Vitamin D binding protein rs7041 genotype alters vitamin D metabolism in pregnant women

Abstract

Research has identified reduced circulating 25-hydroxyvitamin D [25(OH)D] in individuals with the rs7041 (c.1296T>G) T allele in the vitamin D binding protein gene (GC); however, the effects of the T allele on vitamin D biomarkers during pregnancy and lactation are unknown. Thus, we examined the metabolic effects of GC rs7041 on vitamin D biomarkers among third-trimester pregnant (n = 26), lactating (n = 28), and nonpregnant/nonlactating (n = 21) women consuming a single amount of vitamin D (511 IU/d) and related nutrients for 10–12 wk. T allele carriers had less circulating 25(OH)D, regardless of reproductive state [thymine-thymine (TT): 80% of guanine-guanine (GG), P = 0.05; guanine-thymine (GT): 85% of GG, P = 0.1]. Among pregnant women, the T allele attenuated the expected increase in vitamin D binding protein (DBP). Specifically, although GG pregnant women exhibited greater DBP (216%, P < 0.0001) than did GG nonpregnant women, that difference was lessened among GT women, and TT pregnant women did not exhibit greater DBP than TT nonpregnant women. Furthermore, TT pregnant women had greater placental 25(OH)D₃ to 24,25-dihydroxyvitamin D ratios (251% of GG, P = 0.07) and less osteocalcin, a bone formation marker, in the cord blood of their neonates (24% of GT, P = 0.02). Overall, the GC rs7041 genotype modified the effects of pregnancy on maternal and placental vitamin D metabolism, with possible functional consequences for fetal bone development and infant health.—Ganz, A. B., Park, H., Malysheva, O. V., Caudill, M. A. Vitamin D binding protein rs7041 genotype alters vitamin D metabolism in pregnant women.

Mounting evidence suggests that vitamin D inadequacy during pregnancy and lactation is associated with adverse maternal, fetal, and infant bone health outcomes, as well as other adverse maternal outcomes, such as preeclampsia, gestational diabetes, obstructed labor, and infectious disease (1, 2). There have also been reports of adverse fetal outcomes, including aberrations in growth and developmental programming, as well as adverse infant health outcomes (1, 2). Despite increasing awareness of those health risks, vitamin D inadequacy, defined as serum 25-hydroxyvitamin D [25(OH)D] levels < 50 nM by the National Academy of Medicine (Washington, DC, USA), remains highly prevalent among pregnant and lactating women, with estimates ranging from 30 to 96% in different regions worldwide (3).

Serum 25(OH)D concentrations can be influenced by many factors, including UV exposure, seasonal variation, dietary vitamin D intake, body mass index (BMI), age, and race/ethnicity. Genetic factors may also affect serum 25(OH)D levels, and several genome-wide studies have demonstrated that common genetic variants in vitamin D metabolic components are significantly associated with serum 25(OH)D concentrations (4, 5). Thus, it is of interest to ascertain whether genetic factors modify alterations in vitamin D metabolism that occur during pregnancy and lactation and if they contribute to individual differences in vitamin D requirements in those reproductive states.

The vitamin D binding protein (DBP) gene (GC) rs7041 (c.1296T>G) is a single-nucleotide polymorphism that encodes an aspartate to glutamate amino acid change that alters the affinity of DBP for 25(OH)D. The T allele of that genotype has been associated with lower serum 25(OH)D in some, but not all, studies (4, 6–9). Although previous studies have identified differences in circulating 25(OH)D by rs7041 genotype in nonpregnant individuals, the effect of the variant on vitamin D biomarkers, including 25(OH)D in pregnancy and lactation, has not been extensively examined (10).

Taking advantage of a highly controlled, long-term feeding study previously conducted by our research group (11), the present study sought to evaluate the effects of the GC rs7041 genotype on vitamin D metabolism in pregnant, lactating, and nonpregnant/nonlactating women consuming vitamin D intakes approaching the 600 IU recommended dietary allowance.

