of vitamin D status is necessary and must take into consideration inherited characteristics. The identification of high-risk groups will allow for effective prevention and treatment and is a topic of great interest in vitamin D research. Information on genetic variants that affect 25(OH)D levels could be used in Mendelian randomization analyses on vitamin D status and disease outcomes; moreover, these variants have potential as drug targets.
Recent genomewide association studies (GWASs) and candidate gene studies on 25(OH)D have provided important insights into the influence of common genetic variations on vitamin D levels, primarily in Caucasians.[18, 19] However, there are ethnic differences in vitamin D status. Furthermore, there is true population variation, which arises if: (1) different disease-causing alleles predominate in different study populations or variation exists in the degree of linkage disequilibrium between the marker and the disease alleles; or (2) allele frequencies are similar, but the magnitude of the effect of the disease gene is heterogeneous between study settings. Recently, Lu and colleagues screened seven common variants in four genes in a mixed Chinese population from Beijing and Shanghai and showed that different genetic variants might be associated with serum 25(OH)D concentrations, indicating that different genetic backgrounds between these two subpopulations might exist. To investigate the genetic variants affecting 25(OH)D levels in a Chinese population from Shanghai, we screened 96 common variants in 15 genes that regulate the vitamin D metabolic pathway in 2897 Chinese from Shanghai and identified CYP2R1, GC, and DHCR7/NADSYN1 as potential contributors to the variation in serum 25(OH)D levels in this healthy Chinese population.
Materials and Methods
From February 2009 to March 2009, 2243 women (aged 20–96 years) and 865 men (aged 23–94 years)—a total of 3108 healthy Chinese people living in Shanghai—were recruited from several community centers. The study subjects were identical to those included in the Shanghai Osteoporosis Study (SOS). The study subjects were recruited from 10 communities within Shanghai. After stratifying the population of each selected community by age, we randomly sampled participants. Selected participants were called and persuaded to come to the hospital in the corresponding district center. All of the participants were of Han ethnicity. The participants' age, body weight, height, and age at both menarche and amenorrhea were recorded. Height (cm) was measured using a wall-mounted stadiometer. The participants were weighed (kg) while wearing indoor clothing without shoes on a balance-beam scale. Both the stadiometer and the balance-beam scale were regularly calibrated during the study. All of the participants were subjected to blood counts, fasting plasma glucose tests, serum lipid tests, and liver and kidney function tests.
All of the healthy Han subjects included in the present study had (1) normal blood counts, (2) normal results for liver and kidney function tests, and (3) normal values for serum calcium, phosphorus, alkaline phosphatase (ALP), parathyroid hormone (PTH), glucose, insulin, and triglycerides. Participants were excluded from the study if they had diseases deemed to affect vitamin D metabolism, such as cancer, hyperthyroidism, diabetes mellitus, primary hyperparathyroidism, pituitary, or adrenal and rheumatic diseases. Participants who had taken vitamin D and/or calcium supplements within the past 3 months were also excluded. After these exclusions, 3108 participants entered the study. The study was approved by the Ethics Committee of the Shanghai Jiao Tong University Affiliated Sixth People's Hospital. All of the participants signed informed consent forms before entering the study.
Measuring serum 25(OH)D and vitamin D binding protein levels
The serum levels of 25(OH)D were determined using an automated Roche electrochemiluminescence system (E170; Roche Diagnostic GmbH, Mannheim, Germany). The intraassay coefficients of variation (CVs) for 25(OH)D were 5.7% at a level of 25.2 ng/mL, 5.7% at a level of 39.9 ng/mL, and 5.4% at a level of 65.6 ng/mL, respectively. The interassay CVs for 25(OH)D were 9.9% at a level of 25.2 ng/mL, 7.3% at a level of 39.9 ng/mL, and 6.9% at a level of 65.6 ng/mL, respectively. The lower detection limit of 25(OH)D was 4 ng/mL. The serum levels of vitamin D binding protein (DBP) were determined using Human Vitamin D-binding protein enzyme-linked immunosorbent assay (ELISA) Kits (Cusabio Biotech Co., Ltd., Hubei, PR China; Catalog Number: CSB-E11859h). The detection range was 0.156 to 10 mg/L. The intraassay CV was <8% and the interassay CV was <10%. All of the serum samples were taken in the morning (fasting blood) during the winter season (from February 2009 to March 2009).
