Association Between Vitamin D Metabolism Gene Polymorphisms and Risk of Islet Autoimmunity and Progression to Type 1 Diabetes: The Diabetes Autoimmunity Study in the Young (DAISY)

Abstract

Context:

Vitamin D metabolism genes have been associated with type 1 diabetes (T1D) risk; however, these genes have not been investigated for association with the preclinical phase of T1D, islet autoimmunity (IA). Studies of vitamin D metabolism genes may elucidate the role of vitamin D in complex diseases.

Objective:

The objective of the study was to explore the association between seven vitamin D metabolism gene single-nucleotide polymorphisms (SNPs) and the risk of IA and progression to T1D.

Design:

The Diabetes Autoimmunity Study in the Young is a longitudinal, observational study.

Setting:

Newborn screening for human leukocyte antigen, sibling and offspring recruitment, and follow-up took place in Denver, Colorado.

Participants:

A total of 1708 children at increased genetic risk of T1D participated in the study: 148 developed IA and 62 IA-positive children progressed to T1D.

Main Outcome Measures:

IA, defined as positivity for glutamic acid decarboxylase, insulin, or IA-2 autoantibodies on two or more consecutive visits, and T1D, diagnosed by a physician, were the main outcome measures.

Results:

The risk of IA was associated with DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 [hazard ratio 1.36, 95% confidence interval 1.08–1.73 (for each additional minor allele) and hazard ratio 0.59, 95% confidence interval 0.39–0.89 (for A/G compared with the A/A genotype), respectively]. None of the vitamin D SNPs typed was associated with progression to T1D in IA-positive children. Six of the seven SNPs were significantly associated with 25-hydroxyvitamin D levels.

Conclusions:

DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 may play an important role in islet autoimmunity, the preclinical phase of T1D. These findings should be replicated in larger cohorts for confirmation.

Type 1 diabetes (T1D) is an autoimmune disease in which the insulin-producing βcells of the pancreas are destroyed. There is typically a preclinical phase of circulating autoantibodies, called islet autoimmunity (IA), that precedes the clinical diagnosis of T1D.

Vitamin D insufficiency [25-hydroxyvitamin D [25(OH)D] <20 ng/mL or <50 nmol/L] has been associated with a number of extraskeletal chronic disorders (1, 2), including autoimmune diseases, such as T1D, multiple sclerosis, Crohn's disease, and rheumatoid arthritis (2–6). The mechanism by which vitamin D exerts its effects on these diseases is not understood completely, particularly with regard to underlying genetic risk. Determinants of circulating 25(OH)D, the inactive circulating form of vitamin D and an established marker of vitamin D status, include sun exposure and dietary intake. However, reported heritabilities of 25(OH)D of 28.8% (7) to 43% (8) suggest that genetic determinants also play a role (1, 2).

In 2010, a genome-wide association study (GWAS) in approximately 30 000 individuals of European descent identified variants at four loci that were associated with 25(OH)D levels: GC rs2282679, DHCR7 rs12785878, CYP2R1 rs10741657, and CYP24A1 rs6013897 (1). A second GWAS of 25(OH)D levels confirmed the findings with GC, DHCR7, and CYP2R1 (9). These variants are located within or near genes involved in vitamin D transport (GC), cholesterol synthesis (DHCR7), and hydroxylation (CYP2R1 and CYP24A1) (1).

Cooper et al (2) more recently tested genetic variants influencing 25(OH)D metabolism for an association with both circulating 25(OH)D concentrations and T1D. They replicated the associations found in the aforementioned GWAS of the four vitamin D metabolism genes (GC, DHCR7, CYP2R1, and CYP24A1) with 25(OH)D in control subjects (1, 2) and found that CYP27B1, DHCR7, and CYP2R1 were associated with type 1 diabetes (2). CYP27B1 had previously been associated with type 1 diabetes in 2004 (10).

The Diabetes Autoimmunity Study in the Young (DAISY) has been prospectively following up children at increased T1D risk for the development of IA and progression to T1D since 1993. The purpose of this study was to explore associations between seven previously studied vitamin D single nucleotide polymorphisms (SNPs) and the development of IA and progression to T1D in the prospective DAISY cohort.