MATERIALS AND METHODS

Participants and study design

As part of a 10–12-wk controlled-feeding study designed to assess the effect of increased choline intake during pregnancy and lactation, healthy third-trimester pregnant women (n = 26), lactating women 5 wk postpartum (n = 28), and nonpregnant women (n = 21) were randomized to either 480 mg choline/d or 930 mg choline/d. Throughout the controlled-feeding period, all participants consumed 511^+/− 48 IU/d of vitamin D. Of that total, a mean of 311^+/− 48 IU/d was provided by the study diet [estimated using the U.S. Department of Agriculture (USDA) National Nutrient Database for Standard Reference (release 28)], and 200 IU cholecalciferol/d came from a daily prenatal multivitamin supplement (Pregnancy Plus; Fairhaven Health, Bellingham, WA, USA). Mean intakes of calcium and phosphorus were estimated to be 1.6^+/− 0.4-g Ca/d and 1.9^+/− 0.3-g P/d [USDA National Nutrient Database for Standard Reference (release 28)]. All meals were prepared in the Human Metabolic Research Unit (HMRU) at Cornell University, and ≥3 meals (with study supplements) were consumed throughout the week at the HMRU. All other meals and supplements were provided as carry-out and were consumed off site. To ensure compliance during off-site meal consumption, participants were required to complete daily checklists of menu items and supplements and to return food containers and disposables of their supplements on their next visit to the HMRU. Additional details on the study protocol, diet, and supplements are provided in a prior publication (11). The study was approved by institutional review boards at Cornell University and Cayuga Medical Center (Ithaca, NY, USA), and was registered at The U.S. National Library of Medicine database (NCT03051867; http://clinicaltrials.gov) (11–13) All participants provided informed consent.

Sample collection and processing

Fasting blood samples (10 h) were collected at study wk 10 into a serum separator tube (used for serum vitamin D metabolite quantification) and EDTA-coated tubes. Samples were placed on ice immediately, and EDTA tubes were centrifuged (2200 rpm, 15 min, 4°C) to separate plasma (used for quantification of vitamin D metabolites) from buffy coat (used for genotyping). Samples were stored at −80°C until use. Placentas (n = 24) were collected and processed within 90 min of delivery, as previously described (11). Fetal cord blood samples (n = 23) were collected at delivery into EDTA-coated tubes and processed accordingly. Two placentas and 3 cord blood samples were not obtained because 2 participants gave birth at home, and 1 participant gave birth without notifying the research team (11–13).

Genotyping

The commercially available DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) was used to extract DNA from buffy coat samples. Genotyping of GC rs7041 was performed in our facility with Endpoint Genotyping on a LightCycler 480 (Roche Diagnostics USA, Indianapolis, IN, USA) and a commercially available mix (TaqMan Genotyping Master Mix, and Assay Mix; Thermo Fisher Scientific, Waltham). Samples were run in duplicate with in-run standards and a negative water control. Cycling conditions were as follows: 95°C for 10 min, followed by 34 cycles of 92°C for 15 s and 62°C for 90 s.

Quantification of vitamin D metabolites

25(OH)D, C3 epimer of 25-hydroxyvitamin D₃, and 24,25-dihydroxyvitamin D₃

Serum 25(OH)D [the sum of 25(OH)D₂ and 25(OH)D₃], serum C3 epimer of 25-hydroxyvitamin D₃ [3-epi-25(OH)D₃], and plasma 24,25-dihydroxyvitamin D₃ [24,25(OH)₂D] were quantified using an isotope-dilution liquid chromatography mass spectrometry (MS)/MS method, with modifications based on our instrumentation, as previously described (13–15). Assay precision and accuracy were assessed using the National Institute of Standards and Technology (Gaithersburg, MD, USA) Standard Reference Materials and the Vitamin D External Quality Assessment Scheme (Charing Cross Hospital, London, United Kingdom).

1,25-dihydroxyvitamin D₃ and DBP

ELISA kits were used to measure plasma 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃; Immunodiagnostic Systems, The Boldons, United Kingdom] and DBP (R&D Systems; Bio-Techne, Minneapolis, MN, USA).

Free 25-hydroxyvitamin D

Circulating free 25(OH)D was estimated using an equation as previously described (13).

Placental 25(OH)D₃ and 24,25(OH)₂D₃

Both 25(OH)D₃ and 24,25(OH)₂D₃ were extracted from homogenized placental tissues as previously described (16). Methanol (1 ml) was added to 0.4 g of homogenized placental tissues, along with 600 µl of acetonitrile, 1 ml of methyl tertiary butyl ether, and 25 µl of internal standard containing 215 M D₃-25(OH)D₃ (IsoSciences, King of Prussia, PA, USA) and 1.7 M D₆-24,25(OH)2D₃ (Toronto Research Chemicals, Toronto, ON, Canada) A liquid–liquid extraction was followed by a solid-phase extraction on Oasis HLB cartridges (3 cm³/60 mg; Waters, Milford, MA, USA) and subsequent derivatization with 50 µl of 0.1 mg/ml 4-(2-(6,7-dimethoxy-4-methyl-3-oxo-3,4-dihydroquinoxalinyl)ethyl)-1,2,4-triazoline-3,5-dione (DMEQ-TAD; Key Synthesis, Wynnewood, PA, USA) (14, 17). Finally, extracts were resuspended in 110 µl of 60:40 methanol:water, injected onto a pentafluorophenyl column (Accucore PFP 2.1 × 100, 2.6 mm; Thermo Fisher Scientific) with a matching guard column at 45°C, and separated by liquid chromatography–MS/MS (16).