Candidate genes and tag SNP selection
We selected 15 candidate genes according to the following criteria: (1) biological importance in vitamin D metabolism, transportation, or degradation; and (2) evidence of a significant association in previous GWASs. The selected genes were GC, CYP2R1, CYP27A1, CYP27B1, CYP24A1, DHCR7, NADSYN1, CYP3A4, CYP2J2, PTH, ACADSB, CUBN, CYP11A1, CYP1A1, and CYP2C9. The basic characteristics of these 15 genes are shown in Table 1. The roles of these selected genes in the vitamin D cascade are shown in Fig. 1. The detailed functions of the candidate genes are described in the Discussion section.
|Gene||Location||Full name||Selected SNPs|
|GC||4q12–q13||Vitamin D binding protein||15|
|CYP2R1||11p15.2||Vitamin D 25-hydroxylase||5|
|CYP27A1||2q33–qter||Vitamin D(3) 25-hydroxylase||2|
|CYP27B1||12q13.1–q13.3||25 Hydroxyvitamin D-1-alpha hydroxylase||2|
|CYP24A1||20q13||Vitamin D 24-hydroxylase||14|
|NADSYN1||11q13.4||NAD synthetase 1||6|
|CYP3A4||7q21.1||Cytochrome P450 3A4||3|
|CYP2J2||1p31.3–p31.2||Cytochrome P450 2J2||6|
|ACADSB||10q26.13||Acyl-CoA dehydrogenase, short/branched chain||7|
|CYP11A1||15q23–q24||Cytochrome P450 11A1||5|
|CYP1A1||15q24.1||Cytochrome P450 1A1||3|
|CYP2C9||10q24||Cytochrome P450 2C9||4|
For the studied genes, tagging SNPs were selected from the International HapMap Project (http://www.hapmap.org/cgi-perl/gbrowse/hapmap3_B36). The SNPs were selected based on the following criteria: (1) validation status, especially in Chinese; (2) degree of heterozygosity (minor allele frequencies [MAFs] > 0.1); (3) binned by the algorithm such that the pairwise linkage disequilibrium (LD) exceeds a threshold r2 (r2 = 0.05); and (4) requirement for tag-SNPs. SNPs reported in GWASs or potentially functional SNPs in candidate genes were forced into the SNP selection process.
Blood samples were collected from all of the study subjects. Genomic DNA was isolated from peripheral blood leukocytes using the conventional phenol-chloroform extraction method. Genotyping was performed using the high-throughput Sequenom genotyping platform (MassARRAY matrix-assisted laser desorption/ionization-time of flight mass spectroscopy [MALDI-TOF MS] system; Sequenom, San Diego, CA, USA).
Genotype frequencies were tested against the Hardy-Weinberg equilibrium (HWE) using the χ2 test to estimate the laboratory error.
We compared the baseline characteristics of the men and women using the t test for continuous variables. The mean serum 25(OH)D values were computed within each group of homozygous referent (HR), heterozygous (HET), and homozygous variant (HV) genotypes for each SNP. HR was selected as the most common homozygous genotype. The serum 25(OH)D levels were adjusted by the covariates (age, gender, and body mass index [BMI]) using a linear regression approach (SPSS, version 13.0; SPSS Inc., Chicago, IL, USA). The adjusted serum 25(OH)D levels were used for the subsequent data analyses. All of the SNPs that passed the quality control checks were tested for quantitative trait association interaction using the linear model implemented in PLINK (http://pngu.mgh.harvard.edu/purcell/plink/). Out of the four possible models (additive, codominant, dominant, and recessive) in the linear model, we selected the additive model because it generally reflects the additive contribution to risk for complex diseases. The degree of variation in the log-25(OH)D levels imparted by the SNPs was generated from the adjusted general linear regression models as the ratio of the type II sum of squares to the total sum of squares. The genetic risk score (GRS) was calculated by counting the number of CYP2R1-rs10766197 and GC-rs4588 risk alleles. The participants without risk alleles served as the reference group. Logistic regression was used to calculate the odds ratio of vitamin D deficiency according to the GRS. General linear regression was applied for associations between GRS and serum 25(OH)D levels. Vitamin D deficiency is defined as a 25(OH)D level of less than 20 ng/mL. The statistical analyses were performed using SPSS version 11.0 (SPSS Inc., Chicago, IL, USA).