Materials and Methods

Subjects

DAISY is a prospective study composed of two groups of children at increased risk for T1D who were recruited between 1993 and 2004 and are being followed up prospectively for the development of IA and T1D. One group is made up of first-degree relatives of patients with T1D, recruited between birth and 8 years of age (n = 850, 49.8%). The second group consists of infants born at St Joseph's Hospital in Denver, Colorado, whose umbilical cord blood was screened for diabetes-susceptibility genotypes in the human leukocyte antigen (HLA) region (n = 858, 50.2%). Based on their HLA genotype, newborns were categorized into three risk groups determined by the odds of developing T1D by the age of 20 years: high, odds of developing T1D by the age of 20 years, 1:16; moderate, 1:75 in non-Hispanic whites or 1:230 in Hispanics; or low, less than 1:300. All newborns found to be at high risk and a sample of those found to be at moderate risk were asked to participate in the follow-up. The St Joseph's Hospital newborn population is representative of the general population of the Denver metropolitan area. Details of the newborn screening, sibling and offspring recruitment, and follow-up of both cohorts have been published previously (11, 12). Cord blood or the first available blood sample (depending on enrollment group) was sent to Roche Molecular Systems, Inc for PCR-based HLA class II typing. All study protocols were approved by the Colorado Multiple Institutional Review Board, and informed consent was given by parents of all participating children.

Measurement of autoantibodies

Autoantibodies were tested at 9, 15, and 24 months, and annually thereafter, or at their first visit and annually thereafter if the child enrolled after birth. RIAs were used to measure serum autoantibodies to insulin, glutamic acid decarboxylase-65, and IA-2 (BDC512), as previously described (13–16), with rigorous confirmation of all positive and a subset of negative results. The cutoff for positivity was established as the 99th percentile of healthy controls. Children who tested autoantibody positive were put on an accelerated testing schedule of every 3–6 months.

Cases of persistent IA were defined as those children positive for at least one islet autoantibody (IAA, glutamic acid decarboxylase-65, IA-2) on two or more consecutive visits. T1D was diagnosed by a physician and defined according to the criteria for the diagnosis of diabetes by the American Diabetes Association: hemoglobin A_1c of 6.5% or greater or fasting plasma glucose of 126 mg/dL (7.0 mmol/L) or 2-hour plasma glucose of 200 mg/dL or greater (11.1 mmol/L) during an oral glucose tolerance test or in a patient with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose of 200 mg/dL or greater (11.1 mmol/L) (17).

Vitamin D SNP genotyping

DAISY children were genotyped for seven SNPs in genes involved in components of vitamin D metabolism. DHCR7/NADSYN1 rs12785878, CYP2R1 rs10741657, CYP2R1 rs12794714, and CYP24A1 rs6013897 were genotyped based on their genome-wide significance with 25(OH)D concentrations (1). GC rs4588 and GC rs7041 were chosen because they are located in the GC gene that encodes vitamin D binding protein, which is a 52- to 59-kDA protein synthesized in the liver that binds and transports vitamin D and its metabolites [including 25(OH)D and 1,25-dihydroxyvitamin D (1,25[OH]₂D)], and these two SNPs are the most widely studied GC variants (1). CYP27B1 rs4646536 was chosen based on its association with type 1 diabetes (2, 18). Linkage disequilibrium was tested in our population using Haploview version 4.2 as measured by r², with r² = 0.508 for the two GC SNPs and r² = 0.487 for the two CYP2R1 SNPs.

The SNPs were genotyped using the Taqman SNP genotype-based OpenArray platform (Applied Biosystems). Custom-designed, 48-sample arrays and normalized genomic DNA were loaded using the OpenArray AccuFill system, and cycling was performed on a GeneAmp 9700 PCR system (Applied Biosystems), all according to manufacturer protocol. Alleles were analyzed using the OpenArray SNP genotyping analysis software version 1.0.3 and Taqman Genotyper Software 2.0 (Applied Biosystems). All seven SNPs had a 95% call rate or higher.