Biochemical marker of bone formation in neonates (cord blood)

Osteocalcin (OC) was measured in cord plasma obtained at delivery using an ELISA kit (Bio-Techne) (12).

Statistical analysis

The effect of GC rs7041 genotype on vitamin D metabolism was assessed using linear models. Vitamin D metabolites in blood and placenta were modeled as a function of rs7041 genotype. Reproductive status, BMI, season, and ethnicity were included as covariates in the blood metabolite models, and age was considered a covariate and retained in the model at an α-cutoff of 0.05. Choline intake level was considered a covariate but was not significant in any model. We used a 2-level ethnicity/race outcome, white and nonwhite (including Latino, African American, Asian, and mixed race/other) because of the sparsity of nonwhite participants, although grouping these ethnicities did not change our results. Interactions between genotype and reproductive state were retained at an α-cutoff of 0.1. Analysis with and without ethnicity/race in the model yielded the same results. Given that little is known about the factors that influence vitamin D outcomes in the placenta and umbilical cord, the change method was used to identify covariates that influenced outcomes and to construct placental and umbilical cord blood models. BMI, season, ethnicity (white and nonwhite), age, infant sex, parity, and vitamin supplementation status at baseline were evaluated as possible covariates by examining 1 at a time as to whether they changed the estimate of genotype by >10% as compared with a bivariate model of genotype with the outcome of interest. Covariates that passed this test were assembled into an initial model, and a backward selection was performed in which the covariate with the highest P value was eliminated until all covariates included in the model had a significance value of P ≤ 0.2. Standard diagnostic methods were used to assess model assumptions and the fit of the models to the data. Some models had outliers and did not have perfectly normal distributions. 24,25(OH)D₃ met normality criteria after log transformation. ANOVA was determined to be robust enough to deal with all other models, and assessment of models that had outliers yielded normal models when these outliers were excluded and the same results qualitatively in terms of significance. Statistical analysis was performed using the lsmeans package in The R Project for Statistical Computing environment, available from the Comprehensive R Archive Network (18). Pairwise comparisons were performed with Student’s t tests, and reported P values include adjustments for multiple comparisons with Tukey’s test and were considered significant at an α-cutoff of 0.05.

RESULTS

Characteristics of the participants

Baseline characteristics by GC rs7041 genotype are displayed in Table 1.

TABLE 1.

Baseline characteristics

Characteristic	All (n = 74)	GG (n = 18)	GT (n = 36)	TT (n = 20)
BMI, means ± sd (kg/m2)	24.8 ± 5.1	23.3 ± 3.4	26.1 ± 6.1	23.9 ± 3.7
Ethnicity [n (%)]
White	49 (66)	13 (72)	23 (64)	13 (65)
Black	4 (5)	1 (6)	1 (3)	2 (10)
Latina	9 (12)	3 (17)	4 (11)	2 (10)
Asian	6 (8)	1 (6)	2 (6)	3 (15)
Other	6 (8)	0 (0)	6 (17)	0 (0)
Season at study entry (n)
April–September	41	10	19	12
October–March	33	8	17	8
Age, means ± sd (yr)	28.4 ± 5.6	30.3 ± 5.2	27.1 ± 6.2	29.1 ± 4.0
Nonpregnant women (n)	21	6	9	6
Lactating women (n)	27	3	16	8
Pregnant women (n)	26	9	11	6
Parity, primiparas/multiparas	12/14	5/4	3/8	4/2
Infant sex, male/female	19/7	6/3	7/4	6/0
Multivitamin supplement use before study entry (n)	49/74	10/18	26/36	13/20

Pregnant women self-reported prepregnancy BMIs.

25(OH)D

GC rs7041 genotype predicted 25(OH)D₂₊₃, independent of reproductive state (P = 0.04). Individuals with 1 or 2 copies of the T allele exhibited lower serum 25(OH)D₂₊₃ concentrations. Specifically, the rs7041 guanine-guanine (GG) genotype had a 25(OH)D level of 93.5^+/−5.5 nM, vs. heterozygotes [guanine-thymine (GT)] at 79.9^+/−3.8 nM (P = 0.1) and homozygous [thymine-thymine (TT)] genotypes at 75.4^+/−5.1 nM (P = 0.04) (Fig. 1A and Table 2).

Figure 1.