We performed haplotype interaction analysis on the genes containing the significant SNPs. The linkage disequilibrium structure was examined using Haploview 4.2 (http://www.broadinstitute.org/scientific-community/science/programs/medical-and-population-genetics/haploview/haploview) with the Gabriel confidence interval method. PLINK was used to test for the significance of each haplotype (with frequency ≥10%) within the defined blocks.
To control for the familywise error rate, the Bonferroni correction was used to adjust for multiple testing. The results were regarded as statistically significant at a value of p < 0.05.
Of the 3108 participants selected for this study, 192 (6.2%) were excluded because the serum 25(OH)D level could not be measured or was below the lower detection limit, and 119 (3.8%) were excluded because less than 95% of the markers were successfully genotyped across all of the SNPs. This exclusion process left 2897 participants (2304 women and 593 men) for analysis. The clinical characteristics and mean laboratory values of the 2304 women and 593 men are shown in Table 2. Generally, the men had higher BMI and serum 25(OH)D levels than the women. By Pearson correlation analysis, 25(OH)D levels were significantly correlated with age (r = −0.166; p < 0.001) but not with BMI (r = −0.046; p = 0.064). Age and gender accounted for 14.4% of the variation in serum 25(OH)D levels.
|All (n = 2897)||Males (n = 593)||Females (n = 2304)||p|
|Age (years)||46 (32–60)||43 (33–56)||48 (31–61)||0.126|
|<50 years||1550 (53.5%)||366 (61.7%)||1184 (51.4%)|
|≥50 years||1347 (46.5%)||227 (38.3%)||1120 (48.6%)|
|Height (cm)||161.6 ± 8.2||173.1 ± 6.2||159.1 ± 6.2||<0.001|
|Weight (kg)||61.0 ± 10.8||74.8 ± 10.7||58.0 ± 8.1||<0.001|
|BMI (kg/m2)||23.3 ± 3.4||24.9 ± 3.3||23.0 ± 3.3||<0.001|
|Serum 25(OH)D (ng/mL)||19.5 (14.9–23.9)||22.6 (19.0–26.8)||18.5 (14.0–22.9)||<0.001|
|Serum DBP (mg/L)||312 ± 56||308 ± 57||313 ± 56||0.592|
The basic characteristics of the SNPs are listed in Supporting Information Table S1. The MAFs of all of the SNPs examined in this study were comparable to those reported for the HapMap-CHB sample. Two SNPs (rs10493270 in the CYP2J2 gene and rs4646437 in the CYP3A4 gene) were excluded based on the HWE test (p ≤ 0.001). None of the SNPs failed the missingness test (genotyping > 0.05) or the frequency test (MAF < 0.01). After removing the SNPs that did not pass the quality control measures, 94 SNPs, with a mean call rate of 0.998%, remained available for analysis, indicating a very high rate of successful genotyping.