Each SNP was tested for consistency with Hardy-Weinberg proportions using a 1-degree of freedom χ² goodness-of-fit test with a value of P = .05 considered as evidence of a departure from Hardy-Weinberg equilibrium; all seven SNPs were in Hardy-Weinberg equilibrium.

Measurement of plasma 25(OH)D levels

Blood was drawn and kept from light at all times during processing. Plasma was separated immediately, snap frozen in liquid nitrogen, and stored at −70°C until it was sent to the University of Colorado Pediatric Clinical Translational Research Center Core Laboratory, which has a certificate of proficiency from the Vitamin D External Quality Assessment Scheme. 25-Hydroxyvitamin D2 and D3 was measured by RIA (DiaSorin), with a coefficient of variation of 7.5%. Quality control was assessed via an assay of blinded duplicate samples, and excellent agreement (intraclass correlation coefficient of 0.92 in 229 pairs) was observed. The 25(OH)D levels are reported in nanograms per milliliter.

Analysis population

We obtained genetic data on at least one of the seven vitamin D SNPs on 1708 non-Hispanic white children in the DAISY cohort. This included 148 children who developed persistent IA, of whom 62 went on to develop T1D. However, 17 IA cases were positive for autoantibodies on their first clinic visit; these left-censored cases were removed from the development of IA analysis cohort but retained in the progression from IA to T1D cohort. All statistical analyses were limited to non-Hispanic whites in the DAISY cohort.

Statistical analyses

SAS version 9.3 statistical software package (SAS Institute Inc) was used for all statistical analyses.

Analysis of vitamin D SNPs and risk of IA and T1D in IA positive children

SNPs were analyzed for their association both with development of IA and with progression from IA to T1D. For each model, hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression. A clustered time-to-event analysis was performed treating siblings from the same family as clusters, and robust sandwich variance estimates (19) were used for statistical inference. There were 272 families with more than one child enrolled in the study and included in these analyses. Analyses of time to development of IA were adjusted for the HLA-DR genotype (HLA-DR3/4,DQB1*0302 vs other genotypes) and the presence of a first-degree relative with T1D. Analyses of time to progression to T1D were adjusted, in addition, for age at first positive autoantibody visit. The significance threshold was defined as an α of less than .05. Because our analyses were based on a priori hypotheses with SNPs previously found to be associated with 25(OH)D levels and T1D, P values were not corrected for multiple testing.

A priori, we evaluated the linearity assumption for each SNP in the Cox proportional hazards using a contrast of the coefficients from the additive model. CYP27B1 rs4646536 did not meet the linearity assumption and was therefore analyzed as a three-level categorical variable (ie, genotypes). Because the global test for this categorical variable was statistically significant (P = .03), we did two pairwise tests using homozygosity for the major allele as the reference group. The remaining SNPs met the linearity assumption and were therefore analyzed treating the number of minor alleles as a continuous variable with the HR representing an increase in risk for each minor allele. We analyzed each vitamin D SNP in separate adjusted models.

Longitudinal analysis of plasma 25(OH)D levels and vitamin D SNPs

We analyzed the relationship between plasma 25(OH)D levels and individual vitamin D SNPs using a linear mixed modeling approach in a subcohort of DAISY participants. The subcohort is a representative sample of children selected from an eligible pool in DAISY via stratified random sampling based on HLA-DR genotype and family history of type 1 diabetes. Of the 279 non-Hispanic white children selected for the subcohort, 262 children had at least one plasma 25(OH)D measurement and were genotyped for at least one of the vitamin D SNPs and therefore were included in the analyses. The linear mixed models distinguish variability between participants and the variability between repeated measurements over time within participants. Vitamin D intake and caloric intake were measured by annual food frequency questionnaires, as described previously (20). We tested each vitamin D SNP for association with 25(OH)D levels in separate models adjusted for HLA-DR3/4,DQB1*0302, first-degree relative with T1D, season of blood draw, age of study subjects, sex, average daily vitamin D intake from food and supplements, average daily intake of calories, and type of food frequency questionnaire completed [Food Frequency Questionnaire vs Youth Adolescent Questionnaire (17)]. Linear mixed models with a random intercept only, a random slope for age only, and both together were tested for the best fit based on the lowest Akaike information criteria. Linear mixed models with both a random intercept and a random slope for age represented the best fit to the data. The mixed model provides a regression coefficient, a SE, and a P value for each variable to indicate its contribution toward explaining variation in 25(OH)D.