Effect of the GC rs7041 genotype on circulating maternal vitamin D metabolites at study wk 10. The metabolites include: total serum 25(OH)D (A), plasma DBP (B), free 25(OH)D (C), serum 3-epi-25(OH)D₃ (D), and plasma 25(OH)D₃:24,25(OH)₂D₃ (E). Total 25(OH)D was calculated as the sum of 25(OH)D₂ and 25(OH)D₃. Values are least-squared means ± se. Number of participants in each group: GG, n = 18; GT, n = 36; and TT, n = 20. Groups that are not labeled with any of the same letters (a, b, c, d) are significantly different at an α-cutoff of 0.05. DBP and the 25(OH)D₃ to 24,25(OH)₂D₃ ratio exhibited interactions between GC rs7041 genotype and reproductive state. Number of participants in each group for these 2 outcomes: lactating (GG, n = 3; GT, n = 16; and TT, n = 8); nonpregnant (GG, n = 6; GT, n = 9; and TT, n = 6), pregnant (GG, n = 9; GT, n = 11; and TT, n = 6).

TABLE 2.

Circulating vitamin D metabolites among pregnant, lactating, and nonpregnant women at study wk 10

Characteristic	GG	GT	TT	P (GG vs. GT)	P (GG vs. TT)	P (GT vs. TT)
Total 25(OH)D (nM), main effect; P = 0.04	93.5 ± 5.5	79.9 ± 3.8	75.4 ± 5.1	0.1	0.05	0.8
25(OH)D2 (nM), main effect; P = 0.5	1.7 ± 0.4	1.4 ± 0.2	1.2 ± 0.3	0.8	0.5	0.8
25(OH)D3, nM; main effect; P = 0.05	91.8 ± 5.5	78.5 ± 3.8	74.2 ± 5.1	0.1	0.05	0.8
DBP (μg/ml), genotype × RS interaction; P = 0.0002
Nonpregnant	313 ± 32	237 ± 26	125 ± 32	0.6	0.003	0.2
Lactating	415 ± 45	172 ± 21	79 ± 29	0.0004	<0.0001	0.2
P vs. nonpregnant	0.7	0.6	0.97
Pregnant	663 ± 27	383 ± 24	169 ± 33	<0.0001	<0.0001	0.0001
P vs. nonpregnant	<0.0001	0.003	0.99
Free 25(OH)D2,3 (pM), main effect; P = 2 × 10−9	18.5 ± 3.1	24.3 ± 2.1	46.0 ± 2.8	0.3	<0.0001	<0.0001
3-epi-25(OH)D3 (nM), main effect; P = 0.03	3.8 ± 0.5	2.4 ± 0.4	1.9 ± 0.5	0.09	0.02	0.7
25(OH)D3 to 24,25(OH)2D3 ratio; genotype × RS interaction; P = 0.1
Nonpregnant	7.0 ± 0.7	7.6 ± 0.6	6.4 ± 0.7	>0.99	>0.99	0.9
Lactating	9.4 ± 1.0	6.9 ± 0.5	7.4 ± 0.6	0.4	0.7	>0.99
P vs. nonpregnant	0.5	0.99	0.97
Pregnant	9.9 ± 0.6	7.7 ± 0.5	8.4 ± 0.7	0.07	0.8	0.99
P vs. nonpregnant	0.04	>0.99	0.6
1,25(OH)2D3 (pM)	208 ± 29	228 ± 21	204 ± 26	0.8	0.99	0.8
24,25(OH)2D3 (nM)	10.2 ± 1.3	10.1 ± 1.0	10.2 ± 1.2	0.99	>0.99	0.99

Values are least-squared means ± se. Number of participants in each group: GG, n = 18; GT, n = 36; and TT, n = 20. Total 25(OH)D was calculated as the sum of 25(OH)D₂ and 25(OH)D₃. No significant interactions were detected between genotype and reproductive state (RS) for each vitamin D metabolite, unless otherwise indicated. Study wk 10 plasma DBP and the 25(OH)D₃ to 24,25(OH)₂D₃ ratio exhibited interactions between GC rs7041 genotype and RS and comparisons are also presented within genotypes across reproductive states. Number of participants in each group for these 2 outcomes: lactating, GG (n = 3), GT (n = 16), and TT (n = 8); nonpregnant, GG (n = 6), GT (n = 9), and TT (n = 6); pregnant, GG (n = 9), GT (n = 11), and TT (n = 6).

This effect of genotype on 25(OH)D was most prominent in the D₃ component, main effect, P = 0.07. Individuals with 1 or 2 copies of the T allele exhibited lower serum 25(OH)D₃ concentrations. Specifically, rs7041 GG individuals had a level of 91.8^+/−5.5 nM vs. GT heterozygote individuals at 78.5^+/−3.8 nM (P = 0.1) and the homozygous TT group at 74.2^+/−5.1 nM (P = 0.05) (Supplemental Fig. 1B and Table 2). The rs7041 genotype did not significantly predict serum 25(OH)D₂ (P = 0.5) (Supplemental Fig. 1A and Table 2).

DBP

The rs7041 genotype influenced plasma DBP concentrations in our cohort in a manner that depended upon reproductive state (reproductive state × genotype interaction, P = 0.0002).