The SNP genotype frequencies and the mean serum 25(OH)D levels grouped by genotype are presented in Supporting Information Table S2, with the significant results shown in Table 3. Three SNPs, one in the CYP2R1 gene (rs10766197) and two in the GC gene (rs4588 and rs2282679), were identified as being significantly associated with serum 25(OH)D levels. Rs10766197 is located in the promoter of the CYP2R1 gene. Rs2282679 is located in the intron of GC and rs4588 is located in exon 11 of GC, which leads to a Thr/Lys amino acid change at codon 420. The two SNPs in GC are in high LD. The strongest association was observed for GC-rs4588, which accounted for 0.7% of the variation in serum 25(OH)D levels. To analyze the combined effect of CYP2R1-rs10766197 and GC-rs4588 on 25(OH)D levels, we calculated the GRS by counting the number of risk alleles of these two SNPs. Table 4 shows the results for the variants both individually and in combination. As shown in Fig. 2, using a general linear regression model, each additional copy of a risk allele was significantly associated with a 0.12-fold decrease in the log-25(OH)D concentration (p = 3.7 × 10−12). The combined effect of the two SNPs, as measured by GRS, accounted for 1.7% of the variance in the log-25(OH)D concentration.
|Gene||SNP||Mean concentrations of 25(OH)D||Frequency||Beta||p||pa|
|GC||rs4588||17.88||19.55||20.3||0.09832||0.4417||0.46||−0.1635||2.15 × 10−5||0.002016|
|CYP2R1||rs10766197||17.87||18.17||19.48||0.1317||0.4558||0.4125||−0.1484||4.69 × 10−5||0.004411|
|GC||rs2282679||16.87||18.33||19.32||0.09136||0.4364||0.4722||−0.1529||6.60 × 10−5||0.006206|
|Odds ratio (95% CI)||p|
|GC-rs4458 (G)||1.403 (1.245–1.583)||2.7 × 10−8|
|CYP2R1-rs10766197 (A)||1.215 (1.037–1.424)||0.016|
|Number of risk alleles from GC-rs4458 and CYP2R1-rs10766197|
|0||1.0 (Reference)||6.1 × 10−8|
|3 + 4||2.121 (1.586–2.836)|
By haplotype analysis, we found five candidate haplotype blocks, three in GC, one in CYP2R1, and one in DHCR7/NADSYN1 (Fig. 3). The frequencies and beta values of haplotypes with a frequency over 0.1 in each block are shown in Table 5. Among them, haplotype TGA of GC rs705117-rs2282679-rs1491710, haplotype GAGTAC of GC rs842999-rs705120-rs222040-rs4588-rs7041-rs10488854, haplotype CA of GC rs1155563-rs222029, and haplotype AAGA of CYP2R1 rs7936142-rs12794714-rs2060793-rs16930609 were genetic risk factors toward a lower 25(OH)D concentration. In contrary, haplotype TGGGCCC of DHCR7/NADSYN1 rs1790349-rs7122671-rs1790329-rs11606033-rs2276360-rs1629220-rs2282618 were genetic protective factors.
|TGA||0.3083||−0.1529||6.6 × 10−5|
|GAGTAC||0.299||−0.1663||2.0 × 10−5|
In view of the strong association of genetic variants at GC with 25(OH)D concentrations, we also performed a correlative analysis to assess the associations between 25(OH)D and circulating DBP and examined whether these variants were also associated with serum concentrations of DBP. Pearson correlations indicated that 25(OH)D levels were positively correlated with circulating DBP (r = 0.161, p < 0.001). As Fig. 4 shows, both rs4588 and rs2282679 were associated with concentrations of DBP, with the minor alleles (T for rs4588 and G for rs2282679) related to reduced protein concentrations.