Results

Development of islet autoimmunity

We first examined whether vitamin D variants were associated with the development of persistent islet autoimmunity (Table 1). Of a total of 148 IA-positive children in DAISY, 17 had to be excluded from the proportional hazards analyses of IA because they tested autoantibody positive on their first study visits (ie, they were left censored). The mean age at the first IA-positive visit was 5.9 years, and the mean age at last follow-up visit in children who did not develop IA was 9.9 years. IA-positive children were more likely to have the HLA-DR3/4,DQB1*0302 genotype compared with children who did not develop IA.

Table 1.

Characteristics of the Analysis Sample by IA Status and Association of Vitamin D Metabolism SNPs in DAISY Non-Hispanic White Population (n = 1691)

Characteristic	Minor Allele	MAFa	Children Positive for IA (n = 131)	Children Negative for IA (n = 1560)	Adjusted HR and 95% CI	P Value	n
Mean age, yb			5.9 ± 4.2	9.9 ± 5.7	N/A	N/A	N/A
HLA-DR3/4,DQB1*0302			48 (36.6%)	261 (16.7%)	2.86 (1.96, 4.16)	<.0001	1691
First-degree relative with T1D			84 (64.1%)	762 (48.9%)	1.44 (1.00, 2.07)	.05	1691
Sex (female)			68 (51.9%)	735 (47.1%)	1.17 (0.84, 1.65)	.35	1691
DHCR7/NADSYN1 rs12785878c,d	G	0.26	73 (55.7%)e	687 (44.4%)e	1.36 (1.08, 1.73)	.01	1677
GC rs4588c,d	T	0.28	62 (53.5%)e	715 (48.2%)e	1.11 (0.86, 1.44)	.43	1601
GC rs7041c,d	A	0.43	78 (67.2%)e	1027 (69.1%)e	0.88 (0.68, 1.13)	.30	1603
CYP2R1 rs10741657c,d	A	0.40	75 (65.2%)e	940 (64.0%)e	0.99 (0.76, 1.30)	.95	1583
CYP2R1 rs12794714c,d	A	0.43	74 (64.9%)e	987 (67.8%)e	0.95 (0.72, 1.27)	.74	1570
CYP24A1 rs6013897c,d	A	0.22	45 (38.5%)e	577 (38.8%)e	1.10 (0.78, 1.54)	.60	1604
CYP27B1 rs4646536c	G	0.32					1573
A/A			62 (55.9%)	672 (46.0%)	Referent
A/G			36 (32.4%)	638 (43.6%)	0.59 (0.39, 0.89)	.01
G/G			13 (11.7%)	152 (10.4%)	1.03 (0.57, 1.86)	.92

Abbreviation: MAF, minor allele frequency.

^aMAF calculated for children negative for IA.

^bAge at first IA-positive visit in autoantibody-positive children or age at last follow-up in autoantibody-negative children.

^cAdjusted for HLA-DR3/4,DQB1*0302 and first-degree relative with T1D.

^dSNP analyzed treating the number of minor alleles as a continuous variable, with HR representing increase in risk for each additional minor allele.

^en and percentage for individuals with at least one minor allele.

The observed number of IA cases and the number of person-years for each genotype of the vitamin D metabolism SNPs in the DAISY non-Hispanic white population are shown in Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org. Adjusting for HLA-DR3/4,DQB1*0302 and first-degree relative with T1D, development of IA was associated with DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 [HR 1.36, 95% CI 1.08, 1.73 (for each additional minor allele) and HR 0.59, 95% CI 0.39, 0.89 (for A/G compared with A/A genotype), respectively]. When both DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 were included in the model, both SNPs remained significantly associated with development of IA, adjusting for HLA-DR3/4,DQB1*0302 and a first-degree relative with T1D [DHCR7/NADSYN1 rs12785878, HR 1.39, 95% CI 1.07, 1.80 (for each additional minor allele) and CYP27B1 rs4646536, HR 0.59, 95% CI 0.39, 0.89 (for A/G compared with A/A genotype)].