Within each reproductive state, the T allele was associated with lower concentrations of DBP (P < 0.003) (Table 2), with the exception of GT nonpregnant women, who were not significantly different than GG nonpregnant women. Furthermore, the difference in DBP between pregnant and nonpregnant women varied by rs7041 genotypes. Among the women with the GG genotype, GG pregnant women exhibited more than 2 times greater DBP than GG nonpregnant women (663^+/−27 vs. 313^+/−32 μg/ml, P < 0.0001). However, that difference was lessened among the GT heterozygote members to roughly 1.6 times greater DBP levels in GT pregnant women than GT nonpregnant women (383^+/−24 vs. 237^+/−26 μg/ml; P = 0.003). Moreover, no difference in DBP was observed between homozygous TT pregnant and TT nonpregnant women (169^+/−33 vs. 125^+/−32 μg/ml, P = 0.99), suggesting that TT women do not increase DBP during pregnancy (Fig. 1B and Table 2).

Free 25(OH)D

GC rs7041 genotype was a significant predictor of circulating free 25(OH)D (P = 2 × 10⁻⁹). This effect was independent of reproductive state. Individuals with the TT genotype exhibited greater concentrations (46.0^+/−2.8 pM) of free 25(OH)D than did those with the GG (18.5^+/−3.1 pM, P < 0.0001) and GT genotypes (24.3^+/−2.1 pM, P < 0.0001) (Fig. 1C and Table 2).

3-Epi-25(OH)D₃

In addition, genotype predicted serum 3-epi-25(OH)D₃ (P = 0.03). That effect was independent of reproductive state. Additional copies of the T allele were associated with successively lower concentrations of 3-epi-25(OH)D₃. Women with the GG genotype exhibited concentrations of 3.8^+/−0.5 nM, which were greater than GT (2.4^+/−0.4 nM; P = 0.09) and TT (1.9^+/−0.5 nM, P = 0.02) individuals (Fig. 1D and Table 2).

25(OH)D₃ to 24,25(OH)₂D₃ ratio

A genotype by reproductive state interaction tended to predict the 25(OH)D₃ to 24,25(OH)₂D₃ ratio (P = 0.1). Genotypic differences were not observed among nonpregnant and lactating women. However, among pregnant women, GG individuals tended to exhibit greater circulating 25(OH)D₃ to 24,25(OH)₂D₃ ratios than did those with the GT genotype (9.9^+/−0.6 vs. 7.7^+/−0.5; P = 0.07) (Fig. 1E and Table 2).

In addition, the metabolic response to pregnancy appeared to differ among genotypes. Among GG individuals, pregnant women exhibited greater circulating 25(OH)D₃ to 24,25(OH)₂D₃ ratios than nonpregnant women. Specifically, GG pregnant women displayed 25(OH)D₃ to 24,25(OH)₂D₃ ratios of 9.9^+/−0.6, which were significantly greater than nonpregnant GG women (7.0^+/−0.7, P = 0.04). In contrast, pregnant women with the GT and TT genotypes did not exhibit significantly greater 25(OH)D₃ to 24,25(OH)₂D₃ ratios than nonpregnant women with the same genotype. (Fig. 1E and Table 2).

1,25(OH)₂D₃ and 24,25(OH)₂D₃

Genotypic differences in circulating 1,25(OH)₂D₃ and 24,25(OH)₂D₃ were not detected (Table 2).

Placental 24,25(OH)₂D₃

In the placenta, genotype predicted 24,25(OH)₂D₃ (P = 0.04). Additional copies of the T allele were associated with lower concentrations of 24,25(OH)₂D₃. Women with the GG genotype exhibited concentrations of 0.88^+/−0.27 log(nM), which were greater than results from GT (0.06^+/−0.33 log(nM), P = 0.1) and TT (−0.23^+/−0.34 log(nM), P = 0.04) individuals (Fig. 2A and Table 3).

Figure 2.

Effect of GC rs7041 genotype on placental vitamin D metabolites, including 24,25(OH)₂D₃ (A), 25(OH)D₃ (B), and 25(OH)D₃:24,25(OH)₂D₃ (C). Values are least-squared means ± se. Number of participants in each group: GG, n = 8; GT, n = 10; and TT, n = 6. Groups that are not labeled with any of the same letters (a, b) are significantly different at an α-cutoff of 0.05.

TABLE 3.