The present study investigated the association of 15 candidate genes with serum 25(OH)D levels. Figure 1 depicts the placement of these candidate genes within the vitamin D cascade. Humans obtain vitamin D primarily from exposure to sunlight. ACADSB is involved in cholesterol synthesis, and DHCR7/NADSYN1 removes 7-dehydrocholesterol (7-DHC) from the vitamin D pathway, thereby reducing the amount of substrate available for 25(OH)D synthesis. The association of these three genes with vitamin D levels was recently identified in GWASs.[18, 19] Solar ultraviolet B radiation (wavelength, 290–315 nm) penetrates the skin and converts 7-DHC to previtamin D3, which is rapidly converted to vitamin D3. Vitamin D3 is primarily metabolized in the liver to 25(OH)D3 by CYP2R1, but CYP2J2, CYP3A4, CYP27A1, CYP1A1, and CYP2C9 also contribute. Subsequently, 25(OH)D3 is transported by vitamin D binding protein (DBP), encoded by the GC gene, to the kidney. In the proximal tubule cells, cubilin, encoded by the CUBN gene, facilitates the endocytic retrieval of the filtered 25(OH)D3-DBP complex. Once transported into the proximal tubule cells, 25(OH)D3 is metabolized by the enzyme CYP27B1 to its active form, 1,25(OH)2D3. Finally, CYP24A1 catabolizes both 25(OH)D3 and 1,25(OH)2D3 into the biologically inactive, water-soluble calcitroic acid.[31, 32] In addition, the renal production of 1,25(OH)2D3 is tightly regulated by the plasma levels of parathyroid hormone and fibroblast growth factor 23. In the metabolic bypass of vitamin D3, CYP11A1 can hydroxylate vitamin D3, producing 20S-hydroxyvitamin D3 [20(OH)D3] and 20S,23-dihydroxyvitamin D3 [20,23(OH)2D3] as the major metabolites instead of 25(OH)D3, which may also influence the serum 25(OH)D levels.
The identified associations between SNPs in GC and 25(OH)D concentration are consistent with previous studies.[16, 18, 19, 34-41] In addition, we showed that GC variants associated with low 25(OH)D concentrations were strongly related to reduced DBP concentrations, which were also consistent with previous studies.[19, 42] DBP binds to vitamin D sterol metabolites and transports them via the circulation to target organs. Changes in quantity or function of DBP could be accompanied by changes in the relative proportions of free and bound 25(OH)D, with the free proportion being the potential rate-limiting factor for 1,25-(OH)2D production, or the free 25(OH)D itself may be the active form. Analysis of the GC knockout mouse demonstrated that DBP stabilized and maintained serum levels of vitamin D under conditions of variable vitamin D availability, markedly prolonged the serum half-life of 25(OH)D, and slightly prolonged the half-life of vitamin D by slowing its hepatic uptake and increasing the efficiency of its conversion to 25(OH)D in the liver. Verboven and colleagues reported the 2.3-Å crystal structure of DBP in complex with 25-hydroxyvitamin D3, which revealed that the vitamin D binding site is in the N-terminal part of domain I. Because GC-rs2282679 is located in the intron region near actin subdomain III, which is distant from the N-terminal part of domain I, the SNP itself is unlikely to be the disease-causing variant. The possible causal variant is rs4588, which leads to a Thr/Lys amino acid change at codon 420.
CYP2R1 is a member of the CYP2 family that encodes cytochrome P450 proteins. This enzyme is a key vitamin D 25-hydroxylase that hydroxylates vitamin D at the 25-C position to form 25(OH)D in the liver,[25, 46] although many enzymes with in vitro 25-hydroxylase activity have been described. Previous human genetic studies demonstrated that CYP2R1 gene mutations caused vitamin D–dependent rickets type 1B, further indicating the importance of CYP2R1 in vitamin D 25 hydroxylation.[47, 48] However, Tosson and Rose reported a proband and his family with apparent autosomal dominant 25-hydroxylase enzyme deficiency without a mutation in the coding region of the CYP2R1 gene, thereby questioning the importance of different tributary pathways that affect the function of the 25-hydroxylase enzyme. Our results showed that the CYP2R1 gene polymorphism was significantly associated with serum 25(OH)D levels and strongly support the hypothesis that CYP2R1 is the crucial 25-hydroxylase enzyme, rather than CYP2J2, CYP3A4, CYP27A1, CYP1A1, or CYP2C9. The association of CYP2R1 polymorphisms with 25(OH)D levels is supported by previous studies.