We also created an allelic score accounting for the combined effects of DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536. DHCR7/NADSYN1 rs12785878 was treated continuously with an increase in the number of protective major alleles (the A allele), and CYP27B1 rs4646536 was treated dichotomously, with 0 representing individuals with no or two minor alleles and 1 representing individuals with the protective heterozygosity. The score ranged from 0 to 3, with 0 representing individuals with zero DHCR7 rs12785878 A alleles and zero or two CYP27B1 rs4646536 minor alleles and 3 representing individuals with two DHCR7 rs12785878 A alleles and heterozygosity at CYP27B1 rs4646536. The allelic score treated continuously was associated with a decreased risk of islet autoimmunity (HR 0.67, 95% CI 0.54, 0.84).

Progression to T1D in children with IA

We then examined whether vitamin D variants were associated with progression to T1D in IA-positive children (Table 2). Of the 148 IA-positive children in DAISY, 62 developed T1D; the mean age at T1D diagnosis was 8.6 years. The mean age at last follow-up visit in nondiabetic children with IA was 14.0 years. Children who developed T1D were significantly younger when they first tested positive for an autoantibody than IA-positive children who have not progressed to T1D, 3.9 and 7.1 years, respectively. Children who developed T1D were more likely to have the HLA-DR3/4,DQB1*0302 genotype compared with children who did not progress to T1D.

Table 2.

Characteristics of IA-Positive Subjects by T1D Status and Association of Vitamin D Metabolism SNPs in DAISY Non-Hispanic White Population (n = 148)

Characteristic	IA-Positive Children Who Progressed to T1D (n = 62)	IA-Positive Children Who Have Not Progressed to T1D (n = 86)	Adjusted HR and 95% CI	P Value	n
Mean age, ya	8.6 ± 3.9	14.0 ± 4.2	N/A	N/A	N/A
Mean age at first IA-positive visit, y	3.9 ± 3.0	7.1 ± 4.2	0.87 (0.79, 0.96)	.01	148
HLA-DR3/4,DQB1*0302	32 (51.6%)	20 (23.3%)	2.39 (1.47, 3.90)	.001	148
First-degree relative with T1D	45 (72.6%)	56 (65.1%)	1.09 (0.63, 1.89)	.76	148
Sex (female)	29 (46.8%)	46 (53.5%)	1.05 (0.62, 1.76)	.86	148
DHCR7/NADSYN1 rs12785878b,c	35 (56.5%)d	44 (51.2%)d	0.98 (0.69, 1.40)	.92	148
GC rs4588b,c	21 (43.8%)d	47 (57.3%)d	0.68 (0.42, 1.11)	.13	130
GC rs7041b,c	29 (59.2%)d	59 (72.0%)d	0.76 (0.50, 1.18)	.22	131
CYP2R1 rs10741657b,c	31 (62.0%)d	51 (63.8%)d	0.96 (0.68, 1.36)	.80	130
CYP2R1 rs12794714b,c	29 (60.4%)d	56 (69.1%)d	0.92 (0.63, 1.35)	.68	129
CYP24A1 rs6013897b,c	19 (38.8%)d	31 (37.8%)d	1.06 (0.68, 1.63)	.81	131
CYP27B1 rs4646536c					126
A/A	29 (59.2%)	43 (55.8%)	Referent
A/G	16 (32.7%)	25 (32.5%)	0.88 (0.46, 1.70)	.71
G/G	4 (8.2%)	9 (11.7%)	1.21 (0.40, 3.70)	.73

^aAge at T1D diagnosis in diabetic children or age at last follow-up in nondiabetic children.

^bSNP analyzed treating the number of minor alleles as a continuous variable, with HR representing an increase in risk for each additional minor allele.

^cAdjusted for HLA-DR3/4,DQB1*0302, first-degree relative with T1D, and age at first IA-positive visit.

^dn and percentage for individuals with at least one minor allele.

The observed number of T1D cases and number of person-years for each genotype of the vitamin D metabolism SNPs in the DAISY non-Hispanic white population are shown in Supplemental Table 1. There was no evidence of association between any of the vitamin D SNP variants and progression to T1D in IA-positive children, adjusting for HLA-DR3/4,DQB1*0302, first-degree relative with T1D, and age at first autoantibody visit.