Placental vitamin D metabolites and OC in neonatal cord blood at delivery

Characteristic	GG	GT	TT	P (GG vs. GT)	P (GG vs. TT)	P (GT vs. TT)
Placental 24,25(OH)2D3, log(nM); main effect; P = 0.04	0.88 ± 0.27	0.62 ± 0.33	−0.24 ± 0.34	0.1	0.04	0.8
Placental 25(OH)D3 (nM), main effect; P = 0.8	17.9 ± 3.1	18.0 ± 2.4	10.9 ± 3.9	>0.99	0.4	0.3
Placental 25(OH)D3 to 24,25(OH)2D3 ratio; main effect; P = 0.02	8.8 ± 3.7	14.8 ± 2.9	22.1 ± 4.0	0.3	0.07	0.4
Umbilical cord plasma OC (ng/ml), main effect; P = 0.02	49.0 ± 12.3	79.4 ± 11.9	18.8 ± 15.3	0.2	0.3	0.02

Values are least-squared means ± se. Number of participants in each group for placental outcomes: GG, n = 8; GT, n = 10; and TT, n = 6. Number of participants in each group for cord blood plasma OC: GG, n = 8; GT, n = 8; and TT, n = 6.

Placental 25(OH)D₃

Genotypic differences in placental 25(OH)D₃ were not detected. However, TT individuals exhibited nonsignificantly lower placental 25(OH)D₃ than GT and GG individuals (Fig. 2B and Table 3).

Placental 25(OH)D₃ to placental 24,25(OH)₂D₃ ratio

Genotype predicted the 25(OH)D₃ to 24,25(OH)₂D₃ ratio (P = 0.02). Individuals with the TT genotype tended to exhibit greater placental 25(OH)D₃ to 24,25(OH)₂D₃ ratios than did those with the GG genotype (22.1^+/−4.0 vs. 8.8^+/−3.7, P = 0.06) (Fig. 2C and Table 3).

Cord OC

GC rs7041 genotype predicted cord OC (P = 0.02). GG individuals had concentrations of 49.0^+/−12.3 ng/ml, GT individuals had 79.4^+/−11.9 ng/ml, and TT individuals had 18.8^+/−15.3 ng/ml (P = 0.02 compared with GT) (Fig. 3 and Table 3).

Figure 3.

Effect of GC rs7041 genotype on cord blood OC. Values are least-squared means ± se. Number of participants in each group: GG, n = 8; GT, n = 8; and TT, n = 6. Groups that are not labeled with any of the same letters (a, b) are significantly different at an α-cutoff of 0.05.

DISCUSSION

This study evaluated the effects of the GC rs7041 genotype on vitamin D metabolism in pregnant, lactating, and nonpregnant/nonlactating women consuming vitamin D intakes approaching the recommended dietary allowance. Three main findings emerged: 1) GC rs7041 alters circulating vitamin D metabolites in pregnant, lactating, and nonpregnant women; 2) the GC rs7041 T allele is associated with an altered metabolic response to pregnancy and altered placental vitamin D metabolism; and 3) neonates of women with the TT genotype exhibited lower levels of OC, a biochemical marker of bone formation.

GC rs7041 alters vitamin D metabolism in pregnant, lactating, and nonpregnant women

Within our cohort, the GC rs7041 T allele was associated with significantly lower serum total 25(OH)D concentrations, independent of reproductive state. Specifically, we observed lower serum 25(OH)D among heterozygous GT (85% of GG; P = 0.1) and homozygous TT (80% of GG; P = 0.05) individuals across reproductive states (Fig. 1A and Table 2). This finding confirms previous reports of lower concentrations of 25(OH)D in a nonpregnant population with the rs7041 T allele (4, 6–9) and expands on previous results to include pregnant and lactating women. Serum 25(OH)D is the primary indicator of vitamin D status, and women with the rs7041 T allele may be at a greater risk for poor vitamin D status as compared to women with the G allele. Within our cohort, the genotypic effect appeared to be driven by 25(OH)D₃ because we did not detect significant differences in 25(OH)D₂ (Supplemental Fig. 1 and Table 2), which is not surprising given that our study provided cholecalciferol (vitamin D₃) as the dietary supplement. We also observed lower concentrations of 3-epi-25(OH)D₃ in GT (63% of GG; P = 0.09) and TT (50% of GG; P = 0.02) women, consistent with their lower 25(OH)D concentrations (Fig. 1D and Table 2).

In addition, we observed lower DBP concentrations among individuals with the rs7041 T allele. Nonpregnant TT women displayed lower levels of DBP than did nonpregnant GG women; lactating GT and TT women displayed lower concentrations than lactating GG women, and pregnant GT and TT women displayed lower DBP concentrations than pregnant GG women. (Fig. 1B and Table 2).