[18, 19, 41, 50-52] However, the variant in CYP2R1 that is most strongly associated with 25(OH)D levels was not consistent among studies. In two large GWASs in European populations, Ahn and colleagues found that rs2060793 in CYP2R1 was associated with serum 25(OH)D levels, whereas Wang and colleagues reported that rs10741657 was associated with serum 25(OH)D levels. Wjst and colleagues reported that SNP rs10766197 in the CYP2R1 gene was significantly associated with 25(OH)D levels in 872 participants in the German Asthma Family Study. Recently, Lu and colleagues demonstrated that the CYP2R1-rs2060793 variant trended toward an association with serum 25(OH)D levels in the Shanghai subpopulation (p = 0.08) but not in the Beijing subpopulation (p = 0.82), suggesting that genetic variation or environmental factors differ between these two subpopulations. Indeed, heterogeneity was observed in the association of 25(OH)D levels with CYP2R1 in the different cohorts in the GWAS by Ahn and colleagues. In the present study, we found that rs10766197, not rs2060793, was associated with serum 25(OH)D levels in a Chinese population in Shanghai. SNP rs10766197 is located in the promoter region of the CYP2R1 gene. Therefore, studies are necessary to assess whether rs10766197 or unidentified causal variants in LD with it influence the production of CYP2R1.
DHCR7/NADSYN1 is a locus that was identified in recent GWASs on vitamin D deficiency.[18, 19] DHCR7/NADSYN1 removes 7-DHC from the vitamin D pathway, thus reducing the substrate available for 25(OH)D synthesis. Mutations in DHCR7 might confer a competitive advantage to heterozygous carriers because high concentrations of 7-DHC could provide protection against rickets and osteomalacia from hypovitaminosis D. Recently, Lu and colleagues found that the DHCR7-rs1790349 variant had a milder association with serum 25(OH)D levels in the Shanghai subpopulation (p = 0.0363) than in the Beijing subpopulation (p = 0.00017), suggesting that genetic variation or environmental factors differ between these two subpopulations. Therefore, the role of common variants in DHCR7/NADSYN1 in the regulation of vitamin D levels needs to be further studied in the Shanghai population.
Compared with previous studies, our study has several strengths: (1) by conducting this study in a healthy population, the analysis of the genetic impact on serum 25(OH)D levels is not confounded by the potential impact of disease; (2) dense markers and those previously reported in GWASs within 15 candidate genes involved in vitamin D metabolism were selected; and (3) all of the serum samples were taken during the winter season (from February 2009 to March 2009) to minimize the impact of sun exposure. Although participants who had taken vitamin D and/or calcium supplements within 3 months were excluded, the dietary intake of vitamin D and calcium for the participants was not recorded.
Insights gained from studying the circulating levels of vitamin D are likely to have implications in complex diseases, such as osteoporosis, asthma, type 2 diabetes, and cancer. Because the genotype is assigned randomly, the genetic association, unlike the directly observed association of vitamin D intake/status itself, will be less prone to confounding and free from reverse causation because the genotype is not modifiable by disease. In addition, rs10766197 is located in the promoter of the CYP2R1 gene that encodes the key vitamin D 25-hydroxylase. Therefore, with further investigation of the biological mechanisms, this marker may become a potential drug target.
All authors state that they have no conflicts of interest.
This work was supported by the National Natural Science Foundation of China (81070692, 81000360, and 81170803), the Program of the Shanghai Subject Chief Scientist (08XD1403000), STCSM (10DZ1950100), Academic Leaders in Health Sciences in Shanghai (XBR2011014), and Shanghai Leading Talent Plan (051). We gratefully acknowledge the support of the participants in the study.
Authors' roles: Study design: ZLZ and CQZ. Study conduct: ZZ. Data collection: JWH and WZF. Data analysis: ZZ. Data interpretation: ZZ. Drafting of the manuscript: ZZ. Approving final version of the manuscript: ZLZ and CQZ. ZZ takes reponsibiliy for the integrity of the data analysis.