Longitudinal analysis of plasma 25(OH)D levels and vitamin D SNPs

To explore the mechanism behind these genetic associations, we examined whether individual vitamin D variants were associated with plasma 25(OH)D levels in a subcohort of DAISY. Plasma 25(OH)D levels of all visits ranged from 6.7 to 72.0 ng/mL, with a mean of 29.5 ng/mL. A total of 238 of the 2176 subcohort visits, or 10.9%, had inadequate 25(OH)D levels (≤20 ng/mL). Using separate linear mixed models for each vitamin D variant adjusting for HLA-DR3/4,DQB1*0302, first-degree relative with type 1 diabetes, season of blood draw, age of study subjects, sex, vitamin D intake, caloric intake, and type of food frequency questionnaire completed, six of the seven SNPs were associated with plasma 25(OH)D levels (Table 3). The one SNP not associated with plasma 25(OH)D levels was CYP27B1 rs4646536. The mean unadjusted and adjusted 25(OH)D levels for each genotype of the vitamin D metabolism SNPs in the DAISY non-Hispanic white population are shown in Supplemental Table 2.

Table 3.

Plasma Vitamin D Levels Predicted by Vitamin D Metabolism SNPs in DAISY Non-Hispanic White Population (n = 262)

SNP	Unadjusted			Adjusteda			Subjects, n	Visits, n
	Estimate (SE)	95% CI	P Value	Estimate (SE)	95% CI	P Value
DHCR7/NADSYN1 rs12785878b	−1.15 (0.54)	−2.20, −0.09	.03	−1.13 (0.48)	−2.08, −0.19	.02	240	1740
GC rs4588b	−3.45 (0.54)	−4.52, −2.38	<.0001	−3.11 (0.49)	−4.08, −2.14	<.0001	225	1646
GC rs7041b	−2.25 (0.53)	−3.29, −1.22	<.0001	−2.10 (0.47)	−3.03, −1.17	<.0001	227	1651
CYP2R1 rs10741657b	1.97 (0.52)	0.95, 3.00	.0002	2.21 (0.46)	1.30, 3.12	<.0001	225	1638
CYP2R1 rs12794714b	−2.95 (0.51)	−3.95, −1.95	<.0001	−2.82 (0.44)	−3.69, −1.95	<.0001	224	1618
CYP24A1 rs6013897b	−1.93 (0.69)	−3.28, −0.58	.01	−1.59 (0.62)	−2.81, −0.36	.01	226	1649
CYP27B1 rs4646536							221	1595
A/A	Referent			Referent
A/G	1.01 (0.80)	−0.57, 2.59	.21	0.50 (0.74)	−0.95, 1.94	.50
G/G	−0.43 (1.40)	−3.17, 2.31	.76	−1.03 (1.27)	−3.52, 1.45	.42

^aAdjusted for HLA-DR3/4,DQB1*0302, first-degree relative with T1D, season of blood draw, age of study subjects, sex, vitamin D intake, caloric intake, and type of food frequency questionnaire completed.

^bSNP analyzed treating the number of minor alleles as a continuous variable with estimate representing mean difference in 25(OH)D for each additional minor allele.

Discussion

In exploring the associations between seven previously studied vitamin D SNPs and the development of IA and progression to T1D in the prospective DAISY cohort, DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 were associated with development of IA but not associated with progression to T1D in IA-positive children, suggesting these genes play a role early in the development of T1D related to the initial appearance of autoimmunity. Up to this point, these genetic variants had been investigated only in case-control studies using T1D as case status and not in prospective studies with IA as an outcome.