Finally, homozygous TT women exhibited higher free 25(OH)D concentrations (1.9 times that of GT women and 2.5 times that of GG women) (Fig. 1D and Table 2). Along with the evidence of lower total 25(OH)D concentrations (comprising free and DBP-bound forms) in TT women, these data support lower concentrations of DBP-bound 25(OH)D among women with the variant T allele. These findings align with in vitro studies that report differential affinities of 25(OH)D for different DBP isoforms (9, 19). Notably, some studies have reported stronger associations of free 25(OH)D (as compared with DBP-bound vitamin D) with functional outcomes, especially in clinical situations in which DBP is elevated or suppressed. (20) As such, the merits of “free 25(OH)D,” as a marker of vitamin D status, are being considered. (20)

GC rs7041 T allele is associated with an altered metabolic response to pregnancy

Women with the GC rs7041 T allele appear to have a distinct metabolic response to pregnancy as compared to women with the G allele. Previously, DBP concentrations have been reported to increase by 7–152% during pregnancy (2, 21–23). Within our cohort, GG pregnant women exhibited 2.1 times higher plasma DBP concentrations than did GG nonpregnant women. However, that pregnancy-associated increase in DBP appeared to be lessened among heterozygote individuals because GT pregnant women exhibited only 1.6 times higher plasma DBP than shown by GT nonpregnant women, and differences in DBP were not detected between TT pregnant and TT nonpregnant women. Overall, these data suggest that DBP metabolism is modulated by the rs7041 genotype during pregnancy, with no increase for TT pregnant women (Fig. 1B and Table 2). A differential increase in DBP during pregnancy by rs7041 genotype may account for some of the variability among previous reports of increases in DBP during pregnancy. This result should be confirmed in a longitudinal study that compares DBP concentrations during pregnancy to either pre- or postpregnancy concentrations in the same individual. Although it is not clear why TT women fail to increase DBP during pregnancy, the finding indicates that TT women have a different metabolic response to pregnancy than GG women and may differ in the mechanism by which they source vitamin D to the placenta.

The mechanism by which vitamin D enters the placenta is currently unknown; however, there are several possibilities. Placental uptake of vitamin D metabolites could occur preferentially via megalin–cubilin receptors, by simple diffusion in the free form, or in some combination of the 2. DBP of maternal origin has been detected in the human placenta (24), providing evidence that the placenta does, in fact, take up DBP, although it is not known whether this constitutes the major source of placental vitamin D. Moreover, the placental membrane expresses megalin–cubilin receptors (25, 26) [the same receptors that the kidney uses to internalize DBP-bound 25(OH)D], as well as all the necessary components to make the active form of vitamin D (27–29), providing further rationale for placental uptake of DBP-bound 25(OH)D. Among GG individuals, the increase in DBP during pregnancy (a known phenomenon), may transpire to increase the uptake of DBP-bound 25(OH)D by the tissues that express megalin–cubilin receptors, including the placenta. DBP-bound 25(OH)D is more stable and can be taken up across a concentration gradient and, thus, may be favored by the placenta. In contrast, among TT women, the lack of increase in DBP, along with the increase in free 25(OH)D, suggests that, in an environment with compromised DBP concentrations, free 25(OH)D may be a more biologically important form of vitamin D for people with this genotype. Thus, the functional link of vitamin D with DBP and the predominant mechanism of placental uptake of vitamin D during pregnancy may differ by GC rs7041 genotype. Isotope-labeling studies are needed to further investigate this notion.

We also observed differences in the blood 25(OH)D₃ to 24,25(OH)₂D₃ ratio that depended upon genotype and reproductive state. Although GG pregnant women exhibited higher 25(OH)D₃ to 24,25(OH)₂D₃ ratios than GG nonpregnant women, we did not observe the same phenomenon in TT pregnant women as compared with TT nonpregnant women (Fig. 1E and Table 2). Given that the 25(OH)D₃ to 24,25(OH)₂D₃ ratio is an emerging status indicator that accounts for interindividual metabolic differences in catabolizing 25(OH)D to 24,25(OH)₂D, this finding is a second indication [backing up 25(OH)D data] that pregnant women with the T allele may have lower-status vitamin D levels, or more aptly, may not increase their vitamin D status during pregnancy by the same mechanism as GG women. No such differences were detected in the metabolic response to lactation.

GC rs7041 T allele is associated with altered placental vitamin D catabolite concentrations

Interestingly, in line with the notion that TT individuals may source vitamin D differently to the placenta, they also exhibited differences in placental vitamin D metabolites. TT individuals exhibited nonsignificantly lower concentrations of placental 25(OH)D₃, significantly lower 24,25(OH)₂D₃, and tended to exhibit greater placental 25(OH)D₃ to 24,25(OH)₂D₃ ratios as compared with GG individuals (Fig. 2 and Table 3). Substantially higher (nearly 3 times GG) placental 25(OH)D₃ to 24,25(OH)₂D₃ ratios among TT individuals indicates comparatively less vitamin D catabolism relative to 25(OH)D. A reduction in vitamin D catabolism relative to 25(OH)D among TT pregnant women may transpire to secure the placental 25(OH)D pool as a compensatory response to the overall reduced presence and availability of vitamin D in TT individuals during pregnancy.

GC rs7041 TT genotype is associated with a reduced presence of a biochemical marker of bone formation in neonates

The neonates of TT women exhibited lower cord OC as compared with GT individuals (Fig. 3 and Table 3), suggesting that this genetic variant may have functional consequences for fetal bone health and bone development. However, we did not detect differences between TT and GG individuals, which could indicate inadequate power, a heterozygote effect, or a spurious finding. Nonetheless, given the role of OC in regulating glucose metabolism and the links between low DBP and type 1 diabetes (30) and GC rs7041 and diabetes (31), additional studies on the effect of the TT genotype on OC are warranted.

Study limitations

Concerns have been raised about whether the mAb-based assay we used to measure DBP may have distinct immunoreactivity against different isoforms of DBP that result from different genotypes (5). Specifically, previous studies that have used this mAb assay have found lower DBP in certain Gc types, whereas studies that use pAb assays or direct measurement by proteomic methods and MS do not find differences by race or Gc type (32–34). A differential affinity of the DBP assay for the TT genotype, as compared with the GG genotype, limits the extent to which we can draw conclusions from comparisons of DBP concentrations among genotypes. However, it would not be expected to interfere with comparisons within genotypes. Importantly, the finding regarding the apparent lack of increase in DBP during pregnancy among TT individuals is based on comparisons among reproductive groups within genotypes (i.e., TT pregnant vs. TT nonpregnant women). A direct measurement of DBP as well as free 25(OH)D, which was estimated using an equation, would have facilitated comparisons among genotypes.

In addition, these findings represent our study population and cannot necessarily be extrapolated to all individuals with the rs7041 T allele. Combinations of this rs7041 (c.1296T>G; p.Asp416Glu) genotype and an additional single-nucleotide polymorphism rs4588 (c.1307C>A; p.Thr420Lys) result in 3 common isoforms of the vitamin D binding protein: Gc1F (rs7041 T and rs4588 C alleles); Gc1S (rs7041 G and rs4588 C alleles); and Gc2 (rs7041 T and rs4588 A alleles). The rs7041 G allele generally accompanies the rs4588 C allele and generates the Gc1S isoform, whereas the rs7041 T allele can generate either the Gc1F or Gc2 isoform, depending on the individual’s rs4588 genotype. The prevalence of these isoforms varies across racial groups. The Gc1F isoform is more prevalent among black and Asian populations; the Gc1S isoform is most common in whites, but among whites with the GC rs7041 T allele, the Gc2 isoform is more common than the Gc1F isoform (19, 32, 35). Given that our study population was primarily white, it is likely that individuals with the T allele primarily had the Gc2 isoform rather than the Gc1f isoform. Because we did not have data for rs4588 (or a large enough sample size to include additional groups), that homogeneity in race/ethnicity helped our analysis by limiting the variability; however, it also limits the extent to which these results can be extrapolated to other Gc types. All TT pregnant women gave birth to female infants. This was considered in all our statistical models, and infant sex did not have a significant bivariate relationship with any of the placental outcomes or cord OC.

CONCLUSIONS

Our study demonstrates that the GC rs7041 genotype alters vitamin D metabolism as well as the metabolic response to pregnancy and a marker of bone formation in neonates. Notably, although DBP is known to increase markedly during pregnancy, this phenomenon was observed only in individuals with rs7041 GG or GT genotypes; pregnant TT women in our cohort did not exhibit higher DBP concentrations than TT nonpregnant women. The difference in magnitude of DBP changes during pregnancy by genotype may have implications for placental vitamin D transport and the availability of vitamin D to the mother and fetus. Future studies with larger sample sizes are needed to comprehensively characterize the effects of the 3 common Gc isoforms on vitamin D metabolism during pregnancy and the consequences of those metabolic differences for maternal, fetal, and infant health.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

1,25(OH)₂D₃

1,25-dihydroxyvitamin D₃

3-epi-25(OH)D₃

C3 epimer of 25-hydroxyvitamin D₃

24,25(OH)₂D

24,25-dihydroxyvitamin D

25(OH)D

25-hydroxyvitamin D

BMI

body mass index

DBP

vitamin D binding protein

GC

the vitamin D binding protein gene

GG

guanine-guanine

GT

guanine-thymine

HMRU

Human Metabolic Research Unit

LC

liquid chromatography

MS

mass spectrometry

OC

osteocalcin

TT

thymine-thymine

AUTHOR CONTRIBUTIONS

A. B. Ganz, H. Park, and M. A. Caudill designed the research; H. Park, and O. V. Malysheva conducted the research; A. B. Ganz performed genotype assessment of participants, analyzed the data, and wrote the manuscript with input from H. Park; M. A. Caudill had primary responsibility for the final content; and all authors read and approved the final manuscript.