DHCR7/NADSYN1 rs12785878 was recently discovered as a novel marker for association with T1D in a study of seven genes in the vitamin D pathway (2) and is a novel locus for association with vitamin D status (1). DHCR7/NADSYN1 rs12785878 is located in the second intron of the NADSYN1 gene 8 kb 5′ from the transcription initiation site of DHCR7 on chromosome 11q12 and is not predicted to exert functional effects such as transcription factor binding site or splicing modification (21). DHCR7 encodes the rate-limiting enzyme 7-dehydrocholesterol reductase, which converts 7-dehydrocholesterol (7-DHC) to cholesterol in skin, removing the substrate 7-DHC from the synthetic pathway of the production of vitamin D₃, a precursor of 25(OH)D (1, 21–24). Vitamin D₃ is made in the skin when 7-DHC reacts with UV light (21). The G allele of DHCR7/NADSYN1 rs12785878 has been associated with lower levels of 25(OH)D and risk for T1D (2, 21, 23, 24), which we also found with plasma 25(OH)D levels and risk for IA. Our findings that DHCR7 is associated with both vitamin D levels and IA risk suggest the hypothesis that the effect of DHCR7 is mediated by 25(OH)D levels. However, the lack of association between 25(OH)D and IA risk in DAISY that we have reported previously (20) and confirmed in our current investigation (data not shown) suggests that the effect of this DHCR7 variant is not mediated through 25(OH)D levels and that this enzyme may influence diabetes risk via other mechanisms.

CYP27B1 rs4646536 was first found to be associated with T1D in a study of CYP27B1 and CYP24A1 gene polymorphisms (18). rs4646536 is located in intron 6 of the CYP27B1 gene on chromosome 12q13.1-q13.3, which encodes 1α-hydroxylase, the enzyme that converts 25(OH)D into 1,25(OH)₂D. Although CYP27B1 action is downstream of circulating 25(OH)D, causal variants could possibly alter their role in metabolic feedback loops or affect the speed at which 25(OH)D is metabolized. However, the observation that CYP27B1 rs4646536 is not associated with 25(OH)D levels in our cohort, which is similar to what has been seen previously (1, 2), suggests that the action of CYP27B1 in determining risk of IA is not mediated through 25(OH)D levels.

Vitamin D is of interest in the etiology of T1D because of observations that vitamin D supplementation during infancy was inversely associated with T1D (3). In DAISY, however, we recently reported that neither childhood vitamin D intake nor 25(OH)D levels were associated with the risk of IA or progression to T1D in IA-positive children (20). The low prevalence of vitamin D insufficiency [25(OH)D <20 ng/mL or <50 nmol/L] in the DAISY population (10.9%) may have been one reason we did not observe an association between 25(OH)D status and T1D and that these associations may be easier to detect in populations with a higher prevalence of vitamin D insufficiency, as was recently seen in a population of young adults (25). Results from the present study may elucidate the importance of vitamin D metabolism genes that affect the risk of IA, not necessarily through 25(OH)D levels. It is possible that other aspects of the metabolic pathway of vitamin D are more important than 25(OH)D levels in predicting the risk of IA and T1D. Other vitamin D biomarkers, such as 1,25(OH)₂D and vitamin D binding protein, should be investigated. Recently we found an interaction between rs1544410 in the vitamin D receptor gene and rs1893217 in the protein tyrosine phosphatase, nonreceptor type 2 gene associated with progression from IA to T1D but not the risk of IA (26). These findings combined with the current study suggest that vitamin D is acting at both phases in the natural history of T1D.

The major strength of this study is the prospective long-term follow-up from birth of children at an increased risk for T1D, which allowed us to differentiate between genetic risk factors for the appearance of autoimmunity and the subsequent progression to T1D. However, the cost of assembling and following such a unique cohort has limited the number of IA-positive children and children who progress to T1D that we could include in our analysis. Although we would have liked to explore specific, biologically relevant gene environment interactions between 25(OH)D levels (or vitamin D insufficiency) and vitamin D SNPs for an association with the development of IA and the progression to T1D in children with IA, we did not have adequate power. This is the first study to observe an association between DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536 and the development of IA. Therefore, it is important that these associations be investigated in other prospective cohorts to rule out type 1 error and lend further evidence to this association.

In conclusion, two novel intronic variants for the association with T1D, DHCR7/NADSYN1 rs12785878 and CYP27B1 rs4646536, were found to be significantly associated with the appearance of IA. Interestingly, these two variants were not found to be associated with progression to T1D in IA-positive children. These findings may offer insight concerning the complex role of vitamin D in the etiology of T1D.